Software Manual

software-manual

User Manual:

Open the PDF directly: View PDF .
Page Count: 34

Download
Open PDF In Browser	View PDF

Joint Video Experts Team (JVET)
of ITU-T SG16 WP3 and ISO/IEC JTC1/SC29/WG11

Title:
Status:
Purpose:
Author(s):

Source:

Document: JVET-Software Manual

VTM Software Manual
Software AHG working document
Information
Frank Bossen
David Flynn
Xiang Li
Karl Sharman
Karsten Sühring
AHG chairs

frank@bossentech.com
dflynn@blackberry.com
xlxiangli@tencent.com
karl.sharman@eu.sony.com
karsten.suehring@hhi.fraunhofer.de

Abstract
This document is a user manual describing usage of the VTM reference software for the VVC project.
It applies to version 4.0 of the software.

Contents
1

General Information

Installation and compilation
2.1 Build instructions for plain CMake (suggested) . . . . . . . . . . . . . . . . . . . . . .
2.2 Build instructions for make . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
2.3 Tool Installation on Windows . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

3
3
4
4

Using the encoder
3.1 GOP structure table . . . . . .
3.2 Encoder parameters . . . . . .
3.3 Encoder SEI parameters . . .
3.4 Hardcoded encoder parameters

.
.
.
.

5
5
9
20
27

Using the decoder
29
4.1 General . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 29
4.2 Using the decoder analyser . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 30

Block statistics extension
5.1 Usage . . . . . . . . . . .
5.2 Block statistics file formats
5.3 Visualization . . . . . . .
5.4 Adding statistics . . . . . .

.
.
.
.

30
30
31
32
32

.
.
.
.

Supported compilers . . . . . .
GOP structure example . . . . .
File, I/O and source parameters.
Profile and level parameters . . .

.
.
.
.

. 3
. 7
. 9
. 11

List of Tables
1
2
3
4

Date saved: 2019-02-05

5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42

Unit definition parameters . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
Coding structure parameters . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
Motion estimation parameters . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
Mode decision parameters . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
Quantization parameters . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
Slice coding parameters . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
Deblocking filter parameters . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
Coding tools parameters . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
Rate control parameters . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
Encoder debug parameters . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
VUI parameters . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
Range Extensions (Version 2) tool parameters . . . . . . . . . . . . . . . . . . . . .
List of Version 1 and RExt SEI messages . . . . . . . . . . . . . . . . . . . . . . .
Buffering period SEI message encoder parameters . . . . . . . . . . . . . . . . . . .
Picture timing SEI message encoder parameters . . . . . . . . . . . . . . . . . . . .
Recovery point SEI message encoder parameters . . . . . . . . . . . . . . . . . . .
Tone mapping information SEI message encoder parameters . . . . . . . . . . . . .
Frame packing arrangement SEI message encoder parameters . . . . . . . . . . . . .
Display orientation SEI message encoder parameters . . . . . . . . . . . . . . . . .
Green Metadata SEI message encoder parameters . . . . . . . . . . . . . . . . . . .
Structure of pictures information SEI message encoder parameters . . . . . . . . . .
Active parameter sets SEI message encoder parameters . . . . . . . . . . . . . . . .
Decoding unit information SEI message encoder parameters . . . . . . . . . . . . .
Temporal sub-layer zero index SEI message encoder parameters . . . . . . . . . . .
Decoded picture hash SEI message encoder parameters . . . . . . . . . . . . . . . .
Scalable nesting SEI message encoder parameters . . . . . . . . . . . . . . . . . . .
Region refresh information SEI message encoder parameters . . . . . . . . . . . . .
No display SEI message encoder parameters . . . . . . . . . . . . . . . . . . . . . .
Time code SEI message encoder parameters . . . . . . . . . . . . . . . . . . . . . .
Mastering display colour volume SEI message encoder parameters . . . . . . . . . .
Segmented rectangular frame packing arrangement SEI message encoder parameters
Temporal motion-constrained tile sets SEI message encoder parameters . . . . . . .
Chroma resampling filter hint SEI message encoder parameters . . . . . . . . . . . .
Knee function SEI message encoder parameters . . . . . . . . . . . . . . . . . . . .
Colour remapping SEI message encoder parameters . . . . . . . . . . . . . . . . . .
CommonDef.h constants . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
Decoder options . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
Decoder options . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.

11
12
12
12
13
15
15
16
17
18
18
20
20
21
22
22
22
23
24
24
24
24
24
24
25
25
25
25
25
26
26
26
26
27
27
27
29
30

General Information

Reference software is being made available to provide a reference implementation of the HEVC standard being developed by the Joint Video Experts Team (JVET) regrouping experts from ITU-T SG 16
and ISO/IEC SC29 WG11. One of the main goals of the reference software is to provide a basis upon
which to conduct experiments in order to determine which coding tools provide desired coding performance. It is not meant to be a particularly efficient implementation of anything, and one may notice its
apparent unsuitability for a particular use. It should not be construed to be a reflection of how complex
a production-quality implementation of a future VVC standard would be.
This document aims to provide guidance on the usage of the reference software. It is widely suspected to
be incomplete and suggestions for improvements are welcome. Such suggestions and general inquiries
may be sent to the general JVET email reflector on https://lists.rwth-aachen.de/postorius/lists/jvet.lists.
rwth-aachen.de/ (registration required).
2

Date saved: 2019-02-05

Bug reporting
Bugs should be reported on the issue tracker set up at:
https://jvet.hhi.fraunhofer.de/trac/vvc/

Installation and compilation

The software may be retrieved from the GitLab server located at:
https://vcgit.hhi.fraunhofer.de/jvet/VVCSoftware VTM
Table 1 lists the compiler environments and versions for which building the software is tested.
Note that the software makes use of C++11 language features, which may not be available in older
compilers.
Table 1: Supported compilers
Compiler environment

MS Visual Studio
GCC
Xcode/clang

Versions

2015 and 2017
5.4 and 7.3
latest

By default the software is built as 64-bit binaries to be used on a 64-bit OS. This allows the software to
use more than 2GB of RAM.
The software uses CMake to create the needed build files.

2.1

Build instructions for plain CMake (suggested)

Note: A working CMake installation is required for building the software.
CMake generates configuration files for the compiler environment/development environment on each
platform. The following is a list of examples for Windows (MS Visual Studio), macOS (Xcode) and
Linux (make).
Open a command prompt on your system and change into the root directory of this project.
Create a build directory in the root directory:
mkdir build

Use one of the following CMake commands, based on your platform. Feel free to change the commands
to satisfy your needs.
Windows Visual Studio 2015 64 Bit:
cd build
cmake .. -G "Visual Studio 14 2015 Win64"

Then open the generated solution file in MS Visual Studio.
macOS Xcode:
cd build
cmake .. -G "Xcode"

Date saved: 2019-02-05

Then open the generated work space in Xcode.
Linux
For generating Linux Release Makefile:
cd build
cmake .. -DCMAKE_BUILD_TYPE=Release

For generating Linux Debug Makefile:
cd build
cmake .. -DCMAKE_BUILD_TYPE=Debug

Then type
make -j

to build the software.
For more details, refer to the CMake documentation: https://cmake.org/cmake/help/latest/

2.2

Build instructions for make

Note: The build instructions in this section require the make tool and Python to be installed, which are
part of usual Linux and macOS environments. See section 2.3 for installation instruction for Python and
GnuWin32 on Windows.
Open a command prompt on your system and change into the root directory of this project.
To use the default system compiler simply call:
make all

For MSYS2 and MinGW: Open an MSYS MinGW 64-Bit terminal and change into the root directory of
this project.
Call:
make all toolset=gcc

2.3

Tool Installation on Windows

Download CMake: http://www.cmake.org/ and install it.
Python and GnuWin32 are not mandatory, but they simplify the build process for the user.
Python
GnuWin32

https://www.python.org/downloads/release/python-371/
https://sourceforge.net/projects/getgnuwin32/files/getgnuwin32/0.6.30/GetGnuWin32-0.6.3.exe/download

To use MinGW, install MSYS2: http://repo.msys2.org/distrib/msys2-x86 64-latest.exe
Installation instructions: https://www.msys2.org/
Install the needed toolchains:

Date saved: 2019-02-05

pacman -S --needed base-devel mingw-w64-i686-toolchain
,→
mingw-w64-x86_64-toolchain git subversion mingw-w64-i686-cmake
,→
mingw-w64-x86_64-cmake

Using the encoder
TAppEncoder

[--help] [-c config.cfg] [--parameter=value]

Option
--help
-c
--parameter=value

Description
Prints parameter usage.
Defines configuration file to use. Multiple configuration files
may be used with repeated –c options.
Assigns value to a given parameter as further described below.
Some parameters are also supported by shorthand “–opt value”.
These are shown in brackets after the parameter name in the
tables of this document

Sample configuration files are provided in the cfg/ folder. Parameters are defined by the last value
encountered on the command line. Therefore if a setting is set via a configuration file, and then a
subsequent command line parameter changes that same setting, the command line parameter value will
be used.

3.1

GOP structure table

Defines the cyclic GOP structure that will be used repeatedly throughout the sequence. The table should
contain GOPSize lines, named Frame1, Frame2, etc. The frames are listed in decoding order, so Frame1
is the first frame in decoding order, Frame2 is the second and so on. Among other things, the table
specifies all reference pictures kept by the decoder for each frame. This includes pictures that are used
for reference for the current picture as well as pictures that will be used for reference in the future.
The encoder will not automatically calculate which pictures have to be kept for future references, they
must be specified. Note that some specified reference frames for pictures encoded in the very first GOP
after an IDR frame might not be available. This is handled automatically by the encoder, so the reference
pictures can be given in the GOP structure table as if there were infinitely many identical GOPs before the
current one. Each line in the table contains the parameters used for the corresponding frame, separated
by whitespace:
Type: Slice type, can be either I, P or B.
POC: Display order of the frame within a GOP, ranging from 1 to GOPSize.
QPOffset: QP offset is added to the QP parameter to set the final QP value to use for this frame.
QPOffsetModelOff: Offset parameter to a linear model to adjust final QP based on QP + QPoffset.
QPOffsetModelScale: Scale parameter to a linear model to adjust final QP based on QP + QPoffset.
SliceCbQPOffset: The slice-level Cb QP offset.
SliceCrQPOffset: The slice-level Cr QP offset.
QPFactor: Weight used during rate distortion optimization. Higher values mean lower quality
and less bits. Typical range is between 0.3 and 1.
5

Date saved: 2019-02-05

tcOffsetDiv2: In-loop deblocking filter parameter tcOffsetDiv2 is added to the base parameter
LoopFilterTcOffset div2 to set the final tc offset div2 parameter for this picture signalled in the
slice segment header. The final value of tc offset div2 shall be an integer number in the range
−6..6.
betaOffsetDiv2: In-loop deblocking filter parameter betaOffsetDiv2 is added to the base parameter LoopFilterBetaOffset div2 to set the final beta offset div2 parameter for this picture signalled
in the slice segment header. The final value of beta offset div2 shall be an integer number in the
range −6..6.
temporal id: Temporal layer of the frame. A frame cannot predict from a frame with a higher
temporal id. If a frame with higher temporal IDs is listed among a frame’s reference pictures, it is
not used, but is kept for possible use in future frames.
num ref pics active: Size of reference picture lists L0 and L1, indicating how many reference
pictures in each direction that are used during coding.
num ref pics: The number of reference pictures kept for this frame. This includes pictures that
are used for reference for the current picture as well as pictures that will be used for reference in
the future.
reference pictures: A space-separated list of num ref pics integers, specifying the POC of the
reference pictures kept, relative the POC of the current frame. The picture list shall be ordered,
first with negative numbers from largest to smallest, followed by positive numbers from smallest
to largest (e.g. -1 -3 -5 1 3). Note that any pictures not supplied in this list will be discarded
and therefore not available as reference pictures later.
predict: Defines the value of the syntax element inter ref pic set prediction flag. A value of 0
indicates that the reference picture set is encoded without inter RPS prediction and the subsequent
parameters deltaRIdx−1, deltaRPS, num ref idcs and Reference idcs are ignored and do not need
to be present. A value of 1 indicates that the reference picture set is encoded with inter prediction
RPS using the subsequent parameters deltaRIdx−1, deltaRPS, num ref idcs and Reference idcs
in the line. A value of 2 indicates that the reference picture set is encoded with inter RPS but only
the deltaRIdx−1 parameters is needed. The deltaRPS, num ref idcs and Reference idcs values
are automatically derived by the encoder based on the POC and refPic values of the current line
and the RPS pointed to by the deltaRIdx−1 parameters.
deltaRIdx−1: The difference between the index of the curent RPS and the predictor RPS minus
1.
deltaRPS: The difference between the POC of the predictor RPS and POC the current RPS.
num ref idcs: The number of ref idcs to encode for the current RPS. The value is equal to the
value of num ref pics of the predictor RPS plus 1.
reference idcs: A space-separated list of num ref idcs integers, specifying the ref idcs of the inter
RPS prediction. The value of ref idcs may be 0, 1 or 2 indicating that the reference picture is
a reference picture used by the current picture, a reference picture used for future picture or not
a reference picture anymore, respectively. The first num ref pics of ref idcs correspond to the
Reference pictures in the predictor RPS. The last ref idcs corresponds to the predictor picture.
For example, consider the coding structure of Figure 1. This coding structure is of size 4. The pictures
are listed in decoding order. Frame1 shall therefore describe picture with POC = 4. It references picture
0, and therefore has −4 as a reference picture. Similarly, Frame2 has a POC of 2, and since it references
pictures 0 and 4, its reference pictures are listed as -2 2. Frame3 is a special case: even though it
only references pictures with POC 0 and 2, it also needs to include the picture with POC 4, which must
be kept in order to be used as a reference picture in the future. The reference picture list for Frame3
therefore becomes -1 1 3. Frame4 has a POC of 3 and its list of reference pictures is -1 1.

Date saved: 2019-02-05

Figure 1: A GOP structure
B

B
B

POC

Decode Order

Inter RPS prediction may be used for Frame2, Frame3 and Frame4, hence the predict parameter is set
to 1 for these frames. Frame2 uses Frame1 as the predictor hence the deltaRIdx−1 is 0. Similarly for
Frame3 and Frame4 which use Frame2 and Frame3 as predictors, respectively. The deltaRPS is equal
to the POC of the predictor minus the POC of the current picture, therefore the deltaRPS for Frame2 is
4 − 2 = 2, for Frame3 is 2 − 1 = 1 and for Frame4 is 1 − 3 = −2.
In Frame2, reference pictures with POC 0 and 2 are used, so the reference idcs for Frame2 are 1 1
indicating that the reference picture, −4, in Frame1 is still a reference picture in Frame2 and Frame1
is also a reference picture in Frame2. The reference idcs for Frame3 are 1 1 1. The first and second
“1”s indicating that the reference pictures “−2 2” in Frame2 are still reference pictures in Frame3 and
the last “1” indicating that Frame2 is also a reference picture in Frame3. In Frame 4, the reference idcs
are 0 1 1 0. The first “0” indicates that the reference pictures “-1” in Frame 3 is no longer a reference
picture in Frame4. The next two “1”s indicate that the reference pictures “1 3” are now reference pictures
of Frame4. The final “0” indicates that Frame3 is not a reference picture.
In order to specify this to the encoder, the parameters in Table 2 could be used.
Table 2: GOP structure example
Frame1

Frame2

Frame3

Frame4

P
4
1
0.0
0.0
0
0
0.5
0
0
0
1
1
−4
0

B
2
2
0.0
0.0
0
0
0.5
1
0
1
1
2
−2 2
1
0
2
2
11

B
1
3
0.0
0.0
0
0
0.5
2
0
2
1
3
−1 1 3
1
0
1
3
111

B
3
3
0.0
0.0
0
0
0.5
2
0
2
1
2
−1 1
1
0
−2
4
0110

Type
POC
QPOffset
QPOffsetModelOff
QPOffsetModelScale
SliceCbQPOffset
SliceCrQPOffset
QPfactor
tcOffsetDiv2
betaOffsetDiv2
temporal id
num ref pics active
num ref pics
reference pictures
predict
deltaRIdx−1
deltaRPS
num ref idcs
reference idcs

Here, the frames used for prediction have been given higher quality by assigning a lower QP offset.
Also, the non-reference frames have been marked as belonging to a higher temporal layer, to make it
possible to decode only every other frame. Note: each line should contain information for one frame, so
this configuration would be specified as:
Frame1:
Frame2:
Frame3:
Frame4:

P
B
B
B

4
2
1
3

1
2
3
3

0
0
0
0

0.5
0.5
0.5
0.5

0
1
2
2

0
0
0
0

0
1
2
2

1
1
1
1

1
2
3
2

-4
-2
-1
-1

0
2 1 0 2 2 1 1
1 3 1 0 1 3 1 1 1
1 1 0 -2 4 0 1 1 0

Date saved: 2019-02-05

The values of deltaRIdx−1, deltaRPS, num ref idcs and reference idcs of FrameK can be derived from
the POC value of FrameK and the POC, num ref pics and reference pictures values of FrameM , where
K is the index of the RPS to be inter coded and the M is the index of the reference RPS, as follows.
deltaRIdxK − 1 ← K − M − 1 ;
deltaRPSK ← POCM − POCK ;
num ref idcsK ← num ref picsM + 1 ;
for j ← 0 to num ref picsM do
for i ← 0 to num ref idcsK do
if reference picturesM,j + deltaRPSK == reference picturesK,i then
if reference picturesK,i is used by the current frame then
reference idcsK,j = 1;
;
else reference idcsK,j = 2;
;
else
reference idcsK [j] = 0 ;
end
end
end
/* reference picturesM,num ref picsM does not exist and is assumed to be
0
*/
Note: The above (automatic) generation of the inter RPS parameter values has been integrated into the
encoder, and is activated by the value of predict = 2 followed by the value of deltaRIdx−1, only, as
described above.

Date saved: 2019-02-05

3.2

Encoder parameters

Shorthand alternatives for the parameter that can be used on the command line are shown in brackets after the parameter
name.
Table 3: File, I/O and source parameters.
Option

Default

Description

InputFile (-i)

Specifies the input video file.
Video data must be in a raw 4:2:0, or 4:2:2 planar format, 4:4:4 planar format
(Y0 CbCr, RGB or GBR), or in a raw 4:0:0 format.
Note: When the bit depth of samples is larger than 8, each sample is encoded in 2
bytes (little endian, LSB-justified).

BitstreamFile (-b)

Specifies the output coded bit stream file.

ReconFile (-o)

Specifies the output locally reconstructed video file.

SourceWidth (-wdt)
SourceHeight (-hgt)

0
0

Specifies the width and height of the input video in luma samples.

InputBitDepth

Specifies the bit depth of the input video.

MSBExtendedBitDepth

Extends the input video by adding MSBs of value 0. When 0, no extension is applied
and the InputBitDepth is used.
The MSBExtendedBitDepth becomes the effective file InputBitDepth for subsequent
processing.

InternalBitDepth

Specifies the bit depth used for coding. When 0, the setting defaults to the value of
the MSBExtendedBitDepth.
If the input video is a different bit depth to InternalBitDepth, it is automatically converted by:

Pel ∗ 2InternalBitDepth
2MSBExtendedBitDepth
Note: The effect of this option is as if the input video is externally converted to the
MSBExtendedBitDepth and then to the InternalBitDepth and then coded with this
value as InputBitDepth. The codec has no notion of different bit depths.

OutputBitDepth

Specifies the bit depth of the output locally reconstructed video file. When 0, the
setting defaults to the value of InternalBitDepth. Note: This option has no effect on
the decoding process.

InputBitDepthC
MSBExtendedBitDepthC
InternalBitDepthC
OutputBitDepthC

0
0
0
0

Specifies the various bit-depths for chroma components. These only need to be specified if non-equal luma and chroma bit-depth processing is required. When 0, the
setting defaults to the corresponding non-Chroma value.

InputColourSpaceConvert

The colour space conversion to apply to input video. Permitted values are:
UNCHANGED
No colour space conversion is applied
YCbCrToYCrCb
Swap the second and third components
YCbCrtoYYY
Set the second and third components to the values
in the first
RGBtoGBR
Reorder the three components
If no value is specified, no colour space conversion is applied. The list may eventually
also include RGB to YCbCr or YCgCo conversions.

SNRInternalColourSpace

false

When this is set true, then no colour space conversion is applied prior to PSNR calculation, otherwise the inverse of InputColourSpaceConvert is applied.

OutputInternalColourSpace

false

When this is set true, then no colour space conversion is applied to the reconstructed
video, otherwise the inverse of InputColourSpaceConvert is applied.

InputChromaFormat

420

Specifies the chroma format used in the input file. Permitted values (depending on
the profile) are 400, 420, 422 or 444.

Specifies the chroma format to use for processing. Permitted values (depending on
the profile) are 400, 420, 422 or 444; the value of 0 indicates that the value of InputChromaFormat should be used instead.

false

When 0, the PSNR output is a linear average of the frame PSNRs; when 1, additional PSNRs are output which are formed from the average MSE of all the frames.
The latter is useful when coding near-losslessly, where occasional frames become
lossless.

ChromaFormatIDC (-cf)

MSEBasedSequencePSNR

Continued...

Date saved: 2019-02-05

Table 3: File, I/O and source parameters. (Continued)
Option

Default

PrintFrameMSE

false

When 1, the Mean Square Error (MSE) values of each frame will also be output
alongside the default PSNR values.

PrintSequenceMSE

false

When 1, the Mean Square Error (MSE) values of the entire sequence will also be
output alongside the default PSNR values.

SummaryOutFilename

false

Filename to use for producing summary output file. If empty, do not produce a file.

SummaryPicFilenameBase

false

Base filename to use for producing summary picture output files. The actual filenames used will have I.txt, P.txt and B.txt appended. If empty, do not produce a
file.

SummaryVerboseness

false

Specifies the level of the verboseness of the text output.

CabacZeroWordPaddingEnabled

false

When 1, CABAC zero word padding will be enabled. This is currently not the default
value for the setting.

ConformanceWindowMode

Specifies how the parameters related to the conformance window are interpreted
(cropping/padding). The following modes are available:
0
No cropping / padding
1
Automatic padding to the next minimum CU size
2
Padding according to parameters HorizontalPadding and VerticalPadding
3
Cropping according to parameters ConfWinLeft, ConfWinRight, ConfWinTop and ConfWinBottom

HorizontalPadding (-pdx)
VerticalPadding (-pdy)

Specifies the horizontal and vertical padding to be applied to the input video in luma
samples when ConformanceWindowMode is 2. Must be a multiple of the chroma
resolution (e.g. a multiple of two for 4:2:0).

ConfWinLeft
ConfWinRight
ConfWinTop
ConfWinBottom

Specifies the horizontal and vertical cropping to be applied to the input video in luma
samples when ConformanceWindowMode is 3. Must be a multiple of the chroma
resolution (e.g. a multiple of two for 4:2:0).

FrameRate (-fr)

Specifies the frame rate of the input video.
Note: This option only affects the reported bit rates.

FrameSkip (-fs)

Specifies a number of frames to skip at beginning of input video file.

FramesToBeEncoded (-f)

Specifies the number of frames to be encoded (see note regarding TemporalSubsampleRatio). When 0, all frames are coded.

TemporalSubsampleRatio (-ts)

Temporally subsamples the input video sequence. A value of N will skip (N − 1)
frames of input video after each coded input video frame. Note the FramesToBeEncoded does not account for the temporal skipping of frames, which will reduce the
number of frames encoded accordingly. The reported bit rates will be reduced and
VUI information is scaled so as to present the video at the correct speed. The minimum and default value is 1.

FieldCoding

Description

false

When 1, indicates that field-based coding is to be applied.

TopFieldFirst (-Tff)

Indicates the order of the fields packed into the input frame. When 1, the top field is
temporally first.

ClipInputVideoToRec709Range

If 1 then clip input video to the Rec. 709 Range on loading when InternalBitDepth is
less than MSBExtendedBitDepth.

ClipOutputVideoToRec709Range

If 1 then clip output video to the Rec. 709 Range on saving when OutputBitDepth is
less than InternalBitDepth.

EfficientFieldIRAPEnabled

Enable to code fields in a specific, potentially more efficient, order.

HarmonizeGopFirstFieldCoupleEnabled

Enables harmonization of Gop first field couple.

AccessUnitDelimiter

Add Access Unit Delimiter NAL units between all Access Units.

Date saved: 2019-02-05

Table 4: Profile and level parameters
Option

Default

Description

Profile

none

Specifies the profile to which the encoded bitstream complies.
Valid HEVC Ver. 1 values are: none, main, main10, main-still-picture
Valid HEVC Ver. 2 (RExt) values are: main-RExt, high-throughput-RExt,
monochrome, monochrome12, monochrome16, main12, main 422 10, main 422 12, main 444, main 444 10, main 444 12, main 444 16, main intra, main 10 intra,
main 12 intra, main 422 10 intra, main 422 12 intra, main 444 intra, main 444 10 intra, main 444 12 intra, main 444 16 intra.
When main-RExt is specified, the constraint flags are either manually specified, or
calculated via the other supplied settings.
Compatibility flags are automatically determined according to the profile. NB: There
is currently only limited validation that the encoder configuration complies with the
profile, level and tier constraints.

Level

none

Specifies the level to which the encoded bitstream complies. Valid values are: none,
1, 2, 2.1, 3, 3.1, 4, 4.1, 5, 5.1, 5.2, 6, 6.1, 6.2, 8.5
NB: There is currently only limited validation that the encoder configuration complies
with the profile, level and tier constraints.

Tier

main

Specifies the level tier to which the encoded bitsream complies. Valid values are:
main, high.
NB: There is currently only limited validation that the encoder configuration complies
with the profile, level and tier constraints.

MaxBitDepthConstraint

For –profile=main-RExt, specifies the value to use to derive the genwhen 0, use
eral max bit depth constraint flags for RExt profiles;
max(InternalBitDepth, InternalBitDepthC)

MaxChromaFormatConstraint

For –profile=main-RExt, specifies the chroma-format to use for the general profile
constraints for RExt profiles; when 0, use the value of ChromaFormatIDC.

IntraConstraintFlag

false

For –profile=main-RExt, specifies the value of general intra constraint flag to use for
RExt profiles.

OnePictureOnlyConstraintFlag

false

For –profile=main-RExt, specifies the value of general one picture only constraint flag to use for RExt profiles.

LowerBitRateConstraintFlag

true

Specifies the value of general lower bit constraint flag to use for RExt profiles.

ProgressiveSource

false

Specifies the value of general progressive source flag

InterlacedSource

false

Specifies the value of general interlaced source flag

NonPackedSource

false

Specifies the value of general non packed constraint flag

FrameOnly

false

Specifies the value of general frame only constraint flag

Table 5: Unit definition parameters
Option

Default

Description

MaxCUWidth

Defines the maximum CU width.

MaxCUHeight

Defines the maximum CU height.

MaxCUSize (-s)

Defines the maximum CU size.

MaxPartitionDepth (-h)

Defines the depth of the CU tree.

QuadtreeTULog2MaxSize

6
(= log2 (64))

Defines the Maximum TU size in logarithm base 2.

QuadtreeTULog2MinSize

2
(= log2 (4))

Defines the Minimum TU size in logarithm base 2.

QuadtreeTUMaxDepthIntra

Defines the depth of the TU tree for intra CUs.

QuadtreeTUMaxDepthInter

Defines the depth of the TU tree for inter CUs.

Date saved: 2019-02-05

Table 6: Coding structure parameters
Option
IntraPeriod (-ip)

Default

Description
Specifies the intra frame period. A value of −1 implies an infinite period.

−1

DecodingRefreshType (-dr)

Specifies the type of decoding refresh to apply at the intra frame period picture.
0
Applies an I picture (not a intra random access point).
1
Applies a CRA intra random access point (open GOP).
2
Applies an IDR intra random access point (closed GOP).
3
Use recovery point SEI messages to indicate random access.

GOPSize (-g)

Specifies the size of the cyclic GOP structure.

FrameN

Multiple options that define the cyclic GOP structure that will be used repeatedly
throughout the sequence. The table should contain GOPSize elements.
See section 3.1 for further details.

Table 7: Motion estimation parameters
Option

Default

Description

FastSearch

Enables or disables the use of a fast motion search.
0
Full search method
1
Fast search method - TZSearch
2
Predictive motion vector fast search method
3
Extended TZSearch method

SearchRange (-sr)

Specifies the search range used for motion estimation.
Note: the search range is defined around a predictor. Motion vectors derived by the
motion estimation may thus have values larger than the search range.

BipredSearchRange

Specifies the search range used for bi-prediction refinement in motion estimation.

ClipForBiPredMEEnabled

Enables clipping in the Bi-Pred ME, which prevents values over- or under-flowing. It
is usually disabled to reduce encoder run-time.

FastMEAssumingSmootherMVEnabled

Enables fast ME assuming a smoother MV.

HadamardME

true

Enables or disables the use of the Hadamard transform in fractional-pel motion estimation.
0
SAD for cost estimation
1
Hadamard for cost estimation

ASR

false

Enables or disables the use of adaptive search ranges, where the motion search range
is dynamically adjusted according to the POC difference between the current and the
reference pictures.

SearchRange0 = Round SearchRange ∗ ADAPT SR SCALE ∗

MaxNumMergeCand

Specifies the maximum number of merge candidates to use.

DisableIntraInInter

Flag to disable intra PUs in inter slices.

abs(POCcur−POCref)
RateGOPSize

Table 8: Mode decision parameters
Option
LambdaModifierN (-LMN )

LambdaModifierI (-LMI)

Default
1.0

Description
Specifies a value that is multiplied with the Lagrange multiplier λ, for use in the
rate-distortion optimised cost calculation when encoding temporal layer N . If LambdaModifierI is specified, then LambdaModifierI will be used for intra pictures.
N may be in the range 0 (inclusive) to 7 (exclusive).
Specifies one or more of the LambdaModifiers to use intra pictures at each of the
temporal layers. If not present, then the LambdaModifierN settings are used instead.
If the list of values (comma or space separated) does not include enough values for
each of the temporal layers, the last value is repeated as required.
Continued...

Date saved: 2019-02-05

Table 8: Mode decision parameters (Continued)
Option
IQPFactor (-IQF)

Default

Description

-1

Specifies the QP factor to be used for intra pictures during the lambda computation.
(The values specified in the GOP structure are only used for inter pictures). If negative (default), the following equation is used to derive the value:
IQPf actor = 0.57 ∗ (1.0 − M ax(0.5, M in(0.0, 0.05 ∗ s)))
where s = Int(isF ield?(GS − 1)/2 : GS − 1) and GS is the gop size.

ECU

false

Enables or disables the use of early CU determination. When enabled, skipped CUs
will not be split further.

CFM

false

Enables or disables the use of Cbf-based fast encoder mode. When enabled, once
a 2Nx2N CU has been evaluated, if the RootCbf is 0, further PU splits will not be
evaluated.

ESD

false

Enables or disables the use of early skip detection. When enabled, the skip mode will
be tested before any other.

FEN

Controls the use of different fast encoder coding tools. The following tools are supported in different combinations:
a
In the SAD computation for blocks having size larger than 8, only the lines
of even rows in the block are considered.
b
The number of iterations used in the bi-directional motion vector refinement
in the motion estimation process is reduced from 4 to 1.
Depending on the value of the parameter, the following combinations are supported:
0
Disable all modes
1
Use both a & b tools
2
Use only tool b
3
Use only tool a

FDM

true

Enables or disables the use of fast encoder decisions for 2Nx2N merge mode. When
enabled, the RD cost for the merge mode of the current candidate is not evaluated if
the merge skip mode was the best merge mode for one of the previous candidates.

RD-penalty for 32x32 TU for intra in non-intra slices. Enabling this parameter can
reduce the visibility of CU boundaries in the coded picture.
0
No RD-penalty
1
RD-penalty
2
Maximum RD-penalty (no 32x32 TU)

RDpenalty

Table 9: Quantization parameters
Option
QP (-q)
IntraQPOffset

Default

Description

30.0

Specifies the base value of the quantization parameter. If it is non-integer, the QP is
switched once during encoding.

Specifies a QP offset from the base QP value to be used for intra frames.

LambdaFromQpEnable

false

When enabled, the λ, which is used to convert a cost in bits to a cost in distortion
terms, is calculated as:
λ = qpF actor × 2qp+6∗(bitDepthLuma−8)−12 , where qp is the slice QP and
qpF actor is calculated as follows:
= IQF
if IQF >= 0 and slice is a periodic intra slice
= 0.57 × λscale
if slice is a non-periodic intra slice
= value from GOP table
otherwise
where IQF is the value specified using the IntraQPFactor option, and where λscale
is:
1
if LambdaFromQpEnable=true
1.0 − max(0, min(0.5, 0.05 ∗ B)) if LambdaFromQpEnable=false
where B is the number of B frames.
If LambdaFromQpEnable=false, then the λ is also subsequently scaled for non-toplevel hiearchical depths, as follows:
λ = λbase × max(2, min(4, (sliceQP − 12)/6))
In addition, independent on the IntraQPFactor, if HadamardME=false, then for an
inter slice the final λ is scaled by a factor of 0.95.

CbQpOffset (-cbqpofs)
CrQpOffset (-crqpofs)

0
0

Global offset to apply to the luma QP to derive the QP of Cb and Cr respectively.
These options correspond to the values of cb qp offset and cr qp offset, that are
transmitted in the PPS. Valid values are in the range [−12, 12].
Continued...

Date saved: 2019-02-05

Table 9: Quantization parameters (Continued)
Option
LumaLevelToDeltaQPMode

LumaLevelToDeltaQPMaxValWeight

Default

Description

Luma-level based Delta QP modulation.
0
not used
1
Based on CTU average
2
Based on Max luma in CTU

1.0

Weight of per block maximum luma value when LumaLevelToDeltaQPMode=2.

LumaLevelToDeltaQPMappingLuma

Specify luma values to use for the luma to delta QP mapping instead of using default
values. Default values are: 0, 301, 367, 434, 501, 567, 634, 701, 767, 834.

LumaLevelToDeltaQPMappingDQP

Specify DQP values to use for the luma to delta QP mapping instead of using default
values. Default values are: -3, -2, -1, 0, 1, 2, 3, 4, 5, 6.

WCGPPSEnable

Enable the WCG PPS modulation of the chroma QP, rather than the slice, which,
unlike slice-level modulation, allows the deblocking process to consider the adjustment. To use, specify a fractional QP: the first part of the sequence will use
qpc = f loor(QP ) in the following calculation and PPS-0; the second part of the
sequence will use qpc = ceil(QP ) and PPS-1. The chromaQp that is then stored
in the PPS is given as: clip(round(W CGP P SXXQpScale ∗ baseCQp) +
XXQpOf f set) where baseCQp = (W CGP P SChromaQpScale ∗ qpc +
W CGP P SChromaQpOf f set). Note that the slices will continue to have a delta
QP applied.

WCGPPSChromaQpScale

0.0

Scale parameter for the linear chroma QP offset mapping used for WCG content.

WCGPPSChromaQpOffset

0.0

Offset parameter for the linear chroma QP offset mapping used for WCG content.

WCGPPSCbQpScale
WCGPPSCrQpScale

1.0

Per chroma component QP scale factor depending on capture and representation color
space. For Cb component with BT.2020 container use 1.14; for BT.709 material and
1.04 for P3 material. For Cr component with BT.2020 container use 1.79; for BT.709
material and 1.39 for P3 material.

SliceChromaQPOffsetPeriodicity

Defines the periodicity for inter slices that use the slice-level chroma QP offsets, as
defined by SliceCbQpOffsetIntraOrPeriodic and SliceCrQpOffsetIntraOrPeriodic. A
value of 0 disables the periodicity. It is intended to be used in low-delay configurations where an regular intra period is not defined.

SliceCbQpOffsetIntraOrPeriodic
SliceCrQpOffsetIntraOrPeriodic

Defines the slice-level QP offset to be used for intra slices, or once every ’SliceChromaQPOffsetPeriodicity’ pictures.

MaxCuDQPDepth (-dqd)

Defines maximum depth of a minimum CuDQP for sub-LCU-level delta QP. MaxCuDQPDepth shall be greater than or equal to SliceGranularity.

RDOQ

true

Enables or disables rate-distortion-optimized quantization for transformed TUs.

RDOQTS

true

Enables or disables rate-distortion-optimized quantization for transform-skipped
TUs.

SelectiveRDOQ

false

Enables or disables selective rate-distortion-optimized quantization. A simple quantization is use to pre-analyze, whether to bypass the RDOQ process or not. If all
the coefficients are quantized to 0, the RDOQ process is bypassed. Otherwise, the
RDOQ process is performed as usual.

DeltaQpRD (-dqr)

Specifies the maximum QP offset at slice level for multi-pass slice encoding. When
encoding, each slice is tested multiple times by using slice QP values in the range
[−DeltaQpRD, DeptaQpRD], and the best QP value is chosen as the slice QP.

MaxDeltaQP (-d)

Specifies the maximum QP offset at the largest coding unit level for the
block-level adaptive QP assignment scheme.
In the encoder, each largest
coding unit is tested multiple times by using the QP values in the range
[−MaxDeltaQP, MaxDeltaQP], and the best QP value is chosen as the QP value
of the largest coding unit.

dQPFile (-m)
AdaptiveQp (-aq)
MaxQPAdaptationRange (-aqr)
AdaptiveQpSelection (-aqps)

Specifies a file containing a list of QP deltas. The n-th line (where n is 0 for the first
line) of this file corresponds to the QP value delta for the picture with POC value n.
false

Enable or disable QP adaptation based upon a psycho-visual model.

Specifies the maximum QP adaptation range.

false

Specifies whether QP values for non-I frames will be calculated on the fly based on
statistics of previously coded frames.
Continued...

Date saved: 2019-02-05

Table 9: Quantization parameters (Continued)
Option

Default

RecalculateQP...
AccordingToLambda

Description

false

ScalingList

Recalculate QP values according to lambda values. Do not suggest to be enabled in
all intra case.

Controls the specification of scaling lists:
0
Scaling lists are disabled
1
Use default scaling lists
2
Scaling lists are specified in the file indicated by ScalingListFile

ScalingListFile

When ScalingList is set to 2, this parameter indicates the name of the file, which
contains the defined scaling lists. If ScalingList is set to 2 and this parameter is
an empty string, information on the format of the scaling list file is output and the
encoder stops.

MaxCUChromaQpAdjustmentDepth

-1

Specifies the maximum depth for CU chroma QP adjustment; if negative, CU chroma
QP adjustment is disabled.

Table 10: Slice coding parameters
Option
SliceMode

Default
0

SliceArgument
SliceSegmentMode

Description
Controls the slice partitioning method in conjunction with SliceArgument.
0
Single slice
1
Maximum number of CTUs per slice
2
Maximum number of bytes per slice
3
Maximum number of tiles per slice
Specifies the maximum number of CTUs, bytes or tiles in a slice depending on the
SliceMode setting.

SliceSegmentArgument

Enables (dependent) slice segment coding in conjunction with SliceSegmentArgument.
0
Single slice
1
Maximum number of CTUs per slice segment
2
Maximum number of bytes per slice segment
3
Maximum number of tiles per slice segment
Defines the maximum number of CTUs, bytes or tiles a slice segment depending on
the SliceSegmentMode setting.

WaveFrontSynchro

false

Enables the use of specific CABAC probabilities synchronization at the beginning of
each line of CTBs in order to produce a bitstream that can be encoded or decoded
using one or more cores.

TileUniformSpacing

false

Controls the mode used to determine per row and column tile sizes.
0
Each tile column width and tile row height is explicitly set by TileColumnWidthArray and TileRowHeightArray respectively
1
Tile columns and tile rows are uniformly spaced.

NumTileColumnsMinus1
NumTileRowsMinus1

TileColumnWidthArray
TileRowHeightArray

Specifies
the
tile
based
picture
partitioning
geometry
as
NumTileColumnsMinus1 + 1 × NumTileRowsMinus1 + 1 columns
and rows.
Specifies a space or comma separated list of widths and heights, respectively, of each
tile column or tile row. The first value in the list corresponds to the leftmost tile
column or topmost tile row.

Table 11: Deblocking filter parameters
Option

Default

Description

LoopFilterDisable

false

Enables or disables the in-loop deblocking filter.

LFCrossSliceBoundaryFlag

true

Enables or disables the use of in-loop filtering across slice boundaries.
Continued...

Date saved: 2019-02-05

Table 11: Deblocking filter parameters (Continued)
Option
LoopFilterOffsetInPPS

Default

Description

false

If enabled, the in-loop deblocking filter control parameters are sent in PPS. Otherwise, the in-loop deblocking filter control parameters are sent in the slice segment
header. If deblocking filter parameters are sent in PPS, the same values of deblocking
filter parameters are used for all pictures in the sequence (i.e. deblocking parameter
= base parameter value). If deblocking filter parameters are sent in the slice segment
header, varying deblocking filter parameters can be specified by setting parameters
tcOffsetDiv2 and betaOffsetDiv2 in the GOP structure table. In this case, the final
value of the deblocking filter parameter sent for a certain GOP picture is equal to
(base parameter + GOP parameter for this picture). Intra-pictures use the base parameters values.

LoopFilterTcOffset div2

Specifies the base value for the in-loop deblocking filter parameter tc offset div2.
The final value of tc offset div2 shall be an integer number in the range −6..6.

LoopFilterBetaOffset div2

Specifies the base value for the in-loop deblocking filter parameter beta offset div2.
The final value of beta offset div2 shall be an integer number in the range −6..6.

DeblockingFilterMetric

Specifies the use of a deblocking filter metric to evaluate the suitability of deblocking.
If non-zero then LoopFilterOffsetInPPS and LoopFilterDisable must be 0. Currently
excepted values are 0, 1 and 2.

LFCrossSliceBoundaryFlag

true

Enables or disables the use of a deblocking across tile boundaries.

Table 12: Coding tools parameters
Option

Default

Description

AMP

true

Enables or disables the use of asymmetric motion partitions.

SAO

true

Enables or disables the sample adaptive offset (SAO) filter.

TestSAODisableAtPictureLevel

false

Enables the testing of disabling SAO at the picture level after having analysed all
blocks.

SaoEncodingRate

0.75

When ¿0 SAO early picture termination is enabled for luma and chroma.

SaoEncodingRateChroma

0.5

The SAO early picture termination rate to use for chroma (when m SaoEncodingRate
is ¿0). If ¡=0, use results for luma.

SAOLcuBoundary

false

Enables or disables SAO parameter estimation using non-deblocked pixels for LCU
bottom and right boundary areas.

SAOResetEncoderStateAfterIRAP

false

When true, resets the encoder’s SAO state after an IRAP (POC order).

ConstrainedIntraPred

false

Enables or disables constrained intra prediction. Constrained intra prediction only
permits samples from intra blocks in the same slice as the current block to be used
for intra prediction.

FastUDIUseMPMEnabled

true

If enabled, adapt intra direction search, accounting for MPM

FastMEForGenBLowDelayEnabled

true

If enabled use a fast ME for generalised B Low Delay slices

UseBLambdaForNonKeyLowDelayPictures

true

Enables use of B-Lambda for non-key low-delay pictures

TransquantBypassEnable

false

Enables or disables the ability to bypass the transform, quantization and filtering
stages at CU level. This option corresponds to the value of transquant bypass enabled flag that is transmitted in the PPS.
See CUTransquantBypassFlagForce for further details.

Controls the per CU transformation, quantization and filtering mode decision. This
option controls the value of the per CU cu transquant bypass flag.
0
Bypass is searched on a CU-by-CU basis and will be used if the cost is lower
than not bypassing.
1
Bypass is forced for all CUs.
This option has no effect if TransquantBypassEnable is disabled.

false

Enables or disables the use of PCM. The encoder will use cost measures on a CU-byCU basis to determine if PCM mode is to be applied.

CUTransquantBypassFlagForce

PCMEnabledFlag

Continued...

Date saved: 2019-02-05

Table 12: Coding tools parameters (Continued)
Option

Default

Description

PCMLog2MaxSize

5
(= log2 (32))

Specifies log2 of the maximum PCM block size. When PCM is enabled, the PCM
mode is available for 2Nx2N intra PUs smaller than or equal to the specified maximum PCM block size

PCMLog2MinSize

Specifies log2 of the minimum PCM block size. When PCM is enabled, the PCM
mode is available for 2Nx2N intra PUs larger than or equal to the specified minimum
PCM block size.
When larger than PCMLog2MaxSize, PCM mode is not used.

PCMInputBitDepthFlag

true

If enabled specifies that PCM sample bit-depth is set equal to InputBitDepth. Otherwise, it specifies that PCM sample bit-depth is set equal to InternalBitDepth.

PCMFilterDisableFlag

false

If enabled specifies that loop-filtering on reconstructed samples of PCM blocks is
skipped. Otherwise, it specifies that loop-filtering on reconstructed samples of PCM
blocks is not skipped.

WeightedPredP (-wpP)

false

Enables the use of weighted prediction in P slices.

WeightedPredB (-wpB)

false

Enables the use of weighted prediction in B slices.

WPMethod (-wpM)

Sets the Weighted Prediction method to be used.
0
Image DC based method with joint colour component decision.
1
Image DC based method with separate colour component decision.
2
DC + Histogram refinement method (no clipping).
3
DC + Histogram refinement method (with clipping).
4
DC + Dual Histogram refinement method (with clipping).

Log2ParallelMergeLevel

Defines the PPS-derived Log2ParMrgLevel variable.

SignHideFlag (-SBH)

true

If enabled specifies that for each 4x4 coefficient group for which the number of coefficients between the first nonzero coefficient and the last nonzero coefficient along
the scanning line exceeds 4, the sign bit of the first nonzero coefficient will not be
directly transmitted in the bitstream, but may be inferred from the parity of the sum
of all nonzero coefficients in the current coefficient group.

StrongIntraSmoothing (-sis)

true

If enabled specifies that for 32x32 intra prediction block, the intra smoothing when
applied is either the 1:2:1 smoothing filter or a stronger bi-linear interpolation filter.
Key reference sample values are tested and if the criteria is satisfied, the stronger intra
smoothing filter is applied. If disabled, the intra smoothing filter when applied is the
1:2:1 smoothing filter.

TMVPMode

Controls the temporal motion vector prediction mode.
0
Disabled for all slices.
1
Enabled for all slices.
2
Disabled only for the first picture of each GOPSize.

TransformSkip

false

Enables or disables transform-skipping mode decision.

TransformSkipFast

false

Enables or disables reduced testing of the transform-skipping mode decision for
chroma TUs. When enabled, no RDO search is performed for chroma TUs, instead
they are transform-skipped if the four corresponding luma TUs are also skipped.
This option has no effect if TransformSkip is disabled.

Table 13: Rate control parameters
Option

Default

Description

RateControl

false

Rate control: enables rate control or not.

TargetBitrate

Rate control: target bitrate, in bps.

KeepHierarchicalBit

Rate control: 0: equal bit allocation among pictures; 1: fix ratio hierarchical bit
allocation; 2: adaptive hierarchical ratio bit allocation. It is suggested to enable
hierarchical bit allocation for hierarchical-B coding structure.

LCULevelRateControl

true

Rate control: true: LCU level RC; false: picture level RC.

RCLCUSeparateModel

true

Rate control: use LCU level separate R-lambda model or not. When LCULevelRateControl is equal to false, this parameter is meaningless.
Continued...

Date saved: 2019-02-05

Table 13: Rate control parameters (Continued)
Option
InitialQP

Default

Description

Rate control: initial QP value for the first picture. 0 to auto determine the initial QP
value.

RCForceIntraQP

false

Rate control: force intra QP to be equal to initial QP or not.

RCCpbSaturation

false

Rate control: enable target bits saturation to avoid CPB overflow and underflow or
not.

RCCpbSize
RCInitialCpbFullness

Rate control: CPB size, in bps.

0.9

Rate control: ratio of initial CPB fullness per CPB size. (InitalCpbFullness/CpbSize)
RCInitialCpbFullness should be smaller than or equal to 1.

Table 14: Encoder debug parameters
Option

Default

Description

DebugBitstream/DecodeBitstream1

Specifies the first bit stream to be read until a pre-defined switch point is encountered.

DecodeBitstream2

Specifies the second bit stream, to be read after the first random access point after a
QP switch point (specified using SwitchPOC and SwitchQP).

DebugPOC

-1

Specifies a POC, at which a bit stream specified using DebugBitstream or DecodeBitstream1 is no longer read, but rather normal encoding is started.

DebugCTU

-1

When the POC is encountered at which normal encoding is to be resumed, if set, this
option specifies that CTUs up to the specified CTU(in raster scan addressing order
are to be read from the specified bit stream, after which normal encoding is started
the specified CTU.

SwitchPOC

-1

Specifies a POC, at which the specified bit stream is no longer read, but rather normal
encoding is started.

SwitchDQP

Specifies a QP offset to be applied when normal encoding is started as specified by
SwitchPOC.

FastForwardToPOC

When encoding a bit streams, all frames that are not references including transitive
references to the specified POC are skipped.

StopAfterFFtoPOC

false

If enabled, causes the encoder to not encode any frame after the frame specified by
FastForwardToPOC option, in encoding order.

Table 15: VUI parameters
Option

Default

Description

VuiParametersPresent (-vui)

false

Enable generation of vui parameters().

AspectRatioInfoPresent

false

Signals whether aspect ratio idc is present.

AspectRatioIdc

aspect ratio idc

SarWidth

Specifies the horizontal size of the sample aspect ratio.

SarHeight

Specifies the vertical size of the sample aspect ratio.

OverscanInfoPresent

false

Signals whether overscan info present flag is present.

OverscanAppropriate

false

Indicates whether cropped decoded pictures are suitable for display using overscan.
0
Indicates that the decoded pictures should not be displayed using overscan.
1
Indicates that the decoded pictures may be displayed using overscan.

VideoSignalTypePresent

false

Signals whether video format, video full range flag, and colour description present flag are present.

VideoFormat

Indicates representation of pictures.
Continued...

Date saved: 2019-02-05

Table 15: VUI parameters (Continued)
Option

Default

Description

VideoFullRange

false

Indicates the black level and range of luma and chroma signals.
0
Indicates that the luma and chroma signals are to be scaled prior to display.
1
Indicates that the luma and chroma signals are not to be scaled prior to display.

ColourDescriptionPresent

false

Signals whether colour primaries, transfer characteristics and matrix coefficients are
present.

ColourPrimaries

Indicates chromaticity coordinates of the source primaries.

TransferCharateristics

Indicates the opto-electronic transfer characteristics of the source.

MatrixCoefficients

Describes the matrix coefficients used in deriving luma and chroma from RGB primaries.

false

Signals whether chroma sample loc type top field and chroma sample loc type bottom field are present.

ChromaLocInfoPresent
ChromaSampleLocTypeTopField

Specifies the location of chroma samples for top field.

ChromaSampleLocTypeBottomField

Specifies the location of chroma samples for bottom field.

NeutralChromaIndication

false

Indicates that the value of all decoded chroma samples is equal to 1¡¡(BitDepthCr-1).

DefaultDisplayWindowFlag

flag

Indicates the presence of the Default Window parameters.
false
Disabled
true
Enabled

DefDispWinLeftOffset
DefDispWinRightOffset
DefDispWinTopOffset
DefDispWinBottomOffset

Specifies the horizontal and vertical offset to be applied to the input video from the
conformance window in luma samples. Must be a multiple of the chroma resolution
(e.g. a multiple of two for 4:2:0).

FrameFieldInfoPresentFlag

false

Specificies the value of the VUI syntax element ‘frame field info present flag’,
which indicates that pic struct and field coding related values are present in picture
timing SEI messages.

PocProportionalToTimingFlag

false

Specificies the value of the VUI syntax element ‘vui poc proportional to timing flag’, which indicates that the POC value is proportional to the output time with
respect to the first picture in the CVS.

Specificies the value of the VUI syntax element ‘vui num ticks poc diff one minus1’, which specifies the number of clock ticks corresponding to a difference of
picture order count values equal to 1, and is used only when PocProportionalToTimingFlag is true.

NumTicksPocDiffOneMinus

BitstreamRestriction

false

Signals whether bitstream restriction parameters are present.

TilesFixedStructure

false

Indicates that each active picture parameter set has the same values of the syntax
elements related to tiles.

MotionVectorsOverPicBoundaries

false

Indicates that no samples outside the picture boundaries are used for inter prediction.

MaxBytesPerPicDenom

Indicates a number of bytes not exceeded by the sum of the sizes of the VCL NAL
units associated with any coded picture.

MaxBitsPerMinCuDenom

Indicates an upper bound for the number of bits of coding unit() data.

Log2MaxMvLengthHorizontal

Indicate the maximum absolute value of a decoded horizontal MV component in
quarter-pel luma units.

Log2MaxMvLengthVertical

Indicate the maximum absolute value of a decoded vertical MV component in
quarter-pel luma units.

Date saved: 2019-02-05

Table 16: Range Extensions (Version 2) tool parameters
Option

3.3

Default

Description

CostMode

lossy

Specifies the cost mode to use.
lossy
cost = distortion + λ × bits
sequence level lossless cost = distortion/λ + bits.
lossless
As with sequence level lossless, but QP is also set
to 0 (this will be deprecated in the future)
As with sequence level lossless, but QP’=4 is used
mixed lossless lossy
for pre-estimates of transquant-bypass blocks

ExtendedPrecision

false

Specifies the use of extended precision processing flag. Note that unless the HIGH BIT DEPTH SUPPORT macro in TypeDef.h is enabled, all internal bit depths must
be 8 when the ExtendedPrecision setting is enabled. This setting is only valid for the
16-bit RExt profiles.

HighPrecisionPredictionWeighting

false

Specifies the value of high precision prediction weighting flag. This setting is only
valid for the 16-bit or 4:4:4 RExt profiles.

CrossComponentPrediction

false

When true, specifies the use of the cross component prediction tool (4:4:4 processing
only). Version 1 and some Version 2 (RExt) profiles require this to be false.

ReconBasedCrossCPredictionEstimate

false

If true, then when determining the alpha value for cross-component prediction, use
the reconstructed residual rather than the pre-transform encoder-side residual

SaoLumaOffsetBitShift
SaoChromaOffsetBitShift

0
0

Specifies the shift to apply to the SAO parameters. If negative, an estimate will be
calculated based upon the initial QP. Version 1 and some Version 2 (RExt) profiles
require this to be 0.

TransformSkipLog2MaxSize

Specifies the maximum TU size for which transform-skip can be used; the minimum
value is 2. Version 1 and some Version 2 (RExt) profiles require this to be 2.

ImplicitResidualDPCM

false

When true, specifies the use of the implicitly signalled residual RDPCM tool (for
intra). Version 1 and some Version 2 (RExt) profiles require this to be false.

ExplicitResidualDPCM

false

When true, specifies the use of the explicitly signalled residual RDPCM tool (for
intra-block-copy and inter). Version 1 and some Version 2 (RExt) profiles require
this to be false.

ResidualRotation

false

When true, specifies the use of the residual rotation tool. Version 1 and some Version
2 (RExt) profiles require this to be false.

SingleSignificanceMapContext

false

When true, specifies the use of a single significance map context for transformskipped and transquant-bypassed TUs. Version 1 and some Version 2 (RExt) profiles
require this to be false.

GolombRiceParameterAdaptation

false

When true, enable the adaptation of the Golomb-Rice parameter over the course of
each slice. Version 1 and some Version 2 (RExt) profiles require this to be false.

AlignCABACBeforeBypass

false

When true, align the CABAC engine to a defined fraction of a bit prior to coding
bypass data (including sign bits) when coeff abs level remaining syntax elements are
present in the group. This must always be true for the high-throughput-RExt profile,
and false otherwise.

IntraReferenceSmoothing

true

When true, enable intra reference smoothing, otherwise disable it. Version 1 and
some Version 2 (RExt) profiles require this to be true.

Encoder SEI parameters

The table below lists the SEI messages defined for Version 1 and Range-Extensions, and if available, the respective table
that lists the controls within the HM Encoder to include the messages within the bit stream.
Table 17: List of Version 1 and RExt SEI messages
SEI Number

SEI Name

Table number of encoder controls, if available

Buffering period

Table 18

Picture timing

Table 19
Continued...

Date saved: 2019-02-05

Table 17: List of Version 1 and RExt SEI messages (Continued)
SEI Number

SEI Name

Table number of encoder controls, if available

Pan-scan rectangle

(Not handled)

Filler payload

(Not handled)

User data registered by Rec. ITU-T T.35

(Not handled)

User data unregistered

Decoded only

Recovery point

Table 20

Scene information

(Not handled)

Picture snapshot

(Not handled)

Progressive refinement segment start

(Not handled)

Progressive refinement segment end

(Not handled)

Film grain characteristics

(Not handled)

Post-filter hint

(Not handled)

Tone mapping information

Table 21

Frame packing arrangement

Table 22

Display orientation

Table 23

Green Metadata

Table 24

128

Structure of pictures information

Table 25

129

Active parameter sets

Table 26

130

Decoding unit information

Table 27

131

Temporal sub-layer zero index

Table 28

132

Decoded picture hash

Table 29

133

Scalable nesting

Table 30

134

Region refresh information

Table 31

135

No display

Table 32

136

Time code

Table 33

137

Mastering display colour volume

Table 34

138

Segmented rectangular frame packing arrangement

Table 35

139

Temporal motion-constrained tile sets

Table 36

140

Chroma resampling filter hint

Table 37

141

Knee function information

Table 38

142

Colour remapping information

Table 39

143

Deinterlaced field identification

(Not handled)

Table 18: Buffering period SEI message encoder parameters
Option
SEIBufferingPeriod

Default
0

Description
Enables or disables the insertion of the Buffering period SEI messages. This option has no effect if VuiParametersPresent is disabled. SEIBufferingPeriod requires
SEIActiveParameterSets to be enabled.

Date saved: 2019-02-05

Table 19: Picture timing SEI message encoder parameters
Option

Default

SEIPictureTiming

Description
Enables or disables the insertion of the Picture timing SEI messages. This option has
no effect if VuiParametersPresent is disabled.

Table 20: Recovery point SEI message encoder parameters
Option

Default

SEIRecoveryPoint

Description
Enables or disables the insertion of the Recovery point SEI messages.

Table 21: Tone mapping information SEI message encoder parameters
Option

Default

Description

SEIToneMappingInfo

Enables or disables the insertion of the Tone Mapping SEI message.

SEIToneMapId

Specifies Id of Tone Mapping SEI message for a given session.

SEIToneMapCancelFlag

false

Indicates that Tone Mapping SEI message cancels the persistance or follows.

SEIToneMapPersistenceFlag

true

Specifies the persistence of the Tone Mapping SEI message.

SEIToneMapCodedDataBitDepth

Specifies Coded Data BitDepth of Tone Mapping SEI messages.

SEIToneMapTargetBitDepth

Specifies Output BitDepth of Tome mapping function.

SEIToneMapModelId

Specifies Model utilized for mapping coded data into target bit depth range.
0
linear mapping with clipping
1
sigmoidal mapping
2
user-defined table mapping
3
piece-wise linear mapping
4
luminance dynamic range mapping

SEIToneMapMinValue

Specifies the minimum value in mode 0.

SEIToneMapMaxValue

1023

Specifies the maxmum value in mode 0.

SEIToneMapSigmoidMidpoint

512

Specifies the centre point in mode 1.

SEIToneMapSigmoidWidth

960

Specifies the distance between 5the target bit depth in mode 1.

SEIToneMapStartOfCodedInterval

SEIToneMapNumPivots

Array of user-defined mapping table. Default table can be set to the following:
0 12 24 36 48 60 72 84 96 108 120 132 144 156 168 180
192 192 196 204 208 216 220 228 232 240 248 252 260 264
272 276 284 292 292 296 300 304 308 312 320 324 328 332
336 344 348 352 356 360 368 372 376 380 384 388 396 400
404 408 412 420 424 428 432 436 444 444 444 448 452 456
460 464 468 472 476 476 480 484 488 492 496 500 504 508
508 512 516 520 524 528 532 536 540 540 544 548 552 556
560 564 568 572 572 576 580 584 588 592 596 600 604 604
608 612 616 620 624 628 632 636 636 640 644 648 652 656
660 664 668 672 672 672 676 680 680 684 688 692 692 696
700 704 704 708 712 716 716 720 724 724 728 732 736 736
740 744 748 748 752 756 760 760 764 768 768 772 776 780
780 784 788 792 792 796 800 804 804 808 812 812 816 820
824 824 828 832 836 836 840 844 848 848 852 856 860 860
860 864 864 868 872 872 876 880 880 884 884 888 892 892
896 900 900 904 908 908 912 912 916 920 920 924 928 928
932 936 936 940 940 944 948 948 952 956 956 960 964 964
968 968 972 976 976 980 984 984 988 992 992 996 996 1000
1004 1004 1008 1012 1012 1016 1020 1024
0

Specifies the number of pivot points in mode 3.

SEIToneMapCodedPivotValue

Array of coded pivot point in mode 3. A suggested table is:
64 128 256 512 768
Continued...

Date saved: 2019-02-05

Table 21: Tone mapping information SEI message encoder parameters (Continued)
Option

Default

Description

SEIToneMapTargetPivotValue

Array of target pivot point in mode 3. A suggested table is:
48 73 111 168 215

SEIToneMap...
CameraIsoSpeedIdc

SEIToneMap...
CameraIsoSpeedValue

Indicates the camera ISO speed for daylight illumination.

400

SEIToneMap...
ExposureIndexIdc

Specifies the camera ISO speed for daylight illumination of Extended ISO.

SEIToneMap...
ExposureIndexValue

Indicates the exposure index setting of the camera.

400

Specifies the exposure index setting of the cameran of Extended ISO.

SEIToneMapExposure...
CompensationValueSignFlag

Specifies the sign of ExposureCompensationValue.

SEIToneMapExposure...
CompensationValueNumerator

Specifies the numerator of ExposureCompensationValue.

SEIToneMapExposure...
CompensationValueDenomIdc

Specifies the denominator of ExposureCompensationValue.

SEIToneMapRef...
ScreenLuminanceWhite

350

Specifies reference screen brightness setting in units of candela per square metre.

SEIToneMapExtended...
RangeWhiteLevel

800

Indicates the luminance dynamic range.

SEIToneMapNominal...
BlackLevelLumaCodeValue

Specifies luma sample value of the nominal black level assigned decoded pictures.

SEIToneMapNominal...
WhiteLevelLumaCodeValue

235

Specifies luma sample value of the nominal white level assigned decoded pictures.

SEIToneMapExtended...
WhiteLevelLumaCodeValue

300

Specifies luma sample value of the extended dynamic range assigned decoded pictures.

Table 22: Frame packing arrangement SEI message encoder parameters
Option

Default

Description

SEIFramePacking

Enables or disables the insertion of the Frame packing arrangement SEI messages.

SEIFramePackingType

Indicates the arrangement type in the Frame packing arrangement SEI message. This
option has no effect if SEIFramePacking is disabled.
3
Side by Side
4
Top Bottom
5
Frame Alternate

SEIFramePackingInterpretation

Indicates the constituent frames relationship in the Frame packing arrangement SEI
message. This option has no effect if SEIFramePacking is disabled.
0
Unspecified
1
Frame 0 is associated with the left view of a stereo pair
2
Frame 0 is associated with the right view of a stereo pair

SEIFramePackingQuincunx

Enables or disables the quincunx sampling signalling in the Frame packing arrangement SEI messages. This option has no effect if SEIFramePacking is disabled.

SEIFramePackingId

Indicates the session number in the Frame packing arrangement SEI messages. This
option has no effect if SEIFramePacking is disabled.

Date saved: 2019-02-05

Table 23: Display orientation SEI message encoder parameters
Option

Default

SEIDisplayOrientation

Description
Enables or disables the insertion of the Display orientation SEI messages.
0
Disabled
N: 0 < N < (216 − 1)
Enable
display
orientation
SEI
message
with
anticlockwise rotation = N and
display orientation repetition period = 1

Table 24: Green Metadata SEI message encoder parameters
Option

Default

Description

SEIGreenMetadataType

Specifies the type of metadata that is present in the SEI message.
0
Reserved
1
Metadata enabling quality recovery after low-power encoding is present

SEIXSDMetricType

Indicates the type of the objective quality metric.
0
PSNR is used as objective quality metric

Table 25: Structure of pictures information SEI message encoder parameters
Option

Default

SEISOPDescription

Description
Enables or disables the insertion of the Structure of pictures information SEI messages.

Table 26: Active parameter sets SEI message encoder parameters
Option

Default

SEIActiveParameterSets

Description
Enables or disables the insertion of the Active parameter sets SEI messages.

Table 27: Decoding unit information SEI message encoder parameters
Option

Default

SEIDecodingUnitInfo

Description
Enables or disables the insertion of the Decoding unit information SEI messages.
This option has no effect if VuiParametersPresent is disabled.

Table 28: Temporal sub-layer zero index SEI message encoder parameters
Option
SEITemporalLevel0Index

Default
0

Description
Enables or disables the insertion of the Temporal level zero index SEI messages.

Date saved: 2019-02-05

Table 29: Decoded picture hash SEI message encoder parameters
Option

Default

SEIDecodedPictureHash

Description
Enables or disables the calculation and insertion of the Decoded picture hash SEI
messages.
0
Disabled
1
Transmits MD5 in SEI message and writes the value to the
encoder log
2
Transmits CRC in SEI message and writes the value to the
encoder log
3
Transmits checksum in SEI message and writes the value to
the encoder log

Table 30: Scalable nesting SEI message encoder parameters
Option

Default

SEIScalableNesting

Description
Enables or disables the use of the scalable nesting SEI messages.

Table 31: Region refresh information SEI message encoder parameters
Option

Default

SEIGradualDecodingRefreshInfo

Description

Enables or disables the insertion of the Gradual decoding refresh information SEI
messages.

Table 32: No display SEI message encoder parameters
Option
SEINoDisplay

Default
0

Description
When non-zero, generate no-display SEI message for temporal layer N or higher.

Table 33: Time code SEI message encoder parameters
Option
SEITimeCodeEnabled
SEITimeCodeNumClockTs

Default
false
0

Description
When true (non-zero), generate Time code SEI messages.
Number of clock time sets, in the range of 0 to 3 (inclusive).

SEITimeCodeTimeStampFlag

Time stamp flag associated to each time set (comma or space separated list of entries).

SEITimeCodeFieldBasedFlag

Field based flag associated to each time set (comma or space separated list of entries).

SEITimeCodeCountingType

Counting type associated to each time set (comma or space separated list of entries).

SEITimeCodeFullTsFlag

Full time stamp flag associated to each time set (comma or space separated list of
entries).

SEITimeCodeDiscontinuityFlag

Discontinuity flag associated to each time set (comma or space separated list of entries).

SEITimeCodeCntDroppedFlag

Counter dropped flag associated to each time set (comma or space separated list of
entries).

SEITimeCodeNumFrames

Number of frames associated to each time set (comma or space separated list of
entries).

SEITimeCodeSecondsFlag

Flag to signal seconds value presence in each time set (comma or space separated list
of entries).
Continued...

Date saved: 2019-02-05

Table 33: Time code SEI message encoder parameters (Continued)
Option

Default

Description

SEITimeCodeMinutesFlag

Flag to signal minutes value presence in each time set (comma or space separated list
of entries).

SEITimeCodeHoursFlag

Flag to signal hours value presence in each time set (comma or space separated list
of entries).

SEITimeCodeSecondsValue

Seconds value for each time set (comma or space separated list of entries).

SEITimeCodeMinutesValue

Minutes value for each time set (comma or space separated list of entries).

SEITimeCodeHoursValue

Hours value for each time set (comma or space separated list of entries).

SEITimeCodeOffsetLength

Time offset length associated to each time set (comma or space separated list of
entries).

SEITimeCodeTimeOffset

Time offset associated to each time set (comma or space separated list of entries).

Table 34: Mastering display colour volume SEI message encoder parameters
Option

Default

Description

SEIMasteringDisplayColourVolume

false

SEIMasteringDisplayMaxLuminance

10000

Specifies the mastering display maximum luminance value in units of 1/10000 candela per square metre.

SEIMasteringDisplayMinLuminance

Specifies the mastering display minimum luminance value in units of 1/10000 candela per square metre.

0,50000, 0,0, 50000,0

Mastering display primaries for all three colour planes in CIE xy coordinates in increments of 1/50000 (results in the ranges 0 to 50000 inclusive).

16667, 16667

Mastering display white point CIE xy coordinates in normalized increments of
1/50000 (e.g. 0.333 = 16667).

SEIMasteringDisplayPrimaries
SEIMasteringDisplayWhitePoint

When true (non-zero), generate Mastering display colour volume SEI message.

Table 35: Segmented rectangular frame packing arrangement SEI message encoder parameters
Option

Default

SEISegmentedRectFramePacking
SEISegmentedRectFramePackingCancel

Controls generation of segmented rectangular frame packing SEI messages.

false

If true, cancels the persistence of any previous SRFPA SEI message.

Specifies the arrangement of the frames in the reconstructed picture.

SEISegmentedRectFramePackingType
SEISegmentedRectFramePackingPersistence

Description

false

If false the SEI applies to the current frame only.

Table 36: Temporal motion-constrained tile sets SEI message encoder parameters
Option

Default

SEITempMotionConstrainedTileSets

Description

false

When true (non-zero), generates example temporal motion constrained tile sets SEI
messages.

Table 37: Chroma resampling filter hint SEI message encoder parameters
Option
SEIChromaResamplingFilterHint

Default

Description

false

When true (non-zero), generates example chroma sampling filter hint SEI messages.
Continued...

Date saved: 2019-02-05

Table 37: Chroma resampling filter hint SEI message encoder parameters (Continued)
Option

Default

Description

SEIChromaResamplingHorizontalFilterType

Defines the index of the chroma sampling horizontal filter:
0
Unspecified
1
Filters signalled within the SEI message
2
Filters as described by SMPTE RP 2050-1:2012

SEIChromaResamplingVerticalFilterType

Defines the index of the chroma sampling vertical filter:
0
Unspecified
1
Filters signalled within the SEI message
2
Filters as described in the 5/3 filter description of ITU-T
Rec. T.800 — ISO/IEC 15444-1

Table 38: Knee function SEI message encoder parameters
Option

Default

SEIKneeFunctionInfo

Description

false

SEIKneeFunctionId

Enables (true) or disables (false) the insertion of the Knee function SEI messages.

Specifies Id of Knee function SEI message for a given session.

SEIKneeFunctionCancelFlag

false

Indicates that Knee function SEI message cancels the persistance (true) or follows
(false).

SEIKneeFunctionPersistenceFlag

true

Specifies the persistence of the Knee function SEI message.

SEIKneeFunctionInputDrange

1000

Specifies the peak luminance level for the input picture of Knee function SEI messages.

SEIKneeFunctionInputDispLuminance

100

Specifies the expected display brightness for the input picture of Knee function SEI
messages.

SEIKneeFunctionOutputDrange

4000

Specifies the peak luminance level for the output picture of Knee function SEI messages.

SEIKneeFunctionOutputDispLuminance

800

Specifies the expected display brightness for the output picture of Knee function SEI
messages.

SEIKneeFunctionNumKneePointsMinus1

Specifies the number of knee points - 1.

SEIKneeFunctionInputKneePointValue

Array of input knee point. Default table can be set to the following:
600 800 900

SEIKneeFunctionOutputKneePointValue

Array of output knee point. Default table can be set to the following:
100 250 450

Table 39: Colour remapping SEI message encoder parameters
Option

Default

Description

SEIColourRemappingInfoFileRoot (-cri)

3.4

Specifies the prefix of input Colour Remapping Information file. Prefix is completed
by “ x.txt” where x is the POC number. The contents of the file are a list of the SEI
message’s syntax element names (in decoding order) immediately followed by a ‘:’
and then the associated value. An example file can be found in cfg/misc/example colour remapping sei encoder 0.txt.

Hardcoded encoder parameters
Table 40: CommonDef.h constants
Option
ADAPT SR SCALE

Default

Description

Defines a scaling factor used to derive the motion search range is adaptive (see ASR
configuration parameter). Default value is 1.
Continued...

Date saved: 2019-02-05

Table 40: CommonDef.h constants (Continued)
Option

Default

Description

MAX GOP

maximum size of value of hierarchical GOP.

MAX NUM REF

maximum number of multiple reference frames

MAX NUM REF LC

maximum number of combined reference frames

AMVP MAX NUM CANDS

maximum number of final candidates

AMVP MAX NUM CANDS MEM

MRG MAX NUM CANDS

DYN REF FREE

off

dynamic free of reference memories

MAX TLAYER

maximum number of temporal layers

ADAPT SR SCALE

division factor for adaptive search range

EARLY SKIP THRES

1.5

early skip if RD ¡ EARLY SKIP THRES*avg[BestSkipRD]

MAX NUM REF PICS

MAX CHROMA FORMAT IDC

TypeDef.h
Numerous constants that guard individual adoptions are defined within source/Lib/TLibCommon/TypeDef.h.

Date saved: 2019-02-05

4
4.1

Using the decoder
General
TAppDecoder -b str.bin -o dec.yuv [options]

Table 41: Decoder options
Option

Default

Description

(–help)

Prints usage information.

BitStreamFile (-b)

Defines the input bit stream file name.

ReconFile (-o)

Defines reconstructed YUV file name. If empty, no file is generated.

SkipFrames (-s)

Defines the number of pictures in decoding order to skip.

MaxTemporalLayer (-t)

-1

Defines the maximum temporal layer to be decoded. If -1, then all layers are decoded.

TarDecLayerIdSetFile (-l)

Specifies the targetDecLayerIdSet file name. The file would contain white-space
separated LayerId values of the layers that are to be decoded. Omitting the parameter,
or using a value of -1 in the file decodes all layers.

OutputBitDepth (-d)

0
(Native)

Specifies the luma bit-depth of the reconstructed YUV file (the value 0 indicates that
the native bit-depth is used)

OutputBitDepthC

0
(Native)

Defines the chroma bit-depth of the reconstructed YUV file (the value 0 indicates that
the native bit-depth is used)

Enable or disable verification of any Picture hash SEI messages. When this parameter is set to 0, the feature is disabled and all messages are ignored. When set to 1
(default), the feature is enabled and the decoder has the following behaviour:
• If Picture hash SEI messages are included in the bit stream, the same type
of hash is calculated for each decoded picture and written to the log together
with an indication whether the calculted value matches the value in the SEI
message. Decoding will continue even if there is a mismatch.
• After decoding is complete, if any MD5sum comparison failed, a warning is
printed and the decoder exits with the status EXIT FAILURE
• The per-picture MD5 log message has the following formats:
[MD5:d41d8cd98f00b204e9800998ecf8427e,(OK)],
[MD5:d41d8cd98f00b204e9800998ecf8427e,(unk)],
[MD5:d41d8cd98f00b204e9800998ecf8427e,(***ERROR***)]
[rxMD5:b9e1...] where, “(unk)” implies that no MD5 was signalled
for this picture, “(OK)” implies that the decoder agrees with the signalled
MD5, “(***ERROR***)” implies that the decoder disagrees with the
signalled MD5. “[rxMD5:...]” is the signalled MD5 if different.

SEIDecodedPictureHash

OutputDecodedSEIMessagesFilename

When a non-empty file name is specified, information regarding any decoded SEI
messages will be output to the indicated file. If the file name is ’-’, then stdout is used
instead.

SEIColourRemappingInfoFilename

Specifies that the colour remapping SEI message should be applied to the output
video, with the output written to this file. If no value is specified, the SEI message is
ignored and no mapping is applied.

RespectDefDispWindow (-w)

Video region to be output by the decoder.
0
Output content inside the conformance window.
1
Output content inside the default window.

OutputColourSpaceConvert

SEINoDisplay

Specifies the colour space conversion to apply to 444 video. Permitted values are:
UNCHANGED
No colour space conversion is applied
YCrCbToYCbCr
Swap the second and third components
GBRtoRGB
Reorder the three components
If no value is specified, no colour space conversion is applied. The list may eventually
also include RGB to YCbCr or YCgCo conversions.

false

When true, do not output frames for which there is an SEI NoDisplay message.
Continued...

Date saved: 2019-02-05

Table 41: Decoder options (Continued)
Option

Default

ClipOutputVideoToRec709Range

4.2

Description

If 1 then clip output video to the Rec. 709 Range on saving when OutputBitDepth is
less than InternalBitDepth.

Using the decoder analyser

If the decoder is compiled with the macro RExt DECODER DEBUG BIT STATISTICS defined as 1 (either externally, or
by editing TypeDef.h), the decoder will gather fractional bit counts associated with the different syntax elements, producing a
table of the number of bits per syntax element, and where appropriate, according to block size and colour component/channel.
The Linux makefile will compile both the analyser and standard version when the ‘all’ or ‘everything’ target is used (where
the latter will also build high-bit-depth executables).

Block statistics extension

The block statistics extension enables straightforward visualization and statistical analysis of coding tool usage in encoded
bitstreams. The extension enables the reference software encoder and decoder to write out statistics files in a configurable
way, which in turn can be loaded into a suitable YUV player for overlay of the reconstructed YUV sequence, or can be
used for statistical analysis at a selectable scope (e.g. block/picture/sequence level). An example implementation for such
visualization is available with the open-source YUView player (https://github.com/IENT/YUView).

5.1

Usage

The software has to be compiled with the macros ENABLE TRACING and K0149 BLOCK STATISTICS defined as 1. The
statistics can be written by either encoder or decoder.
The extension adds additional trace channels to the “dtrace” functionality of the software. The following trace channels were
added:
D BLOCK STATISTICS ALL All syntax elements are written, no matter whether they are actually encoded or derived.
D BLOCK STATISTICS CODED Tries to write only syntax elements, which have also been encoded.
The following additional encoder options are available (part of “dtrace”). See the file dtrace next.h for more details.
Table 42: Decoder options
Option

Default

Description

TraceFile

File name of the produced trace file.

TraceRule

Specifies which traces should be saved, and for which POCs.

Concrete examples of calls for generating a block statistics file are:
bin/DecoderAppStatic -b str/BasketballDrive_1920x1080_QP37.vvc \
--TraceFile="stats/BasketballDrive_1920x1080_QP37_coded.vtmbmsstats" \
--TraceRule="D_BLOCK_STATISTICS_CODED:poc>=0"
bin/DecoderAppStatic -b str/BasketballDrive_1920x1080_QP37.vvc \
--TraceFile="stats/BasketballDrive_1920x1080_QP37_all.vtmbmsstats" \
--TraceRule="D_BLOCK_STATISTICS_ALL:poc>=0"

Date saved: 2019-02-05

5.2

Block statistics file formats

The trace file will contain a header listing information of all available block statistics. For each statistic it lists a type and a
scale for vectors or range for integers if applicable:
#
#
#
#
#
#
#
#
#

VTMBMS Block Statistics
Sequence size: [832x 480]
Block Statistic Type: PredMode; Flag;
Block Statistic Type: MergeFlag; Flag;
Block Statistic Type: MVL0; Vector; Scale: 4
Block Statistic Type: MVL1; Vector; Scale: 4
Block Statistic Type: IPCM; Flag;
Block Statistic Type: Y_IntraMode; Integer; [0, 73]
Block Statistic Type: Cb_IntraMode; Integer; [0, 73]

Two formats are available for the statistics for each block, a human readable format and a CSV based format. The header
remains the same for both cases.
For both formats each row contains the information for one block statistic. The order of the data is: picture order count
(POC), location of top left corner of the block, size of the block, name of the statistic, and value of the statistic. The
macro BLOCK STATS AS CSV is available in order to choose the required format. The human readable format can also
be easily processed with other software, for example YUView, using regular expressions. The CSV based formats provides
the universal interface required by spreadsheet applications.
The human readable format is based on the format used for the other dtrace statistics. Some examples for this format are:
BlockStat:
BlockStat:
BlockStat:
BlockStat:
BlockStat:
BlockStat:
BlockStat:
BlockStat:
BlockStat:
BlockStat:
BlockStat:
BlockStat:

POC
POC
POC
POC
POC
POC
POC
POC
POC
POC
POC
POC

16
16
16
16
16
16
16
16
16
16
16
16

@(
@(
@(
@(
@(
@(
@(
@(
@(
@(
@(
@(

112,
112,
112,
112,
112,
112,
112,
112,
112,
112,
112,
112,

0)
0)
0)
0)
0)
0)
0)
0)
0)
0)
8)
8)

[
[
[
[
[
[
[
[
[
[
[
[

8x
8x
8x
8x
8x
8x
8x
8x
8x
8x
8x
8x

8]
8]
8]
8]
8]
8]
8]
8]
8]
8]
8]
8]

SkipFlag=1
InterDir=1
MergeFlag=1
MergeIdx=0
MergeType=0
MVPIdxL0=255
MVPNumL0=255
RefIdxL0=0
MVDL0={
0,
0}
MVL0={ -70, 18}
PredMode=0
PartSize=0

Some examples of the CSV based format are:
BlockStat;16;
BlockStat;16;
BlockStat;16;
BlockStat;16;
BlockStat;16;
BlockStat;16;
BlockStat;16;
BlockStat;16;
BlockStat;16;
BlockStat;16;
BlockStat;16;
BlockStat;16;

112;
112;
112;
112;
112;
112;
112;
112;
112;
112;
112;
112;

0;
0;
0;
0;
0;
0;
0;
0;
0;
0;
8;
8;

8;
8;
8;
8;
8;
8;
8;
8;
8;
8;
8;
8;

8;SkipFlag;1
8;InterDir;1
8;MergeFlag;1
8;MergeIdx;0
8;MergeType;0
8;MVPIdxL0;255
8;MVPNumL0;255
8;RefIdxL0;0
8;MVDL0;
0;
0
8;MVL0; -70; 18
8;PredMode;0
8;PartSize;0

Date saved: 2019-02-05

5.3

Visualization

The block statistics can be viewed with YUView, which is freely available under GPLv3: https://github.com/IENT/YUView.
The latest releases and the master branch have the functionality required for viewing the block statistics. YUView assumes
that the file extension of block statistics file is “.vtmbmsstats”. However, if a file is not recognized you can choose from a
list of supported file formats.
Statistics can be overlaid with YUV sequences. Some example snapshots are:

Figure 2: YUView

Figure 3: Motion vectors

5.4

Adding statistics

In order to add further block statistics, do the following:
source/Lib/CommonLib/dtrace blockstatistics.h Add your statistic to the BlockStatistic enum:

Date saved: 2019-02-05

Figure 4: Skip flag

enum class BlockStatistic {
// general
PredMode,
PartSize,
Depth,

Further, add your statistic to the map blockstatistic2description:
static const std::map>
blockstatistic2description =
{
{ BlockStatistic::PredMode,
std::tuple
{"PredMode", BlockStatisticType::Flag, ""}},
{ BlockStatistic::MergeFlag,
std::tuple
{"MergeFlag", BlockStatisticType::Flag, ""}},
{ BlockStatistic::MVL0,
std::tuple
{"MVL0", BlockStatisticType::Vector, "Scale: 4"}},
YOURS

source/Lib/CommonLib/dtrace blockstatistics.cpp All code for writing syntax elements is kept in this file in getAndStoreBlockStatistics. This function is called once for each CTU, after it has been en/decoded. The following macros
have been defined to facilitate writing of block statistics:
DTRACE_BLOCK_SCALAR(ctx,channel,cs_cu_pu,stat_type,val)
DTRACE_BLOCK_SCALAR_CHROMA(ctx,channel,cs_cu_pu,stat_type,val)
DTRACE_BLOCK_VECTOR(ctx,channel,cu_pu,stat_type,v_x,v_y)
DTRACE_BLOCK_AFFINETF(ctx,channel,pu,stat_type,v_x0,v_y0,v_x1,v_y1,v_x2,v_y2)

An example:
DTRACE_BLOCK_SCALAR(g_trace_ctx, D_BLOCK_STATISTICS_ALL,
cu, GetBlockStatisticName(BlockStatistic::PredMode), cu.predMode);

Date saved: 2019-02-05

Block statistics for debugging The statistics can also be used to write out other data, not just syntax elements. Add your
statistics to dtrace blockstatistics.h. Where it should be used the following headers have to be included:
#include "dtrace_next.h"
#include "dtrace_blockstatistics.h"

Date saved: 2019-02-05

Source Exif Data:

File Type                       : PDF
File Type Extension             : pdf
MIME Type                       : application/pdf
PDF Version                     : 1.5
Linearized                      : No
Page Count                      : 34
Page Mode                       : UseOutlines
Author                          : 
Title                           : 
Subject                         : 
Creator                         : LaTeX with hyperref
Producer                        : pdfTeX-1.40.19
Create Date                     : 2019:02:05 18:04:29+01:00
Modify Date                     : 2019:02:05 18:04:29+01:00
Trapped                         : False
PTEX Fullbanner                 : This is pdfTeX, Version 3.14159265-2.6-1.40.19 (TeX Live 2018) kpathsea version 6.3.0

EXIF Metadata provided by EXIF.tools

Software Manual

software-manual

Navigation menu

Versions of this User Manual:

Views

Navigation