H.264 Encoder 2.0 On HDVICP2 And Media Controller Based Platform User’s Guide H264 User

User Manual:

Open the PDF directly: View PDF .
Page Count: 256 [warning: Documents this large are best viewed by clicking the View PDF Link!]

H.264 Encoder 2.0 on HDVICP2 and

Media Controller Based Platform

User’s Guide

Literature Number: SPRUHG3

March 2015

This page is intentionally left blank

IMPORTANT

NOTICE

Texas Instruments Incorporated and its subsidiaries (TI) reserve the right to make corrections, enhancements, improvements and other

changes to its semiconductor products and services per JESD46, latest issue, and to discontinue any product or service per JESD48, latest

issue. Buyers should obtain the latest relevant information before placing orders and should verify that such information is current and

complete. All semiconductor products (also referred to herein as “components”) are sold subject to TI’s terms and conditions of sale supplied

at the time of order acknowledgment.

TI warrants performance of its components to the specifications applicable at the time of sale, in accordance with the warranty in TI’s terms

and conditions of sale of semiconductor products. Testing and other quality control techniques are used to the extent TI deems necessary to

support this warranty. Except where mandated by applicable law, testing of all parameters of each component is not necessarily performed.

TI assumes no liability for applications assistance or the design of Buyers’ products. Buyers are responsible for their products and

applications using TI components. To minimize the risks associated with Buyers’ products and applications, Buyers should provide

adequate design and operating safeguards.

TI does not warrant or represent that any license, either express or implied, is granted under any patent right, copyright, mask work right, or

other intellectual property right relating to any combination, machine, or process in which TI components or services are used. Information

published by TI regarding third-party products or services does not constitute a license to use such products or services or a warranty or

endorsement thereof. Use of such information may require a license from a third party under the patents or other intellectual property of the

third party, or a license from TI under the patents or other intellectual property of TI.

Reproduction of significant portions of TI information in TI data books or data sheets is permissible only if reproduction is without alteration

and is accompanied by all associated warranties, conditions, limitations, and notices. TI is not responsible or liable for such altered

documentation. Information of third parties may be subject to additional restrictions.

Resale of TI components or services with statements different from or beyond the parameters stated by TI for that component or service

voids all express and any implied warranties for the associated TI component or service and is an unfair and deceptive business practice.

TI is not responsible or liable for any such statements.

Buyer acknowledges and agrees that it is solely responsible for compliance with all legal, regulatory and safety-related requirements

concerning its products, and any use of TI components in its applications, notwithstanding any applications-related information or

support that

may be provided by TI. Buyer represents and agrees that it has all the necessary expertise to create and implement safeguards which

anticipate dangerous consequences of failures, monitor failures and their consequences, lessen the likelihood of failures that might cause

harm and take appropriate remedial actions. Buyer will fully indemnify TI and its representatives against any damages arising out of the use of

any TI components in safety-critical applications.

In some cases, TI components may be promoted specifically to facilitate safety-related applications. With such components, TI’s goal is to

help enable customers to design and create their own end-product solutions that meet applicable functional safety standards and

requirements. Nonetheless, such components are subject to these terms.

No TI components are authorized for use in FDA Class III (or similar life-critical medical equipment) unless authorized officers of the parties

have executed a special agreement specifically governing such use.

Only those TI components which TI has specifically designated as military grade or “enhanced plastic” are designed and intended for use in

military/aerospace applications or environments. Buyer acknowledges and agrees that any military or aerospace use of TI

components

which

have not been so designated is solely at the Buyer's risk, and that Buyer is solely responsible for compliance with all legal and regulatory

requirements in connection with such use.

TI has specifically designated certain components as meeting ISO/TS16949 requirements, mainly for automotive use. In any case of use of

non-designated products, TI will not be responsible for any failure to meet ISO/TS16949.

Products Applications

Audio www.ti.com/audio Automotive and Transportation www.ti.com/automotive

Amplifiers amplifier.ti.com Communications and Telecom www.ti.com/communications

Data Converters dataconverter.ti.com Computers and Peripherals www.ti.com/computers

DLP® Products www.dlp.com Consumer Electronics www.ti.com/consumer-apps

DSP dsp.ti.com Energy and Lighting www.ti.com/energy

Clocks and Timers www.ti.com/clocks Industrial www.ti.com/industrial

Interface interface.ti.com Medical www.ti.com/medical

Logic logic.ti.com Security www.ti.com/security

Power Mgmt power.ti.com Space, Avionics and Defense www.ti.com/space-avionics-defense

Microcontrollers microcontroller.ti.com Video and Imaging www.ti.com/video

RFID www.ti-rfid.com

OMAP Applications Processors www.ti.com/omap TI E2E Community e2e.ti.com

Wireless Connectivity www.ti.com/wirelessconnectivity

Mailing Address: Texas Instruments, Post Office Box 655303, Dallas, Texas 75265

Incorporated

This page is intentionally left blank

Preface

Read This First

About This Manual

This document describes how to install and work with Texas Instruments’ (TI) H.264

Encoder implementation on the HDVICP2 and Media Controller Based Platform. It also

provides a detailed Application Programming Interface (API) reference and information on

the sample application that accompanies this component.

TI’s codec implementations are based on the eXpressDSP Digital Media (XDM) standard.

XDM is an extension of the eXpressDSP Algorithm Interface Standard (XDAIS).

Intended Audience

This document is intended for system engineers who want to integrate TI’s codecs with other

software to build a multimedia system based on the HDVICP2 and Media Controller Based

Platform.

This document assumes that you are fluent in the C language, have a good working

knowledge of Digital Signal Processing (DSP), digital signal processors, and DSP

applications. Good knowledge of eXpressDSP Algorithm Interface Standard (XDAIS) and

eXpressDSP Digital Media (XDM) standard will be helpful.

How to Use This Manual

This document includes the following chapters:

 Chapter 1 - Introduction, provides a brief introduction to the XDAIS and

XDM standards. It also provides an overview of the codec and lists its

supported features.

 Chapter 2 - Installation Overview, describes how to install, build, and run

the codec.

 Chapter 3 - Sample Usage, describes the sample usage of the codec.

 Chapter 4 - API Reference, describes the data structures and interface

functions used in the codec.

 Chapter 5 - Frequently Asked Questions, provides answers to few

frequently asked questions related to using this encoder.

 Appendix A - Meta Data Support, explains the meta data support by

encoder.

 Appendix B - Control for Configurable NALU, explains the configurable

NAL unit support by encoder.

Read This First

 Appendix C - Control for User Defined Scaling Matrices, explains the

mechanism of supporting user defined scaling matrices.

 Appendix D - Motion Vector and SAD Access API, describes the method

to access MV and SAD (Analytic Information) data dumped by the encoder.

 Appendix E – Debug Trace Support, describes the method to use H.264

encoder debug and trace mechanism.

 Appendix F – Picture Format, describes the different format of

uncompressed video, which are supported by encoder and the constraints

 Appendix G – Low Latency / Sub Frame Level Synchronization,

describes the method to achieve ultra low latency on the input and output

side of video encoder.

 Appendix H – Long Term Reference Picture Schemes, describes the

method to get long-term reference picture schemes to get error resilient

compressed bit-stream.

 Appendix I – Hierarchical P Structure Coding Scheme, describes the

method of Hierarchical P structure coding scheme to get bit-stream which

has flexibility to have a scalable bitstream in terms of bitrate and framerate

without adding any additional dealy.

 Appendix J – Mapping of Encoding Presets, describes the method to use

extended parameters values that user need to set to meet the exact

behaviour of a particular encoding preset

 Appendix K – Region of Interest Encoding, describes the method to

enable and use the Region of Interest feature.

 Appendix L – Watermarking SEI Message, provides information on the

support for watermarking in this encoder

 Appendix M – N Frame Process Call Support, describes the usage of N

frame processing in single process call.

 Appendix N – Rate Control - High Fidelity Variable Bitrate, provides an

insight to the High Fidelity Variable Bitrate (HF-VBR) Rate Control details of

the encoder.

 Appendix O – Gradual Decoder Refresh (GDR) - an Error resilience

Feature, provides a brief understanding usage details of GDR.

Related Documentation From Texas Instruments

The following documents describe TI’s DSP algorithm standards such as, XDAIS and XDM.

To obtain a copy of any of these TI documents, visit the Texas Instruments website at

www.ti.com.

 TMS320 DSP Algorithm Standard Rules and Guidelines (literature number

SPRU352) defines a set of requirements for DSP algorithms that, if followed,

allow system integrators to quickly assemble production-quality systems

from one or more such algorithms.

Read This First

iii

 TMS320 DSP Algorithm Standard API Reference (literature number

SPRU360) describes all the APIs that are defined by the TMS320 DSP

Algorithm Interface Standard (also known as XDAIS) specification.

 Technical Overview of eXpressDSP - Compliant Algorithms for DSP

Software Producers (literature number SPRA579) describes how to make

algorithms compliant with the TMS320 DSP Algorithm Standard which is

part of TI’s eXpressDSP technology initiative.

 Using the TMS320 DSP Algorithm Standard in a Static DSP System

(literature number SPRA577) describes how an eXpressDSP-compliant

algorithm may be used effectively in a static system with limited memory.

 Using IRES and RMAN Framework Components for C64x+ (literature

number SPRAAI5), describes the IRES interface definition and function

calling sequence.

 eXpressDSP Digital Media (XDM) Standard API Reference (literature

number SPRUEC8)

directory

2.3.2 Installing HDVICP2 library

The HDVICP2 library should be available in the same place as the codec package.

Set a system environment variable named HDVICP2_INSTALL_DIR pointing to

<hdvicp2_directory>\hdvicp20

The test application uses the HDVICP20 library file (ivahd_ti_api_vM3.lib) from

<hdvicp2_directory>\hdvicp20\lib directory

Installation Overview

2-8

2.4 Building and Running the Sample Test Application

The sample test application that accompanies this codec component will run in TI’s Code

Composer Studio development environment. To build and run the sample test application in

Code Composer Studio, follow these steps:

1) Verify that you have installed TI’s Code Composer Studio version Version: 4.2.0.09000

and code generation tools version 4.5.1.

2) Start the code composer studio and set up the target configuration for platform specific

simulator / Emulator

3) Verify that the following codec object libraries exist in \Lib sub-directory

h264enc_ti_host.lib: H.264 encoder library for Ducati

4) Verify that the following codec object libraries exist in \Lib sub- directory (in case of

library based / object code release):

5) Open the Code Composer Studio debug window with the appropriate platform

configuration chosen.

6) Build the sample test application project by gmake

a) Client\Build\TestAppDeviceName\make> gmake -s deps

b) Client\Build\TestAppDeviceName\make> gmake -k -s all

7) All files required for this project are available at the path

\Client\Build\TestAppDeviceName

8) The above step creates an executable file, TestAppEncoder.out in the

\Client\Build\TestAppDeviceName\Out sub-directory.

9) Select Target > Load Program on M3_Video, browse to the \Client\Build\

TestAppDeviceName\Out sub-directory, select the codec executable created in step 6,

and load it into Code Composer Studio in preparation for execution.

10) If you are using sub-system simulator then make sure that iCONT1 and iCONT2 are in

running state, even without loading any program. If you are using platform simulator or

EVM then this step is not needed

11) Select Target > Run on M3_Video window to execute the sample test application.

The sample test application takes the input files stored in the \Client\Test\TestVecs\Input

sub-directory, runs the codec. The reference files stored in the

\Client\Test\TestVecs\Reference sub-directory can be used to verify that the codec is

functioning as expected.

12) On failure, the application exits with the message “Frame encoding failed”.

13) On successful completion, the application displays the information for each frame and

generates output 264 bit-stream in \Client\Test\TestVecs\Output directory. User should

compare with the reference provided in \Client\Test\TestVecs\Reference directory. Both

the 264 bit-stream content should be same to conclude successful execution.

2.5 Configuration Files

This codec is shipped along with:

Installation Overview

2-9

 Encoder configuration file (encoder.cfg) – specifies the configuration

parameters used by the test application to configure the Encoder.

 TestCases.txt – This file has list of config files, these needs to be executed

with parameter (integer) preceding. The meaning of the parameter is below.

 1 – execute the test case

 0 – Terminate the regression

 For multi frame processing in single process call TestCases.txt format

should be like for 4 frame processing in single process call

 4 encoder1.cfg

 encoder2.cfg

 encoder3.cfg

 encoder4.cfg

 0

2.5.1 Encoder Configuration File

The encoder configuration file, encoder.cfg contains the configuration parameters required

for the encoder. The Encoder.cfg file is available in the \Client\Test\TestVecs\Config sub-

directory.

A sample encoder.cfg file is as shown.

# New Input File Format is as follows

# <ParameterName> = <ParameterValue> # Comment

# See configfile.h for a list of supported ParameterNames

################################################################################

# Files

################################################################################

InputFile = "..\..\..\Test\TestVecs\Input\test.yuv"

EncodedFile = "..\..\..\Test\TestVecs\Output\Test.264"

ReconFile = "..\..\..\Test\TestVecs\Output\Test_rec.yuv"

TestFile = "..\..\..\Test\TestVecs\Reference\ref.264"

EncodingPreset = 3 # 3=> XDM_USER_DEFINED(see codec-specific document

# to understand the encoding behaviour).

RateControlPreset = 5 # 1 => Low Delay, 2 => Storage, 3 => Rsvd

# 4 => None, 5 => user defined

MaxInterFrameInterval = 1 # I to P frame distance. 1 indicates no

# B frames. Value >1 indicates presence

# of B frames.

Installation Overview

2-10

Profile = 100 # Profile IDC (66=baseline, 77=main,

# 100=High)

Level = 41 # Level IDC (e.g. 30 = level 3.0)

NumInputUnits = 10 # Number of units of input-data (ex. 10

# Frames to be encoded).

MaxWidth = 1920 # Max Frame width

MaxHeight = 1088 # Max Frame height

################################################################################

# Encoder Control

################################################################################

inputWidth = 176 # Frame width

inputHeight = 144 # Frame height

targetFrameRate = 30000 # Target picture Rate per second * 1000 => For

# 60 fields per second it should be 30000

targetBitRate = 128000 # Target Bit Rate in Bits per second.

intraFrameInterval = 10 # Interval between two consecutive intra frames,

# 0 => Only first frame to be intra coded, 1 =>

# All intra frames, N => One intra #frame and

- # N-1 inter frames, where N > 1

interFrameInterval = 1 # 1 - Only P frames. >1 - Number of B frames

# between two I/P frames.

captureWidth = 176 # Image width to compute image pitch. If

Capture

# Width is > Image Width then use the former for

# image pitch.

captureHeight = 144 # Image width to compute image pitch. If Capture

# Width is > Image Width then use the former for

# image pitch.

Any field in the IVIDENC2_Params structure (see Section 4.2.1.7) can be set in the

Encoder.cfg file using the syntax as shown in the code snippet. If you specify additional fields

in the Encoder.cfg file, ensure that you modify the test application appropriately to handle

these fields.

2.6 Standards Conformance and User-Defined Inputs

To check the conformance of the codec for the default input file shipped along with the

codec, follow the steps as described in Section 2.4.

To check the conformance of the codec for other input files of your choice, follow these

steps:

Installation Overview

2-11

1) Copy the input files to the \Client\Test\TestVecs\Inputs sub-directory.

2) Copy the reference files to the \Client\Test\TestVecs\Reference sub-directory.

14) Edit the configuration file, Encoder.cfg available in the \Client\Test\TestVecs\Config

sub-directory. For details on the format of the Encoder.cfg file, see Section 2.5.1.

2.7 Uninstalling the Component

To uninstall the component, delete the codec directory from your hard disk.

Installation Overview

2-12

This page is intentionally left blank

3-1

Chapter 3

Sample Usage

This chapter provides a detailed description of the sample test application that accompanies

this codec component.

Topic Page

3.1 Overview of the Test Application

3-2

3.2 Frame Buffer Management

3-5

3.3 Handshaking Between Application and Algorithm

3-6

Sample Usage

3-2

3.1 Overview of the Test Application

The test application exercises the IVIDENC2 and extended class of the H.264 Encoder

library. The source files for this application are available in the \Client\Test\Src and

\Client\Test\Inc sub-directories.

Figure 1-1 depicts the sequence of APIs exercised in the sample test application.

The test application is divided into four logical blocks:

 Parameter setup

 Algorithm instance creation and initialization

 Process call

 Algorithm instance deletion

3.1.1 Parameter Setup

Each codec component requires various codec configuration parameters to be set at

initialization. For example, a video codec requires parameters such as video height, video

width, and so on. The test application obtains the required parameters from the Encoder

configuration files.

In this logical block, the test application does the following:

1) Opens the configuration file, listed in TesCases.txt and reads the various

configuration parameters required for the algorithm. For more details on the

configuration files, see Section 2.5.

2) Sets the interface structure based on the values it reads from the configuration file.

3) Does the algorithm instance creation and other handshake via. control methods

4) For each frame reads the input yuv frame into the application input buffer and

makes a process call

5) For each frame dumps out the generated bit-stream into the specified file

3.1.2 Algorithm Instance Creation and Initialization

In this logical block, the test application accepts the various initialization parameters and

returns an algorithm instance pointer. The following APIs implemented by the codec are

called in sequence by ALG_create():

1) algNumAlloc() - To query the algorithm about the number of memory

records it requires.

2) algAlloc() - To query the algorithm about the memory requirement to

be filled in the memory records.

3) algInit() - To initialize the algorithm with the memory structures

provided by the application.

Sample Usage

3-3

A sample implementation of the create function that calls algNumAlloc(), algAlloc(), and

algInit() in sequence is provided in the ALG_create() function implemented in the

alg_create.c file.

After successful creation of the algorithm instance, the test application does resource

allocation for the algorithm. This requires initialization of Resource Manager Module (RMAN)

and grant of required resources (HDVICP2, Tiled memory, and so on). This is implemented

by calling RMAN interface functions in following sequence:

1) numResourceDescriptors() - To understand the number of resources

(HDVICP and buffers) needed by algorithm.

2) getResourceDescriptors() – To get the attributes of the resources.

3) initResources() - After resources are created, application gives

the resources to algorithm through this APII

3.1.3 Process Call

After algorithm instance creation and initialization, the test application does the following:

1) Sets the dynamic parameters (if they change during run-time) by calling the control()

function with the XDM_SETPARAMS command.

2) Sets the input and output buffer descriptors required for the process() function call. The

input and output buffer descriptors are obtained by calling the control() function with the

XDM_GETBUFINFO command.

3) Calls the process() function to encode/decode a single frame of data. The behavior of

the algorithm can be controlled using various dynamic parameters (see Section

4.2.1.8). The inputs to the process function are input and output buffer descriptors,

pointer to the IVIDENC2_InArgs and IVIDENC2_OutArgs structures.

4) When the process() function is called for encoding/decoding a single frame of data, the

software triggers the start of encode/decode. After triggering the start of the

encode/decode frame, the video task can be placed in SEM-pend state using

semaphores. On receipt of interrupt signal at the end of frame encode/decode, the

application releases the semaphore and resume the video task, which does any book-

keeping operations by the codec and updates the output parameters.

Figure 3-1. Process Call with Host Release

Host

System

application

Process call frame n

HDVICP

Tasks

MB level tasks for

frame n

Host Video

Task

Transfer of

tasks at Host

MB level tasks for

frame n+1

Process call frame n+1

Host system

tasks

HDVICP Busy

Sample Usage

3-4

The control() and process() functions should be called only within the scope of the

algActivate() and algDeactivate() XDAIS functions, which activate and deactivate the

algorithm instance respectively. If the same algorithm is in-use between two process/control

function calls, calling these functions can be avoided. Once an algorithm is activated, there

can be any ordering of control() and process() functions. The following APIs are called

in sequence:

5) algActivate() - To activate the algorithm instance.

6) control() (optional) - To query the algorithm on status or setting

of dynamic parameters and so on, using the eight control commands.

7) process() - To call the Encoder with appropriate input/output

buffer and arguments information.

8) control() (optional) - To query the algorithm on status or setting

of dynamic parameters and so on, using the eight available control

commands.

9) algDeactivate() - To deactivate the algorithm instance.

The do-while loop encapsulates frame level process() call and updates the input buffer

pointer every time before the next call. The do-while loop breaks off either when an error

condition occurs or when the input buffer exhausts.

If the algorithm uses any resources through RMAN, then user must activate the resource

after the algorithm is activated and deactivate the resource before algorithm deactivation.

3.1.4 Algorithm Instance Deletion

Once decoding/encoding is complete, the test application must release the resources

granted by the IRES resource Manager interface and delete the current algorithm instance.

The following APIs are called in sequence:

1) getResourceDescriptors() - Free all resources granted by RMAN.

2) algNumAlloc() - To query the algorithm about the number of memory

records it used.

3) algFree() - To query the algorithm to get the memory record

information.

A sample implementation of the delete function that calls algNumAlloc() and algFree() in

sequence is provided in the ALG_delete() function implemented in the alg_create.c file.

After successful execution of the algorithm, the test application frees up the DMA and

HDVICP Resource allocated for the algorithm. This is implemented by calling the RMAN

interface functions in the following sequence:

4) RMAN_freeResources () - To free the resources that were allocated

to the algorithm before process call.

5) RMAN_unregister() – To un-register the HDVICP protocol/resource

manager with the generic resource manager.

Sample Usage

3-5

6) RMAN_exit() - To delete the generic IRES RMAN and release memory.

3.2 Frame Buffer Management

3.2.1 Input Frame Buffer

The encoder has input buffers that stores frames until they are processed. These buffers at

the input level are associated with a buffer input IDs. The IDs are required to track the

buffers that have been processed or locked. The encoder uses this ID, at the end of the

process call, to inform back to application whether it is a free buffer or not. Any buffer given

to the algorithm should be considered locked by the algorithm, unless the buffer is returned

to the application through IVIDENC2_OutArgs->freeBufID[].For more information, see

section 4.2.1.11.

For example, consider the GOP structure for IPPPP frames.

Frame Type

Input ID

Free Buffer ID

As shown in the table, if the input ID for the first frame is 1, the same input ID is returned as

the free buffer ID at the end of the process call. There is no locking of buffers at any point.

Now, consider the GOP structure that has B frames, IBBPBBP.

Frame Type

Input ID

Free Buffer ID

As shown in the table, the first frame input ID (1) is returned as a free buffer ID at the end of

the third process call that is after accumulating buffers for two B frames. For the first two

process calls, free buffer IDs are returned as zero. This initial delay is equal to the number of

B frames.

Since the 4th frame is a P frame, it is returned immediately at the end of the process call.

Then, input IDs, 2 and 3 are returned as free buffers while frames 5 and 6 are being

processed. Hence, if there are two B frames between P frames, the input images for the B

frames are stored and the P frame is encoded first, and then the two B frames are encoded.

This results in two frame period initial delay.

3.2.2 Frame Buffer Format

The frame buffer format to be used for both progressive and interlaced pictures is explained

in Appendix F.

3.2.3 Address Translations

The buffers addresses (DDR addresses) as seen by Media Controller and HDVICP2(VDMA)

will be different. Hence, address translations are needed to convert from one address view to

another. The application needs to implement a MEMUTILS function for this address

translation). An example of the address translation function is shown. The codec will make a

call to this function from the host (Media Controller) library. Therefore, the function name and

Sample Usage

3-6

arguments should follow the example provided below. For a given input address, this function

returns the VDMA view of the buffer (that is, address as seen by HDVICP2).

void *MEMUTILS_getPhysicalAddr(Ptr Addr)

{

return ((void *)((unsigned int)Addr & VDMAVIEW_EXTMEM));

}

Sample setting for the macro VDMAVIEW_EXTMEM is as shown.

#define VDMAVIEW_EXTMEM (0xFFFFFFFF)

3.3 Handshaking Between Application and Algorithm

Application provides the algorithm with its implementation of functions for the video task to

move to SEM-pend state, when the execution happens in the co-processor. The algorithm

calls these application functions to move the video task to SEM-pend state.

Figure 3-2. Interaction Between Application and Codec

Note:

 Process call architecture to share Host resource among multiple

threads.

 ISR ownership is with the Host layer resource manager –

outside the codec.

 The actual codec routine to be executed during ISR is provided

by the codec.

Framework Provided

HDVICP Callback APIs

process()

Application Side

Codec

#include <…/ires_hdvicp.h>

void _MyCodecISRFunction();

MYCODEC::IVIDDEC2::process() {

…. set up for frame decode

HDVICP_configure(h264d, h264d-

>hdvicpHandle,

H264DISRFunction);

HDVICP_wait(h264D, h264d-

>hdvicpHandle);

// Release of HOST

…. End of frame processing

}

void H264DISRFunction(IALG_Handle

handle)

{ H264D_TI_Obj *h264d = (void

*)handle;

HDVICP_done(h264d ,

h264d-

>hdvicpHandle);

}

int _doneSemaphore;

HDVICP_configure(handle,

hdVicpHandle, ISRFunction){

installNonBiosISR(handle,

hdvicpHandle, ISRFunction);

}

HDVICP_wait(handle,

hdVicpHandle){

SEM_pend(_doneSemaphore);

}

HDVICP_done(handle,

hdVicpHandle) {

SEM_post(_doneSemaphore)

}

Sample Usage

3-7

 OS/System related calls (SEM_pend, SEM_post) also outside

the codec.

 Codec implementation is OS independent.

The functions to be implemented by the application are:

 Void HDVICP_configure (IALG_Handle handle,

IRES_HDVICP2_Handle iresHandle, void

(*IRES_HDVICP2_CallbackFxn)(IALG_Handle handle, void

*cbArgs),void *cbArgs)

This function is called by the algorithm to register its ISR function.

The application needs to call this function, when it receives interrupts

pertaining to the video task.

 Void HDVICP_Acquire(IALG_Handle handle,IRES_HDVICP2_Handle

iresHandle, IRES_YieldContext *yieldCtxt,

IRES_HDVICP2_Status *status, Uint32* modeId, Int

lateAcquireArg )

This function is called by the algorithm to acquire the HDVICP2

resource.

 Void HDVICP_Release(IALG_Handle handle, IRES_HDVICP2_Handle

iresHandle)

This function is called by the algorithm to release the HDVICP2

resource.

 Bool HDVICP_wait (void *hdvicpHandle)

This function is called by the algorithm to move the video task to

SEM-pend state. Application should return false if it wants the early

termination of codec.

 Void HDVICP_done (void *hdvicpHandle)

This function is called by the algorithm to release the video task

from SEM-pend state. In the sample test application, these functions are

implemented in hdvicp_framework.c file. The application can implement it

in a way considering the underlying system.

 Bool HDVICP_Reset(IALG_Handle handle, IRES_HDVICP2_Handle

iresHandle)

This function is called by the algorithm to reset the HDVICP2

resource.

Sample Usage

3-8

This page is intentionally left blank

4-1

Chapter 4

API Reference

This chapter provides a detailed description of the data structures and interfaces functions

used in the codec component.

Topic Page

4.1 Symbolic Constants and Enumerated Data Types

4-2

4.2 Data Structures

4-33

4.3 Default and Supported Values of Parameters

4-80

4.4 Interface Functions

4-123

API Reference

4-2

4.1 Symbolic Constants and Enumerated Data Types

This section summarizes all the symbolic constants specified as either #define macros and/or

enumerated C data types. For each symbolic constant, the semantics or interpretation of the

same is also provided.

Table 4-1. List of Enumerated Data Types

Group or Enumeration

Class

Symbolic Constant

Name

Description or Evaluation

IVIDEO_FrameType

For the various IVIDEO_xy_FRAME values, this frame

type is interlaced where both top and bottom fields

are provided in a single frame. The first field is

an x frame, the second field is y field.

IVIDEO_NA_FRAME

Frame type not available

IVIDEO_I_FRAME

IVIDEO_FRAMETYPE_D

EFAULT

Intra coded frame,

Default value.

IVIDEO_P_FRAME

Forward inter coded frame.

IVIDEO_B_FRAME

Bi-directional inter coded

frame.

IVIDEO_IDR_FRAME

Intra coded frame that can be

used for refreshing video

content.

IVIDEO_II_FRAME

Interlaced frame, both fields

are I frames.

IVIDEO_IP_FRAME

Interlaced frame, first field

is an I frame, second field

is a P frame.

IVIDEO_IB_FRAME

Interlaced frame, first field

is an I frame, second field

is a B frame.

IVIDEO_PI_FRAME

Interlaced frame, first field

is a P frame, second field is

a I frame.

IVIDEO_PP_FRAME

Interlaced frame, both fields

are P frames.

IVIDEO_PB_FRAME

Interlaced frame, first field

is a P frame, second field is

a B frame.

IVIDEO_BI_FRAME

Interlaced frame, first field

is a B frame, second field is

an I frame.

API Reference

4-3

Group or Enumeration

Class

Symbolic Constant

Name

Description or Evaluation

IVIDEO_BP_FRAME

Interlaced frame, first field

is a B frame, second field is

a P frame.

IVIDEO_BB_FRAME

Interlaced frame, both fields

are B frames.

IVIDEO_MBAFF_I_FRA

Intra coded MBAFF frame .

IVIDEO_MBAFF_P_FRA

Forward inter coded MBAFF

frame.

IVIDEO_MBAFF_B_FRA

Bi-directional inter coded

MBAFF frame.

IVIDEO_MBAFF_IDR_F

RAME

Intra coded MBAFF frame that

can be used for refreshing

video content.

IVIDENC2_Control

Process based Controls operation for Video encoder

IVIDENC2_CTRL_NONE

IVIDENC2_CTRL_DEFA

ULT

No special control operation

IVIDENC2_CTRL_FORC

ESKIP

Force frame to be skipped.

The encoder should ignore

this operation if the frame

for which the control is

issued is IDR/I frame.

IVIDEO_MetadataType

IVIDEO_METADATAPLA

NE_NONE

Used to indicate no metadata

is requested or available

IVIDEO_METADATAPLA

NE_MBINFO

Used to indicate that MB info

metadata is requested or

available

IVIDEO_METADATAPLA

NE_EINFO

Used to indicate that Error

info metadata is requested or

available

IVIDEO_METADATAPLA

NE_ALPHA

Used to indicate that Alpha

metadata is requested or

available

IVIDEO_ContentType

IVIDEO_CONTENTTYPE

_NA

Frame type is not available.

API Reference

4-4

Group or Enumeration

Class

Symbolic Constant

Name

Description or Evaluation

IVIDEO_PROGRESSIVE

_FRAME

IVIDEO_CONTENTTYPE

_DEFAULT

Progressive video content.

Default value is

IVIDEO_PROGRESSIVE

IVIDEO_INTERLACED

IVIDEO_INTERLACED_

FRAME

Interlaced video content.

IVIDEO_INTERLACED_

TOPFIELD

Interlaced video content, top

field.

IVIDEO_INTERLACED_

BOTTOMFIELD

Interlaced video content,

bottom field.

IVIDEO_RateControlPr

eset

IVIDEO_LOW_DELAY

Constant Bit Rate (CBR)

control for video

conferencing.

IVIDEO_STORAGE

IVIDEO_RATE_CONTRO

L_PRESET_DEFAULT

Variable Bit Rate (VBR)

control for local storage

(DVD) recording,

Default rate control preset

value.

IVIDEO_TWOPASS

Two pass rate control for

non-real time applications.

IVIDEO_NONE

No configurable video rate

control mechanism.

IVIDEO_USER_DEFINE

User defined configuration

using extended parameters.

IVIDEO_SkipMode

IVIDEO_FRAME_ENCOD

IVIDEO_SKIPMODE_DE

FAULT

Input video frame

successfully encoded.

Default skip mode.

IVIDEO_FRAME_SKIPP

Input video frame skipped.

There is no encoded bit-

stream corresponding to the

input frame.

IVIDEO_OutputFrameSt

atus

IVIDEO_FRAME_NOERR

IVIDEO_OUTPUTFRAME

STATUS_DEFAULT

Output buffer is available

(default value).

Default status of the output

frame.

IVIDEO_FRAME_NOTAV

AILABLE

Encoder does not have any

output buffers.

API Reference

4-5

Group or Enumeration

Class

Symbolic Constant

Name

Description or Evaluation

IVIDEO_FRAME_ERROR

Output buffer is available

and corrupted.

For example, if a bit-stream

is erroneous and partially

decoded, a portion of the

decoded image may be

available for display.

Another example is if the

bit-stream for a given frame

decode may be decoded without

error, but the previously

decoded dependant frames were

not successfully decoded.

This would result in an

incorrectly decoded frame.

Not applicable for encoders.

IVIDEO_PictureType

IVIDEO_NA_PICTURE

Frame type not available

IVIDEO_I_PICTURE

IVIDEO_PICTURE_TYP

E_DEFAULT

Intra coded picture.

Default value.

IVIDEO_P_PICTURE

Forward inter coded picture.

IVIDEO_B_PICTURE

Bi-directional inter coded

picture.

IVIDEO_VideoLayout

IVIDEO_FIELD_INTER

LEAVED

Buffer layout is interleaved.

IVIDEO_FIELD_SEPAR

ATED

Buffer layout is field

separated.

IVIDEO_TOP_ONLY

Buffer contains only top

field.

IVIDEO_BOTTOM_ONLY

Buffer contains only bottom

field.

IVIDEO_OperatingMode

IVIDEO_DECODE_ONLY

Decoding mode.

Not applicable for encoders.

IVIDEO_ENCODE_ONLY

Encoding mode.

IVIDEO_TRANSCODE_F

RAMELEVEL

Transcode mode of operation

(encode/decode) that

consumes/generates transcode

information at the frame

level.

IVIDEO_TRANSCODE_M

BLEVEL

Transcode mode of operation

(encode/decode) that

consumes/generates transcode

information at the MB level.

API Reference

4-6

Group or Enumeration

Class

Symbolic Constant

Name

Description or Evaluation

IVIDEO_TRANSRATE_F

RAMELEVEL

Transrate mode of operation

for encoder that consumes

transrate information at the

frame level.

IVIDEO_TRANSRATE_M

BLEVEL

Transrate mode of operation

for encoder, which consumes

transrate information at the

MB level.

Not supported in this version

of H264 Encoder.

IVIDEO_BitRange

IVIDEO_YUVRANGE_FU

Pixel range for YUV is 0-255.

IVIDEO_YUVRANGE_IT

Pixel range for YUV is as per

ITU-T .

IVIDEO_DataMode

IVIDEO_FIXEDLENGTH

Data is exchanged at interval

of fixed size.

IVIDEO_SLICEMODE

Slice mode.

IVIDEO_NUMROWS

Number of rows, each row is

16 lines of video.

IVIDEO_ENTIREFRAME

Processing of entire frame

data.

XDM_AccessMode

XDM_ACCESSMODE_REA

Algorithm read from the

buffer using the CPU.

XDM_ACCESSMODE_WRI

Algorithm writes to the

buffer using the CPU.

XDM_CmdId

XDM_GETSTATUS

Query algorithm instance to

fill Status structure.

XDM_SETPARAMS

Set run-time dynamic

parameters through the

DynamicParams structure.

XDM_RESET

Reset the algorithm. All

fields in the

internal data structures are

reset and all internal

buffers are flushed.

API Reference

4-7

Group or Enumeration

Class

Symbolic Constant

Name

Description or Evaluation

XDM_SETDEFAULT

Restore the algorithm's

internal state to its

original, default values.

The application needs to

initialize the

dynamicParams.size and

status.size fields prior to

calling control() with

XDM_SETDEFAULT. The

algorithm must write to the

status.extendedError field,

and potentially algorithm

specific, extended fields.

XDM_SETDEFAULT differs from

XDM_RESET. In addition to

restoring the algorithm's

internal state, XDM_RESET

also resets any channel

related state.

XDM_FLUSH

Handle end of stream

conditions.

This command forces the

algorithm to output data

without additional input. The

recommended sequence is to

call the control() function

(with XDM_FLUSH) followed by

repeated calls to the

process() function until it

returns an error.

The algorithm should return

the appropriate, class-

specific EFAIL error

(example, ISPHDEC1_EFAIL,

IVIDENC1_EFAIL, and so on),

when flushing is complete.

XDM_GETBUFINFO

Query algorithm instance

regarding its properties of

input and output buffers. The

application only needs

to initialize the

dynamicParams.size, the

status.size, and set any

buffer descriptor fields

(example, status.data) to

NULL prior to calling

control() with

XDM_GETBUFINFO.

API Reference

4-8

Group or Enumeration

Class

Symbolic Constant

Name

Description or Evaluation

XDM_GETVERSION

Query the algorithm's

version.

The result is returned in the

data field of the respective

_Status structure. There is

no specific format defined

for version returned by the

algorithm.

The memory is not allocated

by encoder and needs to be

allocated by user. The

buffer requirement for

holding version number is of

length

IH264ENC_VERSION_LENGTH

XDM_GETCONTEXTINFO

Query a split codec part for

its context needs. Only split

codecs are required to

implement this command.

Not supported in this version

of H264 Encoder.

XDM_GETDYNPARAMSDE

FAULT

Query the algorithm to fill

the default values for the

parameters, which can be

configured dynamically.

To get the current value of

an algorithm instance's

dynamic parameters, it is

recommended that the

algorithm provides them

through the XDM_GETSTATUS

call.

XDM_SETLATEACQUIRE

ARG

Set an algorithm's 'late

acquire' argument. Only

algorithms that utilize the

late acquire IRES feature may

implement this command.

XDM_DataFormat

XDM_BYTE

Big endian stream (default

value)

XDM_LE_16

16-bit little endian stream.

XDM_LE_32

32-bit little endian stream.

XDM_LE_64

64-bit little endian stream.

XDM_BE_16

16-bit big endian stream.

XDM_BE_32

32-bit big endian stream.

XDM_BE_64

64-bit big endian stream.

API Reference

4-9

Group or Enumeration

Class

Symbolic Constant

Name

Description or Evaluation

XDM_ChromaFormat

XDM_CHROMA_NA

Chroma format not applicable.

XDM_YUV_420P

YUV 4:2:0 planar.

XDM_YUV_422P

YUV 4:2:2 planar.

XDM_YUV_422IBE

YUV 4:2:2 interleaved (big

endian).

XDM_YUV_422ILE

YUV 4:2:2 interleaved (little

endian)

XDM_YUV_444P

YUV 4:4:4 planar.

XDM_YUV_411P

YUV 4:1:1 planar.

XDM_GRAY

Gray format.

XDM_RGB

RGB color format.

XDM_YUV_420SP

YUV 4:2:0 chroma semi-planar

format (first plane is luma

and second plane is CbCr

interleaved)

Default value.

XDM_ARGB8888

Alpha plane color format.

XDM_RGB555

RGB555 color format.

XDM_RGB565

RGB565 color format.

XDM_YUV_444ILE

YUV 4:4:4 interleaved (little

endian) color format.

XDM_MemoryType

XDM_MEMTYPE_ROW

XDM_MEMTYPE_RAW

Raw memory type.

XDM_MEMTYPE_TILED8

2D memory in 8-bit container

of tiled memory space.

XDM_MEMTYPE_TILED1

2D memory in 16-bit container

of tiled memory space.

XDM_MEMTYPE_TILED3

2D memory in 32-bit container

of tiled memory space.

XDM_MEMTYPE_TILEDP

AGE

2D memory in page container

of tiled memory space.

API Reference

4-10

Group or Enumeration

Class

Symbolic Constant

Name

Description or Evaluation

XDM_MemoryUsageMode

XDM_MEMUSAGE_DATAS

YNC

Bit-mask to indicate the

usage mode. Bit-0 is Data

Sync mode. If this bit is set

then it means that buffer is

used in data sync mode

XDM_EncodingPreset

XDM_DEFAULT

Default setting of the

algorithm specific creation

time parameters.

XDM_HIGH_QUALITY

Set algorithm specific

creation time parameters for

high quality.

XDM_HIGH_SPEED

Set algorithm specific

creation time parameters for

high speed. In this preset

HDVICP 2.0 utilization is

improved. It is supported for

all profiles.

1.interframeInterval should

be 1 i.e., No B frames.

2. For High profile,

transformBlockSize should be

IH264_TRANSFORM_8x8

3. For Main/Base profiles,

transformBlockSize should be

IH264_TRANSFORM_4x4

XDM_USER_DEFINED

XDM_PRESET_DEFAULT

User defined configuration

using advanced parameters.

Default value.

XDM_HIGH_SPEED_MED

_QUALITY

Set algorithm specific

creation time parameters for

high speed medium quality.

XDM_MED_SPEED_MED_

QUALITY

Set algorithm specific

creation time parameters for

medium speed medium quality.

XDM_MED_SPEED_HIGH

_QUALITY

Set algorithm specific

creation time parameters for

medium speed high quality.

XDM_EncMode

XDM_ENCODE_AU

Encode entire access unit,

including the headers.

Default value.

XDM_GENERATE_HEADE

Encode only header

IVIDENC2_MotionVecto

rAccuracy

IVIDENC2_MOTIONVEC

TOR_PIXEL

Motion vectors accuracy is

only integer pel.

API Reference

4-11

Group or Enumeration

Class

Symbolic Constant

Name

Description or Evaluation

IVIDENC2_MOTIONVEC

TOR_HALFPEL

Motion vectors accuracy is

half pel.

IVIDENC2_MOTIONVEC

TOR_QUARTERPEL

Motion vectors accuracy is

quarter pel.

IVIDENC2_MOTIONVEC

TOR_EIGHTHPEL

Motion vectors accuracy is

one-eighth pel.

IVIDENC2_MOTIONVEC

TOR_MAX

Motion vectors accuracy is

not defined.

XDM_ErrorBit

XDM_PARAMSCHANGE

Bit 8

 1 - Sequence Parameters

Change

 0 - Ignore

This error is applicable for

transcoders. It is set when

some key parameter of the

input sequence changes. The

transcoder returns after

setting this error field and

the correct input sequence

parameters are updated in

outArgs.

XDM_APPLIEDCONCEAL

MENT

Bit 9

 1 - Applied concealment

 0 - Ignore

This error is applicable

for decoders.

It is set when the decoder

was not able to decode the

bit-stream, and the decoder

has concealed the bit-

stream error and produced

the concealed output.

XDM_INSUFFICIENTDA

Bit 10

 1 - Insufficient input

data

 0 - Ignore

This error is applicable for

decoders. This is set when

the input data provided is

not sufficient to produce one

frame of data. This can be

also be set for encoders when

the number of valid samples

in the input frame is not

sufficient to process a

frame.

API Reference

4-12

Group or Enumeration

Class

Symbolic Constant

Name

Description or Evaluation

XDM_CORRUPTEDDATA

Bit 11

 1 - Data

problem/corruption

 0 - Ignore

This error is applicable for

decoders. This is set when

the bit-stream has an error

and not compliant to the

standard syntax.

XDM_CORRUPTEDHEADE

Bit 12

 1 - Header

problem/corruption

 0 - Ignore

This error is applicable for

decoders. This is set when

the header information in the

bit-stream is incorrect. For

example, it is set when

Sequence, Picture, Slice, and

so on are incorrect in video

decoders.

XDM_UNSUPPORTEDINP

Bit 13

 1 – Un-supported

feature/parameter in

input

 0 - Ignore

This error is set when the

algorithm is not able process

a certain input data/bit-

stream format. It can also be

set when a subset of features

in a standard are not

supported by the algorithm.

For example, if a video

encoder only supports 4:2:2

formats, it can set this

error for any other type of

input video format.

XDM_UNSUPPORTEDPAR

Bit 14

 1 - Unsupported input

parameter or

configuration

 0 - Ignore

This error is set when the

algorithm does not support

certain configurable

parameters. For example, if

the video encoder does not

support sliceMode for below

128x80 resolution, it will

return

XDM_UNSUPPORTEDPARAM when the

control function is called

for parameter validation.

API Reference

4-13

Group or Enumeration

Class

Symbolic Constant

Name

Description or Evaluation

XDM_FATALERROR

Bit 15

 1 - Fatal error (stop

encoding)

 0 - Recoverable error

If there is an error, and

this bit is not set, the

error is recoverable.

This error is set when the

algorithm cannot recover from

the current state. It

informs the system not to try

the next frame and possibly

delete the multimedia

algorithm instance.

It implies the codec will not

work when reset.

You should delete the current

instance of the codec.

Note:

The remaining bits that are not mentioned in XDM_ErrorBit are interpreted as:

 Bit 16-32: Used for codec specific error codes.

 Bit 0-7: Codec and implementation specific (see Table 4-4)

The algorithm can set multiple bits to one depending on the error condition.

Table 4-2. H264 Encoder Specific Enumerated Data Types.

Group or Enumeration

Class

Symbolic Constant Name

Description or Evaluation

IH264ENC_Intra4x4Pa

rams

H.264 Encoder slice level control for Intra4x4 modes

IH264_INTRA4x4_NONE

Disable Intra4x4 modes

IH264_INTRA4x4_ISLICES

Enable Intra4x4 modes only

in I Slices

IH264_INTRA4x4_IPBSLICES

IH264_INTRA4x4_DEFAULT

Enable Intra4x4 modes only

in I, P and B Slices.

This is the default setting.

IH264ENC_Level

IH264_LEVEL_10

H.264 Level 1.0

IH264_LEVEL_1b

H.264 Level 1.b

IH264_LEVEL_11

H.264 Level 1.1

IH264_LEVEL_12

H.264 Level 1.2

IH264_LEVEL_13

H.264 Level 1.3

API Reference

4-14

Group or Enumeration

Class

Symbolic Constant Name

Description or Evaluation

IH264_LEVEL_20

H.264 Level 2.0

IH264_LEVEL_21

H.264 Level 2.1

IH264_LEVEL_22

H.264 Level 2.2

IH264_LEVEL_30

H.264 Level 3.0

IH264_LEVEL_31

H.264 Level 3.1

IH264_LEVEL_32

H.264 Level 3.2

IH264_LEVEL_40

H.264 Level 4.0

IH264_LEVEL_41

H.264 Level 4.1

IH264_LEVEL_42

H.264 Level 4.2

IH264_LEVEL_50

H.264 Level 5.0

IH264_LEVEL_51

H.264 Level 5.1

IH264ENC_Profile

Profile identifier for H.264 Encoder

IH264_BASELINE_PROFILE

Baseline profile

IH264_MAIN_PROFILE

Main profile

IH264_EXTENDED_PROFILE

Extended profile

IH264_HIGH_PROFILE

IH264_DEFAULT_PROFILE

High profile.

This is the default setting.

IH264_HIGH10_PROFILE

High 10 profile

IH264_HIGH422_PROFILE

High 4:2:2 profile

IH264ENC_MetadataTy

Meta data type specifc to H.264 encoder

IH264_SEI_USER_DATA_UNREG

ISTERED

H.264 allows inserting SEI

message for any user data,

refer section

D.1.6 of H.264 standard.

By setting this value to any

IVIDENC2_Params::metadataTyp

e[i]

You can provide the data SEI

to be inserted in H.264 bit-

stream

Refer

IH264ENC_MetaDataFormatUserD

efinedSEI for the format of

user data.

API Reference

4-15

Group or Enumeration

Class

Symbolic Constant Name

Description or Evaluation

IH264_REGION_OF_INTEREST

By setting this value to any

IVIDENC2_Params::metadataTyp

e[i]

You can provide region of

interest information for

smart encoding. But this

version of encoder does not

accept Region of

Interest(ROI) information

through

IVIDENC2_Params::metadataTyp

e[i] but instead ROI

information is taken through

IH264ENC_InArgs::

roiInputParams and feature

is controlled

(enabled/disabled) through

parameter

IH264ENC_DynamicParams::

enableROI. Please refer

Appendix K for more details

on ROI.

IH264_USER_DEFINED_SCALIN

GMATRIX

By setting this value to any

IVIDENC2_Params::metadataTyp

e[i]

You can provide scaling

matrices to be used by

encoder. Refer Appendix C

for more details.

IH264ENC_LTRPScheme

IH264ENC_LTRP_NONE

No longterm refernce frame

IH264ENC_LTRP_REFERTO_PER

IODICLTRP

Mark frames as long-term

reference frame with the

period given by LTRPPeriod

of IH264ENC_Params and based

on the frame control

IH264ENC_Control

IH264ENC_LTRP_REFERTOP_PR

OACTIVE

Two long term frames are

supported in this scheme and

long-term index marking and

refernce frame update is

done based the

IH264ENC_Control values

IH264ENC_LTRP_REFERTOP_RE

ACTIVE

Mark frames as long-term

reference frame with the

period given by LTRPPeriod

of IH264ENC_Params. At any

point of time there will be

2 long-term frames and based

on the frame control

IH264ENC_Control

API Reference

4-16

Group or Enumeration

Class

Symbolic Constant Name

Description or Evaluation

IH264ENC_Control

Picture level control

IH264ENC_CTRL_REFER_LONG_

TERM_FRAME

Control to encoder for

referring long term

reference frame

IH264ENC_CTRL_NOWRITE_NOR

EFUPDATE

Control to encoder for

encoding current frame as

non-referencing P frame and

not to update reference

frame for this P frame

IH264ENC_CTRL_WRITE_NOREF

UPDATE

Control to encoder for

encoding current frame as

referencing P frame and not

to update reference frame

for this P frame

IH264ENC_CTRL_NOWRITE_REF

UPDATE

Control to encoder for

encoding current frame as

non-referencing P frame and

to update reference frame

for this P frame

IH264ENC_CTRL_WRITE_REFUP

DATE

Control to encoder for

encoding current frame as

referencing P frame and to

update reference frame for

this P frame

IH264ENC_CTRL_START_GDR

Control to start GDR

activity. Applicable when

intraRefreshMethod is

IH264_INTRAREFRESH_GDR

IH264ENC_PicOrderCo

untType

Picture Order Count Type Identifier

IH264_POC_TYPE_0

IH264_POC_TYPE_DEFAULT

POC type 0.

Default POC type to be used

by encoder.

IH264_POC_TYPE_1

POC type 1

IH264_POC_TYPE_2

POC type 2

IH264ENC_ScalingMat

Preset

Controls the type of scaling matrix picked up by encoder

IH264_SCALINGMATRIX_NONE

Flat scaling matrix: part of

standard (no scaling

matrix).

Default (if profile != HIGH)

API Reference

4-17

Group or Enumeration

Class

Symbolic Constant Name

Description or Evaluation

IH264_SCALINGMATRIX_NORMA

IH264_SCALINGMATRIX_STD_D

EFAULT

For normal contents

(normal contents).

Default (if profile == HIGH)

IH264_SCALINGMATRIX_NOISY

For noisy contents.

IH264_SCALINGMATRIX_USERD

EFINED_SPSLEVEL

Scaling matrices can be

provided at SPS level. See

Appendix C for more details

IH264_SCALINGMATRIX_USERD

EFINED_SPSLEVEL

Scaling matrices can be

provided by at PPS level.

See Appendix C for more

details

IH264ENC_RateContro

lAlgo

These enumerations control the type of rate control

algorithm to be picked up by the encoder. Only useful if

IVIDENC2::rateControlPreset is set as

IVIDEO_USER_DEFINED.

IH264_RATECONTROL_PRC

IH264_RATECONTROL_DEFAULT

Perceptual Rate Control,

controls the QP at MB level

with VBR mode

Default rate control

algorithm.

IH264_RATECONTROL_PRC_LOW

_DELAY

Perceptual Rate Control,

controls the QP at MB level

with CBR (Low delay) mode

IH264ENC_FrameQuali

tyFactor

These enumerations control the quality factor between two

types of frames, I frame quality with respect to P frame.

For example, higher quality factor means I frame quality

is given higher importance compared to P frame.

IH264_QUALITY_FACTOR_1

IH264_QUALITY_FACTOR_DEFA

ULT

Same quality factor between

two types of frame.

It is default quality

factor.

IH264_QUALITY_FACTOR_2

High quality factor to one

frame type between two types

of frame.

IH264_QUALITY_FACTOR_3

Higher quality factor to one

frame type between two types

of frame.

IH264ENC_RateContro

lParamsPreset

These enumerations control the rate control parameters.

This preset controls the USER_DEFINED versus DEFAULT

mode. If you are not aware about the following fields, it

should be set as IH264_RATECONTROLPARAMS_DEFAULT.

API Reference

4-18

Group or Enumeration

Class

Symbolic Constant Name

Description or Evaluation

IH264_RATECONTROLPARAMS_D

EFAULT

Default rate control params.

IH264_RATECONTROLPARAMS_U

SERDEFINED

User defined rate control

params.

Default value.

IH264_RATECONTROLPARAMS_E

XISTING

Keep the rate control params

as existing. This is useful

during control call, if user

does not want to change the

rate control parameters.

IH264ENC_InterCodin

gPreset

These enumerations control the type of inter coding.

IH264_INTERCODING_DEFAULT

Default inter coding params.

IH264_INTERCODING_USERDEF

INED

User defined inter coding

params.

Default value.

IH264_INTERCODING_EXISTIN

Keep inter coding params as

existing. This is useful

during control call, if you

do not want to change the

inter coding params.

IH264_INTERCODING_MED_SPE

ED_HIGH_QUALITY

InterCoding Preset for

Medium speed high quality

encoding

IH264_INTERCODING_HIGH_SP

EED

InterCoding Preset for High

speed encoding. This is

supported only when all the

below conditions (a,b,c,d)

are satisfied

a.enablePartialFrameSkip

should be disabled.

b.intraRefreshMethod should

be default.

c. For High profile,

transformBlockSize should

be IH264_TRANSFORM_8x8

d. For Main/Base profiles,

transformBlockSize should

be IH264_TRANSFORM_4x4

IH264ENC_MeAlgoMode

IH264ENC_MOTIONESTMODE_NO

RMAL

IH264ENC_MOTIONESTMODE_DE

FAULT

Motion estimation algorithm

selection for normal

encoding

API Reference

4-19

Group or Enumeration

Class

Symbolic Constant Name

Description or Evaluation

IH264ENC_MOTIONESTMODE_HI

GH_SPEED

Motion estimation algorithm

selection for high speed

encoding. This is supported

only when all the below

conditions (a,b,c,d,e) are

satisfied.

a)IVIDENC2_DynamicParams::in

terFrameInterval is ‘1’

b) IVIDENC2_DynamicParams ::

mvAccuracy ==

IVIDENC2_MOTIONVECTOR_QUARTE

RPEL

c)IH264ENC_InterCodingParams

:: minBlockSizeP ==

IH264_BLOCKSIZE_16x16

d) For High profile,

transformBlockSize should

be IH264_TRANSFORM_8x8

e) For Main/Base profiles,

transformBlockSize should

be IH264_TRANSFORM_4x4

IH264ENC_IntraCodin

gBias

IH264ENC_INTRACODINGBIAS_

NORMAL

IH264ENC_INTRACODINGBIAS_

DEFAULT

IntraCoding Bias for normal

encoding. No special

restriction on number of

intra macro blocks.

IH264ENC_INTRACODINGBIAS_

HIGH_SPEED

Puts special restriction on

intra macro blocks to limit

it to 12 % of total Mbs in

the picture.

IH264ENC_InterBlock

Size

These enumerations are defined for minimum inter block

size.

IH264_BLOCKSIZE_16x16

IH264_BLOCKSIZE_DEFAULT

16x16 block size. It is also

default block size.

IH264_BLOCKSIZE_8x8

8x8 block size

IH264_BLOCKSIZE_4x4

4x4 Block size

Not supported in this

version of H264 Encoder

IH264ENC_BiasFactor

Control to code the macro block as inter or intra. Also,

for having a macro block use skip MV or regular MV.

IH264_BIASFACTOR_LOW

Low biasing.

API Reference

4-20

Group or Enumeration

Class

Symbolic Constant Name

Description or Evaluation

IH264_BIASFACTOR_MEDIUM

IH264_BIASFACTOR_NORMAL

IH264_BIASFACTOR_DEFAULT

Normal/Med biasing.

Default biasing factor.

IH264_BIASFACTOR_MILD

Mild bias factor

IH264_BIASFACTOR_ADAPTIVE

Adaptive bias factor

IH264_BIASFACTOR_HIGH

High biasing.

IH264ENC_IntraRefre

shMethods

Refresh method type identifier for H.264 Encoder.

IH264_INTRAREFRESH_NONE

IH264_INTRAREFRESH_DEFAUL

Does not forcefully insert

intra macro blocks.

Default intra refresh is

OFF.

IH264_INTRAREFRESH_CYCLIC

_MBS

Inserts intra macro blocks

in a cyclic mode.

Cyclic interval is equal to

intraRefreshRate.

IH264_INTRAREFRESH_CYCLIC

_SLICES

Inserts intra slices (row

based) in a cyclic mode:

Cyclic interval is equal to

intraRefreshRate.

IH264_INTRAREFRESH_RDOPT_

MBS

Position of intra macro

blocks is chosen by encoder,

but the number of forcefully

coded intra macro blocks in

a frame is guaranteed to be

equal to

totalMbsInFrame/intraRefresh

Rate.

API Reference

4-21

Group or Enumeration

Class

Symbolic Constant Name

Description or Evaluation

IH264_INTRAREFRESH_GDR

Instead of a sudden Intra

refresh of entire frame,

the frame is refreshed

gradually over a duration

(which is con figerable) of

frames with refresh

happening by Intra coded

rows scanning from top to

bottom of the scene/picture

This method of intra refresh

mechanism is not supported

for interlaced cases and in

case of ‘B’frames i.e.,

IVIDENC2_Params::inputContentType

should not be

IVIDEO_INTERLACED (no interlace)

and

interFrameInterval == 1 (no ‘B’

frames)

IH264ENC_ChormaComp

onent

These enumerations control the selection of chroma

component to perform chroma intra estimation.

IH264_CHROMA_COMPONENT_CR

_ONLY

IH264_CHROMA_COMPONENT_DE

FAULT

Only Cr component

Default is Only CR

component.

IH264_CHROMA_COMPONENT_CB

_CR_BOTH

Both Cb and Cr component.

IH264ENC_IntraCodin

gPreset

These enumerations control the type of intra coding.

IH264_INTRACODING_DEFAULT

Default intra coding params.

IH264_INTRACODING_USERDEF

INED

User defined intra coding

params.

Default value.

IH264_INTRACODING_EXISTIN

Keep intra coding params as

existing. This is useful

during control call, if you

do not want to change the

inter coding params

API Reference

4-22

Group or Enumeration

Class

Symbolic Constant Name

Description or Evaluation

IH264_INTRACODING_HIGH_SP

EED

Intra coding params for high

speed encoding.

a. Me algo mode should be

IH264ENC_MOTIONESTMODE_HIGH_

SPEED.

b. For High profile,

transformBlockSize should

be IH264_TRANSFORM_8x8.

c. For Main/Base profiles,

transformBlockSize should

be IH264_TRANSFORM_4x4.

IH264ENC_NALUnitTyp

IH264_NALU_TYPE_SPS_WITH_

VUI

Sequence parameter set

having VUI information.

IH264_NALU_TYPE_SLICE

Slice of a non-IDR picture.

IH264_NALU_TYPE_SLICE_DP_

Coded slice data partition

IH264_NALU_TYPE_SLICE_DP_

Coded slice data partition

IH264_NALU_TYPE_SLICE_DP_

Coded slice data partition

IH264_NALU_TYPE_IDR_SLICE

Slice of an IDR picture.

IH264_NALU_TYPE_SEI

Supplemental enhancement

information.

IH264_NALU_TYPE_SPS

Sequence parameter set.

IH264_NALU_TYPE_PPS

Picture parameter set.

IH264_NALU_TYPE_AUD

Access unit delimiter.

IH264_NALU_TYPE_EOSEQ

End of sequence.

IH264_NALU_TYPE_EOSTREAM

End of stream.

IH264_NALU_TYPE_FILLER

Filler data.

IH264_NALU_TYPE_SPS_EXT

Sequence parameter set

extension.

API Reference

4-23

Group or Enumeration

Class

Symbolic Constant Name

Description or Evaluation

IH264_NALU_TYPE_USER_DATA

_UNREGD_SEI

User data un-registered SEI.

IH264ENC_NALUContro

lPreset

These enumerations define the control mechanism for

insertion of

different NALU types at different point in video

sequence.

IH264_NALU_CONTROL_DEFAUL

Default NALU insertion.

IH264_NALU_CONTROL_USERDE

FINED

User defined NALU insertion.

IH264ENC_SliceCodin

gPreset

These enumerations control the type of slice coding.

IH264_SLICECODING_DEFAULT

Default slice coding params.

IH264_SLICECODING_USERDEF

INED

User defined slice coding

params.

Default value.

IH264_SLICECODING_EXISTIN

Keep slice coding params as

existing. This is useful

during control call, if you

do not want to change the

slice coding parameters.

IH264ENC_SliceMode

These enumerations control the mode of slice coding.

IH264_SLICEMODE_NONE

IH264_SLICEMODE_DEFAULT

Single Slice per picture.

This is the default slice

coding mode.

IH264_SLICEMODE_MBUNIT

Slices are controlled based

upon number of macro blocks.

IH264_SLICEMODE_BYTES

Slices are controlled based

on number of bytes.

IH264_SLICEMODE_OFFSET

Slices are controlled based

on user defined offset in

unit of rows.

IH264ENC_StreamForm

These enumerations control the type stream format.

IH264_BYTE_STREAM

IH264_STREAM_FORMAT_DEFAU

Bit-stream contains the

start code identifier.

Default slice coding mode.

IH264_NALU_STREAM

Bit-stream does not contain

the start code identifier.

API Reference

4-24

Group or Enumeration

Class

Symbolic Constant Name

Description or Evaluation

IH264ENC_LoopFilter

Preset

Controls the loop filter preset options

IH264_LOOPFILTER_DEFAULT

Default loop-filtering

params.

IH264_LOOPFILTER_USERDEFI

NED

User defined loop-filtering

params.

IH264ENC_LoopFilter

DisableIDC

Controls H264 loop filter disable options

IH264_DISABLE_FILTER_NONE

IH264_DISABLE_FILTER_DEFA

ULT

Enable filtering of all the

edges.

Default is loop filter

enabled.

IH264_DISABLE_FILTER_ALL_

EDGES

Disable filtering of all the

edges.

IH264_DISABLE_FILTER_SLIC

E_EDGES

Disable filtering of slice

edges.

IH264ENC_SliceGroup

MapType

Map type of slice group.

IH264_INTERLEAVED_SLICE_G

Interleaved slice group.

IH264_DISPERSED_SLICE_GRP

IH264_SLICE_GRP_MAP_DEFAU

Dispersed slice group.

Default value.

IH264_FOREGRND_WITH_LEFTO

VER_SLICE_GRP

ForeGround with Left Over.

IH264_BOX_OUT_SLICE_GRP

Box Out.

IH264_RASTER_SCAN_SLICE_G

Raster Scan.

IH264_WIPE_SLICE_GRP

Wipe slice group.

IH264_EXPLICIT_SLICE_GRP

Explicit Slice group map

type.

IH264ENC_SliceGroup

ChangeDirection

Only valid when sliceGroupMapType is equal to

IH264_RASTER_SCAN_SLICE_GRP, IH264_WIPE_SLICE_GRP, or

IH264_WIPE_SLICE_GRP.

IH264_RASTER_SCAN

IH264ENC_SLICEGROUP_CHANGE

_DIRECTION_DEFAULT

Raster scan order.

Default slice group

direction.

API Reference

4-25

Group or Enumeration

Class

Symbolic Constant Name

Description or Evaluation

IH264_CLOCKWISE

Clockwise (used for box out

FMO parameters).

IH264_RIGHT

Right, used for Wipe FMO

type.

IH264_REVERSE_RASTER_SCAN

Reverse raster scan order.

IH264_COUNTER_CLOCKWISE

Counter clockwise, used for

box out FMO parameters.

IH264_LEFT

Left, used for Wipe FMO

type.

IH264ENC_FMOCodingP

reset

Controls for FMO coding preset

IH264_FMOCODING_NONE

IH264_FMOCODING_DEFAULT

No FMO

Default FMO coding value

IH264_FMOCODING_USERDEFIN

User defined FMO parameters

IH264ENC_EntropyCod

ingMode

Controls the entropy coding type

IH264_ENTROPYCODING_CAVLC

IH264_ENTROPYCODING_DEFAU

CAVLC coding type

IH264_ENTROPYCODING_CABAC

CABAC coding type

IH264ENC_TransformB

lockSize

In H264 intra macro block, transform size depends on the

intra mode, so this applies to inter macro blocks only.

IH264_TRANSFORM_4x4

Transform blocks size is 4x4

IH264_TRANSFORM_8x8

Transform blocks size is 8x8

: Valid for only High

Profile

IH264_TRANSFORM_ADAPTIVE

IH264_TRANSFORM_DEFAULT

Adaptive transform block

size: encoder decides as per

content

IH264ENC_GOPStructu

Type of Group of Pictures

(GOP)

IH264ENC_GOPSTRUCTURE_NON

UNIFORM

IH264ENC_GOPSTRUCTURE_DEF

AULT

Open GOP structure: IBBPBBP

Default

API Reference

4-26

Group or Enumeration

Class

Symbolic Constant Name

Description or Evaluation

IH264ENC_GOPSTRUCTURE_UNI

FORM

Close GOP structure:

BBIBBPBB

IH264ENC_InterlaceC

odingType

Controls the type of interlaced coding

IH264_INTERLACE_PICAFF

PicAFF type of interlace

coding

IH264_INTERLACE_MBAFF

MBAFF type of interlace

coding

IH264_INTERLACE_FIELDONLY

_MRF

Field only coding with

selecting most recent field

as reference

IH264_INTERLACE_FIELDONLY

_ARF

IH264_INTERLACE_DEFAULT

Field only coding where

codec decides the parity of

the field to be used based

on content.

Default setting

IH264_INTERLACE_FIELDONLY

_SPF

Field only coding with

selecting same parity field

as reference.

IH264ENC_VUICodingP

reset

Preset for VUI related

parameters

IH264_VUICODING_DEFAULT

Default VUI Parameters. Note

that Enable/Disable of VUI

is through

nalUnitControlParams

IH264_VUICODING_USERDEFIN

User defined VUI parameters

IH264ENC_StereoInfo

Preset

Preset for StereoInfo parameters

IH264_STEREOINFO_DISABLE

Disable Stereo Video Coding.

IH264_STEREOINFO_ENABLE_D

EFAULT

Enable stereo video coding

in default mode

IH264_STEREOINFO_ENABLE_U

SERDEFINED

Enable stereo video coding

in userdefined mode.

IH264ENC_FramePacki

ngPreset

Preset for Frame packing SEI parameters

IH264_FRAMEPACK_SEI_DISAB

Disable frame packing SEI.

API Reference

4-27

Group or Enumeration

Class

Symbolic Constant Name

Description or Evaluation

IH264_FRAMEPACK_SEI_ENABL

E_DEFAULT

Enable frame packing SEI

coding in default mode

IH264_FRAMEPACK_SEI_USERD

EFINED

Enable frame packing SEI

coding in userdefined mode.

IH264ENC_FramePacki

ngType

Enumerations for Frame Packing arrangement type

IH264_FRAMEPACK_CHECKERBO

ARD

Checker board arrangement of

2 views

IH264_FRAMEPACK_COLUMN_IN

TERLEAVING

Column interleaving

arrangement of 2 views

IH264_FRAMEPACK_ROW_INTER

LEAVING

Row interleaving arrangement

of 2 views

IH264_FRAMEPACK_SIDE_BY_S

IDE

IH264_FRAMEPACK_TYPE_DEFA

ULT

Side by side arrangement of

2 views

IH264_FRAMEPACK_TOP_BOTTO

Top-Bottom arrangement of 2

views

IH264ENC_VideoForma

Video format for VUI parameters

IH264ENC_VIDEOFORMAT_COMP

ONENT

Component video format

IH264ENC_VIDEOFORMAT_PAL

PAL video format

IH264ENC_VIDEOFORMAT_NTSC

NTSC video format

IH264ENC_VIDEOFORMAT_SECA

SECAM video format

IH264ENC_VIDEOFORMAT_MAC

MAC video format

IH264ENC_VIDEOFORMAT_UNSP

ECIFIED

Unspecified video format

IH264ENC_AspectRati

oIdc

Enumeration for aspect ratio

IH264ENC_ASPECTRATIO_UNSP

ECIFIED

Unspecified aspect ratio

IH264ENC_ASPECTRATIO_SQUA

1:1 (square) aspect ratio

API Reference

4-28

Group or Enumeration

Class

Symbolic Constant Name

Description or Evaluation

IH264ENC_ASPECTRATIO_12_1

12:11 aspect ratio

IH264ENC_ASPECTRATIO_10_1

10:11 aspect ratio

IH264ENC_ASPECTRATIO_16_1

16:11 aspect ratio

IH264ENC_ASPECTRATIO_40_3

40:33 aspect ratio

IH264ENC_ASPECTRATIO_24_1

24:11 aspect ratio

IH264ENC_ASPECTRATIO_20_1

20:11 aspect ratio

IH264ENC_ASPECTRATIO_32_1

32:11 aspect ratio

IH264ENC_ASPECTRATIO_80_3

80:33 aspect ratio

IH264ENC_ASPECTRATIO_18_1

18:11 aspect ratio

IH264ENC_ASPECTRATIO_15_1

15:15 aspect ratio

IH264ENC_ASPECTRATIO_64_3

64:33 aspect ratio

IH264ENC_ASPECTRATIO_160_

160:99 aspect ratio

IH264ENC_ASPECTRATIO_4_3

4:3 aspect ratio

IH264ENC_ASPECTRATIO_3_2

3:2 aspect ratio

IH264ENC_ASPECTRATIO_2_1

2:1 aspect ratio

IH264ENC_ASPECTRATIO_EXTE

NDED

Extended aspect ratio

IH264ENC_RoiType

Enumeration for different ROI types

IH264_FACE_OBJECT

ROI is of FACE_OBJECT type

API Reference

4-29

Group or Enumeration

Class

Symbolic Constant Name

Description or Evaluation

IH264_BACKGROUND_OBJECT

ROI is of BACKGROUND_OBJECT

type

IH264_FOREGROUND_OBJECT

ROI is of FOREGROUND_OBJECT

type

IH264_DEFAULT_OBJECT

ROI is of DEFAULT_OBJECT

type

IH264_PRIVACY_MASK

ROI is of PRIVACY_MASK type

IH264ENC_NumTempora

lLayer

IH264_TEMPORAL_LAYERS_1

Only base layer

IH264_TEMPORAL_LAYERS_2

Base layer + Temporal layer

IH264_TEMPORAL_LAYERS_3

Base layer + 2Temporal

layers

IH264_TEMPORAL_LAYERS_4

Base layer + 3Temporal

layers

IH264_TEMPORAL_LAYERS_MAX

Maximum temporal layer

supported

Table 4-3. H264 Encoder Constants

Constant Name

Value

Description of Constant

IVIDENC2_DEFAULTPROFILE

-1

This constant is used when a particular

codec doesn't have a profile, or the

application doesn't know which profile

the codec should use.

IVIDENC2_DEFAULTPLEVEL

-1

This constant is used when a particular

codec doesn't have a level, or the

application doesn't know which profile

the codec should use.

IH264ENC_MAXNUMSLCGPS

Maximum number of slice groups.

IH264ENC_VERSION_LENGTH

Length of the version string. The memory

to get version number is owned by

application.

IH264ENC_MAX_NUM_SLICE_S

TART_OFFSET

Maximum Number of slice start points.

IH264ENC_MAX_SEI_METADTA

_BUFSIZE

0x3FF

Maximum size for

SEI_USER_DATA_UNREGISTERED SEI message.

IH264ENC_MAX_ROI

Maximum number of ROI rectangles

API Reference

4-30

Table 4-4. H.264 Encoder Error Statuses

Group or

Enumeration Class

Symbolic Constant

Name

Description or Evaluation

IH264ENC_ErrorB

IH264ENC_LEVEL_I

NCOMPLAINT_PARAM

ETER

Bit 0 - level non-compliant parameters.

This error is applicable when some

parameters are set, which are not meeting

the limit defined by H.264 standard Table

A-1 Level limits.

The error can be categorized under

following category :

 IH264ENC_LEVEL_INCOMPLAINT_RESOLUTION

: Invalid width/height

 IH264ENC_LEVEL_INCOMPLAINT_HRDBUFSZIE

: Invalid HrdBufferSize

 IH264ENC_LEVEL_INCOMPLAINT_BITRATE :

Invalid Bit Rate

 IH264ENC_LEVEL_INCOMPLAINT_MBSPERSECON

D : Invalid FrameRate/resolution

 IH264ENC_LEVEL_INCOMPLAINT_DPBSIZE :

Invalid DPB size For above 5

situations, only a signal bit (bit-0)

is set as true

IH264ENC_PROFILE

_INCOMPLAINT_CON

TENTTYPE

Bit 1 - Profile in-complaint content

type.

This error is applicable when

IVIDENC2_Params::inputContentType is not

set as IVIDEO_PROGRESSIVE

, and IVIDENC2_Params::profile is set as

IH264_BASELINE_PROFILE.

IH264ENC_PROFILE

_INCOMPLAINT_FMO

_SETTING

Bit 2 - Profile in-complaint FMO setting.

This error is applicable when FMO is

enabled but IVIDENC2_Params::profile is

not set as IH264_BASELINE_PROFILE.

IH264ENC_PROFILE

_INCOMPLAINT_TRA

NSFORMBLOCKSIZE

Bit 3 - Profile in-complaint transform

block size.

This error is set when

IH264ENC_Params::transformBlockSize !=

IH264_TRANSFORM_4x4 &&

IVIDENC2_Params::profile !=

IH264_HIGH_PROFILE.

API Reference

4-31

Group or

Enumeration Class

Symbolic Constant

Name

Description or Evaluation

IH264ENC_PROFILE

_INCOMPLAINT_INT

ERFRAMEINTERVAL

Bit 4 - Profile in-complaint, inter frame

interval.

This error is set when B frames are used

with IH264_BASELINE_PROFILE.

IH264ENC_PROFILE

_INCOMPLAINT_SCA

LINGMATRIXPRESET

Bit 5 - Profile in-complaint scaling

matrix setting.

This error is set when scaling matrix is

used without IH264_HIGH_PROFILE.

IH264ENC_PROFILE

_INCOMPLAINT_ENT

ROPYCODINGMODE

Bit 6 - Profile in-complaint entropy

coding mode setting.

This error is set when cabac is used

without IH264_HIGH_PROFILE/MAIN_PROFILE.

This is create time error

IH264ENC_MAX_BYT

ES_VOILATION_IN_

SLICEMODE_BYTES

Bit 6 - If number of bytes encoded in any

of the slice in the currently encoded

picture is crossing maximum unit size

then this bit will be set.

This is run time error produced during

encoding of a frame

This error bit is shared with

IH264ENC_PROFILE_INCOMPLAINT_ENTROPYCODIN

GMODE.Both Erroneous situations are

mutually exclusive hence the bits are

shared

IH264ENC_MAX_BIT

_RATE_VOILATION

Bit 7 – Max bit rate violation

Under some situations, encoder might not

be able to meet max bit rate. This bit is

set when bits consumed in one unit (1

sec) is more than the allocated as per

the given max bit rate. If the frame rate

is N , and if the max bit rate is

violated in Mth frame than this bit will

get set for frame M to N. (M <= N)

IH264ENC_IMPROPE

R_HDVICP2_STATE

Bit 16 – HDVCIP2 is not in proper state,

before using the HDVICP2, encoder checks

clock setting for all the modules of

HDVICP2 and checks for HDVCIP2 being in

standby state. If not then codec throws

this error

IH264ENC_IMPROPE

R_STREAMFORMAT

Bit 17 - Stream format is not proper.

This error is set when streamFormat is

set as IH264_NALU_STREAM but data synch

is not enabled for put data.

API Reference

4-32

Group or

Enumeration Class

Symbolic Constant

Name

Description or Evaluation

IH264ENC_IMPROPE

R_POCTYPE

Bit 18 - POC type is not proper.

This error is set when POC type 2 is used

in presence of non reference frames.

IH264ENC_IMPROPE

R_DATASYNC_SETTI

Bit 19 - data synch settings are not

proper.

This error is set when encoder is asked

to operate at sub frame level but the

call back function pointer is NULL.

IH264ENC_UNSUPPO

RTED_VIDENC2PARA

Bit 20 - Invalid videnc2 parameters.

This error is set when any parameter of

structure IVIDENC2_Params is not in

allowed range.

IH264ENC_UNSUPPO

RTED_RATECONTROL

PARAMS

Bit 21 - Invalid rate control parameters.

This error is set when any parameter of

structure IH264ENC_RateControlParams is

not in allowed range.

IH264ENC_UNSUPPO

RTED_INTERCODING

PARAMS

Bit 22 - Invalid inter coding parameters.

This error is set when any parameter of

structure IH264ENC_InterCodingParams is

not in allowed range.

IH264ENC_UNSUPPO

RTED_INTRACODING

PARAMS

Bit 23 - Invalid Intra coding parameters.

This error is set when any parameter of

structure IH264ENC_IntraCodingParams is

not in allowed range.

IH264ENC_UNSUPPO

RTED_NALUNITCONT

ROLPARAMS

Bit 24 - Invalid NAL unit coding

parameters.

This error is set when any parameter of

structure IH264ENC_NALUControlParams is

not in allowed range.

IH264ENC_UNSUPPO

RTED_SLICECODING

PARAMS

Bit 25 - Invalid slice coding parameters

This error is set when any parameter of

structure IH264ENC_SliceCodingParams is

not in allowed range

IH264ENC_UNSUPPO

RTED_LOOPFILTERP

ARAMS

Bit 26 - Invalid loop filter related

parameters

This error is set when any parameter of

structure IH264ENC_LoopFilterParams is

not in allowed range

API Reference

4-33

Group or

Enumeration Class

Symbolic Constant

Name

Description or Evaluation

IH264ENC_UNSUPPO

RTED_FMOCODINGPA

RAMS

Bit 27 - Invalid FMO parameters

This error is set when any parameter of

structure IH264ENC_FMOCodingParams is not

in allowed range

IH264ENC_UNSUPPO

RTED_N_FRAME_PRO

CESSCALL_PARAMS

Bit 27 – This error bit is set when

unsupported parameter for N frame process

call is provided to codec.

IH264ENC_DATASYN

CH_RUN_TIME_ERRO

Bit 27 – Error bit to indicate run time

data synch errors mentioned below

 when number of NALs in 1KB of data is

more than 64

This error bit is shared with

IH264ENC_UNSUPPORTED_FMOCODINGPARAMS.

Both Erroneous situations are mutually

exclusive hence the bits are shared

IH264ENC_UNSUPPO

RTED_VUICODxINGP

ARAMS

Bit 28 - Invalid VUI coding parameters

This error is set when any parameter of

structure IH264ENC_VUICodingParams is not

in allowed range

IH264ENC_UNSUPPO

RTED_H264ENCPARA

Bit 29 - Invalid Create time extended

parameters

This error is set when any parameter of

structure IH264ENC_Params is not in

allowed range

IH264ENC_UNSUPPO

RTED_VIDENC2DYNA

MICPARAMS

Bit 30 - Invalid base class dynamic

parameters during control

This error is set when any parameter of

structure IVIDENC2_DynamicParams is not

in allowed range

IH264ENC_UNSUPPO

RTED_H264ENCDYNA

MICPARAMS

Bit 31 - Invalid extended class dynamic

parameters during control

This error is set when any parameter of

structure IH264ENC_DynamicParams

(excluding embedded structures) is not in

allowed range

Table 4-5. H.264 Encoder Extended Error Statuses

Group or

Enumeration Class

Symbolic Constant

Name

Description or Evaluation

API Reference

4-34

Group or

Enumeration Class

Symbolic Constant

Name

Description or Evaluation

IH264ENC_ExtErr

Bits

IH264ENC_EXTERRO

R_ACTIVEREGION

Bit 0 – Active frame region dimensions

are not matching with the encoding frame

dimensions

IH264ENC_EXTERRO

R_ANALYTICINFO_B

UFFERSIZE

Bit 1 – Analytic Info buffer size

provided is less than that of requested

IH264ENC_EXTERRO

R_BITRATE

Bit 2 – target bit rate set is more than

max bit rate

IH264ENC_EXTERRO

R_BITSTREAM_BUFF

ERSIZE

Bit 3 – Output bitstream buffer size

provided is less than that of requested.

And this validation is done only if

IVIDENC2_DynamicParams::ignoreOutbufSizeF

lag is 0. Application has a flexibility

of the not honoring this size by setting

ignoreOutbufSizeFlag to 1

IH264ENC_EXTERRO

R_CAPTUREWIDTH_F

ORCEFRAME_LTRP_Q

PEL

Bit 4 – This bit is set if any of the

below conditions is set

 The size of dynamic params

structure set is not of base class

and nor even extended class

 Encoding dimensions are out of

range

 Target bit rate less than minimum

bt rate or frame rate is less than

 Generate header mode other than

XDM_ENCODE_AU and

XDM_GENERATE_HEADER

 User forced frame other than

IDR/NA frame

 Wrong settings in interframe or

intraFrame intervals

 Incorrect settings in motion

vector accuracy settings

If this bit is set, correspondingly

IH264ENC_EXTERROR_RESOLUTION_BITRATE_FRMI

NTERVAL_GENHEADER is also set

IH264ENC_EXTERRO

R_CONTROLCALL_CM

Bit 5 – Encoder control call is made with

an incorrect command option

IH264ENC_EXTERRO

R_CREATE_ENTROPY

_PROFILE

Bit 6 – Entropy coding mode set is CABAC

for Base line profile which is not

supported

API Reference

4-35

Group or

Enumeration Class

Symbolic Constant

Name

Description or Evaluation

IH264ENC_EXTERRO

R_CREATE_GOPSTRU

CT_LOG2MAX_INTRA

INTERVAL

Bit 7 – This bit is set when

 Parameters transformBlockSize,

entropyCodingMode,

log2MaxFNumMinus4,

maxIntraFrameInterval,

IDRFrameInterval, gopStructure are

out of range.

 LTRP is enabled for interlaced

coding type

If this bit is set, correspondingly

IH264ENC_EXTERROR_CREATE_TXBLKSIZE_ENTROP

Y_POC_LTRP is also set

IH264ENC_EXTERRO

R_CREATE_HPLAYER

Bit 8 – Hierarchal layers set is out of

range i.e., numTemporalLayer set is less

tha 0 or more than 4

IH264ENC_EXTERRO

R_CREATE_HPLAYER

S_BFRAME

Bit 9 – Hierarchal coding is enabled for

'B’ Frames which is not supported

IH264ENC_EXTERRO

R_CREATE_HPLAYER

S_POC

Bit 10 – Picture Order Count(POC) type 1

is enabled for Hierarchal layer coding

which is not supported

IH264ENC_EXTERRO

R_CREATE_HPLAYER

S_REFPICMRKING

Bit 11 - Longterm Referencing (MMCO

commands) is not enabled for Hierarchal

layer coding i.e., referencePicMarking

is other than IH264_LONG_TERM_PICTURE

for Hierarchal layer coding which is not

supported

IH264ENC_EXTERRO

R_CREATE_INTERLA

CE_TYPE

Bit 12 – In interlaced encoding,

interlace coding type set is out of

supported values

IH264ENC_EXTERRO

R_CREATE_LTRP

Bit 13 – LTRP option

(enableLongTermRefFrame) set is out of

allowed ranges

IH264ENC_EXTERRO

R_CREATE_LTRP_HP

Bit 14 – LTRP option

IH264ENC_LTRP_REFERTOP_PROACTIVE is

enabled for Hierarchal layer coding which

is not supported

IH264ENC_EXTERRO

R_CREATE_LTRP_PE

RIOD

Bit 15 - When LTRP is enabled for

Hierarchal layer coding (H-P) then

LTRPPeriod set should be multiple of

subGop length

IH264ENC_EXTERRO

R_CREATE_LVL_DPB

SIZE

Bit 16 – Current DPB size set is more

than the max DPB size allowed in this

level. Current DPB size is calculated

using formaula (maxWidth * maxHeight *

1.5 * max_num_ref_frames)

API Reference

4-36

Group or

Enumeration Class

Symbolic Constant

Name

Description or Evaluation

IH264ENC_EXTERRO

R_CREATE_POC_BFR

AME

Bit 17 – Picture Order Count(POC) type 2

is set for ‘B’ frames which is not

supported

IH264ENC_EXTERRO

R_CREATE_RCDO_PR

OFILE

Bit 18 – RCDO is allowed only for

Baseline profile. Please check these

settings

IH264ENC_EXTERRO

R_CREATE_TXBLKSI

ZE_ENTROPY_POC_L

TRP

Bit 19 – This bit is set when

 Parameters transformBlockSize,

entropyCodingMode,

log2MaxFNumMinus4,

maxIntraFrameInterval,

IDRFrameInterval, gopStructure are

out of range.

 LTRP is enabled for interlaced

coding type

If this bit is set, correspondingly

IH264ENC_EXTERROR_CREATE_GOPSTRUCT_LOG2MA

X_INTRAINTERVAL is also set

IH264ENC_EXTERRO

R_CREATE_TXBLKSI

ZE_PROFILE

Bit 20 – Transform Block size

IH264_TRANSFORM_8x8 is allowed only in

High profile. Please check these settings

IH264ENC_EXTERRO

R_DATASYNC_GETFN

_PTRNULL

Bit 21 – getBufferFxn is a function

pointer used to get buffer. This function

gets called in Data synch flow.

This error bit is set if this function

pointer is NULL

IH264ENC_EXTERRO

R_DATASYNC_MBUNI

T_SLICESIZE

Bit 22 – This bit is set if

 User has enabled call back

notification at slice level and

 Encoder slice mode selected is

IH264_SLICEMODE_MBUNIT and

 Slice Unit size(number of MBs per

slice) set is less than minimum

MBs per slices.

Minimum MBs per slice is

(TotalMbsInPic/63)

IH264ENC_EXTERRO

R_DATASYNC_MODE_

BFRAME

Bit 23 – Data sync, inputDataMode is not

set to IVIDEO_ENTIREFRAME in presence of

‘B’ frames. In ‘B’ cases we allow

processing entire frames and not any mode

API Reference

4-37

Group or

Enumeration Class

Symbolic Constant

Name

Description or Evaluation

IH264ENC_EXTERRO

R_DATASYNC_MODE_

FNPTRNULL

Bit 24 – This bit is set if

 Data sync, outputDataMode is

IVIDEO_FIXEDLENGTH or

IVIDEO_SLICEMODE and putDataFxn is

NULL

 Data sync, inputDataMode is

IVIDEO_NUMROWS and getDataFxn is

NULL

IH264ENC_EXTERRO

R_DATASYNC_MODE_

H241_FNPTRNULL

Bit 25 – In H241 flow, if user is

interested in call back notification at

slice level then he has to provide

function address in getBufferFxn function

pointer else this bit is set

IH264ENC_EXTERRO

R_DATASYNC_MODE_

MINBITRATE

Bit 26 – minBitRate should be 0 if data

sync, outputDataMode is VIDEO_FIXEDLENGTH

or IVIDEO_SLICEMODE

IH264ENC_EXTERRO

R_DATASYNC_OUTPU

TDATAEXCEED

Bit 27 – Encoded output data size has

exceeded the available buffer size

IH264ENC_EXTERRO

R_DATASYNC_UNITS

Bit 28 – Data sync, numOutputDataUnits

are out of allowed ranges in case of

outputDataMode not equal to

IVIDEO_ENTIREFRAME

IH264ENC_EXTERRO

R_DYNAMIC_SRCHCE

NTRE

Bit 29 – GMV, search centre x/y are out

of allowed range

IH264ENC_EXTERRO

R_DYNAMICPARAMS_

PTRNULL

Bit 30 – In encoder control call with

command option XDM_SETPARAMS,

IVIDENC2_DynamicParams structure pointer

is NULL

IH264ENC_EXTERRO

R_EARLYEXIT

Bit 31 – Early exit called because Frame

processing is not completed at IVAHD side

IH264ENC_EXTERRO

R_FIFO_EMPTY_NOP

ROCESS

Bit 32 – In Flush process calls, no more

buffers are locked and hence process call

cannot be made further this point

IH264ENC_EXTERRO

R_FILLERBYTES_NE

GATIVE

Bit 33 – This bit set indicates that the

filler bytes size to be put in bitstream

to meet bit stream constraint is

negative. Filler data negative indiactes

that bits consumed in current unit till

now has crossed the maximum limit.So

inform to the user by seeting appropirate

bit

API Reference

4-38

Group or

Enumeration Class

Symbolic Constant

Name

Description or Evaluation

IH264ENC_EXTERRO

R_FMO_PRESET

Bit 34 – FMO coding preset is other than

default. FMO is not supported in this

release

IH264ENC_EXTERRO

R_FRMPACKING_PRE

SET

Bit 35 – Frame packing preset is not in

allowed range

IH264ENC_EXTERRO

R_FRMPACKING_TYP

E_INPCONTENT

Bit 36 – This bit is set if

 Frame packing is enabled for

interlaced cases

 Frame packing preset is user

defined and frame packing type is

more than 4

(IH264_FRAMEPACK_TOP_BOTTOM)

IH264ENC_EXTERRO

R_FRMRATE_NUMUNI

TSINTICKS

Bit 37 – Level incompliant MBs per second

IH264ENC_EXTERRO

R_GENHEADER_BITS

TREAM_BUFFERSIZE

Bit 38 – In process call with

XDM_GENERATE_HEADER mode, output buffer

size provided should be minimum of 0x100

bytes (to hold SPS and PPS)

IH264ENC_EXTERRO

R_HANDLE_BUFDESC

RIPTORS_PTRNULL

Bit 39 – Pointer of handle or inBufs or

outBufs structure may be NULL

IH264ENC_EXTERRO

R_HIGHSPEED_BFAR

Bit 40 – High speed encoding feature is

not supported for ‘B’ frames. Please

check the settings

IH264ENC_EXTERRO

R_HIGHSPEED_MEAL

GO_TXBLKSIZE_PRO

FILE

Bit 41 – This bit is set if, in High

speed encoding mode

 Transform block size is 4x4 for

High profile case

 Transform block size is 8x8 and

non High profile cases

API Reference

4-39

Group or

Enumeration Class

Symbolic Constant

Name

Description or Evaluation

IH264ENC_EXTERRO

R_HIGHSPEED_PART

IALSKIP_INTRAREF

RESHMETHOD

Bit 42 – This bit is set if, in High

speed encoding mode

 Transform block size is 4x4 for

High profile case

 Transform block size is 8x8 and

non High profile cases

 Partial Frame skip is enabled

 Intra refresh method is other than

default

If this bit is set, correspondingly

IH264ENC_EXTERROR_HIGHSPEED_TXBLKSIZE_PRO

FILE is also set

IH264ENC_EXTERRO

R_HIGHSPEED_TXBL

KSIZE_PROFILE

Bit 43 – This bit is set if, in High

speed encoding mode

 Transform block size is 4x4 for

High profile case

 Transform block size is 8x8 and

non High profile cases

 Partial Frame skip is enabled

 Intra refresh method is other than

default

If this bit is set, correspondingly

IH264ENC_EXTERROR_HIGHSPEED_PARTIALSKIP_I

NTRAREFRESHMETHOD is also set

IH264ENC_EXTERRO

R_INARGS_BASECLA

SS_WATERMARKENAB

Bit 44 – Water Marking feature of the

encoder is enabled and inArgs structure

is base class. If inArgs structure is

base class we cannot access the input key

that has to used in water marking algo.

So please use extended class of inArgs

and pass the input key

IH264ENC_EXTERRO

R_INARGS_CONTROL

Bit 45 – The control parameter passed as

part of inArgs structure is out of

allowed range

IH264ENC_EXTERRO

R_INARGS_OUTARGS

_SIZE

Bit 46 – The size parameter as part of

inArgs or outArgs structure does not

match base class not even extended class

IH264ENC_EXTERRO

R_INARGS_PTRNULL

Bit 47 – inArgs structure pointer passed

in the process call is NULL

IH264ENC_EXTERRO

R_INPCONTENT_TYP

Bit 48 – Input data content type value is

wrong or input data chroma format is not

YUV420

API Reference

4-40

Group or

Enumeration Class

Symbolic Constant

Name

Description or Evaluation

IH264ENC_EXTERRO

R_INPUT_BUFFERID

Bit 49 – inputId as part of inArgs

structure is 0

IH264ENC_EXTERRO

R_INPUTBUF_MEMTY

Bit 50 – memType value set in the plane

descriptors of inBufs structure is

falling out of allowed values

IH264ENC_EXTERRO

R_INPUTBUF_PTR_S

IZE_NULL

Bit 51 - buf value(buffer pointer) or

bufSize.bytes set in the plane

descriptors of inBufs structure is NULL

IH264ENC_EXTERRO

R_INTER_HIGHSPEE

D_MVPERMB

Bit 52 – This bit is set if, in High

speed encoding mode

 ‘B’ frames are enabled

 Motion vector accuracy is not a

Quarter pel (mvAccuracy !=

IVIDENC2_MOTIONVECTOR_QUARTERPEL)

 Motion vectors per macroblock is

not 1 (minBlockSizeP !=

IH264_BLOCKSIZE_16x16)

If this bit is set, correspondingly

IH264ENC_EXTERROR_INTER_HIGHSPEED_QPEL_FR

AMEINTERVAL is also set

IH264ENC_EXTERRO

R_INTER_HIGHSPEE

D_QPEL_FRAMEINTE

RVAL

Bit 53 – This bit is set if, in High

speed encoding mode

 ‘B’ frames are enabled

 Motion vector accuracy is not a

Quarter pel (mvAccuracy !=

IVIDENC2_MOTIONVECTOR_QUARTERPEL)

 Motion vectors per macroblock is

not 1 (minBlockSizeP !=

IH264_BLOCKSIZE_16x16)

If this bit is set, correspondingly

IH264ENC_EXTERROR_INTER_HIGHSPEED_MVPERMB

is also set

IH264ENC_EXTERRO

R_INTER_MVPERMB

Bit 54 – Motion vectors per MB set is not

in allowed range. minBlockSizeP/B is not

equal to IH264_BLOCKSIZE_16x16 and not

even IH264_BLOCKSIZE_8x8

IH264ENC_EXTERRO

R_INTER_PRESET

Bit 55 – interCoding preset value is out

of allowed range

IH264ENC_EXTERRO

R_INTER_SRCHRGN_

SKIPMVBIAS

Bit 56 – searchRangeHorP/B or

searchRangeVerP/B or skipMVCodingBias

values are out of allowed range

API Reference

4-41

Group or

Enumeration Class

Symbolic Constant

Name

Description or Evaluation

IH264ENC_EXTERRO

R_INTERLACE_DATA

LAYOUT

Bit 57 – In interlaced encoding cases,

dataLayout value set as part of inBufs

structure is not IVIDEO_FIELD_INTERLEAVED

and not even IVIDEO_FIELD_SEPARATED

IH264ENC_EXTERRO

R_INTRA_CBCR8X8

Bit 58 – This bit is set if,

 intraCodingPreset is not a default

AND

 chromaIntra8x8Enable is enabled

AND

 chromaComponentEnable value is

falling out of allowed range

Please refer IH264ENC_ChormaComponent for

more details

IH264ENC_EXTERRO

R_INTRA_GDR_BFRA

ME_INPCONTENT_RA

Bit 59 – This bit is set if, in GDR

enabled cases

 Interlaced encoding is enabled

 ‘B’ frames are enabled

 GDR overlap rows is less than 0 or

more than intraRefresh rate

IH264ENC_EXTERRO

R_INTRA_GDR_REFR

ESHRATE

Bit 60 – gdrOverlapRowsBtwFrames is more

than intraRefreshRate

IH264ENC_EXTERRO

R_INTRA_INTER_FR

MINTERVAL

Bit 61 – intraFrame interval is not a

multiple of interFrame interval

IH264ENC_EXTERRO

R_INTRA_LEVEL_MO

Bit 62 - lumaIntra8x8Enable is enabled

for profile other than IH264_HIGH_PROFILE

IH264ENC_EXTERRO

R_INTRA_PRESET

Bit 63 – intraCoding preset is out of

allowed range

IH264ENC_EXTERRO

R_INTRA_REFRESHM

ETHOD

Bit 64 – intraRefreshMethod value is out

of allowed range

IH264ENC_EXTERRO

R_INTRA_REFRESHM

ETHOD_RATE

Bit 65 – intraRefreshRate is less than or

equal to 0 for intraRefreshMethod value

other than default

IH264ENC_EXTERRO

R_INTRA_REFRESHR

ATE

Bit 66 - intraRefreshRate is less than 0

for GDR enabled cases

IH264ENC_EXTERRO

R_IVAHD_BADRESET

Bit 67 – HDVICP reset is not proper

IH264ENC_EXTERRO

R_IVAHD_BADSTATE

Bit 68 – HDVICP is not in standby state

API Reference

4-42

Group or

Enumeration Class

Symbolic Constant

Name

Description or Evaluation

IH264ENC_EXTERRO

R_IVAHD_RELEASE

Bit 69 – HDVICP release fail. This occurs

if HDVICP is not in standby state during

release time

IH264ENC_EXTERRO

R_LEVEL_INPCONTE

Bit 70 – level less than or equal to

IH264_LEVEL_20 for interlaced encoding

IH264ENC_EXTERRO

R_LEVELLIMIT_RES

OLUTION

Bit 71 – encoding

resolutions/frame_dimensions are more

than that of level allowed range

IH264ENC_EXTERRO

R_LOOPFILTER_OFF

ST_LFIDC

Bit 72 – loop filter preset is not a

default and loopfilterDisableIDC,

filterOffsetA/B is/are out of allowed

range or filterOffsetA/B is odd value

IH264ENC_EXTERRO

R_LOOPFILTER_PRE

SET

Bit 73 – loop filter preset is out of

allowed range

IH264ENC_EXTERRO

R_LUMA_INPUTBUF_

MEMTYPE

Bit 74 – memType value set in the plane

descriptors of inBufs structure

corresponding to Luma data is

XDM_MEMTYPE_TILED32 or

XDM_MEMTYPE_TILED16. Allowed value is

only XDM_MEMTYPE_TILED8 for Luma

container

IH264ENC_EXTERRO

R_METADATA_NUMBU

FFERS

Bit 75 – Number of meta data planes

(corresponding to

IH264_USER_DEFINED_SCALINGMATRIX and

IH264_SEI_USER_DATA_UNREGISTERED)provided

is not equal to

inBufs->numMetaPlanes

IH264ENC_EXTERRO

R_METADATABUF_ME

MTYPE

Bit 76 - memType value set in the meta

data plane descriptors of inBufs

structure is not equal to XDM_MEMTYPE_ROW

or XDM_MEMTYPE_TILEDPAGE

IH264ENC_EXTERRO

R_METADATAPLANE_

WGTTABLESIZE

Bit 77 – In case of user defined scaling

matrix, payload_size(bufSize.bytes)

provided as part of meta data plane

descriptors of inBufs structure is less

the size of weightTable

(inBufs->metadataPlaneDesc[index].bufSize.bytes) <

(sizeof(sH264WgtTables_t))

IH264ENC_EXTERRO

R_METADATATYPES

Bit 78 – metadataType[] are other than

the supported values

IH264ENC_EXTERRO

R_MULITCHNL_BFRA

ME_NOTSUPPORTED

Bit 79 – This bit is set if ‘B’ frames

are enabled in multi Frame process call

API Reference

4-43

Group or

Enumeration Class

Symbolic Constant

Name

Description or Evaluation

IH264ENC_EXTERRO

R_MULITCHNL_CHNL

NUMEXCEEDED

Bit 80 – This bit is set if number of

channels to be processed are more that of

maximum supported in multi Frame process

call

IH264ENC_EXTERRO

R_MULITCHNL_DATA

SYNC

Bit 81 – This bit is set if data sync is

enabled in multi Frame process call

IH264ENC_EXTERRO

R_MULITCHNL_FRMP

CK_STEREOIINFO

Bit 82 - This bit is set if

framePackingPreset or stereoInfoPreset is

enabled in multi Frame process call

IH264ENC_EXTERRO

R_MULITCHNL_GENH

EADER_NOTSUPPORT

Bit 83 - This bit is set if frame process

mode is XDM_GENERATE_HEADER in multi

Frame process call

IH264ENC_EXTERRO

R_MULITCHNL_MINB

ITRATE_NOTSUPPOR

TED

Bit 84 - This bit is set if minBitRate

value is more than 0 in multi Frame

process call

IH264ENC_EXTERRO

R_MULITCHNL_MVPE

RMB

Bit 85 – This bit is set if motion

vectors per MB (minBlockSizeP) is not

same for all the channels submitted in

multi Frame process call

IH264ENC_EXTERRO

R_NALU_GOLDENSPS

Bit 86 – In all NALU preset masks,

SPS_WITH_VUI bit is to be set or has to

be reset. Setting of this bit in few

masks and resetting in others is not

allowed

IH264ENC_EXTERRO

R_NALU_PRESET

Bit 87 – NALU control preset is out of

allowed range

IH264ENC_EXTERRO

R_NALU_SPS_VUI

Bit 88 – SEI NALU bit is set in a mask

and SPS_WITH_VUI is not set in that

particular mask

IH264ENC_EXTERRO

R_NOCLEANEXIT

Bit 89 – Early return because M3 has done

early abort

IH264ENC_EXTERRO

R_NUM_INPUT_OUTP

UT_BUFS

Bit 90 – There has to atleast 2 input

buffers (inBufs->numPlanes : 1 Luma and 1

CbCr) and 1 output buffer (outBufs-

>numBufs : bit stream). This bit is set

if this condition is not satisfied

API Reference

4-44

Group or

Enumeration Class

Symbolic Constant

Name

Description or Evaluation

IH264ENC_EXTERRO

R_OUTPUTBUF_MEMT

YPE

Bit 91 – This bit is set if

 output bitstream buffer’s memory

type is not XDM_MEMTYPE_ROW or not

XDM_MEMTYPE_TILEDPAGE

 output buffer’s memory type is out

of allowed range

IH264ENC_EXTERRO

R_OUTPUTBUF_PTR_

SIZE_NULL

Bit 92 – This bit is set if output data

sync is not enabled and bitstream buffer

pointer is NULL or buffer.bytes is 0

IH264ENC_EXTERRO

R_OUTPUTDATASIZE

_EXCEEDED

Bit 93 - Encoded output data size has

exceeded the available buffer size

IH264ENC_EXTERRO

R_PRESET_ENC_RAT

ECTRL_LVL

Bit 94 – This bit is set if

 encodingPreset is out of allowed

range

 rateControlPreset is out of

allowed range

 profile and level out of allowed

range

 data sync mode out of allowed

range

 inputChromaFormat,

inputContentType,operatingMode,

maxInterFrameInterval, maxHeight,

maxWidth and dataEndianness are

out of allowed range

If this bit is set, correspondingly

IH264ENC_EXTERROR_PROFILE_DATASYNC_INPCON

TENT_RES is also set

IH264ENC_EXTERRO

R_PROFILE_BFRAME

Bit 95 – ‘B’ frames are enabled for

baseline profile

IH264ENC_EXTERRO

R_PROFILE_DATASY

NC_INPCONTENT_RE

Bit 96 – This bit is set if

 encodingPreset is out of allowed

range

 rateControlPreset is out of

allowed range

 profile and level out of allowed

range

 data sync mode out of allowed

range

 inputChromaFormat,

inputContentType,operatingMode,

maxInterFrameInterval, maxHeight,

maxWidth and dataEndianness are

out of allowed range

If this bit is set, correspondingly

IH264ENC_EXTERROR_PRESET_ENC_RATECTRL_LVL

is also set

API Reference

4-45

Group or

Enumeration Class

Symbolic Constant

Name

Description or Evaluation

IH264ENC_EXTERRO

R_PROFILE_INPCON

TENT

Bit 97 – interlaced encoding is enabled

for baseline profile

IH264ENC_EXTERRO

R_RATECTRL_BFRAM

EPICSIZE

Bit 98 – rate control, minPicSizeRatioB

value is out of allowed range

IH264ENC_EXTERRO

R_RATECTRL_CBCRQ

PINDEX_INITBUFLV

Bit 99 – rate control,chromaQPIndexOffset

initialBufferLevel are out of allowed

range

IH264ENC_EXTERRO

R_RATECTRL_HRDBU

FFER_LVLEXCEED

Bit 100 – rate control, HRDBufferSize is

falling out of level decided minimum and

maximum limit

IH264ENC_EXTERRO

R_RATECTRL_IFRAM

E_QP

Bit 101 – rate control, qpI is not

falling in the range [qpMinI, qpMaxI]

IH264ENC_EXTERRO

R_RATECTRL_IFRAM

EPICSIZE

Bit 102 – rate control, minPicSizeRatioI

value is out of allowed range

IH264ENC_EXTERRO

R_RATECTRL_IPBFR

AME_QP

Bit 103 – rate control, Qp values set for

‘I’, ‘P’ and ‘B’ frames are falling out

allowed values

IH264ENC_EXTERRO

R_SCLMATRIX_META

DATA

Bit 104 – Scaling matrix preset is user

defined

(IH264_SCALINGMATRIX_USERDEFINED_SPSLEVEL

IH264_SCALINGMATRIX_USERDEFINED_PPSLEVEL)

and meta data is not enabled for user

defined scaling matrix

IH264ENC_EXTERRO

R_RATECTRL_PARAM

SPRESET

Bit 105 – rate control,

rateControlParamsPreset is out of allowed

range

IH264ENC_EXTERRO

R_RATECTRL_PBFRA

ME_QP

Bit 106 – rate control, Qp values for ‘P’

or ‘B’ frames are out of range[minQp,

maxQp]

IH264ENC_EXTERRO

R_RATECTRL_PFRAM

EPICSIZE

Bit 107 – rate control, minPicSizeRatioP

value is out of allowed range

IH264ENC_EXTERRO

R_RATECTRL_PRESE

T_BFRAME_INPCONT

ENT

Bit 108 - IVIDEO_LOW_DELAY

rateControlPreset is not supported for

interlaced encoding and for ‘B’ frame

cases

API Reference

4-46

Group or

Enumeration Class

Symbolic Constant

Name

Description or Evaluation

IH264ENC_EXTERRO

R_RATECTRL_PROFI

LE_SCALINGMTRX

Bit 109 – scaling matrix preset is out of

range or scaling matrix preset is other

than user defined for profile not equal

to IH264_HIGH_PROFILE

IH264ENC_EXTERRO

R_RATECTRL_RCALG

Bit 110 – rate control, rcAlgo is out of

allowed range

IH264ENC_EXTERRO

R_RATECTRL_RCALG

O_INTERLACE_OR_B

FRAME

Bit 111 - RATECONTROL_PRC_LOW_DELAY

rcAlgo is not supported for interlaced

encoding and for ‘B’ frame cases

IH264ENC_EXTERRO

R_RATECTRL_SKIPD

ISTWNDW

Bit 112 – This bit is set if

 rcAlgo is

RATECONTROL_PRC_LOW_DELAY

AND

 skip distribution window length

(skipDistributionWindowLength) or

number of skip frames in

distribution window specified are

out of allowed range

IH264ENC_EXTERRO

R_RATECTRL_VBR

Bit 113 – rate control, with CVBR

settings, VBR duration (VBRDuration) or

VBR sensitivity (VBRsensitivity) are out

of allowed range

IH264ENC_EXTERRO

R_RESOLUTION_BIT

RATE_FRMINTERVAL

_GENHEADER

Bit 114 – This bit is set if any of the

below conditions is set

 The size of dynamic params

structure set is not of base class

and not of extended class

 Encoding dimensions are out of

range

 Target bit rate less than minimum

bt rate or frame rate is less than

 Generate header mode other than

XDM_ENCODE_AU and

XDM_GENERATE_HEADER

 Used forced frame other than

IDR/NA frame

 Wrong settings in interframe or

intraFrame intervals

 Incorrect settings in motion

vector accuracy settings

If this bit is set, correspondingly

IH264ENC_EXTERROR_CAPTUREWIDTH_FORCEFRAME

_LTRP_QPEL is also set

API Reference

4-47

Group or

Enumeration Class

Symbolic Constant

Name

Description or Evaluation

IH264ENC_EXTERRO

R_ROI_COORDINATE

Bit 115 – ROI co-ordinates provided are

not proper. x/y may be less than 0 or

more than frame dimensions. Also please

see that Top.x should be less than

Bottom.x and Top.y should be less than

bottom.y for all co-ordinates

IH264ENC_EXTERRO

R_ROI_NUMBERROIS

Bit 116 – Number of ROI regions set are

out of allowed range

IH264ENC_EXTERRO

R_ROI_PRIORITY

Bit 117 – ROI priority (roiPriority)

provided are out of allowed range. This

check is done with rate control enabled

cases

IH264ENC_EXTERRO

R_ROI_QP

Bit 118 - ROI priority (roiPriority)

provided are out of allowed range. This

check is done with rate control disabled

cases. Here roiPriority holds the Qp

values of the ROI regions

IH264ENC_EXTERRO

R_ROI_TYPE

Bit 119 – ROI types (roiType) provided

are out of allowed range

IH264ENC_EXTERRO

R_SLICE_NONE_DAT

ASYNC

Bit 120 – data sync call back

notification is enabled for every slice

(outputDataMode == IVIDEO_SLICEMODE)and

slice level encoding is disabled

(sliceMode == IH264_SLICEMODE_NONE)

IH264ENC_EXTERRO

R_SLICE_H241_ENT

ROPY_INTERFRAME_

INTERLACE

Bit 121 – This bit is set if H241 is

enabled and

 Width provided is less than that

being supported in H241 cases

 Interlaced encoding is enabled

 Entropy coding mode is CABAC

 ‘B’ frames are enabled

 Slice unit size in bytes

(sliceUnitSize) provided is loess

that being supported

If this bit is set, correspondingly

IH264ENC_EXTERROR_SLICE_H241_WIDTH_SLICES

IZE is also set

IH264ENC_EXTERRO

R_SLICE_H241_STR

MFORMAT_DATASYNC

Bit 122 – stream format IH264_NALU_STREAM

is supported only with default slice mode

and datasync option as IVIDEO_SLICEMODE

API Reference

4-48

Group or

Enumeration Class

Symbolic Constant

Name

Description or Evaluation

IH264ENC_EXTERRO

R_SLICE_H241_WID

TH_SLICESIZE

Bit 123 – This bit is set if H241 is

enabled and

 Width provided is less than that

being supported in H241 cases

 Interlaced encoding is enabled

 Entropy coding mode is CABAC

 ‘B’ frames are enabled

 Slice unit size in bytes

(sliceUnitSize) provided is loess

that being supported

IH264ENC_EXTERRO

R_SLICE_MODE_SIZ

Bit 124 – This bit is set if

 Slice mode is SLICEMODE_MBUNIT and

slice size provided is more than

size of the frame or slice size is

less than 6

 Slice mode is SLICEMODE_OFFSET and

sliceOffset0 is more than

sliceOffset1 or sliceOffset1 is

more than sliceoffset2

IH264ENC_EXTERRO

R_SLICE_PRESET

Bit 125 – slice coding preset is out of

allowed range

IH264ENC_EXTERRO

R_SLICE_STRMFORM

Bit 126 – stream format provided is out

of allowed range

IH264ENC_EXTERRO

R_STATUS_PTRNULL

Bit 127 – status structure pointer passed

in the encoder control call is NULL

IH264ENC_EXTERRO

R_STATUS_SIZE

Bit 128 - The size of status params

structure set is not of base class and

nor even extended class

IH264ENC_EXTERRO

R_STEREO_INPCONT

ENT

Bit 129 – stereo info preset is enabled

for progressive encoding which is not

supported in this version of encoder

IH264ENC_EXTERRO

R_STEREO_PRESET

Bit 130 – stereo info preset set is out

of allowed range

IH264ENC_EXTERRO

R_VERSION_BUFFER

_NULL_OR_SIZE

Bit 131 – data buffer pointer as part of

status structure is NULL. This buffer is

used to place the version number of

encoder. OR The data size as part of

status structure is less than the size

required to place an version number

API Reference

4-49

Group or

Enumeration Class

Symbolic Constant

Name

Description or Evaluation

IH264ENC_EXTERRO

R_VUI_NUMUNITSIN

TICKS

Bit 132 – VUI coding preset is user

defined and the parameter numUnitsInTicks

as part of vuiCodingParams is less than 0

IH264ENC_EXTERRO

R_VUI_PRESET

Bit 133 – VUI coding preset set is out of

allowed range

IH264ENC_NUM_OUT

PUT_BUFS_ANALYTI

CINFO

Bit 134 – Analytic info is enabled and

the buffer to store analytic info is not

provided i.e., ( outBufs->numBufs < 2 AND

enableAnalyticinfo )

API Reference

4-50

4.2 Data Structures

This section describes the XDM defined data structures that are common across codec

classes. These XDM data structures can be extended to define any implementation specific

parameters for a codec component.

4.2.1 Common XDM Data Structures

This section includes the following common XDM data structures:

 XDM2_SingleBufDesc

 XDM2_BufDesc

 XDM1_AlgBufInfo

 IVIDEO1_BufDescIn

 IVIDEO2_BufDesc

 IVIDENC2_Fxns

 IVIDENC2_Params

 IVIDENC2_DynamicParams

 IVIDENC2_Inargs

 IVIDENC2_Status

 IVIDENC2_OutArgs

 XDM_Date

 XDM_Point

 XDM_Rect

 XDM_DataSyncDesc

API Reference

4-51

4.2.1.1 XDM2_SingleBufDesc

║ Description

This structure defines the buffer descriptor for input and output buffers.

║ Fields

Field

Data Type

Input/

Output

Description

*buf

XDAS_Int8

Input

Pointer to the buffer address

memType

XDAS_Int16

Input

Type of memory, See XDM_MemoryType

enumeration in Table 4-1 for more details

usageMode

XDAS_Int16

Input

Memory usage descriptor, this field is set by the owner

of the buffer (typically the application), and read by

users of the buffer (including the algorithm). See

XDM_MemoryUsageMode enumeration for more

details

bufSize

XDM2_BufSize

Input

Buffer size for tile memory/row memory

accessMask

XDAS_Int32

Input

Mask filled by the algorithm, declaring how the buffer

was accessed by the algorithm processor.

If the buffer was not accessed by the algorithm

processor (for example, it was filled through DMA or

other hardware accelerator that does not write through

the algorithm's CPU), then bits in this mask should not

be set.

It is acceptable to set several bits in this mask, if the

algorithm accessed the buffer in several ways.

This mask is often used by the application and/or

framework to manage cache on cache-based

systems.

See XDM_AccessMode enumeration in Table 4-1 for

more details.

4.2.1.2 XDM2_BufDesc

║ Description

This structure defines the buffer descriptor for output buffers.

║ Fields

Field

Data Type

Input/

Output

Description

numBufs

XDAS_Int32

Input

Number of buffers. Must be less than

XDM_MAX_IO_BUFFERS.

Descs[XDM_MAX_IO

_BUFFERS]

XDM2_SingleB

ufDesc

Input

Array of buffer descriptors

API Reference

4-52

4.2.1.3 XDM1_AlgBufInfo

║ Description

This structure defines the buffer information descriptor for input and output buffers. This

structure is filled when you invoke the control() function with the XDM_GETBUFINFO

command.

║ Fields

Field

Data Type

Input/

Output

Description

minNumInBufs

XDAS_Int32

Output

Minimum number of input buffers

minNumOutBufs

XDAS_Int32

Output

Minimum number of output buffers

minInBufSize[XDM_

MAX_IO_BUFFERS]

XDM2_BufSi

Output

Minimum size required for each input buffer

minOutBufSize[XDM

_MAX_IO_BUFFERS]

XDM2_BufSi

Output

Minimum size required for each output buffer

inBufMemoryType[X

DM_MAX_IO_BUFFERS

]

XDAS_Int32

Output

Required memory type for each input buffer.

See XDM_MemoryType enumeration in Table

4-1 for more details.

outBufMemoryType[

XDM_MAX_IO_BUFFER

XDAS_Int32

Output

Required memory type for each output buffer.

See XDM_MemoryType enumeration in Table

4-1 for more details.

minNumBufSets

XDAS_Int32

Output

Minimum number of buffer sets for buffer

management

Note:

For H.264 Encoder, the buffer details are:

 Number of input buffer required is 2 for YUV 420SP chroma

format (memType is XDM_MEMTYPE_TILED8 and

XDM_MEMTYPE_TILED16)

 Number of output buffer required is 1 (Supported memType is

XDM_MEMTYPE_ROW and XDM_MEMTYPE_TILEDPAGE)

 The input buffer sizes (in bytes) for CIF format is:

Y buffer = 352 * 288

UV buffer = 352* 144

 There is no restriction on output buffer size except that it should

contain atleast one frame of encoded data.

 When the input frame buffer that getting encoded by encoder is

not same as capture buffer then encoder still returns the size of

API Reference

4-53

the buffer accessed by him. In these situations application

should take care of proper buffer allocation for input frame buffer

These are the example buffer sizes but you can re-configure

depending on the input format.

4.2.1.4 IVIDEO1_BufDescIn

║ Desciption

This structure defines the buffer descriptor for inputs video buffers.

║ Fields

Field

Data Type

Input/

Output

Description

numBufs

XDAS_Int32

Input

Number of buffers in bufDesc[]

frameWidth

XDAS_Int32

Input

Width of the video frame

frameHeight

XDAS_Int32

Input

Height of the video frame

framePitch

XDAS_Int32

Input

Frame pitch used to store the frame.

This field can also be used to

indicate the padded width.

bufDesc[XDM_MAX_IO_BUFFERS]

XDM1_Singl

eBufDesc

Input

Picture buffers.

4.2.1.5 IVIDEO2_BufDesc

║ Description

This structure defines the buffer descriptor for input and output buffers.

║ Fields

Field

Data Type

Input/

Output

Description

numPlanes

XDAS_Int32

Input/Ou

tput

Number of buffers for video planes

numMetaPlanes

XDAS_Int32

Input/Ou

tput

Number of buffers for metadata

dataLayout

XDAS_Int32

Input/Ou

tput

Video buffer layout, field

interleaved or field separated. See

IVIDEO_VideoLayout

enumeration in Table 4-1 for more

details

planeDesc

[IVIDEO_MAX_NUM_PLANES]

XDM2_Singl

eBufDesc

Input/Ou

tput

Description for video planes

metadataPlaneDesc

[IVIDEO_MAX_NUM_METADATA_PLA

XDM2_Singl

eBufDesc

Input/Ou

tput

Description for metadata planes

API Reference

4-54

Field

Data Type

Input/

Output

Description

NES]

secondFieldOffsetWidth[IVIDE

O_MAX_NUM_PLANES]

XDAS_Int32

Input/Ou

tput

Offset value for second field in

planeDesc buffer (width in pixels)

Valid only if pointer is not NULL.

secondFieldOffsetHeight[IVID

EO_MAX_NUM_PLANES]

XDAS_Int32

Input/Ou

tput

Offset value for second field in

planeDesc buffer (height in lines)

Valid only if pointer is not NULL.

imagePitch[IVIDEO_MAX_NUM_PL

ANES]

XDAS_Int32

Input/Ou

tput

Image pitch for each plane

imageRegion

XDM_Rect

Input/Ou

tput

Decoded image region including

padding/encoder input image (top

left and bottom right).

activeFrameRegion

XDM_Rect

Input/Ou

tput

Actual display region/capture

region (top left and bottom right).

extendedError

XDAS_Int32

Input/Ou

tput

Indicates the error type, if any.

Not applicable for encoders.

frameType

XDAS_Int32

Input/Ou

tput

Video frame types. See

enumeration

IVIDEO_FrameType

enumeration in Table 4-1 for more

details.

Not applicable for encoder input

buffer.

topFieldFirstFlag

XDAS_Int32

Input/Ou

tput

Indicates when the application

(should display)/(had captured) the

top field first. Not applicable for

progressive content.

Not applicable for encoder

reconstructed buffers.

Valid values are XDAS_TRUE and

XDAS_FALSE.

repeatFirstFieldFlag

XDAS_Int32

Input/Ou

tput

Indicates when the first field should

be repeated.

Valid values are XDAS_TRUE and

XDAS_FALSE.

Only applicable for interlaced

content, not progressive.

Not applicable for encoders.

frameStatus

XDAS_Int32

Input/Ou

tput

Video in/out buffer status.

Not applicable for encoder

reconstructed buffers.

Not applicable for encoder

input buffers.

repeatFrame

XDAS_Int32

Input/Ou

Number of times the display

API Reference

4-55

Field

Data Type

Input/

Output

Description

tput

process

needs to repeat the displayed

progressive frame.

This information is useful for

progressive content when the

decoder expects the display

process to repeat the displayed

frame for a certain number of

times. This is useful for pull-down

(frame/field repetition by display

system) support

where the display frame rate is

increased without increasing the

decode frame rate.

Default value is 0.

Not applicable for encoder

reconstructed buffers.

Not required for encoder input

buffer

contentType

XDAS_Int32

Input/Ou

tput

Video content type. See

IVIDEO_ContentType

enumeration in Table 4-1 for more

details.

This is useful when the content

is both interlaced and

progressive. The display

process can use this field to

determine how to render the

display buffer.

chromaFormat

XDAS_Int32

Input/Ou

tput

Chroma format for encoder input

data/decoded output buffer. See

XDM_ChromaFormat

enumeration in Table 4-1 for more

details..

scalingWidth

XDAS_Int32

Input/Ou

tput

Scaled image width for post

processing for decoder.

Not applicable for encoders.

scalingHeight

XDAS_Int32

Input/Ou

tput

Scaled image height for post

processing for decoder.

Not applicable for encoders.

rangeMappingLuma

XDAS_Int32

Input/Ou

tput

Applicable for VC1, set to -1 as

default for other codecs

rangeMappingChroma

XDAS_Int32

Input/Ou

tput

Applicable for VC1, set to -1 as

default for other codecs

enableRangeReductionFlag

XDAS_Int32

Input/Ou

tput

Flag indicating whether to enable

range reduction or not.

Valid values are XDAS_TRUE and

XDAS_FALSE.

Applicable only for VC-1

API Reference

4-56

Figure 4-1 shows IVIDEO2_BufDesc structure with the associated variables.

Figure 4-1. IVIDEO2_BufDesc With Associated Parameters.

IVIDEO_VideoLayout

planeDesc[PLANE_INDEX].bufSize.width,

for memType=tiled

planeDesc[PLANE_IN

DEX].bufSize.height,

for memType=titled

planeDesc[PL

ANE_INDEX].

buf

Video Plane(s)

numPlanes, numMetaPlanes

bottomRight (XBR,YRB)

topLeft (XTL,YTL)

Field Interleaved

Field Separated, top field and

bottom field

secondField

OffsetWidth

secondField

OffsetHeight

bottomRight (X,Y)

activeFrameRegion

imageRegion

topLeft (X,Y)

imagePitch

Video Plane

API Reference

4-57

Note:

The following table provides the number of process calls

that needs to be made for interlaced versus progressive for

different cases.

content

Type

input

Width

input

Heigh

target

Fram

eRate

dataLayou

secondField

OffsetWidth/

Height

IVIDEO

_PROG

RESSI

1920

1088

30000

Ignore

IVIDEO

_INTER

LACED

1920

544

30000

IVIDEO_F

IELD_SEP

ARATED

Non zero

IVIDEO

_INTER

LACED

1920

544

30000

IVIDEO_F

IELD_INT

ERLEAVE

Ignore

IVIDEO

_INTER

LACED

1920

544

30000

IVIDEO_F

IELD_SEP

ARATED

0,0

0: 1920x1080p requires 30 process calls

1: 1920x1080i requires 30 process calls, where each call

accepts two fields in field separated format

2: 1920x1080i requires 30 process calls, where each call

accepts two fields in field interleaved format

3: 1920x1080i requires 60 process calls, where each call

accepts one field

 Co-ordinates of imageRegion and activeFrameRegion

should not be –ve. There is no error check perfromed by

encoder for this case

 bufSize structure of planeDesc doesn’t carry any meaning.

Buffer size is assumed to be sufficient as per width and

height, so it is don’t care

 imagePitch is don’t care if the memType != PAGE and

RAW

 In other cases imagePitch = 0 means same as width and

other values of imagePitch are valid and user

responsibility to provide correct value

4.2.1.6 IVIDENC2_Fxns

║ Description

This structure contains pointers to all the XDAIS and XDM interface functions.

║ Fields

API Reference

4-58

Field

Data Type

Input/

Output

Description

Ialg

IALG_Fxns

Input

Structure containing pointers to all the XDAIS

interface functions.

For more details, see TMS320 DSP Algorithm

Standard API Reference (literature number

SPRU360).

*process

XDAS_Int32

Input

Pointer to the process() function. See section 4.4

for more information

*control

XDAS_Int32

Input

Pointer to the control() function. See section 4.4

for more information

4.2.1.7 IVIDENC2_Params

║ Description

This structure defines the creation parameters for an algorithm instance object. Set this data

structure to NULL, if you are not sure of the values to be specified for these parameters. For

the default and supported values, see Table 4-6.

║ Fields

Field

Data Type

Input/

Output

Description

Size

XDAS_Int32

Input

Size of the base or extended (if being used)

data structure in bytes.

Supported Values:

 sizeof(IVIDENC2_Params)

 sizeof(IH264ENC_Params)

encodingPreset

XDAS_Int32

Input

Preset to control encoder quality. See

XDM_EncodingPreset enumeration in

Table 4-1 for more details.

rateControlPreset

XDAS_Int32

Input

Preset to control rate control selection. See

IVIDEO_RateControlPreset

enumeration in Table 4-1 for more details.

maxHeight

XDAS_Int32

Input

Maximum video height to be supported in

pixels.

maxWidth

XDAS_Int32

Input

Maximum video width to be supported in

pixels.

dataEndianness

XDAS_Int32

Input

Endianness of output data. See

XDM_DataFormat enumeration in Table

4-1 for more details.

API Reference

4-59

Field

Data Type

Input/

Output

Description

maxInterFrameInterval

XDAS_Int32

Input

This is used for setting the maximum number

of B frames between two reference frames.

Distance from I-frame to P-frame:

 1 - No B-frames

 2 - Insert one B-frame.

 3 - Insert two B frames

 N - Insert N-1 B frames between two P

frames.

maxBitRate

XDAS_Int32

Input

This parameter along with the

IVIDENC2_DynamicParams ::

targetBitRate is used to

Enable/Disable High Fidelity Variable

Bitrate (HFVBR/CVBR) Rate Control.

Refer Appendix N HFVBR for more details.

minBitRate

XDAS_Int32

Input

Minimum bit rate for encoding in bits per

second

inputChromaFormat

XDAS_Int32

Input

Chroma format for the input buffer.

See XDM_ChromaFormat enumeration in

Table 4-1 for more details.

inputContentType

XDAS_Int32

Input

Video content type of the buffer being

encoded.

See IVIDEO_ContentType enumeration

in Table 4-1 for more details.

operatingMode

XDAS_Int32

Input

Video coding mode of operation.

See IVIDEO_OperatingMode

enumeration in Table 4-1 for details

Profile

XDAS_Int32

Input

Profile indicator of video encoder. See

IH264ENC_Profile enumeration in Table

4-2 for more details.

Level

XDAS_Int32

Input

Level indicator of video encoder.

See IH264ENC_Level enumeration in

Table 4-2 for details.

inputDataMode

XDAS_Int32

Input

Input data mode.

See IVIDEO_DataMode enumeration in

Table 4-1 for details.

outputDataMode

XDAS_Int32

Input

Output data mode.

See IVIDEO_DataMode enumeration in

Table 4-1 for details.

numInputDataUnits

XDAS_Int32

Input

Number of input slices/rows.

Units depend on the inputDataMode,

such as number of slices/rows/blocks, and

so on.

Ignored if inputDataMode is set to full

frame mode.

API Reference

4-60

Field

Data Type

Input/

Output

Description

numOutputDataUnits

XDAS_Int32

Input

Number of output slices/rows.

Units depend on the outputDataMode,

such as number of slices/rows/blocks, and

so on.

Ignored if outputDataMode is set to full

frame mode.

metadataType[IVIDEO_M

AX_NUM_METADATA_PLANE

XDAS_Int32

Input

Type of the each meta data plane, refer

IVIDEO_MetadataType (or extended

enumeration) for possible values

Note:

The following fields of IVIDENC2_Params data structure are level

dependent:

 maxHeight

 maxWidth

 maxInterFrameInterval

To check the values supported for maxHeight and maxWidth use

the following expression:

maxFrameSizeinMbs >= (maxHeight*maxWidth) / 256;

See Table A.1 – Level Limits in ISO/IEC 14496-10 for the

supported

maxFrameSizeinMbs values.

For example, consider you have to check if the following values

are supported for level 2.0:

 maxHeight = 480

 maxWidth = 720

The supported maxFrameSizeinMbs value for level 2.0 as per

Table A.1 – Level Limits is 396.

Compute the expression as:

maxFrameSizeinMbs >= (480*720) / 256

The value of maxFrameSizeinmbs is 1350 and hence the

condition is not true. Therefore, the above values of maxHeight

and maxWidth are not supported for level 2.0.

See MaxDPB size value by referring to Table A.1 – Level Limits

and make sure currDPBsize <= MaxDPB size

currDPBsize (for 4:2:0 format) =

(maxWidth * maxHeight)* 1.5*(1 +

(maxInterFrameInterval > 1));

API Reference

4-61

 minBitrate need to be at least 10% lower than target bitrate

 minBitrate need to be at least 2 mbps lower than target bitrate

For an example if 22 mbps is target average bitrate, minBitrate should be

19.8 mbps or lower.

For an example if 10 mbps is target average bitrate, minBitrate should be 8

mbps or lower.

4.2.1.8 IVIDENC2_DynamicParams

║ Description

This structure defines the run-time parameters for an algorithm instance object. Set this data

structure to NULL, if you are not sure of the values to be specified for these parameters. For

the default and supported values, see Table 4-6

║ Fields

Field

Data Type

Input/

Output

Description

size

XDAS_Int32

Input

Size of the basic or extended (if being used) data

structure in bytes

inputHeight

XDAS_Int32

Input

Height of input frame in pixels. For interlaced

case, it is height of one field.

inputWidth

XDAS_Int32

Input

Width of input frame in pixels

When the input width is a non-multiple of 16, the

encoder expects the application to pad the input

frame to the nearest multiple of 16 to the right of

the frame. In this case, application should set

inputWidth to actual width but should provide the

padded input YUV data buffer to encoder. The

encoder then sets the difference of the actual

width and padded width as crop information in

the bit-stream

refFrameRate

XDAS_Int32

Input

Reference or input frame rate in fps * 1000. For

example, if the frame rate is 30, set this field to

30000.

targetFrameRate

XDAS_Int32

Input

Target frame rate in fps * 1000. For example, if

the frame rate is 30, set this field to 30000.

targetBitRate

XDAS_Int32

Input

Target bit-rate in bits per second. For example, if

the bit-rate is 2 Mbps, set this field to 2000000.

intraFrameInter

val

XDAS_Int32

Input

Interval between two consecutive intra frames.

For example:

 0 - Only first frame to be intra coded

 1 - No inter frames (all intra frames)

 N - One intra frame and N-1 inter frames,

where N > 1.

API Reference

4-62

Field

Data Type

Input/

Output

Description

generateHeader

XDAS_Int32

Input

Encode entire access unit or only header. See

XDM_EncMode enumeration for details.

captureWidth

XDAS_Int32

Input

If the field is set to:

 0 - Encoded image width is used as pitch.

If encoded image width is non multiple of 16

then it is rounded to next multiple of 16 and

then assigned to pitch.

When an inputWidth is non-multiple of 16,

then the encoder assumes captureWidth as

inputWidth’s next multiple of 16

 Any non-zero value, capture width is used

as pitch (if capture width is greater than

image width).

forceFrame

XDAS_Int32

Input

Force the current (immediate) frame to be

encoded as a specific frame type. See

enumeration IVIDEO_FrameType for more

details

interFrameInter

val

XDAS_Int32

Input

Number of B frames between two reference

frames; that is, the number of B frames between

two P frames or I/P frames. DEFAULT(0).

For example, this field will be:

 0 - To use maxInterFrameInterval.

 1 - Zero B frames between two reference

frames.

 2 - One B frame between two reference

frames.

 3 - Two B frames between two reference

frames. and so on...

mvAccuracy

XDAS_Int32

Input

Pixel accuracy of the motion vector.

See IVIDENC2_MotionVectorAccuracy

enumeration in Table 4-1 for details.

sampleAspectRat

ioHeight

XDAS_Int32

Input

Sample aspect ratio height. This will be

considered by encoder only when

IH264ENC_VUICodingParams::

aspectRatioIdc is

IH264ENC_ASPECTRATIO_EXTENDED

sampleAspectRat

ioWidth

XDAS_Int32

Input

Sample aspect ratio width. This will be

considered by encoder only when

IH264ENC_VUICodingParams::

aspectRatioIdc is

IH264ENC_ASPECTRATIO_EXTENDED

ignoreOutbufSiz

eFlag

XDAS_Int32

Input

Flag to indicate that for bit-stream buffer size,

application needs codec to expect the requested

size or not

Valid values are XDAS_TRUE and

XDAS_FALSE.

API Reference

4-63

Field

Data Type

Input/

Output

Description

putDataFxn

XDM_DataSy

ncPutFxn

Input

Function pointer to produce data at sub-frame

level

putDataHandle

XDM_DataSy

ncHandle

Input

Handle that identifies the data sync FIFO and is

passed as argument to putData calls

getDataFxn

XDM_DataSy

ncPutFxn

Input

Function pointer to receive data at sub-frame

level

getDataHandle

XDM_DataSy

ncHandle

Input

Handle that identifies the data sync FIFO and is

passed as argument to getData calls

getBufferFxn

XDM_DataSy

ncPutFxn

Input

Function pointer to receive buffer at sub-frame

level

getBufferHandle

XDM_DataSy

ncHandle

Input

Handle that identifies the data sync FIFO and is

passed as argument to getBufferFxn calls

lateAcquireArg

XDAS_Int32

Input

Argument used during late acquire, For all

control() commands other than

#XDM_SETLATEACQUIREARG, this field is

ignored and can therefore be set by the caller to

any value. This is a identifier for a channel in

multi channel scenario.

Note:

 The following are the limitations on the parameters of

IVIDENC2_DynamicParams data structure:

inputHeight <= maxHeight

inputWidth <= maxWidth

 See Table A.1 – Level Limits in ISO/IEC 14496-10 for the

supported values of maxMbsPerSecond.

 Use the following expression to calculate FrameSizeinMbs:

FrameSizeinMbs = (inputWidth * inputHeight) / 256;

 Following condition should satisfy

maxMbsPerSecond >= FrameSizeinMbs*targetFrameRate

 A note for inputWidth non-multiple of 16 : When an input width is

a non-multiple of 16, the encoder expects the application to pad

the input frame to the nearest multiple of 16 to the right of the

frame. In this case, application should set inputWidth to actual

width but should provide the padded input YUV data buffer to

encoder. The encoder then sets the difference of the actual

width and padded width as crop information in the bit-stream.

 When inputWidth is non-multiple of 16, the encoder expects

capture width as padded width(nearest multiple of 16). If the

capture width set is 0 and inputWidth is non-multiple of 16, then

API Reference

4-64

the encoder assumes captureWidth as inputWidth’s next

multiple of 16. In all other cases, the capture width provided

through input parameter is used for input frame processing.

 This means that the input frame buffer given to the encoder

should always be multiple of 16. It should not be a problem as it

is possible to configure the capture driver to operate on a

multiple of 16 pitch. We can then direct the encoder to code a

portion of that buffer which could be a non multiple of 16. It is

recommended that the application does appropriate padding so

that the pixels lying outside the non multiple of 16 width are not

uninitialized values. If appropriate padding is not done, it will not

affect encoding functionality but the video quality at the border

MBs may get affected.

 Application need not worry about the padding for non-multiple

inputHeight as this will be taken care by encoder internally.

4.2.1.9 IVIDENC2_Inargs

║ Description

This structure defines the run time input arguments for an algorithm instance object.

║ Fields

Field

Data Type

Input/

Output

Description

Size

XDAS_Int32

Input

Size of the basic or extended (if being used) data

structure in bytes.

inputID

XDAS_Int32

Input

Identifier to attach with the corresponding input

frames to be encoded.

Zero (0) is not a supported inputID. This value is

reserved for cases when there no input buffer is

provided.

This is useful when frames require buffering

(example, B frames) and to support buffer

management.

When there is no re-ordering,

IVIDENC2_OutArgs::freeBufId will be the

same as this inputID field.

control

XDAS_Int32

Input

Encoder control operations, By this parameter various

control operations like forcing a frame to be SKIP can

be achieved, See IVIDENC2_Control and

IH264ENC_Control enumerations for more

details.

API Reference

4-65

4.2.1.10 IVIDENC2_Status

║ Description

This structure defines parameters that describe the status of an algorithm instance object.

║ Fields

Field

Data Type

Input/

Output

Description

Size

XDAS_Int32

Input

Size of the basic or extended (if being used)

data structure in bytes.

extendedError

XDAS_Int32

Output

Extended error code.

See XDM_ErrorBit enumeration in Table

4-1 for details.

Data

XDM1_SingleBuf

Desc

Output

Buffer descriptor for data passing

If this field is not used, the application must

set data.buf to NULL.

This buffer can be used as either input or

output, depending on the command.

The buffer will be provided by the

application, and returned to the application

on return of the

IVIDENC1_Fxns.control()

call. The algorithm must not retain a pointer

to this data.

encodingPreset

XDAS_Int32

Output

Encoding preset.

See XDM_EncodingPreset enumeration

in Table 4-1 for details.

API Reference

4-66

Field

Data Type

Input/

Output

Description

rateControlPreset

XDAS_Int32

Output

Rate control preset.

See IVIDEO_RateControlPreset

enumeration in Table 4-1 for details.

maxInterFrameInte

rval

XDAS_Int32

Output

This is used for setting the maximum

number of B frames between two reference

frames.

Distance from I-frame to P-frame:

 1 - No B-frames

 2 - Insert one B-frame. Not supported

in this version of H264 Encoder

 N - Insert N-1 B frames between two P

frames

inputChromaFormat

XDAS_Int32

Output

Chroma format for the input buffer.

See XDM_ChromaFormat enumeration in

Table 4-1 for details.

inputContentType

XDAS_Int32

Output

Video content type of the buffer being

encoded.

See IVIDEO_ContentType enumeration

in Table 4-1 for details.

operatingMode

XDAS_Int32

Output

Mode of video coding.

See IVIDEO_OperatingMode

enumeration in Table 4-1 for details

profile

XDAS_Int32

Output

Profile indicator of video encoder. See

IH264ENC_Profile enumeration for

details

Level

XDAS_Int32

Output

Level indicator of video encoder.

See IH264ENC_Level enumeration in

Table 4-2 for details.

inputDataMode

XDAS_Int32

Output

Input data mode.

See IVIDEO_DataMode enumeration n

Table 4-1 for details.

outputDataMode

XDAS_Int32

Output

Output data Mode.

See IVIDEO_DataMode enumeration n

Table 4-1 for details.

numInputDataUnits

XDAS_Int32

Output

Number of input slices/rows.

Units depend on the inputDataMode,

such as number of slices/rows/blocks, and

so on.

Ignored if inputDataMode is set to full

frame mode.

API Reference

4-67

Field

Data Type

Input/

Output

Description

numOutputDataUnit

XDAS_Int32

Output

Number of output slices/rows.

Units depend on the outputDataMode,

such as number of slices/rows/blocks, and

so on.

Ignored if outputDataMode is set to full

frame mode.

configurationID

XDAS_Int32

Output

This is based on the codec configuration

and can be used by the framework to

optimize the save/restore overhead of any

resources used.

bufInfo

XDM1_AlgBufInf

Output

Input and output buffer information.

This field provides the application with the

algorithm's buffer requirements. The

requirements may vary depending on the

current configuration of the algorithm

instance.

See XDM1_AlgBufInfo data structure for

details.

encDynamicParams

IVIDENC2_Dynam

icParams

Output

Dynamic parameters in use by encoder.

See IVIDENC2_DynamicParams

enumeration for more details.

In case of extended dynamic parameters,

algorithm can check the size of Status or

DynamicParams and return the

parameters accordingly.

4.2.1.11 IVIDENC2_OutArgs

║ Description

This structure defines the run-time output arguments for an algorithm instance object.

║ Fields

Field

Data Type

Input/

Output

Description

Size

XDAS_Int32

Input

Size of the basic or extended (if being used)

data structure in bytes.

extendedError

XDAS_Int32

Output

Extended error code.

See XDM_ErrorBit enumeration in Table 4-1

for details.

bytesGenerated

XDAS_Int32

Output

The number of bytes generated during the

IVIDENC2_Fxns::process() call.

API Reference

4-68

Field

Data Type

Input/

Output

Description

encodedFrameType

XDAS_Int32

Output

Frame types for video.

See IVIDEO_FrameType enumeration in

Table 4-1 for details.

inputFrameSkip

XDAS_Int32

Output

Frame skipping modes for video.

See IVIDEO_SkipMode enumeration in Table

4-1 for details.

freeBufID[IVIDEO2_M

AX_IO_BUFFERS]

XDAS_Int32

Output

This is an array of input IDs corresponding to the

buffers that have been unlocked in the current

process call.

The first zero entry in array will indicate end of

valid freeBufIDs within the array

Buffers given by application to encoder (through

process call in IVIDEO2_BufDesc # planeDesc)

continue to be owned by the algorithm until they

are released - indicated by the ID being returned

in this freeBuf array.

The buffers released by the algorithm are

indicated by their non-zero ID (previously

provided through IVIDENC2_InArgs#inputID).

A value of zero (0) indicates an invalid ID. The

first zero entry in array will indicate end of valid

freeBufIDs within the array. Hence, the

application can stop searching the array when it

encounters the first zero entry. If no buffer was

unlocked in the process call, freeBufID[0] will

have a value of zero.

reconBufs

IVIDEO2_Buf

Desc

Output

Pointer to reconstruction buffer descriptor.

See IVIDEO2_BufDesc data structure for

more information

These output buffers correspond to

 outBufs->bufs[1]

 outBufs->bufs[2]

 outBufs->bufs[3]

reconBufs.bufDesc[0].buf is equivalent

to outBufs->bufs[1]

reconBufs.bufDesc[1].buf is equivalent

to outBufs->bufs[2]

reconBufs.bufDesc[2].buf is equivalent

to outBufs->bufs[3]

It is optional for encoder to populate this buffer

descriptor. This implementation does not

populate this descriptor.

4.2.1.12 XDM_Date

║ Description

This structure contains the date and time information.

API Reference

4-69

║ Fields

Field

Data Type

Input/

Output

Description

msecsOfDay

XDAS_Int32

Input

Milliseconds of the day

month

XDAS_Int32

Input

Month (0 = January, 11 = December)

dayOfMonth

XDAS_Int32

Input

Day (1 - 31)

dayOfWeek

XDAS_Int32

Input

Day of week (0 = Sunday, 6 = Saturday)

year

XDAS_Int32

Input

Year (since 0)

API Reference

4-70

4.2.1.13 XDM_Point

║ Description

This structure specifies the two dimensional point.

║ Fields

Field

Data Type

Input/

Output

Description

XDAS_Int32

Input

X field of the frame

XDAS_Int32

Input

Y field of the frame

4.2.1.14 XDM_Rect

║ Description

This structure defines the region in the image that is to be encoded.

║ Fields

Field

Data Type

Input/

Output

Description

topLeft

XDM_Point

Input

Top left corner of the frame.

See XDM_Point data structure for details.

bottomRight

XDM_Point

Input

Bottom right corner of the frame.

See XDM_Point data structure for details.

4.2.1.15 XDM_DataSyncDesc

║ Description

This structure provides the descriptor for the chunk of data being transferred in one call to

putData or getData.

║ Fields

Field

Data Type

Input/

Output

Description

size

XDAS_Int32

Input/Ou

tput

Size of this structure

scatteredBlo

cksFlag

XDAS_Int32

Input/Ou

tput

Flag indicating whether the individual data blocks

may be scattered in memory.

*baseAddr

XDAS_Int32

Input/Ou

tput

Base address of single data block or pointer to an

array of data block addresses of size numBlocks.

If scatteredBlocksFlag is set to XDAS_FALSE,

API Reference

4-71

Field

Data Type

Input/

Output

Description

this field points directly to the start of the first block,

and is not treated as a pointer to an array.

If scatteredBlocksFlag is set to XDAS_TRUE,

this field points to an array of pointers to data blocks.

numBlocks

XDAS_Int32

Input/Ou

tput

Number of blocks available

varBlockSize

sFlag

XDAS_Int32

Input/Ou

tput

Flag indicating whether any of the data blocks vary in

size.

Valid values are XDAS_TRUE and XDAS_FALSE.

*blockSizes

XDAS_Int32

Input/Ou

tput

Variable block sizes array.

If varBlockSizesFlag is XDAS_TRUE, this array

contains the sizes of each block.

If varBlockSizesFlag is XDAS_FALSE, this

contains the size of same-size blocks.

Memory for this array (of size numBlocks) has to be

allocated by the caller of the putData API.

API Reference

4-72

4.2.2 H.264 Encoder Data Structures

This section includes the following H.264 Encoder specific extended data structures:

 IH264ENC_Params

 IH264ENC_RateControlParams

 IH264ENC_InterCodingParams

 IH264ENC_IntraCodingParams

 IH264ENC_NALUControlParams

 IH264ENC_SliceCodingParams

 IH264ENC_LoopFilterParams

 IH264ENC_FMOCodingParams

 IH264ENC_DynamicParams

 IH264ENC_Inargs

 IH264ENC_Status

 IH264ENC_OutArgs

 IH264ENC_ProcessParams

 IH264ENC_ProcessParamsList

 IH264ENC_MetaDataFormatNaluInfo

 IH264ENC_MetaDataFormatUserDefinedSEI

 IH264ENC_Fxns

 IH264ENC_VUICodingParams

 IH264ENC_StereoInfoParms

 IH264ENC_FramePackingSEIParams

 IH264ENC_SVCCodingParams

 IH264ENC_ROIInput

API Reference

4-73

4.2.2.1 IH264ENC_Params

║ Description

This structure defines the creation parameters and any other implementation specific

parameters for a H.264 Encoder instance object. The creation parameters are defined in the

XDM data structure, IVIDENC2_Params. For the default and supported values Table 4-14.

║ Fields

Field

Data Type

Input/

Output

Description

videnc2Params

IVIDENC2_Params

Input

See IVIDENC2_Params data structure for

details.

rateControlPara

IH264ENC_RateCo

ntrolParams

Input

Controls all rate control related parameters.

See IH264ENC_RateControlParams

data structure for details.

interCodingPara

IH264ENC_InterC

odingParams

Input

Controls all inter coding related parameters.

See IH264ENC_InterCodingParams

data structure for details.

intraCodingPara

IH264ENC_IntraC

odingParams

Input

Controls all intra coding related parameters.

See IH264ENC_IntraCodingParams

data structure for details.

nalUnitControlP

arams

IH264ENC_NALUCo

ntrolParams

Input

Controls the insertion of different NALUs at

different access points in video sequence.

See IH264ENC_NALUControlParams

data structure for details.

sliceCodingPara

IH264ENC_SliceC

odingParams

Input

Controls all slice coding related parameters.

See IH264ENC_SliceCodingParams

data structure for details.

loopFilterParam

IH264ENC_LoopFi

lterParams

Input

Controls the in-loop filtering process.

See IH264ENC_LoopFilterParams

data structure for details.

fmoCodingParams

IH264ENC_FMOCod

ingParams

Input

Controls the FMO behavior.

See IH264ENC_FMOCodingParams data

structure for details.

vuiCodingParams

IH264ENC_VUICod

ingParams

Input

Controls the VUI parameters coding.

See IH264ENC_VUICodingParams data

structure for details.

stereoInfoParams

IH264ENC_StereoInf

oParams

Input

Controls the stereo video coding.

See IH264ENCStereoInfoParams

data structure for details.

framePackingSEI

Params

IH264ENC_FrameP

ackingSEIParams

Input

Controls the frame packing SEI parameters

for Stereo Video.

See

IH264ENC_FramePackingSEIParams

data structure for details.

API Reference

4-74

Field

Data Type

Input/

Output

Description

svcCodingParams

IH264ENC_SVCCod

ingParams

Input

Controls the SVC coding parameters.

Refer Annex G of the H.264 standard for

more details of SVC and parameters.

interlaceCoding

Type

XDAS_Int8

Input

Controls the type of interlaced coding.

See

IH264ENC_InterlaceCodingType

enumeration in Table 4-2 for more details.

If stereoInfoPreset !=

IH264_STEREOINFO_DISABLE &&

viewSelfContainedFlag == 0

then it gets overridden as

IH264_INTERLACE_FIELDONLY_ARF

If stereoInfoPreset !=

IH264_STEREOINFO_DISABLE &&

viewSelfContainedFlag == 1

then it gets overridden as

IH264_INTERLACE_FIELDONLY_SPF

bottomFieldIntr

XDAS_Int8

Input

Controls the type of coding for second field

for interlaced content

gopStructure

XDAS_Int8

Input

Defines the type of GOP structure, uniform

and non-uniform.

See IH264ENC_GOPStructure

enumeration in Table 4-2 for more details.

entropyCodingMo

XDAS_Int8

Input

Controls the entropy coding type.

See IH264ENC_EntropyCodingMode

enumeration in Table 4-2 for more details.

transformBlockS

ize

XDAS_Int8

Input

Transform block size.

See IH264ENC_TransformBlockSize

enumeration in Table 4-2 for more details.

log2MaxFNumMinu

XDAS_Int8

Input

Limits the maximum frame number in the

bit-stream to (1<<

(log2MaxFNumMinus4 + 4))

Range is 0 to12

picOrderCountTy

XDAS_Int8

Input

Picture order count type.

See IH264ENC_PicOrderCountType

enumeration in Table 4-2 for more details.

IH264_POC_TYPE_1 is not supported

when hierarchical P coding is enabled

(numTemporalLayer > 1).

enableWatermark

XDAS_Int8

Input

Enables or Disables water mark SEI

message in the bit stream

API Reference

4-75

Field

Data Type

Input/

Output

Description

IDRFrameInterva

XDAS_Int32

Input

Interval between two IDR frames, unit of this

parameter is intraFrameInterval

Example:

 0 : Only first I frame as IDR

 1 : All I frames are IDR.

 2 : 1 out of 2 I frames are IDR starting

from first I frame

 N: 1 out of N I frames are IDR starting

from first frame

When (numTemporalLayer > 1)

then IDR frame will reset the temporal

Gop structure and will start a new

temporal Gop structure.

pConstantMemory

XDAS_Int32

Input

This pointer points to the memory area

where constants are located. It has to be in

DDR addressable space by vDMA. This is

useful to allow re-locatable constants for the

applications, which does not use Media

Controller as host. Actual memory

controller/allocator is on another master

processor. If this is set to NULL then

encoder assumes that all constants are

pointed by symbol

H264ENC_TI_ConstData

maxIntraFrameIn

terval

XDAS_Int32

Input

Maximum interval between two consecutive

intra frames. For example:

 0 - Only first frame to be intra coded

 1 - No inter frames (all intra frames)

 N - One intra frame and N-1 inter

frames, where N > 1.

debugTraceLevel

XDAS_UInt32

Input

This parameter configures the codec to

dump a debug trace log

 0 – No Trace is enabled

 1 – Trace Level 1 is enabled

 2 – Trace Level 2 is enabled

 3 – Trace Level 3 is enabled

lastNFramesToLo

XDAS_UInt32

Input

This parameter configures the codec to

maintain a history of last N frames/pictures.

 0 – means only current frame trace is

enabled

 1 – means 1 previous frame trace is

enabled apart from current frame

 N - means N previous frame trace is

enabled apart from current frame

enableAnalytici

nfo

XDAS_Int8

Input

This parameter configures the codec to

expose analytic info like MVs and SAD

parameters

 0 – Disable

 Non-Zero - Enable

API Reference

4-76

Field

Data Type

Input/

Output

Description

enableGMVSei

XDAS_Int8

Input

Enable or disable the TI specific GMV SEI

message in the bit stream

 0 – Disable

 Non-Zero - Enable

constraintSetFl

ags

XDAS_Int8

Input

This parameter controls the values of the

constraint set flags in the bit stream. The

flags that needs to be controlled are

exposed as 4 lower bits of this byte. The 5th

bit is the preset value that tells whether to

use the default values of these flags as set

by encoder or user defined values. The

syntax of these bits is as below (MSB first)

RESVD| RESVD | RESVD | PRESET

|CSF0|CSF1|CSF2

|CSF3

If the PRESET is set to zero then the values

in the CSFX fields are ignored. If PRESET is

1 then encoder takes the values for CSF

fields and codes in the bit stream.

enableRCDO

XDAS_Int8

Input

This parameter is used to enable encoding a

bit stream compliant to Reduced Complexity

Decoding Operations (RCDO) profile

 0 – Disable

 Non-Zero - Enable

enableLongTermR

efFrame

XDAS_Int32

Input

This parameter is used to support long-term

reference frame.

Setting this parameter equal to 1 will instruct

encoder to keep its recent I/IDR frame in its

reference buffer list. So it increases the

DDR foot print by one frame buffer.

LTRPPeriod

XDAS_Int32

Input

This parameter is used to specify the long-

term reference frame marking interval. This

parameter is in use when

enableLongTermRefFrame =

IH264ENC_LTRP_REFERTOP_REACTIVE

IH264ENC_LTRP_REFERTO_PERIODICL

TRP.

numTemporalLaye

XDAS_Int8

Input

This parameter controls the temporal Levels

in bit-stream.

1 - Only base layer available in bit-stream.

2 - Maximum temporal level 1 in bit-stream

3 - Maximum temporal level 2 in bit-stream

4 - Maximum temporal level 3 in bit-stream

API Reference

4-77

Field

Data Type

Input/

Output

Description

referencePicMar

king

XDAS_Int8

Input

This parameter used to control the reference

picture marking for any non-zero value

means Long-term Picture (MMCO

Commands)

0 - Short-term Picture (Sliding Window)

1 - Long-term Picture ( MMCO Commands)

reservedParams[

XDAS_Int32

Input

Some part is kept reserved to add

parameters later without changing the foot

print of interface memory

Note:

Any field from the IH264ENC_Params (excluding

IVIDENC2_Params) structure is useful only when the

encodingPreset field of IVIDENC2_Params data structure is

equal to XDM_USER_DEFINED.

constraintSetFlags: Care must be taken in setting the user

defined values for the constrained set flags. The recommended

settings are:

 Only in the base line profile the value of the CSF3 can be set to 1, If you

want to convey the level as 1b. In all other cases it must be set to 0.

 In base line profile the values of CSF0, CSF1, CSF2 can be set to any

values by application.

 In Main profile, the value of the CSF2 must be set to zero if you want to

enable CABAC. It is recommended that this value is set to zero.

 In High Profile all the value of CSF should be zero as per standard.

4.2.2.2 IH264ENC_RateControlParams

║ Description

This structure controls rate control behavior. For the default and supported values, see Table

4-8.

║ Fields

Field

Data Type

Input/

Output

Description

rateControlPa

ramsPreset

XDAS_Int8

Input

This preset controls the USER_DEFINED versus

DEFAULT mode. If you are not aware about the

fields, it should be set as

IH264_RATECONTROLPARAMS_DEFAULT

scalingMatrix

Preset

XDAS_Int8

Input

The preset controls between default, noisy,

normal and std_default mode. It also allows for

user to provide user defined scaling matrices at

SPS/PPS level. If you are not aware about the

API Reference

4-78

Field

Data Type

Input/

Output

Description

fields, it should be set as

IH264_SCALINGMATRIX_DEFAULT

rcAlgo

XDAS_Int8

Input

This defines the rate control algorithm to be

used. Only useful if

IVIDENC2::rateControlPreset is set as

IVIDEO_USER_DEFINED

qpI

XDAS_Int8

Input

Initial quantization parameter for I/IDR frames.

Valid Range is -1 to 51

-1 indicates auto initialization else Initial QP.

When rateControlPreset =

IVIDEO_NONE, this quantization parameter is

used by the whole video frame/field.

qpMaxI

XDAS_Int8

Input

Maximum quantization parameter for I/IDR

frame(s).

Range is 0 to 51

qpMinI

XDAS_Int8

Input

Minimum quantization parameter for I/IDR

frame(s).

Range is 0 to 51.

qpP

XDAS_Int8

Input

Initial quantization parameter for P frames.

Valid Range is -1 to 51

-1 indicates auto initialization else Initial QP.

When rateControlPreset =

IVIDEO_NONE, this quantization parameter is

used by the whole video frame/field else qpP is

decided by encoder internally. When rate control

is enabled this parameter is used to encode the

initial QP in PPS

qpMaxP

XDAS_Int8

Input

Maximum quantization parameter for inter

frame(s).

Range is 0 to 51.

qpMinP

XDAS_Int8

Input

Minimum quantization parameter for inter

frame(s).

Range is 0 to 51.

qpOffsetB

XDAS_Int8

Input

Offset of B frames Quantization Parameter from

P frames and offset of the layer 1 frame’s

quantization parameter from the base layer in

case of Hierarchical coding.

qpP + qpOffsetB should be in range of

[0,51]

qpMaxB

XDAS_Int8

Input

Maximum quantization parameter for B frame(s).

Range is 0 to 51.

qpMinB

XDAS_Int8

Input

Minimum quantization parameter for B frame(s).

Range is 0 to 51.

API Reference

4-79

Field

Data Type

Input/

Output

Description

allowFrameSki

XDAS_Int8

Input

Controls frame skip.

 0 - Frame can never be skipped

 Non-zero - Frames can be skipped to

achieve target bit-rate

removeExpensi

veCoeff

XDAS_Int8

Input

Flag to remove high frequency expensive co-

efficients.

chromaQPIndex

Offset

XDAS_Int8

Input

Specifies offset to be added to luma Qp for

addressing QpC values table for chroma

components.

Valid value is between -12 and 12, (inclusive)

IPQualityFact

XDAS_Int8

Input

This provides configurality to control I frame

quality with respect to P frame. Higher quality

factor means I frame quality is given higher

importance compared to P frame.

See IH264ENC_FrameQualityFactor data

structure for possible values.

initialBuffer

Level

XDAS_Int32

Input

Initial buffer level for HRD compliance. It informs

that hypothetical decoder can start depending on

the fullness of the HRD buffer.

Initial buffer level should be provided as absolute

value of the buffer size.

HRDBufferSize

XDAS_Int32

Input

Hypothetical reference decoder buffer size. This

size controls the frame skip logic of the encoder.

For low delay applications this size should be

small. This size is in bits.

Maximum value is level dependant and min

value is 4096

minPicSizeRat

ioI

XDAS_Int16

Input

To determine ratio for min picture size.

Allowed values are 1 to 4. Setting this to 0 will

enable encoder to chosen theratio. User has to

specify this value in Q5 format.

minimum picture size is computed in the

following manner,

If minPicSizeRatioI is from 1 to 4

minPicSize = averagePicSize >>

minPicSizeRatioI

If minPicSizeRatioI

is from 5 to 31

minPicSize = averagePicSize *

minPicSizeRatioI

>> Q5

Note that this is guided value to rate control to

determine min picture size and encoder may not

strictly follow this.

maxPicSizeRat

ioI

XDAS_Int16

Input

To determine ratio for max picture size for I

picture.

API Reference

4-80

Field

Data Type

Input/

Output

Description

Upper limit for this parameter in Q5 format

(upperLimitMaxPicSizeRatioI) is equal to

IVIDENC_DynamicParams::intraFrameI

nterval << 5 ;

Allowed values are 2 to 30 & 33 to

upperLimitMaxPicSizeRatioI. Setting

this to 0 and 1 will enable encoder to chosen the

ratio. User has to specify this value in Q5 format.

maximum picture size is computed in the

following manner

If maxPicSizeRatioI is from 2 to 30

maxPicSize = averagePicSize *

maxPicSizeRatioI and

If maxPicSizeRatioI is from 33 to

upperLimitMaxPicSizeRatioI

maxPicSize = averagePicSize *

maxPicSizeRatioI >> Q5

Note that this is guided value to rate control to

determine max picture size and encoder may not

strictly follow this.

minPicSizeRat

ioP

XDAS_Int16

Input

To determine ratio for min picture size for P

pictures. Details similar to

minPicSizeRatioI

maxPicSizeRat

ioP

XDAS_Int16

Input

To determine ratio for max picture size for P

pictures.

Allowed values are 2 to 30 & 33 to 960.

Setting this to 0 and 1 will enable encoder to

chosen the ratio. User has to specify this value in

Q5 format.

Maximum picture size is computed in the

following manner

If maxPicSizeRatioP is from 2 to 30

maxPicSize = averagePicSize *

maxPicSizeRatioP and

If maxPicSizeRatioP is from 33 to 960

maxPicSize = averagePicSize *

maxPicSizeRatioP >> Q5

Note that this is guided value to rate control to

determine max picture size and encoder may not

strictly follow this.

minPicSizeRat

ioB

XDAS_Int16

Input

To determine ratio for min picture size for B

pictures. Details similar to

minPicSizeRatioI

maxPicSizeRat

ioB

XDAS_Int16

Input

To determine ratio for max picture size for B

pictures. Details similar to

maxPicSizeRatioP

API Reference

4-81

Field

Data Type

Input/

Output

Description

enablePRC

XDAS_Int8

Input

Control flag to enable MB level perceptual rate

control

enablePartial

FrameSkip

XDAS_Int8

Input

Control flag to enable partial frame skip. Only

useful with CBR rate control mode

discardSavedB

its

XDAS_Int8

Input

Control Flag to discard saved bits for future

pictures. In VBR ratecontrol mode, the saved bits

in low complexity scenes will be used for future

scene/pictures

With this flag 0, encoder will use saved bits for

future scenes and for any non-zero value

encoder discards the saved bits.

Only useful with VBR ratecontrol mode.

reserved

XDAS_Int8

Input

Some part is maintained as reserved to add

parameters later without changing the foot print

of interface memory

VBRDuration

XDAS_Int32

Input

Duration over which statistics are collected to

switch bit-rate states.Increasing this value will

make VBR wait for longer time before switching

bit-rate state

VBRsensitivit

XDAS_Int8

Input

Specifies the target bitrate used by rate control in

high complexity state.

skipDistribut

ionWindowLeng

XDAS_Int16

Input

Number of frames over which the skip frames

can be distributed

numSkipInDist

ributionWindo

XDAS_Int16

Input

Number of skips allowed within the distribution

window.

enableHRDComp

lianceMode

XDAS_Int8

Input

It controls compliance of H.264 encoder to HRD

model specified in H.264 annexure C trace level

0 : no strict hrd compliance, good for better

quality

Non-Zero : RC behavior complaint to standard

defined HRD

frameSkipThMu

lQ5

XDAS_Int32

Input

Frame skip threshold in Q5 format.

It is computed based on thefollowing equation

frameSkipThreshold = HRDBufferSize –

(frameSkipThMulQ5 * averagePicSize)

>>5

0 : Encoder chosen value

[16,128] : User defined value which will override

encoder chosen value

vbvUseLevelTh

XDAS_Int32

Input

VBV use level in Q5 format.

Vbv buffer use level is computed based on

API Reference

4-82

Field

Data Type

Input/

Output

Description

the following equation vbvUseLevel =

(vbvUseLevelThQ5 * averagePicSize

)>>5

0 : Encoder chosen value

[16,128] : User defined value which will override

encoder chosen value

ReservedRC[3]

XDAS_Int32

Input

Some part is maintained as reserved to add

parameters later without changing the foot print

of interface memory

Note:

 With enablePartialFrameSkip = non-zero, encoder might not

respect the qpMax constraints. Encoded bit-streams might have

macro blocks with QP > qpMax for any picture type.

 In VBR rate control algorithm, with a scene change the frame having

scene change will follw qpMaxI and qpMinI irrespective of frame

type

4.2.2.3 IH264ENC_InterCodingParams

║ Description

This structure contains all the parameters which controls inter MBs coding behavior. For the

default and supported values, see. Table 4-9

║ Fields

Field

Data Type

Input/

Output

Description

interCodingPr

eset

XDAS_Int8

Input

This preset controls the USER_DEFINED versus

DEFAULT mode. If you are not aware about the

fields, it should be set as

IH264_INTERCODING_DEFAULT

searchRangeHo

XDAS_Int16

Input

Horizontal search range for P frames.

Possible values: Non zero, maximum up to 144

searchRangeVe

XDAS_Int16

Input

Vertical search range for P frames.

Possible Values: Non-zero, maximum up to 64

searchRangeHo

XDAS_Int16

Input

Horizontal search range for B frames.

Possible values: Non zero, maximum up to 144

searchRangeVe

XDAS_Int16

Input

Vertical search range for B frames.

Possible values: Non-zero, maximum up to 32

interCodingBi

XDAS_Int8

Input

Bias control for having a macro block coded as

inter or intra

API Reference

4-83

Field

Data Type

Input/

Output

Description

See IH264ENC_BiasFactor enumeration in

Table 4-2 for possible values

skipMVCodingB

ias

XDAS_Int8

Input

Bias control for having a macro block use skip MV

or regular MV.

See IH264ENC_BiasFactor enumeration in

Table 4-2 for possible values

minBlockSizeP

XDAS_Int8

Input

Minimum block size for P frames.

See IH264ENC_InterBlockSize enumeration

in Table 4-2 for possible values

minBlockSizeB

XDAS_Int8

Input

Minimum block size for B frames.

See IH264ENC_InterBlockSize enumeration

in Table 4-2 for possible values

meAlgoMode

XDAS_Int8

Input

Motion Estimation algorithm Mode. See

IH264ENC_MeAlgoMode in Table 4-2 for possible

values.

Note:

 None of the parameters is ignored during run-time.

 meAlgoMode = IH264ENC_MOTIONESTMODE_HIGH_SPEED is

supported only with interFrameInterval = 1, minBlockSizeP =

IH264_BLOCKSIZE_16x16 and mvAccuracy =

IVIDENC2_MOTIONVECTOR_QUARTERPEL

 When meAlgoMode is slected as

IH264ENC_MOTIONESTMODE_HIGH_SPEED in P picture then a

performance customized flow is enabled in encoder for mode decision. In

that flow IH264ENC_IntraCodingParams :: intraRefreshMethod parameter is

ignored. In the same customized flow IH264ENC_RateControlParams ::

enablePartialFrameSkip is also ignored.

4.2.2.4 IH264ENC_IntraCodingParams

║ Description

This structure defines all the operations on H.264 Encoder instance objects. For the default

and supported values, see Table 4-10.

║ Fields

Field

Data Type

Input/

Output

Description

intraCodingPr

eset

XDAS_Int8

Input

This preset controls the user defined versus

default mode. If you are not aware about the

fields, it should be set as

INTRA_CODING_DEFAULT, other wise

API Reference

4-84

Field

Data Type

Input/

Output

Description

INTRA_CODING_USER_DEFINED.

lumaIntra4x4E

nable

XDAS_Int16

Input

This parameter controls the Luma Intra4x4

encoding in video encoder. A bit-field is provided

for each Luma intra4x4 mode as shown:

HOR_UP|VERT_LEFT|HOR_DOWN|VERT_RIGH

T|DIAG_DOWN_RIGHT|DIAG_DOWN_LEFT|DC

|HOR|VER

Set/ reset particular bit to enable/disable that

mode (0=disable, 1=enable) DC (bit-2) is ignored

Bit-10 and above are ignored

lumaIntra8x8E

nable

XDAS_Int16

Input

This parameter controls the Luma Intra8x8

encoding in video encoder. A bit-field is given for

each Luma intra8x8 mode as shown:

HOR_UP|VERT_LEFT|HOR_DOWN|VERT_RIGH

T|DIAG_DOWN_RIGHT|DIAG_DOWN_LEFT|DC

|HOR|VER

Set/ reset particular bit to enable/disable that

mode (0=disable, 1=enable) DC (bit-2)is ignored

For example : 139(decimal) = 0x8B = 010001011

(bits) = HOR, VER, VERT_LEFT are enabled and

DC is always enabled.

Bit-10 and above are ignored

lumaIntra16x1

6Enable

XDAS_Int8

Input

This parameter controls the Luma Intra16x16

encoding in video encoder. A bit-field is given for

each Luma intra16x16 mode as shown:

PLANE|DC|HOR|VER

Set/ reset particular bit to enable/disable that

mode (0=disable, 1=enable) DC (bit-2)is ignored

Bit-4 and above are ignored

chromaIntra8x

8Enable

XDAS_Int8

Input

This parameter controls the chroma Intra8x8

encoding in video encoder. A bit-field is given for

each chroma intra8x8 mode as shown:

PLANE|VER|HOR|DC

Set/ reset particular bit to enable/disable that

mode (0=disable, 1=enable) DC (bit-0) is ignored

Bit-4 and above are ignored

chromaCompone

ntEnable

XDAS_Int8

Input

This parameter controls the chroma intra

prediction search. You can choose to perform

chroma intra estimation for both Cb and Cr

samples or only on Cr samples.

For more details, see

IH264ENC_ChormaComponent enumeration in

API Reference

4-85

Field

Data Type

Input/

Output

Description

Table 4-2.

intraRefreshM

ethod

XDAS_Int8

Input

Mechanism to do intra refresh.

See IH264ENC_IntraRefreshMethods

enumeration in Table 4-2 for possible values

intraRefreshR

ate

XDAS_Int16

Input

Rate at which intra refresh is done.

Case : intraRefreshMethod =

IH264_INTRAREFRESH_CYCLIC_MBS

This rate is specified as One IntraMB per # MBs.

For example if rate is 20, there has to be one intra

MB(s) per 20 Mbs.

Case : intraRefreshMethod =

IH264_INTRAREFRESH_GDR

Intra Refesh Rate is treated as the number of rows

to be intra refreshed per frame.

gdrOverlapRow

sBtwFrames

XDAS_Int16

Input

Defines the Overlap of the Intra Refresh Region

between successive frame in case the

intraRefreshMethod

IH264_INTRAREFRESH_GDR or else treated to

be don't care. Again gdrOverlapRowsBtwFrames

should be less than intraRefreshRate

constrainedIn

traPredEnable

XDAS_Int16

Input

Controls the intra macroblock coding in P slices.

Valid values are 0,non-zero

intraCodingBi

XDAS_Int8

Input

Controls percentage of intra macroblocks. Refer

IH264ENC_IntraCodingBias for supported values.

This control is usefull to tune the HDVICP 2.0

utilization.

Note:

 transformBlockSize is applicable only for inter MBs

 If transformBlockSize == IH264_TRANSFORM_8x8 then encoder

will only use 8x8 transform for INTER coded MBs

 If transformBlockSize == IH264_TRANSFORM_4x4 then encoder

will only use 4x4 transform for INTER coded MBs

 If transformBlockSize == IH264_TRANSFORM_ADAPTIVE then

encoder will decide transform size adaptively at MB-level.

 Intra refresh mechanism IH264_INTRAREFRESH_GDR is not supported for

interlaced sequences and for ‘B’ frame cases.

4.2.2.5 IH264ENC_NALUControlParams

║ Description

API Reference

4-86

This structure contains all the parameters that define the control mechanism for insertion of

different NALU types at different point in video sequence. For the default and supported

values, see Table 4-11.

║ Fields

Field

Data Type

Input/

Output

Description

naluControlPr

eset

XDAS_Int16

Input

This preset controls the user defined versus

default mode. If you are not aware about the

fields, it should be set as

IH264_NALU_CONTROL_DEFAULT other wise

IH264_NALU_CONTROL_USERDEFINED

naluPresentMa

skStartOfSequ

ence

XDAS_Int16

Input

This parameter controls the insertion of different

NALU at start of sequence

A bit-field is given for each NALU type as shown.

14| 13| 12| 11| 10| 9| 8| 7| 6| 5|

4| 3| 2| 1| 0

UD_SEI|SPS+VUI|FILLER|EOSTREAM|EOSE

Q|AUD|PPS|SPS|SEI|IDR_SLICE|SLICE_D

P_C|SLICE_DP_B|SLICE_DP_A|SLICE|

UNSPECIFIED

Set/reset particular bit to enable/disable that

insertion of that NALU (0=disable, 1=enable)

SLICE_DP_A(bit-2), SLICE_DP_B(bit-3),

SLICE_DP_C(bit-4), SPS_EXT(bit-13) is ignored

and assumed to be zero.

EOSEQ(bit-10), EOSTREAM(bit-11) is ignored

and assumed to be zero.

bits 0-5 are ignored See Appendix B for details.

naluPresentMa

skIDRPicture

XDAS_Int16

Input

This parameter controls the insertion of different

NALU at IDR picture

A bit-field is given for each NALU type as shown:

14| 13| 12| 11| 10| 9| 8| 7| 6| 5|

4| 3| 2| 1| 0

UD_SEI|SPS+VUI|FILLER|EOSTREAM|EOSE

Q|AUD|PPS|SPS|SEI|IDR_SLICE|SLICE_D

P_C|SLICE_DP_B|SLICE_DP_A|SLICE|

UNSPECIFIED

Set/ reset particular bit to enable/disable that

insertion of that NALU (0=disable, 1=enable)

SLICE_DP_A(bit-2), SLICE_DP_B(bit-3),

SLICE_DP_C(bit-4), SPS_EXT(bit-13) is ignored

and assumed to be zero

EOSEQ(bit-10), EOSTREAM(bit-11) is ignored

and assumed to be zero

bits 0-5 are ignored See Appendix B for details.

naluPresentMa

skIntraPictur

XDAS_Int16

Input

This parameter controls the insertion of different

NALU at Intra picture(s). A bit-field is given for

each NALU type as shown:

API Reference

4-87

Field

Data Type

Input/

Output

Description

14| 13| 12| 11| 10| 9| 8| 7| 6| 5|

4| 3| 2| 1| 0

UD_SEI|SPS+VUI|FILLER|EOSTREAM|EOSE

Q|AUD|PPS|SPS|SEI|IDR_SLICE|SLICE_D

P_C|SLICE_DP_B|SLICE_DP_A|SLICE|

UNSPECIFIED

Set/ reset particular bit to enable/disable that

insertion of that NALU (0=disable, 1=enable)

SLICE_DP_A(bit-2), SLICE_DP_B(bit-3),

SLICE_DP_C(bit-4), SPS_EXT(bit-13) is ignored

and assumed to be zero

EOSEQ(bit-10), EOSTREAM(bit-11) is ignored

and assumed to be zero

bits 0-5 are ignored See Appendix B for details.

naluPresentMa

skNonIntraPic

ture

XDAS_Int16

Input

This parameter controls the insertion of different

NALU at Non-intra pictures

A bit-field is given for each NALU type as shown:

14| 13| 12| 11| 10| 9| 8| 7| 6| 5|

4| 3| 2| 1| 0

UD_SEI|SPS+VUI|FILLER|EOSTREAM|EOSE

Q|AUD|PPS|SPS|SEI|IDR_SLICE|SLICE_D

P_C|SLICE_DP_B|SLICE_DP_A|SLICE|

UNSPECIFIED

Set/ reset particular bit to enable/disable that

insertion of that NALU (0=disable, 1=enable)

SLICE_DP_A(bit-2), SLICE_DP_B(bit-3),

SLICE_DP_C(bit-4), SPS_EXT(bit-13) is ignored

and assumed to be zero

EOSEQ(bit-10), EOSTREAM(bit-11) is ignored

and assumed to be zero.

bits 0-5 are ignored See Appendix B for details.

naluPresentMa

skEndOfSequen

XDAS_Int16

Input

This parameter controls the insertion of different

NALU at end of sequence

A bit-field is given for each NALU type as shown:

14| 13| 12| 11| 10| 9| 8| 7| 6| 5|

4| 3| 2| 1| 0

UD_SEI|SPS+VUI|FILLER|EOSTREAM|EOSE

Q|AUD|PPS|SPS|SEI|IDR_SLICE|SLICE_D

P_C|SLICE_DP_B|SLICE_DP_A|SLICE|

UNSPECIFIED

Set/ reset particular bit to enable/disable that

insertion of that NALU (0=disable, 1=enable)

Except bit-11 and bit-12, rest all bits are ignored

and assumed to be zero.

See Appendix B for details See Appendix B for

details.

API Reference

4-88

4.2.2.6 IH264ENC_SliceCodingParams

║ Description

This structure contains all the parameters which controls slice encoding. For the default and

supported values, see Table 4-12.

║ Fields

Field

Data Type

Input/

Output

Description

sliceCodingPr

eset

XDAS_Int8

Input

This preset controls the user defined versus

default mode. If you are not aware about the

fields, it should be set as

IH264_SLICECODING_DEFAULT

sliceMode

XDAS_Int16

Input

This defines the control mechanism to split a

picture in slices. It can be a> Single slice per

picture. b> MB based. c> Bytes based(H241) or

d> Offset based in units of rows.

Restriction for H241: The sliceMode 2 is

supported only if the frame width is more than 128

and sliceUnitSize value should be greater

than 256.

sliceUnitSize

XDAS_Int32

Input

 If sliceMode ==

IH264_SLICEMODE_MBUNIT, then this

parameter informs the number of macro

blocks in one slice

 If sliceMode ==

IH264_SLICEMODE_BYTES, then this

parameter informs the number of bytes in one

slice

 If sliceMode ==

IH264_SLICEMODE_OFFSET, then this

parameter informs the number of offset

information provided by user. Actual offset are

provided with sliceRowStartNumber

parameter.

sliceStartOff

set[IH264ENC_

MAX_NUM_SLICE

_START_OFFSET

]

XDAS_Int8

Input

Row numbering is assumed to start from 0.

Entries in this array must have numbers in

ascending order. First slice of the picture is always

starting from 0th row of the picture, so 0th entry is

the offset of second slice in picture.

 Example 1: sliceStartOffset[0] = 25 ,

sliceStartOffset [1] = 30,

sliceStartOffset [2] = 40 will result

into 4 slices starting from row# 0, 25, 30 and

 Example 2: sliceStartOffset [0] = 25

, sliceStartOffset [1] = 70,

sliceStartOffset [2] = 60 is invalid

 Example 3: sliceStartOffset [0] = 25

, sliceStartOffset [1] = 50,

sliceStartOffset [2] = 100 will result

into 3 slices starting from row# 0, 25 and 50 (if

API Reference

4-89

Field

Data Type

Input/

Output

Description

number of rows in picture < (100 + 1))

streamFormat

XDAS_Int8

Input

Controls the type of stream: byte stream format or

NALU format

See IH264ENC_StreamFormat enumeration in

enumeration in Table 4-2 for possible values

4.2.2.7 IH264ENC_LoopFilterParams

║ Description

This structure contains all the parameters, which controls loop filtering operations. For the

default and supported values, see Table 4-13.

║ Fields

Field

Data Type

Input/

Output

Description

loopfilterPre

set

XDAS_Int8

Input

This preset controls the user defined versus

default mode. If you are not aware about the

fields, it should be set as

IH264_SLICECODING_DEFAULT

loopfilterDis

ableIDC

XDAS_Int8

Input

Controls H.264 loop filter disabling options

filterOffsetA

XDAS_Int8

Input

Alpha offset for loop filter

Range is [-12, 12] even number

filterOffsetB

XDAS_Int8

Input

Beta offset for loop filter

Range is [-12, 12] even number

4.2.2.8 IH264ENC_FMOCodingParams

║ Description

This structure contains all the parameters which controls FMO operations. For the default

and supported values, see Table 4-14.

║ Fields

Field

Data Type

Input/

Output

Description

fmoCodingPres

XDAS_Int8

Input

This preset controls the user defined versus

default mode. If you are not aware about the

fields, it should be set as

IH264_SLICECODING_DEFAULT

API Reference

4-90

Field

Data Type

Input/

Output

Description

numSliceGroup

XDAS_Int8

Input

Total number of slice groups.

Valid values are [0,8]

sliceGroupMap

Type

XDAS_Int8

Input

Type of slice group. See

IH264ENC_SliceGroupMapType enumeration in

Table 4-2 for possible values.

sliceGroupCha

ngeDirectionF

lag

XDAS_Int8

Input

Only valid when sliceGroupMapType is equal

to IH264_RASTER_SCAN_SLICE_GRP,

IH264_WIPE_SLICE_GRP or

IH264_WIPE_SLICE_GRP.

See

IH264ENC_SliceGroupChangeDirection

enumeration in Table 4-2 for possible values

sliceGroupCha

ngeRate

XDAS_Int8

Input

Only valid when sliceGroupMapType is equal

to IH264_RASTER_SCAN_SLICE_GRP,

IH264_WIPE_SLICE_GRP or

IH264_WIPE_SLICE_GRP

Valid values are : [0, factor of number of Mbs in a

row]

sliceGroupCha

ngeCycle

XDAS_Int16

Input

Only valid when sliceGroupMapType is equal

to IH264_RASTER_SCAN_SLICE_GRP,

IH264_WIPE_SLICE_GRP or

IH264_WIPE_SLICE_GRP

Valid values can be 0 to

numMbsRowsInPicture, also constrained by

sliceGroupChangeRate*sliceGroupChan

geCycle < totalMbsInFrame

sliceGroupPar

ams[MAXNUMSLC

GPS]

XDAS_Int16

Input

This field is useful when sliceGroupMapType

is equal to either

IH264_INTERLEAVED_SLICE_GRP or

IH264_FOREGRND_WITH_LEFTOVER_SLICE_

GRP

In case of IH264_INTERLEAVED_SLICE_GRP,

the i-th entry in this array is used to specify the

number of consecutive slice group macro blocks to

be assigned to the i-th slice group in raster scan

order of slice group macro block units.

Valid values are 0 to totalMbsInFrame again

constrained by sum of all the elements should not

exceed totalMbsInFrame

In case of

IH264_FOREGRND_WITH_LEFTOVER_SLICE_

GRP:

 First entry in the array specify the start

position of foreground region in terms of

macro block number. Valid values are [0,

totalMbsInFrame-1].

 Second entry in the array specifies the end

API Reference

4-91

Field

Data Type

Input/

Output

Description

position of foreground region in terms of

macro block number. Valid values are [0,

totalMbsInFrame-1] with following constrains:

endPos > startPos &&

endPosmbsInOneRow >

startPosmbsInOneRow

4.2.2.9 IH264ENC_DynamicParams

║ Description

This structure defines the run-time parameters and any other implementation specific

parameters for a H.264 Encoder instance object. The run-time parameters are defined in the

XDM data structure, IVIDENC2_DynamicParams. For the default and supported values, see

Table 4-20.

║ Fields

Field

Data Type

Input/

Output

Description

videnc2DynamicParams

IVIDENC2_Dynami

cParams

Input

See IVIDENC2_DynamicParams

data structure for details.

rateControlParams

IH264ENC_RateCo

ntrolParams

Input

Controls all rate control related

parameters. Only few are supported to

be changed as part control call.

See

IH264ENC_RateControlParams

data structure for more details.

interCodingParams

IH264ENC_InterC

odingParams

Input

Controls all inter MB coding related

parameters. Only few are supported to

be changed as part control call.

See

IH264ENC_InterCodingParams

data structure for more details

intraCodingParams

IH264ENC_IntraC

odingParams

Input

Controls all intra coding related

parameters. Only few are supported to

be changed as part control call.

See

IH264ENC_IntraCodingParams

data structure for details.

API Reference

4-92

Field

Data Type

Input/

Output

Description

sliceCodingParams

IH264ENC_SliceC

odingParams

Input

Controls all slice coding related

parameters. Only few are supported to

be changed as part control call. See

IH264ENC_SliceCodingParams

data structure for more details.

sliceGroupChangeCycle

XDAS_Int32

Input

Only valid when

sliceGroupMapType is equal to

IH264_RASTER_SCAN_SLICE_GRP,

IH264_WIPE_SLICE_GRP or

IH264_WIPE_SLICE_GRP

Valid values can be 0 to

numMbsRowsInPicture, also

constrained by

sliceGroupChangeRate*sliceG

roupChangeCycle <

totalMbsInFrame

Only valid when

sliceGroupMapType is equal to

IH264_RASTER_SCAN_SLICE_GRP.

Valid values are : [0, factor of number

of Mbs in a row]

searchCenter

XDM_Point

Input

Search center for motion estimation.

XDM_Point.x == 0x7FFF means

ignore searchCenter

enableStaticMBCount

XDAS_Int8

Input

Flag to indicate enable/disable of H.241

defined Static MB count

 0 – Disable

 Non-Zero - Enable

enableROI

XDAS_Int32

Input

Flag to Enable/Disable ROI coding.

 Non-Zero – enable ROI coding.

 0 – disable ROI coding.

Default value = 0.

ROI will be automatically disabled in

case of full frame skip and for skip

macroblocks.

reservedDynParams[3]

XDAS_Int32

Input

Some part is maintained as reserved to

add parameters later without changing

the foot print of interface memory

Note:

Any field from the IH264ENC_DynamicParams excluding

IVIDENC2_DynamicParams)structure is useful only when the

encodingPreset field of IVIDENC2_Params data structure is

equal to XDM_USER_DEFINED.

API Reference

4-93

4.2.2.10 IH264ENC_InArgs

║ Description

This structure defines the run-time input arguments for H.264 Encoder instance object.

║ Fields

Field

Data Type

Input/

Output

Description

videnc2InArgs

IVIDENC2_InArgs

Input

See IVIDENC2_InArgs data

structure for details

processId

XDAS_Int32

Input

processId in InArgs was kept to

ease the acquire time optimization

in application code. In N channel

case, acquire is happening for last

channel and this (processId) as

argument is passed into acquire

call. This will make application to

understand that for which process

call, acquire has been made. With

this information application can

optimize the time spent in acquire.

Like, it might have happened that

from last call of acquire, HDVICP2

became unavailable to any further

process call(s). In this scenario

application will get to know that

HDVICP2 was not given to

somebody else from last process

call, and hence it can do some

optimization in acquire routine.

roiInputParams

IH264ENC_RoiInput

Input

This is to pass the ROI related data

to the algorithm.

See IH264ENC_RoiInput data

structure (Section 4.2.2.11) for

details.

inputKey

XDAS_UInt32

Input

This parameter along with the few

important properties of a frame are

used to generate the encrypted

key. If watermarking is enabled

then this encrypted key would be

inserted in the form of user data

unregistered SEI message in the

encoded stream

4.2.2.11 IH264ENC_RoiInput

║ Description

API Reference

4-94

This structure defines the run-time ROI related input information for H.264 Encoder instance

object.

║ Fields

Field

Data Type

Input/

Output

Description

listROI

[IH264ENC_MAX_ROI]

XDM_Rect

Input

This gives the location of each ROI

in terms of top left and bottom right

(x,y) co-ordinates.

roiType[IH264ENC_MAX

_ROI]

XDAS_Int8

Input

Type of each ROI. The supported

types are FACE_OBJECT,

BACKGROUND_OBJECT,

FOREGROUND_OBJECT,

DEFAULT_OBJECT,and

PRIVACY_MASK .

numOfROI

XDAS_Int8

Input

Number of ROIs in the current

frame.

roiPriority[IH264ENC

_MAX_ROI]

XDAS_Int32

Input

Holds the priority information of

each ROI. Valid values include all

integers between -8 and 8,

inclusive. A higher value means

that more importance will be given

to the ROI compared to other

regions. In other words, it

determines the number of bits

given to ROI.

This parameter holds the mask

color information if ROI is of type

privacy mask.

In fixed Qp mode, This filed holds

the Qp value of specified ROI.

4.2.2.12 IH264ENC_Status

║ Description

This structure defines parameters that describe the status of the H.264 Encoder and any

other implementation specific parameters. The status parameters are defined in the XDM

data structure, IVIDENC2_Status.

Note:

This encoder supports a maximum of 36 ROIs in a frame i.e.,

IH264ENC_MAX_ROI is 36.

Overlapping of ROIs of different ROI type is not allowed.

If the ROI is detected as FACE_OBJECT, then a guard band is added

around it. For all other ROI types, no guard band is added.

API Reference

4-95

║ Fields

Field

Data Type

Input/

Output

Description

Videnc2Status

IVIDENC2_Status

Output

See IVIDENC2_Status data structure for

details.

Status of the h264 encoder along with error

information, if any.

rateControlPar

ams

IH264ENC_RateCo

ntrolParams

Output

See IH264ENC_RateControlParams

data structure for details.

interCodingPar

ams

IH264ENC_InterC

odingParams

Output

See IH264ENC_InterCodingParams

data structure for details.

intraCodingPar

ams

IH264ENC_IntraC

odingParams

Output

See IH264ENC_IntraCodingParams

data structure for details.

nalUnitControl

Params

IH264ENC_NALUCo

ntrolParams

Output

See IH264ENC_NALUControlParams

data structure for details.

sliceCodingPar

ams

IH264ENC_SliceC

odingParams

Output

See IH264ENC_SliceCodingParams

data structure for details.

loopFilterPara

IH264ENC_LoopFi

lterParams

Output

See IH264ENC_LoopFilterParams

data structure for details.

fmoCodingParam

IH264ENC_FMOCod

ingParams

Output

See IH264ENC_FMOCodingParams data

structure for details.

vuiCodingParam

IH264ENC_VUICod

ingParams

Output

See IH264ENC_VUICodingParams data

structure for details.

stereoInfoPara

IH264ENC_Stereo

InfoParams

Output

See IH264ENC_StereoInfoParams

structure for details.

framePackingSE

IParams

IH264ENC_FrameP

ackingSEIParams

Output

See

IH264ENC_FramePackingSEIParams

structure for details.

svcCodingParam

IH264ENC_SVCCod

ingParams

Output

See IH264ENC_SVCCodingParams

structure for details.

interlaceCodin

gType

IH264ENC_Interl

aceCodingType

Output

See IH264ENC_InterlaceCodingType

enumeration in Table 4-2 for details.

bottomFieldInt

XDAS_Int8

Output

Controls the type of coding for bottom field

for interlaced content

gopStructure

IH264ENC_GOPStr

ucture

Output

See IH264ENC_GOPStructure

enumeration in Table 4-2 for details

entropyCodingM

ode

IH264ENC_Entrop

yCodingMode

Output

See IH264ENC_EntropyCodingMode

enumeration in Table 4-2 for details.

API Reference

4-96

Field

Data Type

Input/

Output

Description

transformBlock

Size

IH264ENC_Transf

ormBlockSize

Output

See IH264ENC_TransformBlockSize

enumeration in Table 4-2 for details.

log2MaxFNumMin

us4

XDAS_Int8

Output

Limits the maximum frame number in the bit-

stream to (1<< (log2MaxFNumMinus4 + 4))

Range is 0 to 12.

picOrderCountT

ype

IH264ENC_PicOrd

erCountType

Output

See IH264ENC_PicOrderCountType

enumeration in Table 4-2 for details.

enableWatermar

XDAS_Int8

Output

This Parameter tells if WaterMark SEI

messages is enabled or disabled in

bitstream

0 – Disable, Non-Zero - Enable

IDRFrameInterv

XDAS_Int32

Output

Interval betweenw two IDR frames, it should

be and integer multiple of

intraFrameInterval

maxIntraFrameI

nterval

XDAS_Int32

Output

Maximum Interval between two consecutive

intra frames. For example:

 0 - Only first frame to be intra coded

 1 - No inter frames (all intra frames)

N - One intra frame and N-1 inter frames,

where N > 1.

debugTraceLeve

XDAS_UInt32

Output

Level of trace

lastNFramesToL

XDAS_UInt32

Output

Number of previous pictures for which trace

is available

enableAnalytic

info

XDAS_Int8

Output

This parameter configures the codec to

expose analytic info like MVs and SAD

parameters

 0 – Disable

 Non-Zero – Enable

enableGMVSei

XDAS_Int32

Output

Enable or disable the TI specific GMV SEI

message in the bit stream

 0 – Disable

 Non-Zero - Enable

API Reference

4-97

Field

Data Type

Input/

Output

Description

constraintSetF

lags

XDAS_Int8

Output

This parameter controls the values of the

constraint set flags in the bit stream. The

flags that needs to be controlled are

exposed as 4 lower bits of this byte. The 5th

bit is the preset value that tells whether to

use the default values of these flags as set

by encoder or user defined values. The

syntax of these bits is as below (MSB first)

RESVD| RESVD | RESVD | PRESET

|CSF0|CSF1|CSF2

|CSF3

If the PRESET is set to zero then the values

in the CSFX fields are ignored. If PRESET is

1 then encoder takes the values for CSF

fields and codes in the bit stream.

enableRCDO

XDAS_Int8

Output

This parameter is used to enable encoding a

bit stream compliant to Reduced Complexity

Decoding Operations (RCDO) profile

 0 – Disable

 Non-Zero – Enable

enableLongTerm

RefFrame

XDAS_Int8

Output

This parameter is used to support long-term

reference frame.

Setting this parameter equal to 1 will instruct

encoder to keep its recent I/IDR frame in its

reference buffer list. So it increases the DDR

foot print by one frame buffer.

LTRPPeriod

XDAS_Int32

Output

This parameter is used to specify the long-

term reference frame marking interval. This

parameter is in use when

enableLongTermRefFrame =

IH264ENC_LTRP_REFERTOP_REACTIVE

IH264ENC_LTRP_REFERTO_PERIODICL

TRP.

searchCenter

XDM_Point

Output

See XDM_Point data structure for details.

enableStaticMB

Count

XDAS_Int8

Output

Flag to indicate enable/disable of H.241

defined Static MB count

 0 – Disable

 Non-Zero - Enable

extMemoryDebug

TraceAddr

XDAS_UInt32

Output

Address in external memory where the trace

data is available

numTemporalLay

XDAS_Int8

Output

This parameter controls the temporal levels

in bit-stream.

referencePicMa

rking

XDAS_Int8

Output

This parameter used to control the reference

picture marking.

API Reference

4-98

Field

Data Type

Input/

Output

Description

extMemoryDebug

TraceSize

XDAS_UInt32

Output

Size of the trace data

enableROI

XDAS_Int8

Output

Flag to indicate enable/disable ROI coding.

 Non-Zero – enable ROI coding.

 0 – disable ROI coding.

extErrorCode[I

H264ENC_EXTERR

OR_NUM_MAXWORD

XDAS_UInt32

Output

This parameter carries the sub extended

error bits set by encoder if any. In

Configuration failure or run time error ,

application/user can look into these bits to

correct his settings. By default this

parameter carries a value of 0 but if there

exists an error in encoder then a bit(s)

corresponding to the error(s) will be set.

Please see enum

IH264ENC_ExtErrBits for more details.

4.2.2.13 IH264ENC_OutArgs

║ Description

This structure defines the run-time output parameters for the H.264 Encoder instance object.

║ Fields

Field

Data Type

Input/

Output

Description

videnc2OutArg

IVIDENC2_OutArg

Output

See IVIDENC2_OutArgs data structure for

details.

bytesGenerate

dBotField

XDAS_Int32

Output

Number of bytes generated for bottom field during

the IVIDENC2_Fxns::process()

call. This field is updated only in case of

contentType = Interlaced and both the fields

are provided to codec in single process call

vbvBufferLeve

XDAS_Int32

Output

This variable tells the buffer level at the end of

every picture coding from decoder perspective.

The value populated in this variable is latest for

every process call

numStaticMBs

XDAS_Int32

Output

Number of static MBs (defined by H241) in

encoded picture during the

IVIDENC2_Fxns::process()

call. This field is valid only if

dynamicParams.enableStaticMBCount is set to

non-zero.

temporalId

XDAS_Int32

Output

This parameter carries the temporal layer Id of

current frame in Hierarchical encoding.

 If the value of

API Reference

4-99

Field

Data Type

Input/

Output

Description

IH264ENC_Params::numTemporalLaye

r parameter is 1

(IH264_TEMPORAL_LAYERS_1, base layer

encoding) then its value is 0 for P-pictures

and 1 for B-pictures.

 If the value of

IH264ENC_Params::numTemporalLaye

r parameter is more than 1 (H-P encoding)

then this parameter holds the temporal layer

id of the current picture.

 In case of interlace, both the fields will have

the same temporal id.

 If the value of

IH264ENC_SVCCodingParams::svcExt

ensionFlag is set

(IH264_SVC_EXTENSION_FLAG_ENABLE)

, then the bit-stream will have the

SSPS,prefix-NALU. The temporal_id value

encoded in the prefix-NALU and the value of

this parameter are same.

control

XDAS_Int32

Output

Encoder control operations. Most of the times it is

IVIDENC2_InArgs::control. But there are certain

cases when it is not same as

IVIDENC2_InArgs::control, hence it is advisable to

look at this output information.

extErrorCode[

IH264ENC_EXTE

RROR_NUM_MAXW

ORDS]

XDAS_UInt32

Output

This parameter carries the sub extended error bits

set by encoder if any. In Configuration failure or

run time error , application/user can look into

these bits to correct his settings. By default this

parameter carries a value of 0 but if there exists

an error in encoder then a bit(s) corresponding to

the error(s) will be set. Please see enum

IH264ENC_ExtErrBits for more details.

Note:

Interpretation of bytesGenerated field depends upon usage of

base/extended class.

If Base class of OutArgs only:

 outArgs->bytesGenerated will have bytes generated of a frame for

progressive case

 outArgs->bytesGenerated will have sum of bytes generated for both

field if single process call is made for both the fields (interlaced case)

 outArgs->bytesGenerated will have bytes generated for each field if

single process call is made for each field (interlaced case)

If Extended class of OutArgs only:

 outArgs->bytesGenerated will have bytes generated of a frame for

progressive case

 outArgs->bytesGenerated will have sum of bytes generated for both

field if single process call is made for both the fields and

API Reference

4-100

outargsextended->bytesGeneratedBottomField will have bytes

generated for bottom field (interlaced case)

 outArgs->bytesGenerated will have bytes generated for each field if

single process call is made for each field (interlaced case)

4.2.2.14 IH264ENC_ProcessParams

║ Description

This structure defines the container for holding the channel information.

║ Fields

Field

Data Type

Input/

Output

Description

handle

IVIDENC2_Handle

Input

Handle for the channel.

inBufs

IVIDEO2_BufDesc *

Input

Input buffers for the channel.

outBufs

XDM2_BufDesc *

Output

Output buffers for the channel.

inArgs

IVIDENC2_InArgs *

Input

Input arguments for the channel.

outArgs

IVIDENC2_OutArgs *

Output

Output arguments for the channel.

4.2.2.15 IH264ENC_ProcessParamsList

║ Description

This structure defines the container for holding the N channel information.

║ Fields

Field

Data Type

Input/

Output

Description

numEntries

XDAS_Int32

Input

Number of channels in the given container.

enableErrorCheck

XDAS_Int32

Output

Checks the non supported features in N

channel scenario

processParams []

IH264ENC_Proces

sParams

Input

Array holding the process parameters. The

array has a maximum of

IH264ENC_MAX_LENGTH_PROCESS_LIST

(24) elements.

4.2.2.16 IH264ENC_MetaDataFormatNaluInfo

║ Description

This structure defines the format of meta data used to provide information about slice.

║ Fields

API Reference

4-101

Field

Data Type

Input/

Output

Description

naluSize

XDAS_Int32

Output

Size of each NAL Unit

4.2.2.17 IH264ENC_MetaDataFormatUserDefinedSEI

║ Description

This structure defines the format of meta data used to provide information about macro-

block.

║ Fields

Field

Data Type

Input/

Output

Description

Size

XDAS_Int32

Input

Size of the payload

payload[IH264

ENC_MAX_SEI_M

ETADTA_BUFSIZ

XDAS_Int8

Input

Payload buffer holding the user

defined SEI

4.2.2.18 IH264ENC_Fxns

║ Description

This structure defines all the operations on H.264 Encoder instance objects.

║ Fields

Field

Data Type

Input/

Output

Description

Ividenc

IVIDENC2_Fxns

Output

See IVIDENC2_Fxns data structure for details.

processMulti

XDAS_Int32

*fnPtr(IH264ENC

_ProcessParamsL

ist

*processList)

Output

Function pointer to the multi-channel process call

definition.

4.2.2.19 IH264ENC_VUICodingParams

║ Description

This structure contains all the parameters, which controls VUI parameters. Refer Annex E of

the H.264 standard for more details of VUI and parameters

║ Fields

API Reference

4-102

Field

Data Type

Input/

Output

Description

vuiCodingPreset

XDAS_Int8

Input

This preset controls the USER_DEFINED

versus DEFAULT mode. If you are not aware

about the fields, it should be set as

IH264_VUICODING_DEFAULT

aspectRatioInfoPres

entFlag

XDAS_UInt

Input

This controls the insertion of aspect ratio

information in VUI part of bit-stream

 zero : No aspect ratio related information is

transmitted

 non-zero : aspect ratio related information

is transmitted

aspectRatioIdc

XDAS_UInt

Input

Encoder inserts aspectRatioIdc as it is in

the bit-stream. It is user's responsibility

to input appropriate value.

See Table E-1 of H264 standard or enum

IH264ENC_AspectRatioIdc for valid

values.

When aspectRatioIdc ==

IH264ENC_ASPECTRATIO_EXTENDED(255

), encoder will look at

IVIDENC2_DynamicParams::sampleAsp

ectRatioHeight and

IVIDENC2_DynamicParams::sampleAsp

ectRatioWidth and use them as

sar_height and sar_width

respectively. aspectRatioIdc is left to user

to provide correct value.

if aspectRatioInfoPresentFlag ==0

then encoder ignores this parameter

videoSignalTypePres

entFlag

XDAS_UInt

Input

This controls the insertion of video signal type in

VUI part of bit-stream

 zero : No video signal related information is

transmitted.

 non-zero : video signal related information

is transmitted.

videoFormat

XDAS_UInt

Input

This controls the video format type in VUI part of

bit-stream. Encoder inserts

videoFormat(lower 3 bits) as it is in the bit-

stream. It is user's responsibility to provide

appropriate value of this.

See Table E-2 H264 standard or enum

IH264ENC_VideoFormat for valid values.

videoFullRangeFlag

XDAS_UInt

Input

This controls the video full range flag in VUI part

of bit-stream.

 zero: video range is not full{0, 255}

non-zero: video range is full

API Reference

4-103

Field

Data Type

Input/

Output

Description

timingInfoPresentFl

XDAS_UInt

Input

This controls the insertion of timing info related

parameters in VUI part of bit-stream

 zero: timing information is present

 non-zero: timing information is not present

hrdParamsPresentFla

XDAS_UInt

Input

This controls the insertion of HRD parameters in

VUI part of bit-stream

 zero: HRD Parameters are present

 non-zero: HRD Parameters are not present

numUnitsInTicks

XDAS_UInt3

Input

This controls the insertion of

numUnitsInTicks parameters in VUI part of

bit-stream

Valid values are [1, targetFrameRate]

 If this parameter is set by user then the

targetFrameRate has multiplication

factor of numUnitInTicks instead of

1000

4.2.2.20 IH264ENC_StereoInfoParams

║ Description

This structure contains all the parameters, which controls stereo video coding. Refer Annex

D of the H.264 standard for more details of Stereo Video Coding and parameters.

║ Fields

Field

Data Type

Input/

Output

Description

stereoInfoPreset

XDAS_UInt8

Input

This preset controls the Enable/Disable of

stereo videoc coding.

if enabled then USER_DEFINED or DEFAULT

mode. If user wants stereo video coding and

not aware about the fields, it should be set as

IH264_STEREO_ENABLE_DEFAULT

topFieldIsLeftViewFlag

XDAS_UInt

Input

This controls top field in video coded sequence

as a left view or right view.

 zero : Top field is Left View

 non-zero: Top field is Right view

viewSelfContainedFlag

XDAS_UInt

Input

This controls the Left/Right view should refer

Left view or Right view.

 Zero

 Leftview can refer to Leftview or Rightview.

 Right view can refer to Leftview or

Rightview.

 Non-zero

 Leftview can refer only to Leftview

 Rightview can refer only to Rightview

API Reference

4-104

4.2.2.21 IH264ENC_FramePackingSEIParams

║ Description

This structure contains all the parameters, which controls Frame Packing SEI. Refer Annex D

of the H.264 standard for more details of Frame Packing SEI Coding and parameters.

║ Fields

Field

Data Type

Input/

Output

Description

framePackingPreset

XDAS_UInt

Input

This Preset controls the Enable/Disable of Frame

packing SEI message encoding. If its enable

then controls the USER_DEFINED vs DEFAULT

mode. If User is not aware about following fields,

it should be set as

IH264_FRAMEPACK_SEI_ENABLE_DEFAULT

 0: Frame packing SEI is Disabled

(IH264_FRAMEPACK_SEI_DISABLE)

 1: Default Frame packing SEI parameters

(IH264_FRAMEPACK_SEI_ENABLE_DEF

AULT)

 2: User defined Frame packing SEI

information pamameters

(IH264_FRAMEPACK_SEI_USERDEFINE

When Frame packing SEI coding is enabled then

input content type (coding type) should be

Progressive coding.

framePackingType

XDAS_UInt

Input

Indicates that frame packing arrangement type

Refer IH264ENC_FramePackingType for

possible values

frame0PositionX

XDAS_UInt

Input

location of the upper left sample of frame 0 (Left

view) in horizontal direction

Note: Only the lower 4 bits are considered

frame0PositionY

XDAS_UInt

Input

location of the upper left sample of frame 0 (Left

view) in vertical direction

Note: Only the lower 4 bits are considered

frame1PositionX

XDAS_UInt

Input

location of the upper left sample of frame 1

(Right view) in horizontal direction

Note: Only the lower 4 bits are considered

frame1PositionY

XDAS_UInt

Input

location of the upper left sample of frame 1

(Right view) in vertical direction

Note: Only the lower 4 bits are considered

reservedByte

XDAS_UInt

Input

Value of

frame_packing_arrangement_reserved

_byte syntax element

API Reference

4-105

4.2.2.22 IH264ENC_SVCCodingParams

║ Description

This structure contains all the parameters, which controls SVC Coding and parameters.

║ Fields

Field

Data Type

Input/

Output

Description

svcExtensionFlag

XDAS_UInt

Input

This parameter configures the codec to put SVC

extensions in the bit-stream. For normal H.264

operation, this Flag needs to be ZERO (default

value). For Encoder instance to encode SSPS,

Prefix-NALU, Coded Slice in the bit-stream, this

flag needs to be set.

 0:IH264_SVC_EXTENSION_FLAG_DISABL

E - Disables all SVC features/syntaxes and

rest of the structure is not read/respected.

 1:IH264_SVC_EXTENSION_FLAG_ENABL

E - Encodes the required SVC related

syntaxes of the layer for which H.264 Codec

has been instantiated.

 2:IH264_SVC_EXTENSION_FLAG_ENABL

E_WITH_EC_FLEXIBILITY - Encodes the

required SVC related syntaxes of the layer

for which H.264 Codec has been

instantiated. This mode is used to generate

the bitstream which is compatible to TI-SVC

decoder and which will make work easy of

TI-SVC decoder’s Error Concelment by

putting info no_inter_layer_pred_flag in the

svc-bitstream slice header.

dependencyID

XDAS_UInt

Input

This parameter tell whether the current instance

is for Base layer or for enhancement layer and

also conveys Layer ID Info. This field is

respected only when svcExtensionFlag is set.

For configuring the encoder instance for BL then

this parameter should be ZERO. For

configuring the encoder instance for EL, this

parameter should hold the value of the layer ID

qualityID

XDAS_UInt

Input

This parameter tells Quality ID of the layer that

the current instance of encoder is going to

encode. This field is respected only when

svcExtensionFlag is set. For configuring the

encoder instance for BL then this parameter

should be ZERO

enhancementProfileID

XDAS_UInt

Input

This parameter conveys the enhancement

encoder instance like what should be the profile

ID to be encoded in the Sub-Sequence

Parameter Set (SSPS).This parameter is dont

care when, the svcExtensionFlag is not

set.Possible values are

IH264SVC_BASELINE_PROFILE (83) or

IH264SVC_HIGH_PROFILE (86)

API Reference

4-106

Field

Data Type

Input/

Output

Description

layerIndex

XDAS_UInt

Input

This parameter conveys the enhancement

encoder instance like what should be the

pic_parameter_set_id and seq_parameter_set_id

to be encoded in the Picture Parameter Set

(PPS) and Sub-Sequence Parameter

Set (SSPS). layerIndex is don’t care or treated to

be ZERO when svcExtensionFlag is not enabled

refLayerDQId

XDAS_Int8

Input

This parameter conveys the the DQ Id of the

ReferenceLayer.

API Reference

4-107

4.3 Default and Supported Values of Parameters

This section provides the default and supported values for the following data structures:

 IVIDENC2_Params

 IVIDENC2_DynamicParams

 IH264ENC_RateControlParams

 IH264ENC_InterCodingParams

 IH264ENC_IntraCodingParams

 IH264ENC_NALUControlParams

 IH264ENC_SliceCodingParams

 IH264ENC_LoopFilterParams

 IH264ENC_FMOCodingParams

 IH264ENC_VUICodingParams

 IH264ENC_StereoInfoParams

 IH264ENC_FramePackingSEIParams

 IH264ENC_SVCCodingParams

 IH264ENC_Params

 IH264ENC_DynamicParams

Table 4-6. Default and Supported Values for IVIDENC2_Params

Field

Default Value

Supported Value

Size

sizeof(IH264ENC_Pa

rams)

 sizeof(IVIDENC2_Params)

 sizeof(IH264ENC_Params)

ENCODINGPRESET

XDM_DEFAULT

 XDM_DEFAULT

 XDM_HIGH_SPEED

 XDM_USER_DEFINED

 XDM_HIGH_SPEED_MED_QUALITY 1

 XDM_MED_SPEED_HIGH_QUALITY 2

rateControlPreset

IVIDEO_STORAGE

 IVIDEO_STORAGE

 IVIDEO_NONE

 IVIDEO_USER_DEFINED

 IVIDEO_RATECONTROLPRESET_DEFAULT

 IVIDEO_LOW_DELAY

maxHeight

1088

[80, 4096] if contentType is

IVIDEO_PROGRESSIVE

[80, 2048]: if contentType is

IVIDEO_INTERLACED

API Reference

4-108

Field

Default Value

Supported Value

maxWidth

1920

[96, 4352]

dataEndianness

XDM_BYTE

maxInterFrameInte

rval

[1,31] if contentType is

IVIDEO_PROGRESSIVE

[1, 16]: if contentType is

IVIDEO_INTERLACED

When(numTemporalLayer > 1), then

maxInterFrameInterval =

interFrameInterval = 1

maxBitRate

-1

Any Value, see Appendix N HFVBR .

minBitRate

Any Value, See Notes as part of section 4.2.1.7

inputChromaFormat

XDM_YUV_420SP

inputContentType

IVIDEO_PROGRESSIVE

 IVIDEO_PROGRESSIVE

 IVIDEO_PROGRESSIVE_FRAME

 IVIDEO_INTERLACED

 IVIDEO_INTERLACED_FRAME

operatingMode

IVIDEO_ENCODE_ONLY

Profile

IH264_HIGH_PROFILE

 IH264_BASELINE_PROFILE

 IH264_MAIN_PROFILE

 IH264_HIGH_PROFILE

 IVIDENC2_DEFAULTPROFILE

Level

IH264_LEVEL_40

 IH264_LEVEL_10

 IH264_LEVEL_1b

 IH264_LEVEL_11

 IH264_LEVEL_12

 IH264_LEVEL_13

 IH264_LEVEL_20

 IH264_LEVEL_21

 IH264_LEVEL_22

 IH264_LEVEL_30

 IH264_LEVEL_31

 IH264_LEVEL_32

 IH264_LEVEL_40

 IH264_LEVEL_41

 IH264_LEVEL_42

 IH264_LEVEL_50

 IH264_LEVEL_51

 IVIDENC2_DEFAULTLEVEL

inputDataMode

IVIDEO_ENTIREFRAME

 IVIDEO_ENTIREFRAME

 IVIDEO_NUMROWS

outputDataMode

IVIDEO_ENTIREFRAME

 IVIDEO_ENTIREFRAME

API Reference

4-109

Field

Default Value

Supported Value

 IVIDEO_FIXEDLENGTH

 IVIDEO_SLICEMODE

 When minBitRate != 0 then only

IVIDEO_ENTIREFRAME is supported

numInputDataUnits

Ignored and assumed to be 1

numOutputDataUnit

[1,64]

metadataType[IVID

EO_MAX_NUM_METADA

TA_PLANES]

IVIDEO_METADATAPLA

NE_NONE

 IVIDEO_METADATAPLANE_NONE

 IH264_USER_DEFINED_SCALINGMATRIX

 IH264_SEI_USER_DATA_UNREGISTERED

Note:

 XDM_HIGH_SPEED_MED_QUALITY parameters are same as

XDM_DEFAULT

 XDM_MED_SPEED_HIGH_QUALITY is same as XDM_DEFAULT except

the change in minBlockSizeP and minBlockSizeB to

IH264_BLOCKSIZE_8x8 instead of IH264_BLOCKSIZE_16x16

 For low delay rate control options maxInterFrameInterval can not be

more than 1 and contentType can not be IVIDEO_INTERLACED

Table 4-7. Default and Supported Values for IVIDENC2_DynamicParams

Field

Default Value

Supported Value

size

sizeof(IH264ENC_Dyna

micParams)

 sizeof(IVIDENC2_DynamicParams)

 sizeof(IH264ENC_DynamicParams)

inputHeight

1088

[80, 4096] if contentType is

IVIDEO_PROGRESSIVE

[80, 2048]: if contentType is

IVIDEO_INTERLACED

inputWidth

1920

[96, 4352]

refFrameRate

30000

Ignore

targetFrameRate

30000

Valid Values as per Level Limit

targetBitRate

12000000

Valid Values (> 16*1024) as per Level Limit

intraFrameInter

val

Any value >=0

generateHeader

XDM_ENCODE_AU

XDM_GENERATE_HEADER

captureWidth

1920

>= inputWidth

API Reference

4-110

Field

Default Value

Supported Value

forceFrame

IVIDEO_NA_FRAME

IVIDEO_IDR_FRAME

interFrameInter

val

[1,31] if contentType is

IVIDEO_PROGRESSIVE

[1, 16]: if contentType is

IVIDEO_INTERLACED

mvAccuracy

IVIDENC2_MOTIONVECTO

R_QUARTERPEL

IVIDENC2_MOTIONVECTOR_QUARTERPEL

IVIDENC2_MOTIONVECTOR_PIXEL

sampleAspectRat

ioHeight

Any value, only lower 16 bits are considered by

encoder

sampleAspectRat

ioWidth

Any value, only lower 16 bits are considered by

encoder

ignoreOutbufSiz

eFlag

XDAS_TRUE

[0,non-zero]

*putDataFxn

NULL

Valid function pointer, NULL

putDataHandle

Any Value

*getDataFxn

NULL

Valid function pointer, NULL

getDataHandle

Any Value

getBufferFxn

Valid function pointer, NULL

getBufferHandle

NULL

Valid function pointer, NULL

lateAcquireArg

IRES_HDVICP2_UNKNOWN

LATEACQUIREARG (-1)

Any Value

Table 4-8. Default and Supported Values for IH264ENC_RateControlParams

Field

Default Value

Supported Value

rateControlParams

Preset

IH264_RATECONTROLP

ARAMS_DEFAULT

 IH264_RATECONTROLPARAMS_DEFAULT

 IH264_RATECONTROLPARAMS_USERDEFI

NED

 IH264_RATECONTROLPARAMS_EXISTING

scalingMatrixPres

IH264_SCALINGMATRI

X_NORMAL(if

profile == HIGH)

IH264_SCALINGMATRI

X_NONE(if profile

!= HIGH)

 IH264_SCALINGMATRIX_NONE

 IH264_SCALINGMATRIX_NORMAL

 IH264_SCALINGMATRIX_NOISY

 IH264_SCALINGMATRIX_STD_DEFAULT

 IH264_SCALINGMATRIX_USERDEFINED_

SPSLEVEL

 IH264_SCALINGMATRIX_USERDEFINED_

PPSLEVEL

API Reference

4-111

Field

Default Value

Supported Value

rcAlgo

IH264_RATECONTROL_

DEFAULT

 IH264_RATECONTROL_DEFAULT

 IH264_RATECONTROL_PRC

 IH264_RATECONTROL_PRC_LOW_DELAY

qpI

[-1,51]

qpMaxI

[0,51]

qpMinI

[0,51]

qpP

[-1,51]

qpMaxP

[0,51]

qpMinP

[0,51]

qpOffsetB

The value of (qpP + qpOffsetB) should

be in range of [0,51]

qpMaxB

[0,51]

qpMinB

[0,51]

allowFrameSkip

Not supported – don’t care

removeExpensiveCo

eff

0,non-zero

chromaQPIndexOffs

[-12,12]

IPQualityFactor

IH264_QUALITY_FACT

OR_DEFAULT

Ignore

initialBufferLeve

Equal to

HRDBufferSize

 Any value between –(2^31 -10^8) to (2^31 -

10^8)

HRDBufferSize

2*targetBitRate

for VBR Rate

Control

½*targetBitRate

for CBR

RateControl

Any value which is level compliant

minPicSizeRatioI

[0,4] & [5,31]

maxPicSizeRatioI

640

[0,30] & [33 to

(IVIDENC_DynamicParams::intraFrameIn

terval * 32 )] , except 31 and 32

minPicSizeRatioP

[0,4] & [5,31]

API Reference

4-112

Field

Default Value

Supported Value

maxPicSizeRatioP

[0,960] except 31 and 32

minPicSizeRatioB

[0,4]&[5,31]

maxPicSizeRatioB

[0,960] except 31 and 32

enablePRC

[0, non-zero]

enablePartialFram

eSkip

[0, non-zero]

discardSavedBits

[0, non-zero]

reserved

VBRDuration

[0,3600]

VBRsensitivity

[0,8]

skipDistributionW

indowLength

[0,10]

numSkipInDistribu

tionWindow

[0,10]

enableHRDComplian

ceMode

[0, non-zero]

frameSkipThMulQ5

0 & [16,128]

vbvUseLevelThQ5

0 & [16,128]

Note:

For low delay rate control options maxInterFrameInterval can

not be more than 1 and contentType can not be

IVIDEO_INTERLACED

Table 4-9. Default and Supported Values for IH264ENC_InterCodingParams

Field

Default Value

Supported Value

interCodingPreset

IH264_INTERCODING_

DEFAULT

 IH264_INTERCODING_DEFAULT

 IH264_INTERCODING_USERDEFINED

 IH264_INTERCODING_EXISTING

 IH264_INTERCODING_MED_SPEED_HIGH

_QUALITY

 IH264_INTERCODING_HIGH_SPEED

API Reference

4-113

Field

Default Value

Supported Value

searchRangeHorP

144

[16,144]

searchRangeVerP

[16,64]

searchRangeHorB

144

[16,144]

searchRangeVerB

[16,32]

interCodingBias

IH264_BIASFACTOR_D

EFAULT

Ignore

skipMVCodingBias

IH264_BIASFACTOR_D

EFAULT

 IH264_BIASFACTOR_DEFAULT

 IH264_BIASFACTOR_LOW

 IH264_BIASFACTOR_MILD

 IH264_BIASFACTOR_ADAPTIVE

minBlockSizeP

IH264_BLOCKSIZE_DE

FAULT

 IH264_BLOCKSIZE_16x16

 IH264_BLOCKSIZE_DEFAULT

 IH264_BLOCKSIZE_8x8

minBlockSizeB

IH264_BLOCKSIZE_DE

FAULT

 IH264_BLOCKSIZE_16x16

 IH264_BLOCKSIZE_DEFAULT

 IH264_BLOCKSIZE_8x8

meAlgoMode

IH264ENC_MOTIONEST

MODE_DEFAULT

 IH264ENC_MOTIONESTMODE_DEFAULT

 IH264ENC_MOTIONESTMODE_HIGH_SPEE

Note:

 minBlockSizeP and minBlockSizeB should be same if there

is B frame.

Table 4-10. Default and Supported Values for IH264ENC_IntraCodingParams

Field

Default Value

Supported Value

intraCodingPreset

IH264_INTRACODING

_DEFAULT

 IH264_INTRACODING_DEFAULT

 IH264_INTRACODING_USERDEFINED

 IH264_INTRACODING_EXISTING

 IH264_INTRACODING_HIGH_SPEED

lumaIntra4x4Enabl

0xFF if (profile

IH264_HIGH_PROFIL

0x0 if (profile

IH264_HIGH_PROFIL

[0x000, 0x1FF]

API Reference

4-114

Field

Default Value

Supported Value

inputContentType

IVIDEO_PROGRESSIV

0x1F if (profile

IH264_HIGH_PROFIL

inputContentType

IVIDEO_PROGRESSIV

lumaIntra8x8Enabl

0x0 if (profile

IH264_HIGH_PROFIL

0xFF if (profile

IH264_HIGH_PROFIL

inputContentType

IVIDEO_PROGRESSIV

0x1F if (profile

IH264_HIGH_PROFIL

inputContentType

IVIDEO_PROGRESSIV

[0x000, 0x1FF]

lumaIntra16x16Ena

ble

0xF

[0x0, 0xF]

chromaIntra8x8Ena

ble

0xF

[0x0, 0xF]

chromaComponentEn

able

IH264_CHROMA_COMP

ONENT_DEFAULT

 IH264_CHROMA_COMPONENT_CR_ONLY

 IH264_CHROMA_COMPONENT_CB_CR_BOTH

intraRefreshMetho

IH264_INTRAREFRES

H_DEFAULT

 IH264_INTRAREFRESH_DEFAULT

 IH264_INTRAREFRESH_CYCLIC_MBS

 IH264_INTRAREFRESH_GDR

intraRefreshRate

>=0, effective only intraRefreshMethod !=

API Reference

4-115

Field

Default Value

Supported Value

IH264_INTRAREFRESH_DEFAULT

gdrOverlapRowsBtw

Frames

<= intraRefreshRate

constrainedIntraP

redEnable

Zero, non-zero

intraCodingBias

IH264ENC_INTRACOD

INGBIAS_DEFAULT

 IH264ENC_INTRACODINGBIAS_DEFAULT

 IH264ENC_INTRACODINGBIAS_HIGH_SPE

Table 4-11. Default and Supported Values for IH264ENC_NALUControlParams

Field

Default Value

Supported Value

naluControlPreset

IH264_NALU_CONTRO

L_DEFAULT

 IH264_NALU_CONTROL_DEFAULT

 IH264_NALU_CONTROL_USERDEFINED

naluPresentMaskStartOf

Sequence

0x01A0

See Appendix B for more details

naluPresentMaskIDRPic

ture

0x01A0

See Appendix B for more details

naluPresentMaskIntraPi

cture

0x0002

See Appendix B for more details

naluPresentMaskNonIntr

aPicture

0x0002

See Appendix B for more details

naluPresentMaskEndOf

Sequence ;

0x0C00

See Appendix B for more details

Table 4-12. Default and Supported Values for IH264ENC_SliceCodingParams

Field

Default Value

Supported Value

sliceCodingPreset

IH264_SLICECODING

_DEFAULT

 IH264_SLICECODING_DEFAULT

 IH264_SLICECODING_USERDEFINED

 IH264_SLICECODING_EXISTING

sliceMode

IH264_SLICEMODE_D

EFAULT

 IH264_SLICEMODE_NONE

 IH264_SLICEMODE_MBUNIT

 IH264_SLICEMODE_OFFSET

 IH264_SLICEMODE_BYTES

sliceUnitSize

[6,number_of_mbs_in_picture]: when sliceMode

== IH264_SLICEMODE_MBUNIT

[256, Any Number]: when sliceMode ==

API Reference

4-116

Field

Default Value

Supported Value

IH264_SLICEMODE_BYTES

Ignore if sliceMode !=

IH264_SLICEMODE_MBUNIT && sliceMode

!= IH264_SLICEMODE_BYTES

sliceStartOffset[

IH264ENC_MAX_NUM_

SLICE_START_OFFSE

{0, 0, 0}

Increasing order: Any Value >=0 .

streamFormat

IH264_STREAM_FORM

AT_DEFAULT

IH264_BYTE_STREAM

IH264_NALU_STREAM

Note:

 sliceMode == IH264_SLICEMODE_BYTES is only supported

under below conditions:

Width >= 128 pixels

inputContentType != IVIDEO_INTERLACED

entropyCodingMode != IH264_ENTROPYCODING_CABAC

interFrameInterval == 1 (No ‘B’ Frames)

streamFormat==IH264_NALU_STREAM is only supported when

outputDataMode == IVIDEO_SLICEMODE with sub frame level

communications

Table 4-13. Default and Supported Values for IH264ENC_LoopFilterParams

Field

Default Value

Supported Value

loopfilterPreset

IH264_LOOPFILTER_

DEFAULT

 IH264_LOOPFILTER_DEFAULT

 IH264_LOOPFILTER_USERDEFINED

loopfilterDisable

IDC

IH264_DISABLE_FIL

TER_DEFAULT

 IH264_DISABLE_FILTER_NONE

 IH264_DISABLE_FILTER_ALL_EDGES

 IH264_DISABLE_FILTER_SLICE_EDGES

filterOffsetA

[-12, 12] even value

filterOffsetB

[-12, 12] even value

Table 4-14. Default and Supported Values for IH264ENC_FMOCodingParams

Field

Default Value

Supported Value

fmoCodingPreset

IH264_FMOCODING_D

EFAULT

IH264_FMOCODING_NONE

API Reference

4-117

Field

Default Value

Supported Value

numSliceGroups

Ignore

sliceGroupMapType

IH264_SLICE_GRP_M

AP_DEFAULT

Ignore

sliceGroupChangeD

irectionFlag

IH264ENC_SLICEGRO

UP_CHANGE_DIRECTI

ON_DEFAULT

Ignore

sliceGroupChangeR

ate

Ignore

sliceGroupChangeC

ycle

Ignore

sliceGroupParams[

MAXNUMSLCGPS]

{0 ,0}

Ignore

Table 4-15. Default and Supported Values for IH264ENC_VUICodingParams

Field

Default Value

Supported Value

vuiCodingPreset

IH264_VUICODING_D

EFAULT

IH264_VUICODING_DEFAULT

IH264_VUICODING_USERDEFINED

aspectRatioInfoPr

esentFlag

0,non-zero

aspectRatioIdc

[0,255]: No Error Check, user is responsible to

provide correct value

videoSignalTypePr

esentFlag

0,non-zero

videoFormat

IH264ENC_VIDEOFOR

MAT_NTSC

 IH264ENC_VIDEOFORMAT_COMPONENT

 IH264ENC_VIDEOFORMAT_PAL

 IH264ENC_VIDEOFORMAT_NTSC

 IH264ENC_VIDEOFORMAT_SECAM

 IH264ENC_VIDEOFORMAT_MAC

 IH264ENC_VIDEOFORMAT_UNSPECIFIED

videoFullRangeFla

0,non-zero

timingInfoPresent

Flag

0,non-zero

hrdParamsPresentF

lag

0,non-zero

numUnitsInTicks

1000

[1,targetFrameRate]

API Reference

4-118

Table 4-16. Default and Supported Values for IH264ENC_StereoInfoParams

Field

Default Value

Supported Value

stereoInfoPreset

IH264_STEREOINFO_

DISABLE

 IH264_STEREOINFO_DISABLE

 IH264_STEREOINFO_ENABLE_DEFAULT

 IH264_STEREOINFO_ENABLE_USERDEFIN

topFieldIsLeftViewFlag

0,non-zero

viewSelfContainedFlag

0,non-zero

Table 4-17. Default and Supported Values for IH264ENC_FramePackingSEIParams

Field

Default Value

Supported Value

framePackingPreset

IH264_FRAMEPACK_S

EI_DISABLE

 IH264_FRAMEPACK_SEI_DISABLE

 IH264_FRAMEPACK_SEI_ENABLE_DEFAUL

 IH264_FRAMEPACK_SEI_USERDEFINED

framePackingType

IH264_FRAMEPACK_T

YPE_DEFAULT

 IH264_FRAMEPACK_CHECKERBOARD

 IH264_FRAMEPACK_COLUMN_INTERLEAVI

 IH264_FRAMEPACK_ROW_INTERLEAVING

 IH264_FRAMEPACK_SIDE_BY_SIDE

 IH264_FRAMEPACK_TOP_BOTTOM

frame0PositionX

[0,15]

frame0PositionY

[0,15]

Frame1PositionX

[0,15]

Frame1PositionY

[0,15]

reservedByte

[0,255]

Table 4-18. Default and Supported Values for IH264ENC_SVCCodingParams

Field

Default Value

Supported Value

svcExtensionFlag

IH264_SVC_EXTENSI

ON_FLAG_DISABLE

 IH264_SVC_EXTENSION_FLAG_DISABLE

 IH264_SVC_EXTENSION_FLAG_ENABLE

 IH264_SVC_EXTENSION_FLAG_ENABLE_W

ITH_EC_FLEXIBILITY

dependencyID

[0,255]

qualityID

[0,255]

API Reference

4-119

Field

Default Value

Supported Value

enhancementProfil

eID

[0,255]

layerIndex

[0,255]

refLayerDQId

[0,255]

Table 4-19. Default and Supported Values for IH264ENC_Params

Field

Default Value

Supported Value

videnc2Params

See Table 4-6. Default and Supported Values for IVIDENC2_Params

rateControlParams

See Table 4-8. Default and Supported Values for

IH264ENC_RateControlParams

interCodingParams

See Table 4-9. Default and Supported Values for

IH264ENC_InterCodingParams

intraCodingParams

See Table 4 9. Default and Supported Values for

IH264ENC_IntraCodingParams

nalUnitControlPar

ams

See Table 4-11. Default and Supported Values for

IH264ENC_NALUControlParams

sliceCodingParams

See Table 4-12. Default and Supported Values for

IH264ENC_SliceCodingParams

loopFilterParams

See Table 4-13. Default and Supported Values for IH264ENC_LoopFilterParams

fmoCodingParams

See Table 4-14. Default and Supported Values for

IH264ENC_FMOCodingParams

vuiCodingParams

See

Table 4-15. Default and Supported Values for IH264ENC_VUICodingParams

stereoInfoParams

See

Table 4-16. Default and Supported Values for IH264ENC_StereoInfoParams

framePackingSEIPa

rams

See Table 4-17. Default and Supported Values for

IH264ENC_FramePackingSEIParams

svcCodingParams

See Table 4 17. Default and Supported Values for

IH264ENC_SVCCodingParams

interlaceCodingTy

IH264_INTERLACE_F

IELDONLY_ARF

 IH264_INTERLACE_FIELDONLY

 IH264_INTERLACE_FIELDONLY_MRF

 IH264_INTERLACE_FIELDONLY_ARF

 IH264_INTERLACE_DEFAULT

 IH264_INTERLACE_FIELDONLY_SPF

bottomFieldIntra

0, non-zero

API Reference

4-120

Field

Default Value

Supported Value

gopStructure

IH264ENC_GOPSTRUC

TURE_NONUNIFORM

 IH264ENC_GOPSTRUCTURE_NONUNIFORM

 IH264ENC_GOPSTRUCTURE_DEFAULT

 IH264ENC_GOPSTRUCTURE_UNIFORM

entropyCodingMode

IH264_ENTROPYCODI

NG_CAVLC(if

Profile ==

BASELINE)

IH264_ENTROPYCODI

NG_CABAC(if

Profile !=

BASELINE)

 IH264_ENTROPYCODING_CABAC

 IH264_ENTROPYCODING_CAVLC

transformBlockSiz

IH264_TRANSFORM_A

DAPTIVE (if

Profile == HIGH)

IH264_TRANSFORM_4

x4 (if Profile !=

HIGH)

 IH264_TRANSFORM_4x4

 IH264_TRANSFORM_8x8

 IH264_TRANSFORM_ADAPTIVE

log2MaxFNumMinus4

[0,12]

picOrderCountType

IH264_POC_TYPE_0

 IH264_POC_TYPE_0

 IH264_POC_TYPE_1

 IH264_POC_TYPE_2

enableWatermark

[0, non-zero]

IDRFrameInterval

Any value

pConstantMemory

NULL

NULL, Valid Address pointing to constants in DDR

maxIntraFrameInte

rval

Any value >= 0

debugTraceLevel

Zero, non-zero (all non-zero values are considered

as same level)

lastNFramesToLog

Any Value

enableAnalyticinf

Zero, non-zero

enableGMVSei

Zero, non-zero

constraintSetFlag

Zero, non-zero

enableRCDO

Zero, non-zero

enableLongTermRef

Frame

IH264ENC_LTRP_NON

 IH264ENC_LTRP_NONE

 IH264ENC_LTRP_REFERTO_PERIODICLTR

API Reference

4-121

Field

Default Value

Supported Value

 IH264ENC_LTRP_REFERTOP_PROACTIVE

 IH264ENC_LTRP_REFERTOP_REACTIVE

enableLongTermRefFrame should be

IH264ENC_LTRP_NONE, when

interFrameInterval > 1

LTRPPeriod

Zero, nonZero positive value

When numTemporalLayer > 1, it should

be multiple of Sub-Gop length

numTemporalLayer

 IH264_TEMPORAL

_LAYERS_1

 [IH264_TEMPORAL_LAYERS_1 ,

IH264_TEMPORAL_LAYERS_4]

referencePicMarki

 IH264_LONG_TER

M_PICTURE

 IH264_SHORT_TERM_PICTURE

 IH264_LONG_TERM_PICTURE

reservedParams[3]

0,0,0,0

Ignore

Note:

1) When numTemporalLayer > 1, then enableLongTermRefFrame

should not be IH264ENC_LTRP_REFERTOP_PROACTIVE.

2) When numTemporalLayer > 1, referencePicMarking should be

always IH264_LONG_TERM_PICTURE.

3) When numTemporalLayer > 1, picOrderCountType should not be

IH264_POC_TYPE_1.

API Reference

4-122

Table 4-20. Default and Supported Values for IH264ENC_DynamicParams

Field

Default Value

Supported Value

videnc2DynamicParam

See Table 4-7. Default and Supported Values for IVIDENC2_DynamicParams

rateControlParams

See Table 4-8. Default and Supported Values for

IH264ENC_RateControlParams

interCodingParams

See Table 4-9. Default and Supported Values for

IH264ENC_InterCodingParams

intraCodingParams

See Table 4-9. Default and Supported Values for

IH264ENC_InterCodingParams

sliceCodingParams

See Table 4-12. Default and Supported Values for

IH264ENC_SliceCodingParams

sliceGroupChangeCyc

Ignore

searchCenter

{0x7FFF,0x7FFF}

(-64,64), 0x7FFF --> ignore user provided gMV

and use internal

enableStaticMBCount

Zero, non-zero

enableROI

Zero, non-zero

reservedDynParams[3

]

Ignore

API Reference

4-123

4.4 Interface Functions

This section describes the Application Programming Interfaces (APIs) used in the H.264

Encoder. The APIs are logically grouped into the following categories:

 Creation – algNumAlloc(), algAlloc()

 Initialization – algInit()

 Control – control()

 Data processing – algActivate(), process(), processMulti(),

algDeactivate()

 Termination – algFree()

You must call these APIs in the following sequence:

1) algNumAlloc()

2) algAlloc()

3) algInit()

4) algActivate()

5) process()/processMulti()

6) algDeactivate()

7) algFree()

control() can be called any time after calling the algInit() API.

algNumAlloc(), algAlloc(), algInit(), algActivate(), algDeactivate(), and

algFree() are standard XDAIS APIs. This document includes only a brief description for the

standard XDAIS APIs. For more details, see TMS320 DSP Algorithm Standard API

Reference (literature number SPRU360).

API Reference

4-124

4.4.1 Creation APIs

Creation APIs are used to create an instance of the component. The term creation could

mean allocating system resources, typically memory.

║ Name

algNumAlloc() – determine the number of buffers that an algorithm requires

║ Synopsis

XDAS_Int32 algNumAlloc(Void);

║ Arguments

Void

║ Return Value

XDAS_Int32; /* number of buffers required */

║ Description

algNumAlloc() returns the number of buffers that the algAlloc() method requires. This

operation allows you to allocate sufficient space to call the algAlloc() method.

algNumAlloc() may be called at any time and can be called repeatedly without any side

effects. It always returns the same result. The algNumAlloc() API is optional.

For more details, see TMS320 DSP Algorithm Standard API Reference (literature number

SPRU360).

║ See Also

algAlloc()

API Reference

4-125

║ Name

algAlloc() – determine the attributes of all buffers that an algorithm requires

║ Synopsis

XDAS_Int32 algAlloc(const IALG_Params *params, IALG_Fxns

**parentFxns, IALG_MemRec memTab[]);

║ Arguments

IALG_Params *params; /* algorithm specific attributes */

IALG_Fxns **parentFxns;/* output parent algorithm

functions */

IALG_MemRec memTab[]; /* output array of memory records */

║ Return Value

XDAS_Int32 /* number of buffers required */

║ Description

algAlloc() returns a table of memory records that describe the size, alignment, type, and

memory space of all buffers required by an algorithm. If successful, this function returns a

positive non-zero value indicating the number of records initialized.

The first argument to algAlloc() is a pointer to a structure that defines the creation

parameters. This pointer may be NULL; however, in this case, algAlloc() must assume

default creation parameters and must not fail.

The second argument to algAlloc() is an output parameter. algAlloc() may return a

pointer to its parent’s IALG functions. If an algorithm does not require a parent object to be

created, this pointer must be set to NULL.

The third argument is a pointer to a memory space of size

nbufs * sizeof(IALG_MemRec) where, nbufs is the number of buffers returned by

algNumAlloc() and IALG_MemRec is the buffer-descriptor structure defined in ialg.h.

After calling this function, memTab[] is filled up with the memory requirements of an

algorithm.

For more details, see TMS320 DSP Algorithm Standard API Reference (literature number

SPRU360).

║ See Also

algNumAlloc(), algFree()

API Reference

4-126

4.4.2 Initialization API

Initialization API is used to initialize an instance of the algorithm. The initialization parameters

are defined in the IVIDENC2_Params structure (see section 4.2 for details).

║ Name

algInit() – initialize an algorithm instance

║ Synopsis

XDAS_Int32 algInit(IALG_Handle handle, IALG_MemRec

memTab[], IALG_Handle parent, IALG_Params *params);

║ Arguments

IALG_Handle handle; /* algorithm instance handle*/

IALG_memRec memTab[]; /* array of allocated buffers */

IALG_Handle parent; /* handle to the parent instance */

IALG_Params *params; /* algorithm initialization

parameters */

║ Return Value

IALG_EOK; /* status indicating success */

IALG_EFAIL; /* status indicating failure */

║ Description

algInit() performs all initialization necessary to complete the run time creation of an

algorithm instance object. After a successful return from algInit(), the instance object is ready

to be used to process data.

The first argument to algInit() is a handle to an algorithm instance. This value is initialized

to the base field of memTab[0].

The second argument is a table of memory records that describe the base address, size,

alignment, type, and memory space of all buffers allocated for an algorithm instance. The

number of initialized records is identical to the number returned by a prior call to

algAlloc().

The third argument is a handle to the parent instance object. If there is no parent object, this

parameter must be set to NULL.

The last argument is a pointer to a structure that defines the algorithm initialization

parameters.

For more details, see TMS320 DSP Algorithm Standard API Reference (literature number

SPRU360).

Since there is no mechanism to return extended error code for unsupported parameters, this

version of encoder returns IALG_EOK even if some parameter unsupported is set. But

subsequence control/process call it returns the detailed error code

║ See Also

algAlloc(), algMoved()

API Reference

4-127

4.4.3 Control API

Control API is used for controlling the functioning of the algorithm instance during run-time.

This is done by changing the status of the controllable parameters of the algorithm during

run-time. These controllable parameters are defined in the Status data structure (see

section 4.2 for details).

║ Name

control() – change run time parameters and query the status

║ Synopsis

XDAS_Int32 (*control) (IVIDENC2_Handle handle,

IVIDENC2_Cmd id, IVIDENC2_DynamicParams *params,

IVIDENC2_Status *status);

║ Arguments

IVIDENC2_Handle handle; /* algorithm instance handle */

IVIDENC2_Cmd id; /* algorithm specific control commands*/

IVIDENC2_DynamicParams *params /* algorithm run time parameters */

IVIDENC2_Status *status /* algorithm instance status parameters */

║ Return Value

IALG_EOK; /* status indicating success */

IALG_EFAIL; /* status indicating failure */

XDM_EUNSUPPORTED; /* status indicating parameters not

supported*/

║ Description

This function changes the run time parameters of an algorithm instance and queries the

algorithm’s status. control() must only be called after a successful call to algInit()

and must never be called after a call to algFree().

The first argument to control() is a handle to an algorithm instance.

The second argument is an algorithm specific control command. See XDM_CmdId

enumeration for details.

The third and fourth arguments are pointers to the IVIDENC2_DynamicParams and

IVIDENC2_Status data structures respectively.

Note:

If you are using extended data structures, the third and fourth

arguments must be pointers to the extended DynamicParams and

Status data structures respectively. Also, ensure that the size

field is set to the size of the extended data structure. Depending

on the value set for the size field, the algorithm uses either basic

or extended parameters.

API Reference

4-128

║ Preconditions

The following conditions must be true prior to calling this function; otherwise, its operation is

undefined.

 control() can only be called after a successful return from algInit()

and algActivate().

 If algorithm uses DMA resources, control() can only be called after a

successful return from DMAN3_init().

 handle must be a valid handle for the algorithm’s instance object.

 params must not be NULL and must point to a valid

IVIDENC2_DynamicParams structure.

 status must not be NULL and must point to a valid IVIDENC2_Status

structure.

 If a buffer is provided in the status->data field, it must be physically

contiguous and owned by the calling application.

║ Postconditions

The following conditions are true immediately after returning from this function.

 If the control operation is successful, the return value from this operation is

equal to IALG_EOK; otherwise it is equal to either IALG_EFAIL or an

algorithm specific return value. If status or handle is NULL then codec

returns IALG_EFAIL

 If the control command is not recognized or some parameters to act upon

are not supported, the return value from this operation is not equal to

XDM_EUNSUPPORTED.

 The algorithm should not modify the contents of params. That is, the data

pointed to by this parameter must be treated as read-only.

 If a buffer was provided in the status->data field, it is owned by the calling

application.

║ Example

See test application file, TestAppEncoder.c available in the \Client\Test\Src sub-directory.

║ See Also

algInit(), algActivate(), process()

API Reference

4-129

4.4.4 Data Processing API

Data processing API is used for processing the input data.

║ Name

algActivate() – initialize scratch memory buffers prior to processing.

║ Synopsis

void algActivate(IALG_Handle handle);

║ Arguments

IALG_Handle handle; /* algorithm instance handle */

║ Return Value

Void

║ Description

algActivate() initializes any of the instance’s scratch buffers using the persistent memory

that is part of the algorithm’s instance object.

The first (and only) argument to algActivate() is an algorithm instance handle. This

handle is used by the algorithm to identify various buffers that must be initialized prior to

calling any of the algorithm’s processing methods.

For more details, see TMS320 DSP Algorithm Standard API Reference. (Literature number

SPRU360).

║ See Also

algDeactivate()

API Reference

4-130

║ Name

process() – basic encoding/decoding call

║ Synopsis

XDAS_Int32 (*process)(IVIDENC2_Handle handle,

IVIDEO2_BufDesc *inBufs, XDM2_BufDesc *outBufs,

IVIDENC2_InArgs *inargs, IVIDENC2_OutArgs *outargs);

║ Arguments

IVIDENC2_Handle handle; /* algorithm instance handle */

IVIDEO2_BufDesc *inBufs; /* algorithm input buffer

descriptor */

XDM2_BufDesc *outBufs; /* algorithm output buffer

descriptor */

IVIDENC2_InArgs *inargs /* algorithm runtime input

arguments */

IVIDENC2_OutArgs *outargs /* algorithm runtime output

arguments */

║ Return Value

IALG_EOK; /* status indicating success */

IALG_EFAIL; /* status indicating failure */

║ Description

This function does the basic encoding/decoding. The first argument to process() is a

handle to an algorithm instance.

The second and third arguments are pointers to the input and output buffer descriptor data

structures respectively (see IVIDEO2_BufDesc and XDM_BufDesc data structure for details).

The fourth argument is a pointer to the IVIDENC2_InArgs data structure that defines the run

time input arguments for an algorithm instance object.

The last argument is a pointer to the IVIDENC2_OutArgs data structure that defines the run

time output arguments for an algorithm instance object.

Note:

If you are using extended data structures, the fourth and fifth

arguments must be pointers to the extended InArgs and

OutArgs data structures respectively. Also, ensure that the size

field is set to the size of the extended data structure. Depending

on the value set for the size field, the algorithm uses either basic

or extended parameters.

║ Preconditions

The following conditions must be true prior to calling this function; otherwise, its operation is

undefined.

API Reference

4-131

 process() can only be called after a successful return from algInit()

and algActivate().

 If algorithm uses DMA resources, process() can only be called after a

successful return from DMAN3_init().

 handle must be a valid handle for the algorithm’s instance object.

 Buffer descriptor for input and output buffers must be valid.

 Input buffers must have valid input data.

 inBufs->numBufs indicates the total number of input

 Buffers supplied for input frame, and conditionally, the encoders MB data

buffer.

 inArgs must not be NULL and must point to a valid IVIDENC2_InArgs

structure.

 outArgs must not be NULL and must point to a valid IVIDENC2_OutArgs

structure.

 inBufs must not be NULL and must point to a valid IVIDEO1_BufDescIn

structure.

 inBufs->bufDesc[0].bufs must not be NULL, and must point to a valid

buffer of data that is at least inBufs->bufDesc[0].bufSize bytes in

length.

 outBufs must not be NULL and must point to a valid XDM_BufDesc

structure.

 outBufs->buf[0] must not be NULL and must point to a valid buffer of

data that is at least outBufs->bufSizes[0] bytes in length.

 The buffers in inBuf and outBuf are physically contiguous and owned by

the calling application.

║ Postconditions

The following conditions are true immediately after returning from this function.

 If the process operation is successful, the return value from this operation is

equal to IALG_EOK; otherwise it is equal to either IALG_EFAIL or an

algorithm specific return value.

 After successful return from process() function, algDeactivate() can be

called.

 The algorithm must not modify the contents of inArgs.

 The algorithm must not modify the contents of inBufs, with the exception of

inBufs.bufDesc[].accessMask. That is, the data and buffers pointed to

by these parameters must be treated as read-only.

 The algorithm must appropriately set/clear the

IVIDEO2_BufDescIn::bufDesc[].accessMask field in inBufs to

indicate the mode in which each of the buffers in inBufs were read. For

example, if the algorithm only read from inBufs.bufDesc[0].buf using

API Reference

4-132

the algorithm processor, it could utilize #XDM_SETACCESSMODE_READ to

update the appropriate accessMask fields. The application may utilize these

returned values to manage cache.

 The buffers in inBufs are owned by the calling application.

║ Example

See test application file, TestAppEncoder.c available in the \Client\Test\Src sub-directory.

║ See Also

algInit(), algDeactivate(), control()

Note:

 A video encoder or decoder cannot be preempted by any other

video encoder or decoder instance. That is, you cannot perform

task switching while encode/decode of a particular frame is in

progress. Pre-emption can happen only at frame boundaries

and after algDeactivate() is called.

 The input data is an uncompressed video frame in one of the

format defined by inputChromaFormat of IVIDENC2_Params

structure. The encoder outputs H.264 compressed bit-stream in

the little-endian format.

 outBufs->bufs[0] may contain the encoded data buffer. See

IVIDENC2_OutArgs.encodedBufs for more details.

 outBufs->bufs[1], outBufs->bufs[2], and outBufs-

>bufs[3] are used when providing reconstruction buffers.

API Reference

4-133

║ Name

processMulti() – N channel video encoding call

║ Synopsis

XDAS_Int32 (*processMulti) (IH264ENC_ProcessParamsList *processList);

║ Arguments

IH264ENC_ProcessParamsList *processList; /* Container for N channels.

Each channel structure(IH264ENC_ProcessParams) contains handle, *inBufs,

*outBufs, *inArgs, *outArgs */

║ Return Value

IALG_EOK; /* status indicating success */

IALG_EFAIL; /* status indicating failure */

║ Description

This function does the basic H264 video encoding for N channels. The argument to

processMulti() is a container for N channels. The structure IH264ENC_ProcessParams

contains five parameters. The first parameter is a handle to an algorithm instance.

The second and third parameters are pointers to the input and output buffer descriptor data

structures respectively (see IVIDEO2_BufDesc and XDM_BufDesc data structure for

details).

The fourth parameter is a pointer to the IVIDENC2_InArgs data structure that defines the

run time input arguments for an algorithm instance object.

The last parameter is a pointer to the IVIDENC2_OutArgs data structure that defines the run

time output arguments for an algorithm instance object.

Note:

If you are using extended data structures, the fourth and fifth

arguments must be pointers to the extended InArgs and

OutArgs data structures respectively. Also, ensure that the size

field is set to the size of the extended data structure. Depending

on the value set for the size field, the algorithm uses either basic

or extended parameters.

║ Preconditions

The following conditions must be true prior to calling this function; otherwise, its operation is

undefined.

 processMulti() can only be called after a successful return from

algInit() and algActivate().

 If algorithm uses DMA resources, processMulti() can only be called after

a successful return from DMAN3_init().

 handle must be a valid handle for the algorithm’s instance object.

 Buffer descriptor for input and output buffers must be valid.

API Reference

4-134

 Input buffers must have valid input data.

 inBufs->numBufs indicates the total number of input

 Buffers supplied for input frame, and conditionally, the encoders MB data

buffer.

 inArgs must not be NULL and must point to a valid IVIDENC2_InArgs

structure.

 outArgs must not be NULL and must point to a valid IVIDENC2_OutArgs

structure.

 inBufs must not be NULL and must point to a valid IVIDEO1_BufDescIn

structure.

 inBufs->bufDesc[0].bufs must not be NULL, and must point to a valid

buffer of data that is at least inBufs->bufDesc[0].bufSize bytes in

length.

 outBufs must not be NULL and must point to a valid XDM_BufDesc

structure.

 outBufs->buf[0] must not be NULL and must point to a valid buffer of

data that is at least outBufs->bufSizes[0] bytes in length.

 The buffers in inBuf and outBuf are physically contiguous and owned by

the calling application.

║ Postconditions

The following conditions are true immediately after returning from this function.

 If the process operation is successful, the return value from this operation is

equal to IALG_EOK; otherwise it is equal to either IALG_EFAIL or an

algorithm specific return value.

 After successful return from processMulti() function, algDeactivate()

can be called.

 The algorithm must not modify the contents of inArgs.

 The algorithm must not modify the contents of inBufs, with the exception of

inBufs.bufDesc[].accessMask. That is, the data and buffers pointed to

by these parameters must be treated as read-only.

 The algorithm must appropriately set/clear the

IVIDEO2_BufDescIn::bufDesc[].accessMask field in inBufs to

indicate the mode in which each of the buffers in inBufs were read. For

example, if the algorithm only read from inBufs.bufDesc[0].buf using

the algorithm processor, it could utilize #XDM_SETACCESSMODE_READ to

update the appropriate accessMask fields. The application may utilize these

returned values to manage cache.

 The buffers in inBufs are owned by the calling application.

║ Example

See test application file, TestAppEncoder.c available in the \Client\Test\Src sub-directory.

║ See Also

API Reference

4-135

algInit(), algDeactivate(), control()

Note:

 A video encoder or decoder cannot be preempted by any other

video encoder or decoder instance. That is, you cannot perform

task switching while encode/decode of a particular frame is in

progress. Pre-emption can happen only at frame boundaries

and after algDeactivate() is called.

 The input data is an uncompressed video frame in one of the

format defined by inputChromaFormat of IVIDENC2_Params

structure. The encoder outputs H.264 compressed bit-stream in

the little-endian format.

 outBufs->bufs[0] may contain the encoded data buffer. See

IVIDENC2_OutArgs.encodedBufs for more details.

 outBufs->bufs[1], outBufs->bufs[2], and outBufs-

>bufs[3] are used when providing reconstruction buffers.

API Reference

4-136

║ Name

algDeactivate() – save all persistent data to non-scratch memory

║ Synopsis

Void algDeactivate(IALG_Handle handle);

║ Arguments

IALG_Handle handle; /* algorithm instance handle */

║ Return Value

Void

║ Description

algDeactivate() saves any persistent information to non-scratch buffers using the

persistent memory that is part of the algorithm’s instance object.

The first (and only) argument to algDeactivate() is an algorithm instance handle. This

handle is used by the algorithm to identify various buffers that must be saved prior to next

cycle of algActivate() and processing.

For more details, see TMS320 DSP Algorithm Standard API Reference (literature number

SPRU360).

║ See Also

algActivate()

API Reference

4-137

4.4.5 Termination API

Termination API is used to terminate the algorithm instance and free up the memory space

that it uses.

║ Name

algFree() – determine the addresses of all memory buffers used by the algorithm

║ Synopsis

XDAS_Int32 algFree(IALG_Handle handle, IALG_MemRec

memTab[]);

║ Arguments

IALG_Handle handle; /* handle to the algorithm instance */

IALG_MemRec memTab[]; /* output array of memory records */

║ Return Value

XDAS_Int32; /* Number of buffers used by the algorithm */

║ Description

algFree() determines the addresses of all memory buffers used by the algorithm. The

primary aim of doing so is to free up these memory regions after closing an instance of the

algorithm.

The first argument to algFree() is a handle to the algorithm instance.

The second argument is a table of memory records that describe the base address, size,

alignment, type, and memory space of all buffers previously allocated for the algorithm

instance.

For more details, see TMS320 DSP Algorithm Standard API Reference (literature number

SPRU360).

║ See Also

algAlloc()

4-138

This page is intentionally left blank

5-1

Chapter 5

Frequently Asked Questions

This chapter provides answers to few frequently asked questions related to using this

encoder.

5.1 Release Package

Question

Answer

Can this codec release be

used on any HDVICP2 and

Media Controller based

platform?

Yes, you can use it on any HDVICP2 and Media Controller based platforms

(eg DM816x, DM814x). The Test application shipped along with this release

is meant for a particular platform. Before using it to different platform, you

need to ensure that the addresses provided in linker command file are taken

care. In addition, the HDVICP2 related addresses through HDVICP IRES

interface should be provided correctly.

5.2 Code Build and Execution

Question

Answer

Build error saying that

code/data memory section

is not sufficient for

placement

Make sure that project settings are not changed from the released package.

Change in debug options for compilation may make code/data memory size

insufficient for placement.

Application returns an error

saying “Cannot open input

file “….YUV” while running

the host test app

Make sure that input YUV path is given correctly. If the application is

accessing YUVs from network, ensure that the network connectivity is stable.

5.3 Issues with Tools/FC Version

Question

Answer

What tools are required to run

the standalone codec?

To run the codec on standalone setup, you need Framework

components, Code Composer Studio, ARM compiler tools (CG tools).

If you are running on the simulator, then the correct version of the

Platform specific CSP is needed (See section 2.2 for more details.)

Which simulator version should

I use for this release?

Code Composer Studio (CCSv4) version 4.2.0.09000 has to be installed.

Netra simulator CSP version 0.7.1 (or newer) has to be installed after

installing Code Composer Studio,

This release can be obtained by software updates on CCSV4. Please

make sure that following site is listed as part of “Update sites to visit”

http://software-

dl.ti.com/dsps/dsps_public_sw/sdo_ccstudio/CCSv4/Updates/NET

RA/site.xml

Frequently Asked Questions

5-2

Question

Answer

What CG tools version is used

for this release?

CG tools version 4.5.1 is used for this release.

What if the application is using

different CG tools version?

The memory layout of the interface data structures does not change with

different version of compilers(if bit-fields are not used). In addition, it does

not change the mechanism of generating signature for functions. This

version can be used even if the application is with different CG tools

because no bit-fields are used in interface.

Is this encoder integrated with

codec engine, if yes with which

version?

Yes, this encoder is integrated with Codec Engine version 3.20.00.16

5.4 Algorithm Related

Question

Answer

What XDM interface does

codec support?

Codec supports XDM IVIDENC2 interface

What are the profiles supported

in this version of encoder?

This version of encoder supports baseline, main and high profiles. FMO

feature is not supported for baseline profile.

What is the maximum level

supported by this encoder?

The encoder supports the level up to 5.1

What is the maximum bit rate

supported?

Maximum bit rate depends upon the level setting. This version supports

the maximum bit rate of 50 mbps (Base Line and Main profile Level 4.2)

and 62.5 mbps (High Profile Level 4.2) for 30 fps case.

To achieve the real time performance with CABAC bit rate should be less

than 25 Mbps for 30fps case

Can I encode with bit rate or any

other parameter (example

resolution) more than specified

in level 4.2?

Yes. Video encoder will return a non-fatal level incompliance error, but still

it continues encoding. It is not guaranteed to achieve real time

performance for bit rates higher than specified.

Can I encode with level higher

than 4.2

Yes. Functional point of view encoder support Level 5.1 (with contarsints

on resolution not be higher than 4352x4096). But performance is not

guaranteed to be real time

Can I reduce DDR footprint of

encoder?

Yes. DDR foot print is majorly dependent on maxWidth and maxHeight

parameters and also dependent on whether long term reference frame is

enabled or not.

What stream formats are

supported in this version of

encoder?

This version supports byte-stream and NALU format

What are the input frame

formats supported?

Can I encode YUV 422 input

format buffer?

This version supports only YUV420 semi-planar input format only.

No. other formats than YUV420 semi-planar are not supported

Frequently Asked Questions

5-3

Question

Answer

What is granularity of the

process call?

The encoder supports only frame level encoding API. However, it also

supports data sync APIs for output bit stream, which is a call back to the

application for data synchronization.

What are the resolutions

supported?

The encoder supports all resolutions until up to 4352x4096. The minimum

resolution supported is 96x80. Width and height should be multiple of 2.

Encoder asks few buffers in

TILED memory, can I override

the encoder’s request and

provided buffers in different

space?

Yes, you can over ride the encoder’s request but with below constraints

 TILED PAGE can be overridden by RAW

 TILED8, TILED16 can be overridden by TILED PAGE, RAW

 TILED16 can be overridden by TILED8, RAW, TILED PAGE

Encoder requires large amount

of memory to compress bit-

streams. The encoder does not

require the same amount

memory after compression. Can

this memory usage be reduced?

Yes, you need to set ignoreOutBufSizeFlag = XDAS_TRUE &&

getBufferFxn = Valid Function Pointer

If the application is not capable of providing memory at run time with

codec’s request by getBufferFxn then it can point to a dummy function

which returns -1.

Can I change bit-rate, frame

rate, resolution at run time

Yes

Will change in above parameters

result in a IDR insertion

Change in resolution will result in IDR insertion. Change in bit rate may

insert IDR if HRD parameters are coded as part of bit-stream. Similarly if

timing info related parameters are coded in bit-stream then it can cause

insertion of IDR by doing change in frame rate

Does the encoder support B

frame encoding?

In what order does encoder

expect the frames, encode order

or capture order?

How the delay is controlled?

Yes, encoder supports B frame encoding. It accepts the frames in capture

order and internally processes them in encoder order.

Encoder has a mechanism to lock and free the input buffer, based on this

it has a initial delay to produce the bit stream, which is equivalent to

number of B frames getting encoded. Subsequent process call should

produce the compressed bit-stream and also frees up a buffer.

How many continuous B frames

can I have? Is there a

performance/quality impact?

In case of progressive content maximum 31 continuous B frames can be

produced. With interlaced content maximum 32 B fields can be produced

Quality is not tuned for more than two B frames so for motion sequences it

is not advised to have more than two B frames

Performance is impacted slightly; this is because if B frames are more than

two then some information related to buffers are stored in external memory

compared to internal memory because of limited DTCM. Hence, it affects

the performance.

Does the encoder support meta

data input/output?

Yes, this version of the encoder supports reading in meta data for user

data unregistered SEI and user defined scaling matrix. For more details on

how this data is written See A and C.

Does this version of H264

Encoder expose motion vectors

for a frame to the application?

Yes

Can encoder take the motion

vectors given externally for

encoding or say in a transcode

scenario?

Frequently Asked Questions

5-4

Question

Answer

Can codec do frame rate

conversion?

No, refFrameRate and targetFrameRate needs to be same.

Does this version of encoder

support interlaced coding?

Yes, this version of H.264 Encoder supports interlaced coding with field

only coding. MBAFF and PICAFF are not supported. However, controls to

decide parity of reference field are given to user, like SPF, MRF, ARF.

In case of interlaced, will single

encode (Process call),

encode both the fields?

Encoder allows both fields processing in single process call as well

process call per field.

Does Algorithm support sub-

frame level communication

mechanism for low-delay

applications?

Yes. It has the mechanism for sub-frame level communication for both

input and output buffers.

Does this version of encoder

support encoding multiple slices

in a frame?

Yes, slices can be generated bases upon number of macro blocks per

slice, number of bytes per slice and also based upon the row start offset in

a frame

Is there a limit on number of

slices supported per frame by

encoder?

Yes, encoder can generate one slice per 6 macroblocks not below that

when configured in sliceMode = IH264_SLICEMODE_MBS. When

sliceMode = IH264_SLICEMODE_MBS, it can allow minimum value of

bytes per slice as 256

Does Algorithm support H.241

based packetization (slice cap/

maxBitsPerSlice) feature ?

Yes.

For a given configuration why

performance is poorer incase of

H.241 enabled compared to

without H.241?

Incase of H.241, for every slice boundaries encoder needs to flush and

restart the pipeline to meet the strict restriction on the bytes generated for

slices. The performance becomes poorer as the number of slices

generated per frame is higher (in other words if bytes/ slice is very low).

In case of interlaced, can bottom

field come first in bit stream?

Yes. A sequence can look like this: BF, TF, BF, TF, BF, TF…. Encoder

allows accepting the information as top field is first field or not

For Interlace content how the

YUV data are expected, is it

interleaved or field separated

Both format is supported, interleaved and field separated.

Can frac-pel refinement of

motion vectors be disabled?

Yes

Can the encoder give multiple

Motion vector for a macro block?

Yes

Is there any performance

difference between 1MV and

4MV per macro block?

Yes, please refer the data sheet for the impact on performance

Frequently Asked Questions

5-5

Question

Answer

What is the behavior of Codec

on cache properties of input and

output buffer

All input and output buffer of encoder are read/written by DMA. So codec

assumes that all input data is valid in DDR memory before feeding in to

encoder. Also outout of encoder is guaranteed to be in DDR.

Hecne the parameters like InBufs :

IVIDEO2_BufDesc.planeDesc[idx].usageMode and OuBufs

: XDM2_BufDesc.Descs[0].usageMode are don’t care

However for the trace and debug related buffers produced by encoder it is

not true. There are some buffers for which data can be in cache memory

and cache write back from application side will be needed, refer Appendix

E for more details

What is rateControlPreset

and

rateControlParamsPreset

? What is the difference between

these two?

rateControlPreset control the rate control algorithm

(IH264ENC_RateControlParams ::rcAlgo) but

rateControlParamsPreset controls the other associated parameters

specified in IH264ENC_RateControlParams structure. When

rateControlPreset is user defined then only

IH264ENC_RateControlParams ::rcAlgo is resepected otherwise

it is controlled by rateControlPreset. But even if

rateControlPreset is not user defined other parameters of

IH264ENC_RateControlParams structure are possible to be user

controllable by setting rateControlParamsPreset as user defined

Does the encoder support multi-

channel operation?

Yes.

What is granularity of the

process call?

The encoder supports only frame level encoding API. However, it supports

data sync APIs for sub frame level data exchange between Application

and Encoder, both at input and output side. Refer Appendix for more

information.

Does a Luma buffer and

corresponding Chroma buffer

needs to be contiguous in

memory?

Can the encoder generate

headers only?

Yes, have a control call of XDM_GENERATE_HEADER before the

particular process call.

Does encoder support skipping

of frames?

Yes, encoder will encode requested frame as all macro block as skipped MB.

Please refer to user guide for further details

What is the benefit of asking a

frame as skip

It can help to balance the performance or bitrate in certain situations. The frame

being asked to be coded as skip consumes very less MHZ of the HDVICP2. It can

finish the entire frame/field processing in less than 5 MHz

Is it possible to configure the

stream format (Byte stream vs

NAL stream format) at frame

level run-time?

No, it can be configured only at create time

How to use interlaced encoding

in H.264 encoder

You need to configure contentType as IVIDEO_INTERLACED

and provide the pointers to field buffers

appropriatelty during process call

Frequently Asked Questions

5-6

Question

Answer

How to change resolution

dynamically?

You need to make a control call of encoder with XDM_SETPARAMS

command. At this time configure the inputWidth and inputHeight parameter

indicating the new resolution. Subsequnt process call we start assuming

the newly configured resolution.

How to change frame rate,

bitrate or any other dynamic

parameter dynamically?

You need to make a control call of encoder with XDM_SETPARAMS

command. At this time configure the appropriate parameter with new

value. Subsequnt process call we start assuming the newly configured

resolution.

How to force Intra frames in

H.264 encoder

You need to make a control call of encoder with XDM_SETPARAMS

command and forceFrame = IVIDEO_IDR_FRAME. Subsequnt process

call will be coded as IDR frame. The effect of this control call is only for

one frame and subsequent frame will be coded as per defined gop

structure

How to generate SPS and PPS

headers in bit-stream?

Refer Appendix B. If you want dynamically before certain frames then use

XDM_GENERATE_HEADER

How to insert user data SEI

message in H.264 bitstream

Refer Appendix A

How to insert picture timing SEI

message?

Refer Appendix B

What is the latency of the

codec?

This encoder is designed for low latency applications hence it can take

uncompressed data with a minimum unit of 1 MB row (16 lines) and can

provide compressed bit-stream out with a minumum unit of 1 slice

(compressed unit used for packets).

Now based upon what is the slice rate - one can compute the latency at

which compressed data will be available at encoder output

Example - assume each frame has 20 slices then each slice is available at

the output of the encoder at (33 ms / 20 + 0.3 ms intial overhead) time

interval =~ 2 ms

So you should be able to compute the latency for you application based

upon slice rate.

Can H.264 encoder do all Intra

frames as IDR encoding?

Yes.

Can H.264 encoder do all Intra

frames encoding?

Yes, H.264 encoder can do all intra frames encoding. One has to set

IntraFrameInterval with correct value. If you want to reduce DDR foot print

for this use case then configure create time parameter

maxIntraFrameInterval = 1

How many channels of H264

Encoder can be supported?

Given the standalone data for each resolution in data sheet, please do the

math yourself accounting for the MHz clock of HDVICP2 and DDR

Bandwidth on the SoC.

Can the encoder be run on any

OS?

Yes.

Encoder implementation is independent of Operating System.

Only necessity is that the component interacting with encoder has to be

VIDENC2interface compliant.

Frequently Asked Questions

5-7

Question

Answer

How will the application know

when to stop calling process

function after applying

XDM_FLUSH?

When application puts the encoder in flush mode by calling control with

XDM_FLUSH, encoder starts encoding locked frames if any. Application

needs to call process in a do-while loop till the encoder return

XDM_INSUFFICIENTDATA error.

How to enable Hierarchical P –

coding?

Configure the parameter numTemporalLayer to a value greater than 1.

Current version of encoder supports upto a maximum of 4 temporal layers.

How to enable Watermark

feature in Enocder?

It can be enabled by setting any non-zero value to the parameter

enableWatermark as a part of IH264ENC_Params . And the key is

passed to encoder through inputKey as a part of IH264ENC_InArgs.

How many watermark SEI

messages are inserted in the

stream for interlace case?

Only one SEI message is inserted for a pair of fields

Do we need to pass two input

keys(one key per field) in

interlaced coding?

No. Enocder accepts only one key for a pair of fields in interlaced coding.

In case of 60 process call encoding, the key fed for the second field is

considered. For more details, refer 'Appendix L'.

Does encoder generates

watermark SEI message when a

process call is with made

XDM_GENERATE_HEADER

enabled?

No. There will not be any SEI message for this scenario. And the input Key

passed to encoder in this process call is ignored.

Does encoder supports all the

features for resolution more than

2kx2k?

Yes.

Suppose the user has

configured MaxWidth/Maxheight

more than 2048x2048 and

actual encoding

inputWidth/inputHeight are less

than 2048x2048. Any

precautions or features

unsupport for this configuration?

Yes, there are few suggestions to the user in this configuration,

1. There would be performance degradtion of about 12 to 15MHz for 30

frames HD resolution.

2.If it is a prior known that the dynamic change in resolution never goes

beyond 2048x2048, then always configure maxWidthxmaxHeight to

2048x2048 for better performance.

Can the user give separate ROI

Input parameters for each field

in Interlaced Cases?

No. Both the fields will use the same ROI Input parameters. For more

details, refer 'Appendix K'.

Any performance degradation

with ROI enabled?

Yes, Around 4MHz overhead for 30 frames encoding of 1080p resolution

input with 1MV and around 10MHz for 30 frames encoding of 1080p

resolution streams with 4MV enabled. n streams with 4MV enabled.

Performance also depends on the number of Roi regions in a frame.

How can user specify the colour

for privacy masked region?

Set IH264ENC_RoiInput->roiPriority[] with the Y,Cb,Cr

combination.

The data type of this parameter is a integer(4 bytes) array.

0th byte = Cr ,

1st byte = Cb ,

2nd byte = Y and 3rd byte is ignored.

By default all these values are zero(grey colour).

Frequently Asked Questions

5-8

5.5 Trouble Shooting

Question

Answer

Encoder generates an output bit

stream which has garbage

frames?

Please check whether the input YUV given to encoder is proper or not.

Encoder expects/supports the YUV NV12 format only.

In the encoded bit stream luma

information looks proper but not

chroma information.

Please check the input YUV format fed into encoder. Encoder supports

only YUV NV12 format.

Codec misbehaves or hangs

when sliceMode =

IH264_SLICEMODE_BYTES

This is a known shortcoming in the simulator. This feature has been

verified on hardware.

In the first process call, I am

getting the error as

IH264VDEC_ERR_HDVICP2_I

MPROPER_STATE

Before HDVICP2 is given to codec, HDVICP2 has to be in standby mode.

Other wise this error will show up. So check the HDVICP2_Reset functionality

used in the Application side. For sample flow and implementation, refer Test

Application in the release package. Note that in some configurations of

simulator, reset might not be needed.

The encoder gives error during

creation, what could be the

reason?

The create call failure is due to non-availability of the memory requested

by the codec.

The XDM control call fails, what

could be the reason?

The following are few of reasons for the error:

 If create time parameter is not set properly then encoder returns

back during subsequent process/control call with detailed error code

 Encoder is called with un-supported dynamic parameter.

The process call returns error,

what are the possible reasons?

The following are few of reasons for the error:

 The input or output pointers are null

 The input or output buffer sizes are not sufficient or incorrect

 Creation/control time failure

 Run time error occurred during encoding of the frame

A-1

Appendix A

Meta Data Support

This appendix explains the meta data support by encoder. Encoder supports multiple meta

data as consumer as well as producer.

Topic Page

A.1 Control Parameter to Enable/Disable Metadata

A-2

A.2 Format of meta data

A-2

A.3 Steps to enable a meta data with Example

A-3

Meta Data Support

A-2

A.1 Control Parameter to Enable/Disable Metadata

This feature can be enabled/disabled through create time parameters

IVIDENC2_Params::metadataType[IVIDEO_MAX_NUM_METADATA_PLANES]. There can be

maximum 3(IVIDEO_MAX_NUM_METADATA_PLANES) meta data planes possible to be

supported with one instance of encoder.

Each element of metadataType[] array can possibly take following enumerated values. For

supported values with this version of encoder, please refer Table 4.5.

Enumeration

Value

IVIDEO_METADATAPLANE_NONE

-1

IVIDEO_METADATAPLANE_MBINFO

IVIDEO_METADATAPLANE_EINFO

IVIDEO_METADATAPLANE_ALPHA

IH264_SEI_USER_DATA_UNREGISTERED

256

IH264_REGION_OF_INTEREST

257

IH264_USER_DEFINED_SCALINGMATRIX

258

If user wants to pass user defined scaling matrix via meta data plane 2 then

IVIDENC2_Params::metadataType[2] should be set to

IH264_USER_DEFINED_SCALINGMATRIX.

If user don’t want to use any meta data plane then all the entries of

IVIDENC2_Params::metadataType[] should be set to IVIDEO_METADATAPLANE_NONE

A.2 Format of meta data

Format of Each meta data that is supported has to be defined by the encoder. The format for

each supported meta data is explained below:

A.2.1 SEI_USER_DATA_UNREGISTERED

For this purpose encoder allows only one meta data (not multiple units of this meta data).

The format is as shown below:

Size (32-bit)

Payload[size]

The maximum value of size can be 1023 bytes. Encoder only reads the lower 10-bits of the

size field

A.2.2 MBINFO

Format of this meta data is yet to be defined

Meta Data Support

A-3

A.2.3 ROI

Please refer Appendix K for the details related to format of this meta data

A.2.4 USER_DEFINED_SCALINGMATRIX

Please refer Appendix C for the details related to format of this meta data.

A.3 Steps to enable a meta data with Example

The way to pass meta data to encode is through inBufs to the encoder during process call.

The way to get meta data from encode is through outBufs of the encoder during process

call.

When application request the buffer information through control call with XDM_GETBUFINFO,

encoder considers IVIDENC2_Params::metadataType[] array to count the buffers

required at input/output level. For each meta data one additional buffer is required. If for

some metadata size is not known by encoder then it should return size =-1 so that

application can allocate as per its knowledge. Same way for some meta-data application

might not provide the size to codec through XDM2_SingleBufDesc.bufSize.bytes, in that

case application can set it to -1. The meta data which has size set to -1 should have first

word (32-bit) of meta data as size and properly updated.

For Example: User want to insert SEI_USER_DATA_UNREGISTERED meta data at each IDR

picture, the following steps should be followed

1) Create the encoder object with

IVIDENC2_Params::metadataType[IDX_SEI_METADATA] =

IH264_SEI_USER_DATA_UNREGISTERED

2) Also the user data un-registered SEI bit in the NAL unit mask for IDR picture should be

set IH264ENC_SET_NALU(naluPresentMaskIDRPicture,

USER_DATA_UNREGD_SEI)

3) Call Control function with XDM_GETBUFINFO. Encoder should return one additional

input buffer as required, size of the buffer will be -1 as encoder doesn't know the size

Application should have a memory allocated for this meta data and pass on to the encoder

via

As mentioned in section A.2.1 this meta-data format includes size field, so encoder will read

size from the actual meta data buffer and utilize the buffer.

ppBuffer points to this buffer in memory

Size (32-bit)

Payload[size]

Here IDX_SE_METADATA can be any value 0 to 2 (IVIDEO_MAX_NUM_METADATA_PLANES-1).

IVIDEO2_BufDesc *inBufs->numMetaPlanes = 1 ;

inBufs->metadataPlaneDesc[IDX_SEI_METADATA].buf = pBuffer ;

inBufs->metadataPlaneDesc[IDX_SEI_METADATA].bufSize.bytes = -1

Meta Data Support

A-4

This page is intentionally left blank

B-1

Appendix B

Control for Configurable NALU

This appendix explains the configurable NAL unit support by encoder. This is to help the

application to decide the position of few key NAL units at different position in video sequence

Topic Page

B.1 Position in Video Sequence

B-2

B.2 NAL Units in H.264 Video Sequence

B-2

B.3 Control masks

B-2

B.4 End of Sequence Identification

B-4

B.5 Erroneous Situations

B-4

Control for Configurable NALU

B-2

B.1 Position in Video Sequence

There are five main positions in a video sequence

Start of the Sequence

I frame

IDR frame

End of Sequence

All other positions which are not identified by above 4 positions (non Intra frame

positions)

B.2 NAL Units in H.264 Video Sequence

There are following possible NAL units in H.264 encoder.

1) IH264_NALU_TYPE_UNSPECIFIED

2) IH264_NALU_TYPE_SLICE

3) IH264_NALU_TYPE_SLICE_DP_A

4) IH264_NALU_TYPE_SLICE_DP_B

5) IH264_NALU_TYPE_SLICE_DP_C

6) IH264_NALU_TYPE_IDR_SLICE

7) IH264_NALU_TYPE_SEI

8) IH264_NALU_TYPE_SPS

9) IH264_NALU_TYPE_PPS

10) IH264_NALU_TYPE_AUD

11) IH264_NALU_TYPE_EOSEQ

12) IH264_NALU_TYPE_EOSTREAM

13) IH264_NALU_TYPE_FILLER

This version defines one more NALU which is modified version of SPS nal unit, It is

SPS having VUI

14) IH264_NALU_TYPE_SPS_WITH_VUI

B.3 Control masks

Encoder defines a control mask for each position in video sequence. Hence, it has following

masks as part of IH264ENC_NALUControlParams, which can be configured as creation

time

1) naluPresentMaskStartOfSequence

Control for Configurable NALU

B-3

2) naluPresentMaskIDRPicture

3) naluPresentMaskIntraPicture

4) naluPresentMaskNonIntraPicture

5) naluPresentMaskEndOfSequence

Each of the mask is 14-bit mask with following bit allocation

Bit-0,1,2,3,4,5 are ignored in each mask

Since IDR picture is also an Intra picture, for IDR picture the nalUnitMask used by encoder

is Oring of naluPresentMaskIDRPicture and naluPresentMaskIntraPicture.

Similarly, naluPresentMaskStartOfSequence also considers the properties of IDR picture

SEI (bit 6): This bit control the insertion of following SEI messages in video sequence. For

details of these SEI messages refer Appendix D of H.264 standard

 timing_info_sei

 buffering_period_sei: This SEI is put only at IDR frames even if it is

enabled for other positions in video sequence

 stereo_video_info_sei :This SEI is put only when stereoInfoPreset is

enabled

If you want to encode with rateControlPreset == IVIDEO_NONE then

nal_hrd_parameters_present_flag and vcl_hrd_parameters_present_flag will be

false. Hence, buffering period SEI message will not be present.

SPS (bit 7): This bit controls the insertion of SPS in the video sequence. For a start of

sequence SPS is must so encoder internally assumes this bit as 1 for

naluPresentMaskStartOfSequence

PPS (bit 8): This bit controls the insertion of PPS in the video sequence. For a start of

sequence, PPS is a must, hence, the encoder internally assumes this bit as 1 for

naluPresentMaskStartOfSequence

AUD (bit 9): This bit controls the insertion of access unit delimiter NAL unit

EOSEQ (bit 10): This bit controls the insertion of end of sequence NAL unit. This bit is

ignored for all the NAL unit masks except naluPresentMaskEndOfSequence.

USER_DATA_UNREGD_SEI

SPS_VUI

FILLER

EOSTREAM

EOSEQ

AUD

PPS

SPS

SEI

IDR_SLICE

SLICE_DP_C

SLICE_DP_B

SLICE_DP_A

SLICE

UNSPECIFIED

Control for Configurable NALU

B-4

EOSTREAM (bit 11): This bit controls the insertion of end of stream NAL unit. This bit is

ignored for all the NAL unit masks except naluPresentMaskEndOfSequence.

FILLER(bit 12): This bit informs encoder to insert filler data. It is encoder’s decision to put

filler data or not based upon the constant bit rate need

SPS_VUI(bit 13): This bit informs encoder to insert SPS data along with VUI (Video usability

Information).

USER_DATA_UNREGD_SEI (bit 14): This bit controls the insertion of user data

unregistered SEI. To insert SEI some additional information has to be provided by user, refer

Appendix A for more details

Bit-13 supersedes bit-3

B.4 End of Sequence Identification

Encoders don’t know in general that this is the end of sequence position in video. Hence

encoder except user to put it into flush mode to identify end of sequence.

When encoder is in flush mode it stops excepting input via process call and processes the

buffers which it internally have (In case of B frame there are delays in producing the output

hence encoder has some buffers unprocessed). Being in flush mode encoder knows that all

the input buffers are exhausted or not and hence can decide the end of sequence

So when there is only P frames (no B frames) and still user want encoder to use

naluPresentMaskEndOfSequence, he/she should call a control method with XDM_FLUSH

B.5 Erroneous Situations

Following are the situations, which are erroneous, for each of the situation encoder returns

IH264ENC_UNSUPPORTED_NALUNITCONTROLPARAMS error code

1) If user want to encode the SEI, it is necessary to have SPS with VUI. So if user has

configured SEI bit as 1 for some position in video sequence and there is no naluMask

prior to that position having SPS_VUI enabled then encoder returns error

For example:

naluPresentMaskStartOfSequence (bit-13 is 0, bit-6 is 1): This is

erroneous situation

naluPresentMaskStartOfSequence (bit-13 is 1, bit-6 is 0) and

naluPresentMaskNonIntraPicture (bit-13 is 0, bit-6 is 1): This is not

erroneous situation

2) If Bit-13 (SPS + VUI bit) in naluPresentMaskStartOfSequence is 0 then it should be 0 in

all the remaining mask

3) If Bit-13 (SPS + VUI bit) in naluPresentMaskStartOfSequence is 1 then it should be 1 in

all the mask which contains Bit-7 (SPS bit) as 1

4) If stereoInfoPreset is enabled (Stereo Video Coding) then inputcontenttype should be

Interlaced.

Control for User Defined Scaling Matrices

C-5

Appendix C

Control for User Defined Scaling

Matrices

This appendix explains the mechanism of supporting user defined scaling matrices.

Following operations are performed at different stages:

1) Creation time

2) Control time

3) Process level

C.1 Creation Time

The following parameters should be set during creation of encoder

1) IVIDENC2_Params::metadataType[IDX_SCALINGMTX_METADATA] =

IH264_USER_DEFINED_SCALINGMATRIX, here IDX_SCALINGMTX_METADATA can

be any value between 0 to IVIDEO_MAX_NUM_METADATA_PLANES - 1.

2) IH264ENC_Params::IH264ENC_RateControlParams:: scalingMatrixPreset

to be set as IH264_SCALINGMATRIX_USERDEFINED_SPSLEVEL or

IH264_SCALINGMATRIX_USERDEFINED_PPSLEVEL

IH264_SCALINGMATRIX_USERDEFINED_SPSLEVEL means that encoder will generate scaling

matrices in bit-stream for each SPS

IH264_SCALINGMATRIX_USERDEFINED_PPSLEVEL means that encoder will generate

scaling matrices in bit-stream for each PPS

typedef enum

{

IH264_SCALINGMATRIX_NONE = 0 , //!<

Flat Scaling matrix : part of standard (NO Scaling Matrix)

IH264_SCALINGMATRIX_NORMAL = 1 , //!<

For normal contents

IH264_SCALINGMATRIX_DEFAULT = IH264_SCALINGMATRIX_NORMAL,

//!< default = IH264_SCALINGMATRIX_NORMAL (if profile == HIGH) &

IH264_SCALINGMATRIX_NONE (if profile!=HIGH).

IH264_SCALINGMATRIX_NOISY = 2 , //!<

For noisy contents

IH264_SCALINGMATRIX_STD_DEFAULT = 3 , //!<

Default Scaling Matrix provided by H.264 standard

IH264_SCALINGMATRIX_USERDEFINED_SPSLEVEL = 4 , //!< User

defined SM at SPS level

Control for User Defined Scaling Matrices

C-6

IH264_SCALINGMATRIX_USERDEFINED_PPSLEVEL = 5 , //! User defined

SM at PPS level

}

C.2 Control Time

Call Control function with XDM_GETBUFINFO. Encoder should return one additional input

buffer as required. Size of the buffer will be 896 bytes

C.3 Process level

Application should have memory allocated for this meta data and pass on to the encoder

via IVIDEO2_BufDesc *inBufs->numMetaPlanes = 1

inBufs->metadataPlaneDesc[IDX_SCALINGMTX_METADATA].buf =

pBuffer ;

inBufs->metadataPlaneDesc[index].bufSize.bytes = 896

If application want to provide the size as part of meta data then it should set inBufs-

>metadataPlaneDesc[index].bufSize.bytes = -1 otherwise encoder will read the size

from metadataPlaneDesc[index].bufSize.bytes field.

Index of metadataPlaneDesc follows these rules:

Note:

 Encoder assumes the availability of payload during processing of entire

sequence, if it is user defined.

 Encoder assumes that for each process call the payload is provided.

C.3.1 Format of Payload

typedef struct

{

U16 wgt4x4[2][3][2][4][4];

//[Intra(0)/Inter(1)][Y(0)/Cb(1)/Cr(2)][Inv(0)/Fwd(1)][4][4]

U16 wgt8x8[2][2][8][8];

//[intra(0)/inter(1)][inv(0)/Fwd(1)][8][8]

} sH264WgtTables_t ;

Comments in above structure explain the usage of each dimension. For example, forward

Intra Chroma Cr component is pointed by sH264WgtTables_t::

wgt4x4[0][2][1][4][4];

Inv/Fwd are explained below:

 Inv: This means the actual scaling matrices which decoder derives after

decoding from the bit-stream

Control for User Defined Scaling Matrices

C-7

 Fwd: This is a derived value from Inv data which is used by encoder in

Forward path, it is 1/Inv value in Q.18 format and only lower 16 bits are

considered (upper two bits are always 0)

Example:

Inv =

{{16, 16, 16, 16}, {16, 16, 16, 16},{16, 16, 16, 16},{16, 16,

16, 16}};

Fwd =

{

MIN((0x40000 + 16/2)/16, 0xFFFF), MIN((0x40000 + 16/2)/16,

0xFFFF), MIN((0x40000 + 16/2)/16, 0xFFFF), MIN((0x40000 +

16/2)/16, 0xFFFF)

{

MIN((0x40000 + 16/2)/16, 0xFFFF), MIN((0x40000 + 16/2)/16,

0xFFFF), MIN((0x40000 + 16/2)/16, 0xFFFF), MIN((0x40000 +

16/2)/16, 0xFFFF)

{

MIN((0x40000 + 16/2)/16, 0xFFFF), MIN((0x40000 + 16/2)/16,

0xFFFF), MIN((0x40000 + 16/2)/16, 0xFFFF), MIN((0x40000 +

16/2)/16, 0xFFFF)

{

MIN((0x40000 + 16/2)/16, 0xFFFF), MIN((0x40000 + 16/2)/16,

0xFFFF), MIN((0x40000 + 16/2)/16, 0xFFFF), MIN((0x40000 +

16/2)/16, 0xFFFF)

};

C.3.2 Constraints on Payload Data

Each value has to be an unsigned 16-bit value. As per formula to compute forward matrix

value, the minimum value for scaling matrix weight in inverse path is 4.

Maximum value for scaling matrix weight in inverse path is 255

There is no error check performed for the values of scaling matrices and the behavior is not

defined for non-supported values.

Control for User Defined Scaling Matrices

C-8

This page is intentionally left blank

Motion Vector and SAD Access API

D-9

Appendix D

Motion Vector and SAD Access API

This section describes the method to access MV and SAD (Analytic Information) data

dumped by the encoder.

D.1 Description

The Motion Vector and SAD Access API is a part of the XDM process() call, used by the

application to encode a frame. A parameter enabledAnalyticinfo is provided as a part of

create time parameters, which can be set or reset at a frame level during create-time. Setting

this flag to non-zero value indicates that the analytic info is needed. When this parameter is

set to non-zero value, the process() call returns the motion vector and SAD data in the buffer

provided by the application.

For every macro block, the data returned is 10 bytes, a signed horizontal displacement

component (signed 16-bit integer) and a vertical displacement component (signed 16-bit

integer) in L0 and L1 direction and SAD (16-bit integer).

The following sequence should be followed for Analytic Info access:

15) In the create time parameters, set the flag to access analytic data

/* Enable MV access */

createParams ->enableAnalyticinfo = 1;

16) Allocate output buffers and define the output buffer descriptors

/* Output Buffer Descriptor variables */

XDM2_BufDesc outputBufDesc;

/* Get the input and output buffer requirements for the

codec */

control(.., XDM_GETBUFINFO, extn_dynamicParams, ..);

If Analytic info access is enabled in step1, this call returns the output buffer

info as numBufs =2, along with the minimal buffer sizes.

/* Initialize the output buffer descriptor */

outputBufDesc.numBufs = 2;

/* Stream Buffer */

outputBufDesc.descs[0].buf = streamDataPtr; //pointer to

H264 bit-stream

outputBufDesc.descs[0].bufSize.bytes =

status.videnc2Status.bufInfo.minOutBufSize[0].bytes;

/* MV & SAD Buffer */

Motion Vector and SAD Access API

D-10

outputBufDesc.descs[1].buf = Output_Buffer_Base_Addr;

//pointer to MV and SAD data

outputBufDesc.descs[1].bufSize.bytes =

status.videnc2Status.bufInfo.minOutBufSize[1].bytes;

17) Call frame encode API

/* Process call to encode 1 frame */

process(.. ,.. , outputBufDesc, .. );

After this call, the buffer outputBufDesc.descs[1].buf will have SAD

and Motion vector data. The data format of this buffer will be like,

AnalyticHeaderInfo

Data (SAD and MV)

Define a structure:

struct AnalyticHeaderInfo

{

U32 NumElements;

ElementInfo elementInfoField0SAD;

ElementInfo elementInfoField1SAD;

ElementInfo elementInfoField0MVL0;

ElementInfo elementInfoField0MVL1;

ElementInfo elementInfoField1MVL0;

ElementInfo elementInfoField1MVL1;

} ;

Where as

NumElements -> Total number of elements in the buffer

(As of now SAD ,MV in L0 direction and MV in L1

direction for each field in case of interlace content)

ElementInfo is

typedef struct

{

/*Starting position of data from the buffer base

address*/

U32 StartPos;

/* No. of bytes to jump from the current position to

get the next data of this element group */

U32 Jump;

Motion Vector and SAD Access API

D-11

/* Number of data elements in this group */

U32 Count;

}ElementInfo;

The data format will differ for each frame type; there can be four different formats as,

1. Process call which generates one P frame/field

2. Process call which generates two P fields

3. Process call which generates one B frame/field

4. Process call which generates two B fields

Note: The data present in the shaded boxes of the below figures are don’t

care values.

Process call, which generates one P frame/field:

Figure D-1. Data format of Analytic Information in case of P frame/field.

Process call, which generates two P fields:

NumElements = 2

elementInfoField0SAD

elementInfoField1SAD

elementInfoField0MVL0

elementInfoField0MVL1

elementInfoField1MVL0

elementInfoField1MVL1

SAD and MV Data

NumElements = 4

elementInfoField0SAD

elementInfoField1SAD

elementInfoField0MVL0

elementInfoField0MVL1

elementInfoField1MVL0

elementInfoField1MVL1

SAD and MV Data

Motion Vector and SAD Access API

D-12

Figure D-2. Data format of Analytic Information in case of two P fields.

Process call, which generates one B frame/field:

Figure D-3. Data format of Analytic Information in case of B frame/field

Process call, which generates two B fields:

Figure D-4. Data format of Analytic Information in case of two B fields.

NumElements = 3

elementInfoField0SAD

elementInfoField1SAD

elementInfoField0MVL0

elementInfoField0MVL1

elementInfoField1MVL0

elementInfoField1MVL1

SAD and MV Data

NumElements = 6

elementInfoField0SAD

elementInfoField1SAD

elementInfoField0MVL0

elementInfoField0MVL1

elementInfoField1MVL0

elementInfoField1MVL1

SAD and MV Data

Motion Vector and SAD Access API

D-13

D.2 Example Usage

For example, data in output buffer dumped by the codec for a progressive - B frame is as shown

below,

Figure D-5. MV and SAD data dump by codec in case of

progressive B frames.

To get the MVL0 data for all macroblocks, the application should have code as,

S16 *Src = (U32)Output_Buffer_Base_Addr +

elementInfoMVL0->StartPos;

U32 Jump = elementInfoMVL0->Jump;

S16 *MVL0 = Addr_to_store_MV_inL0 ;

Jump = Jump / sizeof(S16);

for (i = 0; i < elementInfoMVL0->Count; i = i++)

{

* MVL0 ++ = Src[i * Jump]; // To get MVx

* MVL0 ++ = Src[((i *Jump) + 1)]; //To get MVy

}

NumElements

elementInfoField0SAD

elementInfoField1SAD(Don’t care)

elementInfoField0MVL0

elementInfoField0MVL1

elementInfoField1MVL0(Don’tcare)

elementInfoField1MVL1(Don’tcare)

SAD_MB0,

SAD_MB1,

……..

SAD_MBm-1

[MVxL0,MVyL0,MVxL1,MVyL1]MB0,

[MVxL0,MVyL0,MVxL1,MVyL1]MB1,

……….

[MVxL0,MVyL0,MVxL1,MVyL1]MBm-1

Output_Buffer_Base_Addr

elementInfoSAD->startPos

elementInfoMVL0->startPos

AnalyticHeaderInfo

elementInfoMVL1->startPos

Motion Vector and SAD Access API

D-14

Note:

 The motion vectors are with quaterpel resolution.

 SAD = ABS(Ref(i,j) – Src(i,j)) where, Ref is the macro block of the

reference region and Src is the macro block of the source image.

 The motion vectors seen in the encoded stream is based on the best

coding decision, which is a combination of motion estimation and

mode decission. The MV buffer returns the results of the motion

estimation in quaterpel resolution (lowest SAD) and this may be

different from the motion vectors seen in the bit-stream. More details

are given below :

Some macro blocks in a P-frame may be coded as Intra macro

blocks based on the post motion estimation decisions. In this case, the

motion vectors computed in the motion estimation stage (assuming

that this macro block is inter) is returned.

Due to the post motion estimation decisions for some macro blocks,

the actual motion vector encoded may be forced to skip MV. In this

case, the non-skip motion vector available after the motion estimation

is returned.

For I-frames, motion vectors and SAD are not present in the buffer.

E-1

Appendix E

Debug Trace Support

This appendix explains the Debug Trace support details on encoder. This is to help the

application to get the trace data generated by Encoder from external memory

E.1 Debug Trace design in Encoder

Encoder has “debugTraceLevel” interface to select the debug trace level. When

“debugTraceLevel” is set to zero then Encoder will not generate any trace data.

Otherwise, it will generate the trace data in external memory. Encoder has support to log last

N frame’s debug trace data which is controlled by “lastNFramesToLog” interface parameter.

If the encoder is requested to generate debug trace data then encoder will request for

external memory to store these trace data. The size of this memory depends on the

“lastNFramesToLog” parameter. If the size for one process calls trace data is A bytes then

the total bytes requested will be

Total size (X) = (1 + lastNFramesToLog)* A bytes.

Each instance of this trace buffer has two sections of trace data. First section (B) is written in

external memory by Media Controller through cache and other section (C) is written by

HDVICP2.0 using DMA.

Here size of both B and C are aligned to cache line size, which is 32 bytes. Codec will not do

any cache related operation at any point of time. Since the section B is written by Media

Controller through cache, cache write back needs to be performed for this section. If

application has programmed the lastNFramesToLog values as N, then the cache write

back needs to be performed N+1 times. HDVICP2.0 will write section C using DMA, so the

cache write back is not required for this section.

Debug Trace Support

E-2

E.1.1 Steps to utilize debug trace support in H264 encoder

Create encoder with following settings

IH264ENC_Params.debugTraceLevel = 1;

IH264ENC_Params.lastNFramesToLog = N; (example: 10)

Then make a control call with “XDM_GETSTATUS” command to get the following parameters

from codec

Debug trace level used by codec

IH264ENC_Status. debugTraceLevel

Number of frames for which log is available

IH264ENC_Status. lastNFramesToLog

Base address of trace data in external memory

IH264ENC_Status.extMemoryDebugTraceAddr

Total size of trace buffer in external memory

IH264ENC_Status.extMemoryDebugTraceSize

Size of trace buffer for one process call (A) is

A = IH264ENC_Status.extMemoryDebugTraceSize / (IH264ENC_Status.

lastNFramesToLog + 1)

Cache write back operation needs to be performed before reading this data from external

memory.

Pseudo code for cache write back

ddrAddress = IH264ENC_Status.extMemoryDebugTraceAddr;

totalNumFrames = (IH264ENC_Status.lastNFramesToLog + 1);

for(i = 0; i < totalNumFrames i++)

{

CacheWriteBack(ddrAddress,B);

ddrAddress += A;

}

Definition of “CacheWriteBack” function

CacheWriteBack(void * address, int size);

Here

address : Start address for write back operation

Size : length in bytes

F-1

Appendix F

Picture format

This appendix explains the picture format details for encoder. Encoder expects the input

uncompressed picture to be in NV12 format.

F.1 NV12 Chroma Format

NV12 is YUV 420 planar with 2 separate planes, one for Y, one for U and V interleaved.

Luma Plane

Y0,0

Y0,1

Y1,0

Y1,1

Chroma Plane

U0,0

V0,0

U1,0

V1,0

WIDTH

HEIGHT

HEIGHT/2

Picture format

F-2

F.2 Progressive and Interlaced Format

F.2.1 Progressive Format

ACTIVE REGION (LUMA)

picLumaBufferAddr

frameHeight

ACTIVE REGION (CHROMA)

picChromaBufferAddr

frameWidth

imagePitch[0]

0,0 Y

0,1 Y

0,2 Y

0,3

2,0 Y

2,1 Y

2,2 Y

2,3

0,0 V

0,0 U

0,1 V

0,1

1,0 V

1,0 U

1,1 V

1,1

imagePitch[1]

frameWidth

frameHeight/2

ActiveRegion: Data to be encoded

Extra region beyond the ActiveRegion may be allocated by application due to imagePitch

constraints.

Both luma and chroma buffers can be allocated independently and both can have their pitch

different

Picture format

F-3

F.2.2 Interlaced Format

ACTIVE REGION

TOP FIELD (Luma)

0,0 Y

0,1 Y

0,2 Y

0,3

2,0 Y

2,1 Y

2,2 Y

2,3

ACTIVE REGION

TOP FIELD (Chroma)

0,0 V

0,0 U

0,1 V

0,1

2,0 V

2,0 U

2,1 V

2,1

ACTIVE REGION

BOTTOM FIELD (Luma)

1,0 Y

1,1 Y

1,2 Y

1,3

3,0 Y

3,1 Y

3,2 Y

3,3

ACTIVE REGION

BOTTOM FIELD (Chroma)

1,0 V

1,0 U

1,1 V

1,1

3,0 V

3,0 U

3,1 V

3,1

picLumaTopBufferAddr

picLumaBottomBufferAddr

picChromaTopBufferAddr

picChromaBottomBufferAddr

framewidth

frameWidth

imagePitch[0]

imagePitch[1]

imagePitch[0]

frameHeight / 4

frameHeight / 2

ActiveRegion:

Data to be encoded

Extra region beyond the ActiveRegion may be allocated by application due to imagePitch

constraints.

Both luma and chroma buffers can be allocated independently and both can have their pitch

different

The figure shown is for the case when field data is separate, Encoder also supports filed

interleaved data format where both filed are interleaved in memory.

Picture format

F-4

F.3 Constraints on Parameters

- imagePitch need to comply with following constraints

- imagePitch shall be greater or equal to the Width (passed by the application host).

- imagePitch is “don’t care” if the buffer is in TILED8, TILED16 or TILED32 region

- Buffer Addresses need to comply with following constraints

- addresses shown as picLumaBufferAddr in figures shouldn’t point to any region

which is not TILED8 or RAW/TILED PAGE

- The addresses shown as picChromaBufferAddr in figures shouldn’t point to any

region which is not TILED8, TILED16 or RAW/TILED PAGE

- In interlaced picture for field interleaved case the luma and chroma buffer must be in

RAW buffer

Constraints on resolutions are defined as below

Progressive:

- Minimum frameWidth = 96

- Minimum frameHeight = 80

- Maximum frameWidth = 4352

- Maximum frameHeight = 4096

- frameWidth shall be multiple of 2

- frameHeight shall be multiple of 2

Interlaced:

- Minimum frameWidth = 96

- Minimum (frameHeight/2) = 80

- Maximum frameWidth = 4352

- Maximum (frameHeight/2) = 4096

- frameWidth shall be multiple of 2

- frameHeight shall be multiple of 4.

G-1

Appendix G

Low Latency / Sub Frame Level

Synchronization

This appendix explains the details of H264 encoder’s low latency features and how to

exercise them.

G.1 Description

Most of the TI Video Codec interfaces prior to IVIDENC2 and IVIDDEC3 allow frame level

data communication capabilities. A user can configure the codec to encode/decode a

complete frame but not any sub-frame level data communications. If at all any then it is via

codec’s extended interface.

This appendix explains the sub-frame level data communication capabilities of video codec

using data synch call backs defined with IVIDENC2 interface

G.2 H.264 Encoder Input with sub frame level synchronization

H.264 encoder allows accepting partial frames for the application on input side and can start

encoding. This section explains the IVIDENC2 interface details which help to achieve the sub

frame level communications on input side of a video encoder.

Table 21 , Table 22 and Table 23 explain the creation, control and handshake parameters

related to sub frame level data communication for input data of video encoder respectively.

Details column is a generic column and “valid values” column is specific to video encoder

input.

Table 21 Creation time parameter related to sub frame level data communication for input

data of video encoder

Parameter

Name

Details

Valid values

IVIDENC2_Pa

rams::inputDa

taMode

Defines the mode of accepting

the input frame.

IVIDEO_ENTIREF

RAME

entire frame data is

given to encoder

IVIDEO_NUMRO

Frame data is given in

unit of Number of mb

rows, each mb row is

16 lines of video

IVIDENC2_Pa

rams::numInp

utDataUnits

Unit of input data

Don’t care. As the information about the data

can be available during sub frame level

communication

Low Latency / Sub Frame Level Synchronization

G-2

Table 22 Dynamic parameters related to sub frame level data communication for input

data of video encoder

Parameter

Name

Details

Valid values

IVIDENC2_Dy

namicParams:

:getDataFxn

This function pointer is provided

by the app/framework to the

video encoder. The encoder

calls this function to get partial

video buffer(s) from the

app/framework.

Apps/frameworks that support

datasync should set this to non-

NULL.

Any non-NULL value if inputDataMode !=

IVIDEO_ENTIREFRAME

IVIDENC2_Dy

namicParams:

:getDataHandl

It defines the handle to be used

while requesting data to

application. This is a handle

which the codec must provide

when calling getDataFxn.

Apps/frameworks that support

datasync should set this to non-

NULL. For an algorithm, this

handle is read-only; it must not

be modified when calling the

app-

registered IVIDENC2_DynamicP

arams.getDataFxn(). The

app/framework can use this

handle to differentiate callbacks

from different algorithms.

Any Value

Table 23 Handshake parameters related to sub frame level data communication for input

data of video encoder

Parameter

Name

Details

Valid values

XDM_DataSy

ncDesc::size

Size of the XDM_DataSyncDesc structure

Sizeof(XDM_DataSyncDesc)

XDM_DataSy

ncDesc::

scatteredBloc

ksFlag

Flag indicating whether the individual data

blocks may be scattered in memory.

Note that each individual block must be

physically contiguous.

Valid values are XDAS_TRUE and

XDAS_FALSE.

If set to XDAS_FALSE, the baseAddr field

points directly to the start of the first block, and

is not treated as a pointer to an array.

If set to XDAS_TRUE, the baseAddr array

must contain the base address of each

individual block.

Don’t care as buffer is

assumed to be contiguous

XDM_DataSy

ncDesc::base

Addr

Base address of single data block or pointer to

an array of data block addresses of size

numBlocks.

If scatteredBlocksFlag is set to XDAS_FALSE,

this field points directly to the start of the first

block, and is not treated as a pointer to an

array.

If scatteredBlocksFlag is set to XDAS_TRUE,

this field points to an array of pointers to the

data blocks.

Don’t care since it is

assumed to be contigous yuv

buffer and initial address is

via inbuf at process call.

Low Latency / Sub Frame Level Synchronization

G-3

XDM_DataSy

ncDesc::num

Blocks

Number of data blocks

Any Value. If <= zero then

codec assumes no data

provided and does call back

to App again. The unit of this

is number of row.

XDM_DataSy

ncDesc::varBl

ockSizeFlag

Flag indicating whether any of the data blocks

vary in size.

Don’t care, as unit of size is

one row

XDM_DataSy

ncDesc::block

Sizes

Variable block sizes array.

Don’t care Since unit is

assumed to be multiple of

number of rows which is

indicated by numBlocks.

If application, want to use video encoder to operate with sub frame on input side

It should create the video encoder with IVIDENC2_Params::inputDataMode =

IVIDEO_NUMROWS.

It should also make a control call with IVIDENC2_DynamicParams::getDataFxn = non-NULL;

to use sub frame level data communication, control call is mandatory.

It should provide the base address of the input buffer during process call

It should provide all the data availability via getDataFxn call back, during process call the

input buffer is assumed to be data-less

Constraint

In presence of B frame, IVIDENC2_Params::inputDataMode = IVIDEO_NUMROWS is an

erroneous case

IVIDENC2_DynamicParams::getDataFxn == NULL && IVIDENC2_Params::inputDataMode

== IVIDEO_NUMROWS is an erroneous situation and codec returns error during process

call.

G.3 H.264 Encoder Output with sub frame level synchronization

H.264 encoder allows providing partial compressed bit-stream to the application on output

side. This section explains the IVIDENC2 interface details, which help to achieve the sub

frame level communications on output side of a video encoder.

Table 24, Table 25 explain the creation and control parameters related to sub frame level

data communication for output data of video encoder respectively.

Details column is a generic column and “valid values” column is specific to video encoder

output.

Table 24 Creation time parameter related to sub frame level data communication for

output data of video encoder

Parameter

Name

Details

Valid values

IVIDENC2_Pa

rams::outputD

ataMode

Defines the mode of providing

the output data.

IVIDEO_ENTIREFRAME

Entire frame bit-stream

is given out by the

encoder

IVIDEO_FIXEDLENGTH

bit-stream is provided

by encoder after a

fixed length of bytes.

The length has to be

Low Latency / Sub Frame Level Synchronization

G-4

multiple of 1K

IVIDEO_SLICEMODE

bit-stream is provided

by encoder after

producing a single(or

more) number of NAL

Units

IVIDENC2_Pa

rams::numOut

putDataUnits

Unit of output data

Don’t care if outputDataMode ==

IVIDEO_ENTIREFRAME

Any positive value if outputDataMode !=

IVIDEO_ENTIREFRAME

if outputDataMode ==

IVIDEO_FIXEDLENGTH then it indicates the

basic unit of size (in multiple of 1K) at which

encoder should inform the application.

Eg: Here 4 means that encoder should

inform after producing every 4*1024 bytes to

application

if outputDataMode == IVIDEO_SLICEMODE

then it indicates the basic unit of slices at

which encoder should produce the bit-

stream.

Eg: Here 5 means that after encoding a set

of 5 NALUs,

encoder should inform to application

Table 25 Dynamic parameters related to sub frame level data communication for output

data of video encoder

Parameter

Name

Details

Valid values

IVIDENC2_Dy

namicParams:

:putDataFxn

This function pointer is provided

by the app/framework to the

video encoder. The encoder

calls this function when data has

been put in output buffer. It is to

inform the app/framework.

Apps/frameworks that support

datasync should set this to non-

NUL

Any non-NULL value if outputDataMode !=

IVIDEO_ENTIREFRAME

IVIDENC2_Dy

namicParams:

:putDataHandl

It defines the handle to be used

while informing data availability

to application. This is a handle

which codec must provide when

calling putDataFxn.

Apps/frameworks that support

datasync should set this to non-

NULL. For an algorithm, this

handle is read-only; it must not

be modified when calling the

app-

registered IVIDENC2_DynamicP

arams.putDataFxn(). The

app/framework can use this

handle to differentiate callbacks

from different algorithms.

Any Value

To simplify the codec implementation, the information sharing by codec to application

happens at a quantum of 1K byte data. In this document, each 1K byte is referred as page.

Low Latency / Sub Frame Level Synchronization

G-5

If application, want to use video encoder to operate with sub frame on output side

It should create the video encoder with IVIDENC2_Params::outputDataMode =

IVIDEO_SLICEMODE or IVIDEO_FIXEDLENGTH.

It should also make a control call with IVIDENC2_DynamicParams::putDataFxn = non-NULL;

to use sub frame level data communication, control call is mandatory.

It should provide the base address and available space of the output buffer during process

call

Erroneous case

IVIDENC2_DynamicParams::putDataFxn == NULL && IVIDENC2_Params::outputDataMode

!= IVIDEO_ENTIREFRAME is an erroneous situation and codec returns error during process

call.

If outPutDataMode == IVIDEO_SLICE and multiple slices are not enabled (sliceMode ==

IH264_SLICEMODE_NONE), encoder returns error

(IH264ENC_UNSUPPORTED_SLICECODINGPARAMS) during create time

If numOutputDataUnits > 64 or numOutputDataUnits < 0 with outputDataMode !=

IVIDEO_ENTIREFRAME is an erroneous situation and code returns error

(IH264ENC_IMPROPER_DATASYNC_SETTING) during create time

If Number of B frame > 0 && inputDataMode != IVIDEO_ENTIREFRAME, encoder returns

error (IH264ENC_IMPROPER_DATASYNC_SETTING) at create time

If minBitRate > 0 && outputDataMode != IVIDEO_ENTIREFRAME, encoder returns error

(IH264ENC_UNSUPPORTED_VIDENC2PARAMS) at create time

If outPutDataMode == IVIDEO_SLICE and sliceMode = IH264_SLICEMODE_BYTES, then

encoder expects getBufferFxn to be implemeneted by application. So outPutDataMode ==

IVIDEO_SLICE && sliceMode = IH264_SLICEMODE_BYTES && getBufferFxn == NULL) is

erroneous condition and encoder returns IH264ENC_IMPROPER_DATASYNC_SETTING

error during control call

G.3.1 H.264 Encoder mechansim to accpet partial buffer and non contiguous

buffer on output side

Before unserstanding, the interface related to different outputDataMode, it is important to

understand about the interface, which allows encoder to accept non-contiguous memory

With IVIDENC2 interface video encoder can work with a situation when it has not been

provided complete bit-stream buffer to it during process call. Application can provide non

contiguous chunks of memory with some size constraints to encoder and it can produce the

bit-stream in these buffers.

It is achieved by IVIDENC2_DynamicParams::getBufFxn() interface.

To get the encoder working with partial output buffer, there is no specific creation time

parameter.

Control call is mandatory and application need to provide a valid function pointer as

IVIDENC2_DynamicParams::getBufFxn.

Application also need to set IVIDENC2_DynamicParams::ignoreOutbufSizeFlag as

true to prevent encoder reporting error

Low Latency / Sub Frame Level Synchronization

G-6

Table 26 and Table 27 explain the control and handshake parameters related to sub frame

level data communication to handle partial output buffer by video encoder respectively.

Details column is a generic column and “valid values” column is specific to video encoder.

Following points should be noticed to use video encoder with partial buffer on output side

getBuf is independent of outputDataMode or inputDataMode. It is only meant for codec to

ask application for a buffer, if encoder has exhausted for output bit-stream

During process call the initial stream address and size are provided by application. No

constraint on this information and encoder consumes this buffer space

During data synch (via getBuf) codec can accept a multiple non contiguous buffers from

application each of them has to be multiple of 2K. (only exception here is when encoder is

congiured to work with outputDataMode = IVIDEO_SLICE_MODE and outputDataMode ==

IVIDEO_SLICE_MODE. With this case encoder can accept any size which is >=

sliceUnitSize )

if scatteredBlocksFlag is non zero then Maximum number of blocks provided by user should

be 8. If application provides more than 8 block then codec will just accept 8 blocks and rest of

the blocks will be ignored (constraint)

If scatteredBlocksFlag flag is zero than there is no limit on numBlocks.

If the function pointer IVIDENC2_DynamicParams::getBufFxn provided is null then encoder

will first consume the buffer provided in process call (by writing the bit stream data), if that

buffer is exhausted then encoder has to do proper pipe down and come out from the process

call with error (XDM_INSUFFICIENT_DATA).

Table 26 Dynamic parameters related to accept partial buffer for output bit-stream

Parameter

Name

Details

Valid values

IVIDENC2_Dy

namicParams:

:getBufFxn

This function pointer is provided

by the app/framework to the

video encoder. The encoder

calls this function to get partial

bit-stream buffer(s) from the

app/framework.

Apps/frameworks that support

datasync should set this to non-

NULL.

Any non-NULL value to use partial buffer

for bit-stream space

IVIDENC2_Dy

namicParams:

:getDataHandl

This is a handle which the codec

must provide when calling the

app-registered

IVIDENC2_DynamicParam.getB

ufferFxn(). Apps/frameworks

that don't support datasync

should set this to NULL. For an

algorithm, this handle is read-

only; it must not be modified

when calling the app-registered

IVIDENC2_DynamicParams.get

BufferFxn(). The app/framework

can use this handle to

differentiate callbacks from

different algorithms.

Any Value

Low Latency / Sub Frame Level Synchronization

G-7

Table 27 Handshake parameters related to accept partial buffer for output bit-stream

Parameter

Name

Details

Valid values

XDM_DataSy

ncDesc::size

Size of the XDM_DataSyncDesc structure

sizeof(XDM_DataSyncDesc)

XDM_DataSy

ncDesc::

scatteredBloc

ksFlag

Flag indicating whether the individual data

blocks may be scattered in memory.

Note that each individual block must be

physically contiguous.

Valid values are XDAS_TRUE and

XDAS_FALSE.

If set to XDAS_FALSE, the baseAddr field

points directly to the start of the first block, and

is not treated as a pointer to an array.

If set to XDAS_TRUE, the baseAddr array

must contain the base address of each

individual block.

XDAS_TRUE or

XDAS_FALSE

XDM_DataSy

ncDesc::base

Addr

Base address of single data block or pointer to

an array of data block addresses of size

numBlocks.

If scatteredBlocksFlag is set to XDAS_FALSE,

this field points directly to the start of the first

block, and is not treated as a pointer to an

array.

If scatteredBlocksFlag is set to XDAS_TRUE,

this field points to an array of pointers to the

data blocks.

non-NULL, if NULL then

again call back.

If baseAddress[i] is NULL

then again call back (where

i=0 to numBlock -1 when

scatteredBlocksFlag is non-

zero)

XDM_DataSy

ncDesc::num

Blocks

Number of data blocks

Any Value. If <= zero then

codec assumes no data

provided and does call back

to App again.

<=8 if scatteredBlocksFlag !=

if scatteredBlocksFlag != 0

then values higher than 8 are

assumed to be 8

XDM_DataSy

ncDesc::varBl

ockSizeFlag

Flag indicating whether any of the data blocks

vary in size.

XDAS_TRUE or

XDAS_FALSE

XDM_DataSy

ncDesc::block

Sizes

Variable block sizes array.

non-NULL. If it is NULL then

again call back

definition of blockSize[i] is

different for different

situations as mentioned

below

- For sliceMode =

IH264_SLICEMODE_BYTES

and outputDataMode ==

IVIDEO_SLICE_MODE it

should hold a value >=

sliceUnitSize

- For other situations it

should hold a value which is

multiple of 2K

If application doesn’t

Obey these restrictions then

the behavior is undefined

Low Latency / Sub Frame Level Synchronization

G-8

totalBlockSize =

SUM(blockSizes[0] to

blockSizes[numBlocks-1]) if

varBlockSizesFlag is non

zero.

totalBlockSize = numBlocks *

blockSizes[0] if

varBlockSizesFlag is zero

if totalBlocksSize is 0 the call

back again

G.3.2 H.264 Encoder behavior with outputDataMode as IVIDEO_SLICEMODE

Table 28 explains the handshake parameters for sub frame level data communication with

outputDataMode = IVIDEO_SLICEMODE

Communication point by codec to application about data availability is one of the below

whichever is later

numof slices( numOutputDataUnit) is encoded i.e. if in the current page ,numOfSlice >=

numOutputDataUnit then make a putData call.

Minimum 1K of data is encoded i.e. numOfSlices exceeds numOutputDataUnit in the first

page cross itself.

Note that communication point is always at page cross over except at the last call (end of

process) where bit stream can end at any point in the page.

Incase of outputDataMode = IVIDEO_SLICEMODE, following points should be noted

numOutputDataUnit is the frequency after which codec will inform to App. So in

IVIDEO_SLICE_MODE, lets outputDataUnit is 8 then after 8 slice codec has to make

putData call.

This encoder implementations has constraint of limiting maximum allowed value of

outputDataUnit as 64

Let’s say numOutputDataUnit is 64, and in one page codec generates 63 slices and in the

next page it generated again 64 slices, in this case codec will inform all the 127 slices. So

maximum generated value by encoder for numBlocks is 127

Bit-stream can be non-contiguous at NAL boundaries, if the encoder is configured to

generate NAL Units of fixed length (sliceMode == IH264_SLICEMODE_BYTES). In this case

after each NALU completion, encoder moves to next NALU’s start address even there are

few bytes left in the previous buffer(packet)

If the encoder is configured to generate slices based upon macroBlockPerSlice (sliceMode

==IH264_SLICEMODE_MBS or sliceMode == IH264_SLICEMODE_OFFSET) then the bit-

stream is assumed to be contiguous in memory, hence it is user’s responsibility to provide

the bit-stream address during data synch calls(XDM_DataSyncDesc::baseAddr) to be in

continuation of the earlier bit-stream address provided to encoder

Application provides buffer size and address for bit-stream during process call, both of them

are honored and consumed by encoder until it needs more space to write bit-stream (refer

getBuf interface of video encoder for more details)

Low Latency / Sub Frame Level Synchronization

G-9

All data availability is informed via data synch calls, while process exit the bytesGenerated

indicates the total sum (not the size of last chunk)

Table 28 Handshake parameters related to sub frame level data communication for output

data of video encoder (outputDataMode = IVIDEO_SLICEMODE)

Parameter

Name

Details

Valid values

XDM_DataSy

ncDesc::size

Size of the XDM_DataSyncDesc structure

sizeof(XDM_DataSyncDesc)

XDM_DataSy

ncDesc::

scatteredBloc

ksFlag

Flag indicating whether the individual data

blocks may be scattered in memory.

Note that each individual block must be

physically contiguous.

Valid values are XDAS_TRUE and

XDAS_FALSE.

If set to XDAS_FALSE, the baseAddr field

points directly to the start of the first block, and

is not treated as a pointer to an array.

If set to XDAS_TRUE, the baseAddr array

must contain the base address of each

individual block.

Flag indicating whether the

individual data slices may be

scattered in memory.

Constraint: None

XDM_DataSy

ncDesc::base

Addr

Base address of single data block or pointer to

an array of data block addresses of size

numBlocks.

If scatteredBlocksFlag is set to XDAS_FALSE,

this field points directly to the start of the first

block, and is not treated as a pointer to an

array.

If scatteredBlocksFlag is set to XDAS_TRUE,

this field points to an array of pointers to the

data blocks.

This field points directly to

the start of the data for the

active transaction

XDM_DataSy

ncDesc::num

Blocks

Number of data blocks

Any Value and it is the

number of slices generated

till the point of putData call. If

outputDataUnit is 7, in the

page cross over which would

be the communication point

and it generated 8 slices,

then numbBlocks is 8 and all

8 slices will be informed to

App.

Codec can generate

following possible values of

numBblocks

1 <= numBlocks <= 127

XDM_DataSy

ncDesc::varBl

ockSizeFlag

Flag indicating whether any of the data blocks

vary in size.

XDAS_TRUE or

XDAS_FALSE(slice sizes are

not constant most of the

time)

XDM_DataSy

ncDesc::block

Sizes

Variable block sizes array.

If varBlockSizesFlag is

XDAS_TRUE, this array

contains the sizes of each

slice. So total slice size is

sum of (blockSizes[0] to

blockSizes[numBlocks -1].

If varBlockSizesFlag is

XDAS_FALSE, this contains

the size of same-size slices.

So total data given by

encoder to app would be

(numBlocks * blocSizes[0])

Low Latency / Sub Frame Level Synchronization

G-10

G.3.3 H.264 Encoder behavior with outputDataMode as IVIDEO_FIXEDLENGTH

Table 29 explains the handshake parameters for sub frame level data communication with

outputDataMode = IVIDEO_FIXEDLENGTH

Communication point by codec to application about data availability is one of the below

whichever is earlier

1 K Bytes * numOutputDataUnit of data is encoded.

if 64 non-continuous blocks have been generated by encoder.

Note that communication point is always at page cross over except at the last call (end of

process) where bit stream can end at any point in the page.

Incase of outputDataMode = IVIDEO_FIXEDLENGTH, following points should be noted

numOututDataUnit is the frequency after which codec will inform to App. so in

IVIDEO_FIXED_LENGTH, lets outputDataUnit is 10 then after 10 page cross over (which is

communication point to app) in SL2 bitstream space, codec will make putData call. if

numOutputDataUnit is 10, and initial bitstream buffer size given in process call is 0.5 KB,

then codec will put a putData call after 9.5 KB of encoding, not after 10.5 KB.

Application provides buffer size and address for bit-stream during process call, both of them

are honored and consumed by encoder until it needs more space to write bit-stream (refer

getBuf interface of video encoder for more details)

All data availability is informed via data synch calls, while process exit the bytesGenerated

indicates the total sum (not the size of last chunk)

Table 29 Handshake parameters related to sub frame level data communication for output

data of video encoder (outputDataMode = IVIDEO_FIXEDLENGTH)

Parameter

Name

Details

Valid values

XDM_DataSy

ncDesc::size

Size of the XDM_DataSyncDesc structure

sizeof(XDM_DataSyncDesc)

XDM_DataSy

ncDesc::

scatteredBloc

ksFlag

Flag indicating whether the individual data

blocks may be scattered in memory.

Note that each individual block must be

physically contiguous.

Valid values are XDAS_TRUE and

XDAS_FALSE.

If set to XDAS_FALSE, the baseAddr field

points directly to the start of the first block, and

is not treated as a pointer to an array.

If set to XDAS_TRUE, the baseAddr array

must contain the base address of each

individual block.

Flag indicating whether the

individual data block may be

scattered in memory.

XDAS_TRUE or

XDAS_FALSE

XDM_DataSy

ncDesc::base

Addr

Base address of single data block or pointer to

an array of data block addresses of size

numBlocks.

If scatteredBlocksFlag is set to XDAS_FALSE,

this field points directly to the start of the first

block, and is not treated as a pointer to an

array.

If scatteredBlocksFlag is set to XDAS_TRUE,

this field points to an array of pointers to the

data blocks.

Base address of single data

block or pointer to an array of

block addresses of size

numBlocks.

If scatteredBlocksFlag is set

to XDAS_FALSE, this field

points directly to the start of

the first block, and is not

treated as a pointer to an

array.

If scatteredBlocksFlag is set

to XDAS_TRUE, this field

Low Latency / Sub Frame Level Synchronization

G-11

points to an array of pointers

to the data blocks i.e. from

baseAddr[0] to

baseAddr[numBlocks-1]

XDM_DataSy

ncDesc::num

Blocks

Number of data blocks

It is the number of blocks

generated till the point of

putData call.

Codec can generate

following possible values of

numBblocks

1 <= numBlocks <= 64

XDM_DataSy

ncDesc::varBl

ockSizeFlag

Flag indicating whether any of the data blocks

vary in size.

Flag indicating whether any

of the data blocks vary in

size. Valid values

XDAS_TRUE or

XDAS_FALSE

XDM_DataSy

ncDesc::block

Sizes

Variable block sizes array.

If varBlockSizesFlag is

XDAS_TRUE, this array

contains the sizes of each

block. So total data size or

bitstream is sum of

(blockSizes[0] to

blockSizes[numBlocks -1].

If varBlockSizesFlag is

XDAS_FALSE, this contains

the size of same-size data

blocks.So total data given by

encoder to app would be

(numBlocks * blocSizes[0])

Low Latency / Sub Frame Level Synchronization

G-12

This page is intentionally left blank

H-1

Appendix H

Long Term Reference Picture Schemes

This appendix explains the usage details of long-term reference picture support.

H.1 Description

Most of the applications which are sensitive to errors over network need error resilient

features in the vide encoder. Long-term reference picture allows encoder to prevent

propogation of erros in few temporal frames in past. Most of the multi-way real time

communication systems can get benifited with this error resiliency feature in video encoders.

H.2 Supported Schemes and Usage

This version of encoder supports following mechanisms for long-term reference picture,

1. Periodic Long term Reference picture

2. Proactive Long term Referencing

3. Reactive Long term Referencing

H.2.1 Periodic Long term Referencing

This scheme allows encoder to get instructed to refer to last marked long term reference picture.

Pictures are marked as long term reference picture based on the given period. To enable this

scheme following operations should be performed at create time

IH264ENC_Params::enableLongTermRefFrame should be set to

IH264ENC_LTRP_REFERTO_PERIODICLTRP.

IH264ENC_Params::LTRPPeriod should be set to long term reference picture marking interval i.e,

interval between two consecutive long term reference pictures.

This will cause encoder instance memory to be higher by one frame than in normal operation. If

normal operation is requiring 2 reference frames, now it will be require 3 reference frames to be

stored.

At process level IVIDENC2_InArgs::control should be set to

IH264ENC_CTRL_REFER_LONG_TERM_FRAME for the desired frame to refer to lastly marked

long term reference picture, so user has flexibility for each encoding frame to inform to encoder to

use lastly marked long term reference picture. The below picture explains how long term reference

pictures are marked based on LTRP Period and IntraframeInterval.

Long Term Reference Picture Schemes

H-2

Figure H-1. Marking of Long Term reference picture

Next two figures explains the usages of this feature in a 2 way video transmission system

Figure H-2. Long term referencing to I/IDR - 2 way video transmission system.

Long Term Reference Picture Schemes

H-3

H.2.2 Proactive Long term Referencing

This scheme allows encoder to be instructed to refer to create a structure with which it is

reactive to errors. In previous scheme, an action is taken after the error gets introduced and

recognized from the receiver, where as in this scheme it allows to create a bit-stream gop

structure which is recoverable in presense of errors.

To enable this scheme following operations should be performed:

At create time IH264ENC_Params::enableLongTermRefFrame should be set to

IH264ENC_LTRP_REFERTOP_PROACTIVE. This will not cause encoder instance memory

to be higher than normal operation. For IPPP.. kind of sequence 2 frame buffers will be

required as in normal scenarion. In normal scenarion when there is no long term referencing

is enabled than one buffer is used to write the reconstructed data for current frame that is

being encoded and another buffer is used as reference frame for current frame. In this kind of

long term referencing scheme, among the two buffers only one buffer will be reference frame,

another buffer will be kept for future usages. User can control which frame to be

reconstructed and for which frame reference frame needs to be changed.

At process level IVIDENC2_InArgs::control should be set to either of the 4 values based

upon the need

IH264ENC_CTRL_NOWRITE_NOREFUPDATE

IH264ENC_CTRL_WRITE_NOREFUPDATE

IH264ENC_CTRL_NOWRITE_REFUPDATE

IH264ENC_CTRL_WRITE_REFUPDATE

Figure H-3. Proactive Long term Referencing.

In the above figure, while encoding 5th frame user has given control to write the reconstructed

frame, so that is why in this process call 5th frame got reconstructed and placed in one of the

buffer, Non reference frame is always flushed if needed for storing new reconstructed frame.

In the above example till the encoding of 9th frame, frame number ‘1’ was used as reference.

In 10th frame encoding user has given the control about to update the refrence, so from this

frame onward 5th frame will be used as reference.

One can acheieve different GOP structure with different values of IVIDENC2_InArgs::control

at frame level

Long Term Reference Picture Schemes

H-4

If one sets IVIDENC2_InArgs::control = IH264ENC_CTRL_WRITE_REFUPDATE then

below GOP structure is achieved. This is equivalent to normal gop structure with no long

term referencing

Based upon the feedback from receiver, one can achieve dynamically either one of the below

gop structure.

Figure H-4.GOP structure in LTRP.

With this control, user can achieve unchained P frame in which all P frame refers to only I

frame. This GOP structure is useful for storage thinning over the time. In video security

domain, if the content is aged it can be thnned by removing any of the P frames by video

editing in below GOP structure.

Since the control is provided at picture level, it gives a lot of flexibility to application for getting

desired gop structure dynamically. Below table provides the value of control field to achieve

both situations.

Long Term Reference Picture Schemes

H-5

H.2.3 Reactive Long term Referencing

The operation of this scheme is similar to Long term Referencing to PeriodicLTRP scheme (H.2.1).

This scheme differs by maintaining two recent long term reference pictures at any point of time.

When encoder gets control command to refer to long term picture, it will refer to old long term

reference picture from the recently marked 2 long term reference picture. Pictures are marked as

long term reference picture based on the given period

To enable this scheme following operations should be performed at create time

IH264ENC_Params::enableLongTermRefFrame should be set to

IH264ENC_LTRP_REFERTOP_REACTIVE.

IH264ENC_Params::LTRPPeriod should be set to long term reference picture marking interval i.e,

interval between two consecutive long term reference pictures.

This will cause encoder instance memory to be higher by two frames than in normal operation. If

normal operation is requiring 2 reference frames, now it will be require 4 reference frames to be

stored.

Next two figures explains the usages of this feature in a 2 way video transmission system

Frame

If P5 is well received from receiver

If P5 is not well received from receiver

IVIDENC

2_InArgs:

:control

IH264ENC_CTRL_WRITE_REFUPDATE

IH264ENC_CTRL_NOWRITE_NOREFUPDATE

IH264ENC_CTRL_WRITE_NOREFUPDATE

IH264ENC_CTRL_NOWRITE_NOREFUPDATE

IH264ENC_CTRL_WRITE_REFUPDATE

IH264ENC_CTRL_WRITE_NOREFUPDATE

IH264ENC_CTRL_NOWRITE_NOREFUPDATE

Long Term Reference Picture Schemes

H-6

Figure H-5. Reactive Long term Referencing - 2 way video transmission system.

I-1

Appendix I

Hierarchical P structure Coding Scheme

This appendix explains the usage details of Hierarchical P structure coding.

I.1 Description

Hierarchical P structure allows additional flexibility to have a scalable bit-stream in terms of

bit rate and frame rate without adding any additional delay. With this structure, pictures are

coded in different temporal layers, where a picture only refers to pictures belonging to layers

below it for temporal prediction

Below figure I.1 shows the temporalLayer 4 structure coding (i.e from temporal layer 0 to 3).

Figure I-1. Hierarchical P structure coding for Temporal Layer 4 with layer numbers

By removing the higher layer pictures, one can achieve a decodable bit-stream with lesser

frame rate and hence lesser bitrate as well. In above example by removing the pictures from

layer 3, the resultant bit-stream is of half the frame rate and almost half the bit-rate of original

bit-stream

I.2 Supported Schemes and Usage

This version of encoder supports following mechansims for Hierarchical P structure

coding.The fig I.1 Hierarchical P structural coding can be generated using two referencing

schemes which are mentioned below.

I.2.1 Long term Referencing (MMCO Commands)

This scheme get enabled when referencePicMarking == 1.In this mode of

operation,DPB management is done by long term frames. LongTerm pictures have

longTermIndex correspoding to their layer.MMCO Commands get used for this operation.

Hierarchical P structure Coding Scheme

I-2

Figure I-2. Hierarchical P structure coding for Temporal Layer 4 with MMCO Commands

For interlaced coding, LT 0 and LT1 are toggled for base layer and LT2 and LT3 used for

higher(Enhancement) layers. If interlaceCodingType is coded using MRF(Most recent

field referencing scheme) then top layer will not be Non-Reference frame.

I.2.2 Short term Referencing (Sliding Window)

This scheme get enabled when referencePicMarking == 0.In this mode of

operation,DPB management is done by short term frames (Sliding Window) which is the

default DPB management in H264 Decoders.

I.3 Comparison of Referencing scheme

Short Term Referencing

Long Term Referencing

DPB management done using Sliding

window

DPB management done using MMCO

Commands

The overall DPB buffer requirement at the

decoder end will be higher

More efficient in DPB buffer requirement

The temporal layers cannot be identified

unless informed through SVC syntax or

though some external means.

LongTermIndex (LT) can be used to

identify the various temporal layers in

absence of SVC syntax

The decoding technique is relatively

simpler.

The decoding technique is relatively

complex

J-1

Appendix J

Mapping of Encoding Presets

J.1 Description:

User provided extended parameters of interface structure are taken in account only when

encodingPreset is XDM_USER_DEFINED. When encodingPreset is not

XDM_USER_DEFINED then encoder selects the appropriate values for extended parameter

as per the encodingPreset selected. This appendix explains the extended parameters values

that user need to set to meet the exact behavior of a particular encodingPreset like

XDM_HIGH_SPEED, XDM_MED_SPEED_HIGH_QUALITY etc. This table is useful when

user wants to change a particular parameter without modifying the other default values for a

particular preset.

encodingPreset

=XDM_MED_SPEED_HIGH_QU

ALITY

encodingPreset

=XDM_USER_DEFINED

rateControlParams.rateC

ontrolPreset

IH264_RATECONTROLPARAMS

_DEFAULT

IH264_RATECONTROLPARAM

S_DEFAULT

intercodingParams.inter

CodingPreset

IH264_INTERCODING_DEFAUL

IH264_INTERCODING_MED_S

PEED_HIGH_QUALITY

intraCodingParams.intra

CodingPreset

IH264_INTRACODING_DEFAUL

nalUnitControlParams.n

alUnitCodingPreset

IH264_NALU_CONTROL_DEFA

ULT

IH264_NALU_CONTROL_DEFA

ULT

sliceCodingParams.slice

CodingPreset

IH264_SLICECODING_DEFAUL

loopFilterParams.loopFilt

erPreset

IH264_LOOPFILTER_DEFAULT

fmoCodingParams.fmoC

odingPreset

IH264_FMOCODING_DEFAULT

vuiCodingParams.vuiCo

dingPreset

IH264_VUICODING_DEFAULT

stereoInfoParams.stereo

InfoPreset

IH264_STEREOINFO_DISABLE

framePackingSEIParam

s.framePackingSEIPrese

IH264_FRAMEPACK_SEI_DISA

BLE

IH264_FRAMEPACK_SEI_DISA

BLE

svcCodingParams.

svcExtensionFlag

IH264_SVC_EXTENSION_FLAG

_DISABLE

IH264_SVC_EXTENSION_FLA

G_DISABLE

interlaceCodingType

IH264_INTERLACE_DEFAULT

bottomFieldIntra

gopStructure

IH264ENC_GOPSTRUCTURE_D

EFAULT

IH264ENC_GOPSTRUCTURE_

DEFAULT

entropyCodingMode

IH264_ENTROPYCODING_DEF

AULT

IH264_ENTROPYCODING_DEF

AULT

transformBlockSize

IH264_TRANSFORM_DEFAULT

Mapping of Encoding Presets

J-2

log2MaxFNumMinus4

picOrderCountType

IH264_POC_TYPE_DEFAULT

enableWatermark

IDRFrameInterval

pConstantMemory

NULL

maxIntraFrameInterval

0x7FFFFFFF

debugTraceLevel

lastNFramesToLog

enableAnalyticinfo

enableGMVSei

constraintSetFlags

enableRCDO

enableLongTermRefFra

LTRPPeriod

numTemporalLayer

IH264_TEMPORAL_LAYERS_1

referencePicMarking

IH264_LONG_TERM_PICTURE

encodingPreset

=XDM_HIGH_SPEED

encodingPreset

=XDM_USER_DEFINED

rateControlParams.rateC

ontrolPreset

IH264_RATECONTROLPARAMS

_DEFAULT

IH264_RATECONTROLPARAM

S_DEFAULT

intercodingParams.inter

CodingPreset

IH264_INTERCODING_DEFAUL

IH264_INTERCODING_HIGH_S

PEED

intraCodingParams.intra

CodingPreset

IH264_INTRACODING_DEFAUL

IH264_INTRACODING_HIGH_S

PEED

nalUnitControlParams.n

alUnitCodingPreset

IH264_NALU_CONTROL_DEFA

ULT

IH264_NALU_CONTROL_DEFA

ULT

sliceCodingParams.slice

CodingPreset

IH264_SLICECODING_DEFAUL

loopFilterParams.loopFilt

erPreset

IH264_LOOPFILTER_DEFAULT

fmoCodingParams.fmoC

odingPreset

IH264_FMOCODING_DEFAULT

vuiCodingParams.vuiCo

dingPreset

IH264_VUICODING_DEFAULT

stereoInfoParams.stereo

InfoPreset

IH264_STEREOINFO_DISABLE

framePackingSEIParam

s.framePackingSEIPrese

IH264_FRAMEPACK_SEI_DISA

BLE

IH264_FRAMEPACK_SEI_DISA

BLE

svcCodingParams.

svcExtensionFlag

IH264_SVC_EXTENSION_FLAG

_DISABLE

IH264_SVC_EXTENSION_FLA

G_DISABLE

interlaceCodingType

IH264_INTERLACE_DEFAULT

bottomFieldIntra

gopStructure

IH264ENC_GOPSTRUCTURE_D

EFAULT

IH264ENC_GOPSTRUCTURE_

DEFAULT

entropyCodingMode

IH264_ENTROPYCODING_DEF

AULT

IH264_ENTROPYCODING_DEF

AULT

transformBlockSize

IH264_TRANSFORM_DEFAULT

log2MaxFNumMinus4

Mapping of Encoding Presets

J-3

picOrderCountType

IH264_POC_TYPE_DEFAULT

enableWatermark

IDRFrameInterval

pConstantMemory

NULL

maxIntraFrameInterval

0x7FFFFFFF

debugTraceLevel

lastNFramesToLog

enableAnalyticinfo

enableGMVSei

constraintSetFlags

enableRCDO

enableLongTermRefFra

LTRPPeriod

numTemporalLayer

IH264_TEMPORAL_LAYERS_1

referencePicMarking

IH264_LONG_TERM_PICTURE

Table 30 Prameter Mapping for various encoding presets

Note :

 When encoding preset is XDM_HIGH_SPEED then

IH264ENC_IntraCodingParams :: intraRefreshMethod is ignored.

 When encoding preset is XDM_HIGH_SPEED then

IH264ENC_RateControlParams :: enablePartialFrameSkip is ignored.

Mapping of Encoding Presets

J-4

This page is intentionally left blank

K-1

Appendix K

Region of Interest Encoding

This appendix explains the usage details of Region of Interest (ROI) support.

K.1 Description

Region of Interest (ROI) encoding allows encoding of specified regions in a frame with higher

quality as compared to other regions (“regions of non-interest”). This approach helps to

maximize the perceptual quality in ROIs and reduce overall bit-rate for the same perceived

quality.

K.2 Usage of ROI feature

K.2.1 Enabling ROI Support

The following parameter is added in IH264ENC_DynamicParams for enabling ROI support.

XDAS_Int32 enableROI;

Set it to non-zero value to enable ROI encoding. Set it to 0 to disable ROI encoding. Default

value is 0 (i.e., ROI is disabled).

ROI will be automatically disabled in case of full frame skip, and for skip macroblocks.

The following parameter has been added in IH264ENC_Status so that application can get to

know the status of enableROI parameter used for encoding.

XDAS_Int8 enableROI;

If the value of this variable is zero, then ROI is disabled. Otherwise, ROI is enabled.

K.2.2 Interface to Encoder

IH264ENC_RoiInput

║ Description

IH264ENC_RoiInput parameter is added in IH264ENC_InArgs to enable the application to

pass ROI-related information to the encoder.

║ Fields

Field

Data Type

Input/

Output

Description

listROI[IH264

ENC_MAX_ROI]

XDM_Rect

Input

The location of each ROI in terms of top left and

bottom right (x,y) co-ordinates.

The encoder supports a maximum of 36 ROIs in a

frame i.e., IH264ENC_MAX_ROI is 36.

Region of Interest Encoding

K-2

Field

Data Type

Input/

Output

Description

roiType[IH264

ENC_MAX_ROI]

XDAS_Int8

Input

Type of each ROI. The supported types are

FACE_OBJECT, BACKGROUND_OBJECT,

FOREGROUND_OBJECT,DEFAULT_OBJECT,

PRIVACY_MASK

numOfROI

XDAS_Int8

Input

Number of ROIs in the current frame.

roiPriority[I

H264ENC_MAX_R

OI]

XDAS_Int32

Input

Priority information of each ROI. Valid values

include all integers between -8 and 8, inclusive. A

higher value means that more importance will be

given to the ROI compared to other regions. In

other words, it determines the number of bits given

to ROI.

If the ROI type is PRIVACY_MASK then the mask

details are provided through this parameter.

If ROI is enabled in fixed Qp mode, This filed

holds the Qp of specified ROI.

Note:

 If the ROI is detected as FACE_OBJECT, then a guard band is added

around it. For all other ROI types, no guard band is added.

 If the roiType is set as PRIVACY_MASK then roiPriority will

specify the color of mask in 32 bits as mentioned below

Bits 0-7 : Cb value

Bits 8-15 : Cr value

Bits 16-31: Luma Value

Default value is “0” for GRAY color.

In case of Multiple masking regions in a frame, all masking regions will

use only the color of first masking region by ignoring other regions

colors given by user.

 In case of ROI + Multiple Privacy Masking scenario, privacy masking

region will be given first follwed by other ROI regions.

 In case of overlapping of ROIs with same type and different priority, the

priority used for the overlapped ROI will be the latest priority given by

the user.

 For example, if ROI region 3 (i.e., region corresponding to

listROI[2], roiType[2] and roiPriority[2]) and ROI region 5 (i.e.,

region corresponding to listROI[4], roiType[4] and roiPriority[4])

overlap, the value specified through roiPriority[4] will be used to

encode the overlapping MBs.

L-1

Appendix L

Watermarking SEI Message

L.1 Brief Description

With the rapid development of Internet technology, media data are used more and more

widely. This makes media data not only easy to be transmitted, but also easy to be copied

and spread out. Thus, the legal issue arises that some media data should be protected from

unauthorized users or operations.

Watermarking is a mechanism to add identity to a bitstream to help decoder to identify the

media content. For video security applications, it has become a de-facto requirement to

prevent tampering with video.

HDVICP2 H.264 codecs support a watermarking scheme at no loss in performance.

The proposed watermarking mechanism in HDVICP2 H.264 codec is shown in the following

figure.

Figure L-1. WaterMarking Mechanism

H.264 Encoder

Encoder accepts a 32-bit key

Encoder encrypts the key using the properties of bit-stream which can be obtained on

decoder side as well

Encrypted Key = fn(input_key, bit-stream parameters)

Encoder inserts the encrypted key in the form of user data unregistered SEI message in the

encoded stream

H.264 Decoder

Decodes the encrypted key in the form of user data unregistered SEI message

WaterMarking SEI Message

L-2

Decrypts the key using the properties of bit-stream

Provides the decrypted 32-bit key

System

Feeds the key on encoder, gets the key from decoder and compares them to identify content

tampering

Note

If the system has TI provided H.264 decoder on HDVICP2, it has capability to decrypt

the key.

For the system which doesn’t use H.264 decoder from TI, the decoder needs to dump

out few properties of bit-stream to decrypt the key

L.2 Usage of watermarking feature

L.2.1 Enabling Watermark Support

An additional create time flag has been added to enable watermarking feature. The following parameter

has been added to IH264ENC_Params.

enableWatermark

Set to non-zero value to enable encrypting of

watermark input key.

Watermarking would be disabled if set to 0.

L.2.2 Passing the input key to the Encoder

The following parameter has been added to IH264ENC_InArgs.

inputKey

32-bit input key. If the fields of a frame are

encoded in separate process calls, the input

key fed along with the later process call would

be used.

The encrypted key is inserted as part of a user data unregistered SEI message in the next process

call before encoding the next frame. In case of interlaced field coding, the encrypted key is

computed and inserted at the end of the second field. Only one encrypted key is inserted

for a pair of fields.

For the last frame, the watermark SEI message would be inserted at the end of MB

data.

Note:

 In case of interlaced, only the properties of the second field (in

decoding order) are considered.

 The number of bytes encoded for FillerDataBytes_NALU are not

accounted in the Total Bytes Generated (TBG)

 If the XDM_EncMode of the process call is

XDM_GENERATE_HEADER, then the inputKey fed in this particular

WaterMarking SEI Message

L-3

process call is ignored by the encoder.

L.3 Watermarking utilization with non-TI decoder

There are lot of applications, which might not use TI provided HDVICP2 based H.264

decoder and need to decrypt the key. This section helps to clarify the method of decryption

for such situations

L.3.1 Stream format for Encrypted Key

The encoder inserts the encrypted key in form of user data unregistered SEI message in the

encoded stream. The user data unregistered SEI message is identified as watermark

message by comparing the first 16 bytes of data with the hexadecimal pattern “e88345e0-

61ce-11e1-9ca8-0002a5d5c51b”. If the SEI data matches with the string, we take next 4

bytes of data as watermark encrypted key. Below figure shows the bit-stream content of SEI

message. In this figure the K3_K2_K1_K0 is the encrypted key being transmitted as part of

SEI message.

0x01 0x00 0x00 0x00

Bit Position

0xE8 0x14 0x05 0x06

0x61 0xE0 0x45 0x83

0x9C 0xE1 0x11 0xCE

0xA5 0x02 0x00 0xA8

K0 0x1B 0xC5 0xD5

K3 K2 K1

Base Address of User

Defined SEI

0x00

0x04

0x08

0x0C

0x10

0x14

0x18

0x1C

1624

By using this information, decoder should be able to decode the encrypted key as part of bit

stream. Now user has to decrypt the key.

L.3.2 Decyption of Encrypted Key

Using the properties of the bit stream like total bytes generated, number of slices, picture

type, etc., along with the encrypted key from the bit stream, one can retrieve the watermark

input key that was fed to the encoder.

For this, the following method is defined:

WaterMarking SEI Message

L-4

U32 RetrieveInputKey(

U32 encryptedKey,

U32 totalFrameBytes,

U32 diffIntraToNonIntraMBs,

U32 numSlices,

U32 *pSliceSizes,

U32 poc,

U32 numMBsInFrame,

U32 picType

);

Parameter Name

Data type

Input /

Output

Description

encryptedKey

XDAS_Int32

Input

The encrypted key parsed out from the

SEI message in the encoded stream

totalFrameBytes

XDAS_Int32

Input

Total bytes generated (till end of MB

data)

diffIntraToNonIntra

MBs

XDAS_Int32

Input

Absolute difference of number of intra

MBs and number of non-intra MBs

numSlices

XDAS_Int32

Input

Number of slices

pSliceSizes

XDAS_Int32

Input

Pointer to an array containing the

number of MBs in each slice

Poc

XDAS_Int32

Input

Picture Order Count

numMBsInFrame

XDAS_Int32

Input

Number of MBs in a frame

picType

XDAS_Int32

Input

Picture Type (values used correspond

to Table 7-6 in the H.264 standard

document)

Note

All the above statistics are accumulated over one frame in progressive cases and only of the

second field (decoding order) in interlaced cases.

Return Value

32-bit decrypted key is returned. The decrypted key is expected to match with the input key

fed to the encoder.

TI will supply implementation of RetrieveInputKey in a library form so user can use along with

his/her decoder.

M-1

Appendix M

N Frame Process Call Support

Encoder can support N frames processing in single process call. In this method user has to

provide all the necessary input parameter (like handle, InArgs, outArgs, InBufs, outBuf) for

each frame thorough newly defined XDM API for mullti frame process call. This method is

useful in reducing the thread overhead at Media Controller. This support can be utilized to

encode either N frames from single channel or one-one from N channels. Encoder is fully

unawared of the association between the frames.

M.1 Max value of numChannels (N):

Max value of number of frames that can be processed in single process call depends on max

input width among all the inputs.

Table 31 : Maximum number of channels supported for various resolutions

InputWidth

Max number of

channels (IPP

seq)

2048

1920

1280

720

640

352

176

M.2 Limitations when using N channel frame processing:

Followings are the limitation of N frame process call support

No B frames. IH264ENC_Params :: IVIDENC2_Params :: maxInterFrameInterval should be

one.

No DataSync / Low latency feature. IVIDENC2_Params :: outputDataMode/InputDataMOde

should be IVIDEO_ENTIREFRAME

No features such as ROI, FramePackSEI and StereoVideoSEI

N Frame Process Call Support

M-2

All channels should have same MV type (1mv or 4mv).

Minimum bit rate support will not be supported.

Insertion of End of stream and End of sequence will not be supported.

Encoding only Header in a process call is not supported i.e., XDM_EncMode sholud not be

XDM_GENERATE_HEADER.

In this method HDVICP acquire is done when first frame processing starts and release is

done after all the frame processing is finshed. All the acquire and release is done with the

last frame handle.

M.3 XDM interface for Multi Channel process call

#define IH264ENC_MAX_LENGTH_PROCESS_LIST (24)

typedef struct

{

IVIDENC2_Handle handle;

IVIDEO2_BufDesc *inBufs;

XDM2_BufDesc *outBufs;

IVIDENC2_InArgs *inArgs;

IVIDENC2_OutArgs *outArgs;

} IH264ENC_ProcessParams;

typedef struct

{

XDAS_Int32 numEntries;

XDAS_Int32 enableErrorCheck;

IH264ENC_ProcessParams

processParams[IH264ENC_MAX_LENGTH_PROCESS_LIST];

} IH264ENC_ProcessParamsList ;

typedef struct IH264ENC_Fxns

{

IVIDENC2_Fxns ividenc;

XDAS_Int32 (*processMulti)

(IH264ENC_ProcessParamsList *processList);

N Frame Process Call Support

M-3

} IH264ENC_Fxns;

New processMulti API has been be defined for this purpose. If for a channel, incorrect

parameter is passed then that particular channel will be skipped for encoding and reaming

channels will be encoded. And appropriate error bit will be set for skipped channel.

M.4 Steps to achieve N frame processing in single process call

Populate all the input parmeters like ( handle, inBufs, outBufs, inArgs, outArgs) for every

frame of input data.

Prepare the instance of the data type IH264ENC_ProcessParamsList.

Call the newly defined API processMulti with address of IH264ENC_ProcessParamsList as a

single argument.

After return from the multi process call, utilize the information updated by codec for each

frame in corresponding outArgs and outBufs.

Backward compatibility is maintiained after supporting N frame process call. Older API for

process call can be used for single frame processing in a process call.

N Frame Process Call Support

M-4

This page is intentionally left blank

N-1

Appendix N

Rate Control - High Fidelity Variable

Bitrate

This appendix provides an insight to the High Fidelity Variable Bitrate (HF-VBR) Rate Control

details of the encoder.

N.1 Description

The Rate Control algorithm (RC) in an encoder is required to

Ensure that the overall bits generated is meets the target bit-rate specified to the encoder.

Ensure that the overall perceptual video quality is maximized.

Conventional rate control algorithms are designed to achieve same average bitrate for all

durations of the video sequence High Fidelity Variable Bitrate (HF-VBR) Rate control

algorithm is a new rate control which adapts the instantaneous bit-rate to change at different

times based on the complexity of video at that point in time. HF-VBR reacts to the

instantaneous video complexity in the below way

Use higher-than-average bitrate in highly complex segment of video.

Use lower-than-average bitrate when in simple segment of video.

Figure N-1. Graph representing HF-VBR reaction to the video complexity.

RC-HFVBR

N-2

HF-VBR rate control allows the bitrate to change based on the complexity of the scene. The

rate control takes two inputs viz. 1. targetBitrate and 2. maxBitrate. For scene with normal

complexity, the RC operates at targetBitrate. When the scene complexity increases, the RC

increases the operating bitrate to a higher value. However it is not allowed to exceed

maxBitrate. In a longer duration, the overall bitrate achieved will be targetBitrate. HF-VBR

rate control is specially suited for video surveillance where one would intend to encode with

better quality when there is an increase in scene complexity

Note: HF-VBR: High Fidelity Variable Bitrate Rate control scheme is designed to achieve

targetBitrate in longer duration of time. Hence if one observes the overall bitrate (for long

duration), it will always be same as targetBitrate. It will go nearer to maxBitrate for short

duration in case the complexity increases.. The complexity estimate done by HF-VBR is

based on previous history. If you start and stop the video recording with high complex video

thoughout, HF-VBR will not treat it differently. This condition will be treated like normal VBR

and you will see the instantaneous bitrate does not go above the targetBitrate for the whole

duration. Hence in lab test, one has to be careful when inferring at results. They should let

video get recorded with static sequence for few seconds before changing to complex video

(like hand movement before camera etc).

N.2 Parameters & Configuration

The below mentioned parameters along with rcAlgo determine the behavior of rate control.

maxBitrate

This parameter limits the maximum bitrate which the rate control can achieve during the high

complexity duration of the video. The value of maxBitRate must be at least 1.5 times

targetBitRate. Only then HF-VBR: High Fidelity Variable Bitrate control is turned ON in the

encoder.

VBRDuration

This parameter is applicable to HF-VBR: High Fidelity Variable Bitrate. The time interval (in

seconds) during which encoder collects statistics related to the complexity of the video to

vary the instantaneous bitrate. Larger value of this parameter results in the rate control

algorithm reacting to complexity changes slowly. Allowed values are 1-3600 only

If VBRDuration is not set and only maxBitrate value is set to a value that is atlaest 1.5 times

the target bit rate then VBRDuration is taken by the encoder to be 8sec.

VBRsensitivity

This parameter is applicable to HF-VBR: High Fidelity Variable Bitrate. It controls the

sensitivity of the HF-VBR algorithm towards the complexity of video. It can take any value

from 0 to 8. A lower value signifies that maxBitrate will be used for very complex scene, in

case complexity increase is not high, HF-VBR will choose a bitrate between target bitrate and

maxBitrate. If set to higher value say 8, rate control tries to achieve maxBitrate even for small

complexity increase. Since complexity is a subjective measure, it is recommended to tune

this parameter based on user expectation.

RC-HFVBR

N-3

N.3 How to specify RC mode

rateControlPreset = IVIDEO_STORAGE (2)

rateControlPreset = IVIDEO_USER_DEFINED (5)

rcAlgo = IH264_RATECONTROL_PRC (0)

and

maxBitRate >= (1.5 x targetBitRate )

RC-HFVBR

N-4

This page is intentionally left blank

O-1

Appendix O

Gradual Decoder Refresh (GDR)- an

Error resilience Feature

This appendix provides a brief understanding of GDR and explains the usage details of GDR.

O.1 Description

GDR is a mechanism, which creates a bit-stream having spatially coded macroblocks (intra

macroblocks) in moving overlapped ( or non-overlapped) region of the picture. These intra

macrblocks should cover the entire picture region over few pictures.

O.1.1 There are two schemes in GDR

Vertical GDR

Horizontal GDR

Figure O-1. GDR Schemes.

Vertical GDR is implemented in HDVICP2 H.264 encoder.

GDR

O-2

Refreshed region is the portion which is error free in a given picture. In above figure,

refreshed region is outlined with red box. The number of pictures required to refresh the

decoder are mostly referred as GDR period. In above figure horizontal GDR has GDR Period

of 4 and Veritcal GDR has GDR Period of 3. Above figure shows, non overlapped GDR

scheme, but one can also have overlapped GRD scheme in which the intra region of Frame

N and Frame N+1 overlaps.

O.1.2 Constraints on Intra and Inter MBs in a GDR’d region

In Refreshed Region intra and inter MBs need to have below constraints

Intra MBs should only refer to spatial region which is part of refreshed region

Inter MBs should only refer the temporal region which is part of refreshed region in that

picture

O.2 Parameters and Configuration

O.2.1 Enabling GDR feature

Set intraCodingPreset as IH264_INTRACODING_USERDEFINED (1) and

Set intraRefreshMethod as IH264_INTRAREFRESH_GDR (4)

This method of intra refresh mechanism is not supported for 1)interlaced

sequences and 2) ‘B’ frame cases.

User can set GDR to be started at any picture during encoding session. To

start GDR, user need to set below parameter

IVIDENC2_InArgs::control = IH264ENC_CTRL_START_GDR

O.2.2 GDR control Parameters

Below are the contorl parameter to allow user to set GDR period and the control for

overlapped portion

intraRefreshRate

This parameter is treated/interpreted as the number of rows to be

intra refreshed per frame.

Supported values are [0, NonZero]

gdrOverlapRows

This parameter is the number of rows overlap between successive

GDR frames and value should be less than intraRefreshRate.

Supported values are [0, NonZero].

Based on network condition/feedback Encoder would be commanded through interface to

start GDR at any frame

H.264 Encoder 2.0 On HDVICP2 And Media Controller Based Platform User’s Guide H264 User

Navigation menu

Versions of this User Manual:

Views

Navigation