SPU2 Overview Manual
User Manual: Pdf
Open the PDF directly: View PDF 
.
Page Count: 80
| Download | |
| Open PDF In Browser | View PDF |
SPU2 Overview Copyright © 2002 Sony Computer Entertainment Inc. All Rights Reserved. SCE Confidential SCE CONFIDENTIAL SPU2 Overview Version 6.0 © 2002 Sony Computer Entertainment Inc. Publication date: April 2002 Sony Computer Entertainment Inc. 1-1, Akasaka 7-chome, Minato-ku Tokyo 107-0052 Japan Sony Computer Entertainment America 919 East Hillsdale Blvd. Foster City, CA 94404, U.S.A. Sony Computer Entertainment Europe 30 Golden Square London W1F 9LD, U.K. The SPU2 Overview is supplied pursuant to and subject to the terms of the Sony Computer Entertainment PlayStation® license agreements. The SPU2 Overview is intended for distribution to and use by only Sony Computer Entertainment licensed Developers and Publishers in accordance with the PlayStation® license agreements. Unauthorized reproduction, distribution, lending, rental or disclosure to any third party, in whole or in part, of this book is expressly prohibited by law and by the terms of the Sony Computer Entertainment PlayStation® license agreements. Ownership of the physical property of the book is retained by and reserved by Sony Computer Entertainment. Alteration to or deletion, in whole or in part, of the book, its presentation, or its contents is prohibited. The information in the SPU2 Overview is subject to change without notice. The content of this book is Confidential Information of Sony Computer Entertainment. and PlayStation® are registered trademarks, and GRAPHICS SYNTHESIZERTM and TM EMOTION ENGINE are trademarks of Sony Computer Entertainment Inc. All other trademarks are property of their respective owners and/or their licensors. ® © SCEI -2- SCE CONFIDENTIAL SPU2 Overview Version 6.0 About This Manual The "SPU2 Overview" describes the functions, configuration, and sound generation mechanisms of the SPU2, the sound processor for the PlayStation 2. - Chapter 1 "Overview of SPU2" describes the configuration and main functions of the SPU2 and the format of waveform data used as a sound source. - Chapter 2 "Sound Generation" describes voice generation using waveform data as a sound source, sound data stream input/output, mixing, and effects. - Chapter 3 "Register List" describes the registers that control the SPU2. - Chapter 4 "Appendix" shows the volume variation rates for values of the envelope rate parameters. Changes Since Release of 5th Edition Since release of the 5th Edition of the SPU2 Overview Manual, the following changes have been made. Note that each of these changes is indicated by a revision bar in the margin of the affected page. Ch. 2: Sound Generation • Section 2.2.6. Monaural Output, has been added on page 29. • A correction has been made to the description following Figure 2-12 Sound Data Output on page 31. Ch. 4: Appendix • A correction has been made to the “Time” heading in the Exponential Decrement Mode table on page 76. • A correction has been made to the “0.1110” row in the Linear Decrement Mode table on page 78. © SCEI -3- SCE CONFIDENTIAL SPU2 Overview Version 6.0 (This page is left blank intentionally) © SCEI -4- SCE CONFIDENTIAL SPU2 Overview Version 6.0 Glossary Term EE EE Core COP0 COP1 COP2 GS GIF IOP SBUS VPU (VPU0/VPU1) VU (VU0/VU1) VIF (VIF0/VIF1) VIFcode SPR IPU word qword Slice Packet Transfer list Tag DMAtag GS primitive Context GIFtag Display list Definition Emotion Engine. CPU of the PlayStation 2. Generalized computation and control unit of EE. Core of the CPU. EE Core system control coprocessor. EE Core floating-point operation coprocessor. Also referred to as FPU. Vector operation unit coupled as a coprocessor of EE Core. VPU0. Graphics Synthesizer. Graphics processor connected to EE. EE Interface unit to GS. Processor connected to EE for controlling input/output devices. Bus connecting EE to IOP. Vector operation unit. EE contains 2 VPUs: VPU0 and VPU1. VPU core operation unit. VPU data decompression unit. Instruction code for VIF. Quick-access data memory built into EE Core (Scratchpad memory). EE Image processor unit. Unit of data length: 32 bits Unit of data length: 128 bits Physical unit of DMA transfer: 8 qwords or less Data to be handled as a logical unit for transfer processing. A group of packets transferred in serial DMA transfer processing. Additional data indicating data size and other attributes of packets. Tag positioned first in DMA packet to indicate address/size of data and address of the following packet. Data to indicate image elements such as point and triangle. A set of drawing information (e.g. texture, distant fog color, and dither matrix) applied to two or more primitives uniformly. Also referred to as the drawing environment. Additional data to indicate attributes of GS primitives. A group of GS primitives to indicate batches of images. © SCEI -5- SCE CONFIDENTIAL SPU2 Overview Version 6.0 (This page is left blank intentionally) © SCEI -6- SCE CONFIDENTIAL SPU2 Overview Version 6.0 Contents 1. Overview of SPU2................................................................................................................................................................... 9 1.1. Features of SPU2 ........................................................................................................................................................... 10 1.1.1. Core.......................................................................................................................................................................... 10 1.1.2. Sound Data Input Function.................................................................................................................................. 10 1.1.3. Voice Processing Function ................................................................................................................................... 10 1.1.4. Sound Data Output Function............................................................................................................................... 11 1.1.5. Digital Effect Processing....................................................................................................................................... 11 1.1.6. Local Memory......................................................................................................................................................... 11 1.1.7. External Output...................................................................................................................................................... 12 1.2. Local Memory ................................................................................................................................................................ 13 1.2.1. Data Allocated in Local Memory ......................................................................................................................... 13 1.2.2. Addressing in Local Memory................................................................................................................................ 14 1.2.3. Interrupt by Access ................................................................................................................................................ 14 1.3. Waveform Data Format................................................................................................................................................ 15 1.3.1. Waveform Data Block ........................................................................................................................................... 15 1.3.2. Endpoint.................................................................................................................................................................. 15 1.3.3. Loop Processing ..................................................................................................................................................... 16 1.4. Reset ................................................................................................................................................................................ 17 2. Sound Generation ................................................................................................................................................................. 19 2.1. Voice Processing ............................................................................................................................................................ 20 2.1.1. Sound Sources......................................................................................................................................................... 20 2.1.2. Pitch Transformation............................................................................................................................................. 22 2.1.3. Pitch Modulation.................................................................................................................................................... 23 2.1.4. Envelope.................................................................................................................................................................. 24 2.1.5. Volume .................................................................................................................................................................... 25 2.1.6. Key-On/Key-Off................................................................................................................................................... 26 2.1.7. Mixing Switch ......................................................................................................................................................... 27 2.2. Sound Data Input Processing....................................................................................................................................... 28 2.2.1. Sound Data Input Area ......................................................................................................................................... 28 2.2.2. Volume Processing................................................................................................................................................. 28 2.2.3. Mixing with Voice Output .................................................................................................................................... 29 2.2.4. Bypass Processing (CORE0) ................................................................................................................................ 29 2.2.5. 32-bit Sound Data Input (CORE1) ..................................................................................................................... 29 2.2.6. Monaural Output.................................................................................................................................................... 29 2.3. Sound Data Output Processing ................................................................................................................................... 30 2.4. Mixing.............................................................................................................................................................................. 32 2.5. Digital Effect Processing .............................................................................................................................................. 33 2.5.1. Signal Flow for Digital Effect Processing........................................................................................................... 33 2.5.2. Work Area for Digital Effect Processing ............................................................................................................ 33 2.5.3. Effect Volume ........................................................................................................................................................ 34 © SCEI -7- SCE CONFIDENTIAL SPU2 Overview Version 6.0 2.6. Master Volume ...............................................................................................................................................................35 2.7. Interrupt Processing.......................................................................................................................................................36 3. Register List ............................................................................................................................................................................39 3.1. Classification of Registers..............................................................................................................................................40 3.2. Registers in Pairs.............................................................................................................................................................41 VOLL / VOLR : Voice volume......................................................................................................................................42 PITCH : Pitch when sound is generated........................................................................................................................44 ADSR1 / ADSR2 : Envelope..........................................................................................................................................45 ENVX : Current value of envelope ................................................................................................................................46 VOLXL / VOLXR : Current value of volume.............................................................................................................47 PMON0 / PMON1 : Pitch modulation specification..................................................................................................48 NON0 / NON1 : Voice allocation to noise generator................................................................................................49 VMIX* : Mixing specification of voice output .............................................................................................................50 MMIX : Output specification after voice mixing..........................................................................................................51 IRQAH / IRQAL : Interrupt address specification.....................................................................................................52 KON0 / KON1 : Key-on specification.........................................................................................................................53 KOF0 / KOF1 : Key-off specification..........................................................................................................................54 TSAH / TSAL : Transfer start address..........................................................................................................................55 SSAH / SSAL : Starting address of waveform data .....................................................................................................56 LSAXH / LSAXL : Address of loop point ...................................................................................................................57 NAXH / NAXL : Address of waveform data to be read next...................................................................................58 ESAH / ESAL : Starting address in the work area for effect processing .................................................................59 EEAH : End address in the work area for effect processing......................................................................................60 ENDX0 / ENDX1 : Endpoint passing flag .................................................................................................................61 MVOLL / MVOLR : Master volume ............................................................................................................................62 EVOLL / EVOLR : Return volume of effect ..............................................................................................................64 AVOLL / AVOLR : Volume for external input ..........................................................................................................65 BVOLL / BVOLR : Volume for sound data input......................................................................................................66 MVOLXL / MVOLXR : Current value of master volume.........................................................................................67 4. Appendix.................................................................................................................................................................................69 4.1. Rate Parameter Table.....................................................................................................................................................70 4.1.1. +Lin Mode...............................................................................................................................................................71 4.1.2. –Lin Mode ...............................................................................................................................................................72 4.1.3. +Exp Mode (Normal phase).................................................................................................................................73 4.1.4. +Exp Mode (Reverse phase).................................................................................................................................74 4.1.5. –Exp Mode ..............................................................................................................................................................75 4.1.6. Decay Rate (DR) .....................................................................................................................................................76 4.1.7. Sustain Level (SL) ...................................................................................................................................................77 4.1.8. –Lin Mode for Release Rate (RR).........................................................................................................................78 4.1.9. –Exp Mode for Release Rate (RR) .......................................................................................................................79 © SCEI -8- SCE CONFIDENTIAL SPU2 Overview Version 6.0 1. Overview of SPU2 This chapter provides an overview of the SPU2. © SCEI -9- SCE CONFIDENTIAL SPU2 Overview Version 6.0 1.1. Features of SPU2 The SPU2 is a sound synthesis processor, which is composed of two cores. The SPU2 also contains local memory and external I/O. The two cores have the ability to: • Reproduce sound data input successively from the host. • Process voices. • Output voice-processed sound data to the host successively. • Perform digital effect processing. The following sections describe the SPU2 components and core functions. 1.1.1. Core The cores (CORE0 and CORE1) are the basic components of the SPU2, each having a sound generation function with 24 voices. They operate at a frequency of 36.864 MHz, and have a sound generation resolution of 48 kHz. The unit of processing, 1/48000 second, is represented as 1Ts. When setting the same register successively (e.g. when varying the pitch of a sound consecutively to realize a portamento), write operations to the register must be at least 1Ts apart. (Write operations to some registers must be at least 2 Ts apart. For details, refer to the Description for each register in "3. Register List".) If the register is written in less than the specified time interval (less than 1 Ts when not specified), the SPU2 operations become indeterminate, and expected results cannot be obtained. This produces serious effects, particularly on registers working as a switch, such as key-on or key-off. CORE0 and CORE1 are functionally equal and operate independently. They are connected in such a way that the output from CORE0 is input to CORE1 and the final mixed sound is output from CORE1. 1.1.2. Sound Data Input Function The sound data input function processes 16-bit or 32-bit data strings transferred successively from the host to the SPU2 as sound data, and outputs them by mixing with the voice-processed output. CORE0 and CORE1 each have one stereo input channel. The input buffer is a reserved area in the local memory. To transfer data smoothly, it uses a double buffer function, which requests a data transfer when half of the area is processed. For details, refer to "2.2. Sound Data Input Processing". 1.1.3. Voice Processing Function Sound in the SPU2 is generated in units of voices. Each core has 24 voices, so the whole SPU2 can generate 48 voices. Each voice has waveform data compressed by ADPCM as a sound source. After pitch transformation, pitch modulation and envelope processing, the outputs of each voice are mixed and become the final sound output. © SCEI -10- SCE CONFIDENTIAL SPU2 Overview Version 6.0 Sound Source The sound source is waveform data that has been compressed by ADPCM, and is decoded (decompressed) by hardware at a sampling rate of 48 kHz. The noise generator of each core can be used as a sound source as well. Pitch Transformation Sound can be generated by varying the pitch of the sound source within the range of -12 octaves to +2 octaves. Pitch Modulation The pitch of the sound source can be modulated by using the crest value of another voice. Envelope Processing The envelope, which controls volume variation from key-on to key-off, is specified with five parameters: Attack Rate, Decay Rate, Sustain Rate, Sustain Level and Release Rate. For Attack Rate, Sustain Rate and Release Rate, non-linear variation can be specified. Voice Volume Volume can be set for the L channel and R channel of each voice. Constant, linear and exponential variation curves can be selected. Mixing/Switching 24 voices/stereo output (48 channels in total) are synthesized into 2 stereo units in each core. Each voice can be added to the output or not. Refer to "2.1. Voice Processing" for details of voice processing. 1.1.4. Sound Data Output Function Mixed sounds (2 stereo units per core) and a specified two-channel voice can be output to the host successively. This allows the sound data generated by the SPU2 to be processed by the host processor. An output buffer is reserved in the local memory. In order to transfer data smoothly, it uses a double buffer function, which requests a data transfer when half of the area is processed. Refer to “2.3. Sound Data Output Processing for details. 1.1.5. Digital Effect Processing Digital effects such as reverb, echo and delay can be applied to the mixed sounds. Although digital effects can be processed independently in each core, the effects in CORE1 can be reapplied to the final output from CORE0. The work area for digital effect processing is in the local memory. Refer to "2.5. Digital Effect Processing" for details. 1.1.6. Local Memory The SPU2 has 2 MBytes (16 Mbits) of local memory for its exclusive use. This memory is used as a buffer for sound data I/O, a waveform data area for voice processing, and a work area for digital effect processing. The remaining memory can be used freely. © SCEI -11- SCE CONFIDENTIAL SPU2 Overview Version 6.0 It is possible to access the local memory from the host via DMA transfer or single transfer. Sound generation never stops when transferring data, but it might not be correctly executed when overwriting waveform data being used by the SPU2. To avoid this, an interrupt can be generated for the host when each core accesses a specific address in the local memory. Refer to "1.2. Local Memory" for details. 1.1.7. External Output The SPU2 adopts the following functions as methods of outputting final sounds. • Digital output through S/PDIF (Sony/Philips Digital Interface) • Analog output through D/A converter © SCEI -12- SCE CONFIDENTIAL SPU2 Overview Version 6.0 1.2. Local Memory The local memory is memory for the SPU2’s exclusive use; it is used as a data I/O buffer with the host and as a work area of the SPU2. 1.2.1. Data Allocated in Local Memory The local memory is divided into the following four areas. Host Memory Host SPU2 Local Memory Sound Data Output Area Sound Data Input Area Address (in short-word units) 00 0000 00 1FFF 00 2000 00 27FF 00 2800 Waveform Data Area Work Area for Digital Effect Delay Processing 0X XXXX 0? FFFF 0y yyyy 0F FFFF (2 MB) Figure 1-1 Memory Allocation and Addressing in Local Memory Sound Data Input Area This is the sound data input buffer from the host; its address is fixed. Sound data is successively written from the host, and the written data is sequentially read via hardware and processed as sound data by the SPU2. Sound Data Output Area This is the sound data output buffer from the SPU2 to the host; its address is fixed. Sound data generated in the SPU2 is written successively, and is readable from the host sequentially. © SCEI -13- SCE CONFIDENTIAL SPU2 Overview Version 6.0 Digital Effect Work Areas The work areas used by the cores for digital effect delay processing are in 2 locations. The start address of each location can be set freely, but the end address has restrictions on alignment. 1.2.2. Addressing in Local Memory The local memory is configured in 16-bit units; addresses are allocated every 16 bits (short word). Each address is specified with a 32-bit value, of which 22 bits are enabled. The lower 20 bits of the 22 bits show a range of 2 Mbytes, and the upper 2 bits are set to 00. Since the SPU2 registers are configured in 16-bit units, each register which performs addressing is a pair of high and low registers. 16 bits 00 0000 15 Address H 5 0 128 KB XX Higher Address Register XX 0000 XX YYYY 128 KB XX FFFF 2 MB 15 Address L Lower Address Register 0 YYYY 0F FFFF Figure 1-2 Addressing 1.2.3. Interrupt by Access When one of the cores accesses a specific address in the local memory, an interrupt can be generated for the host. The address for generating an interrupt can be set, one per core, with the IRQAH and IRQAL registers. Interrupts generated by both cores are detected at the same time on the host. A function provided by the sound library represents which core has generated the interrupt. CORE0 [IRQAH/L] Local Memory Interrupt to Host Arbitrary Processing Figure 1-3 Interrupt by Process Access (Example: Specification in CORE0) For details, refer to "2.7. Interrupt Processing". © SCEI -14- SCE CONFIDENTIAL SPU2 Overview Version 6.0 1.3. Waveform Data Format The waveform data that becomes the sound source of each voice is in a format unique to the SPU2, by adopting ADPCM as a compression method. 1.3.1. Waveform Data Block Waveform data is configured in units of 16-byte blocks, each of which includes a 16-bit header and 28 4-bit samples. The following attributes are included in the header. Field LOOP/START LOOP LOOP/END DECODE Bit Position 10 9 8 7:0 15 11 Header 7 Loop Flag SD3 SD7 SD11 SD15 SD19 SD23 SD27 Sound Sample Data 7 short words Loop Flag Contents Loop point information Loop existence/non-existence Endpoint information Parameter for decoding 15 SD2 SD6 SD10 SD14 SD18 SD22 SD26 11 (reserved) 10 L O O P / S T A R T 3 0 Decode Parameter SD1 SD0 SD5 SD4 SD9 SD8 SD13 SD12 SD17 SD16 SD21 SD20 SD25 SD24 9 L O O P 8 L O O P / E N D Figure 1-4 1 Block of Waveform Data 1.3.2. Endpoint The number of blocks of waveform data is arbitrary. By setting the LOOP/END bit of the header to 1, the endpoint is specified, showing the position where the waveform data ends. When sound generation reaches the block specified as the endpoint, the last sample data of the block is processed, and then sound generation moves to the block which has the loop point specification immediately before (i.e. the block shown by the LSAXH/L register). When no loop is specified, the last sample data of the block is processed, and then muting is applied to the voice in process by hardware. As a result, sound generation of the voice stops. © SCEI -15- SCE CONFIDENTIAL SPU2 Overview Version 6.0 1.3.3. Loop Processing By setting the LOOP/START bit of the header to 1, the loop point is specified. The header address is maintained in the LSAXH/L register when sound generation moves to this block, and sound generation moves to the first sample data of this block after processing the endpoint block. Only 1 loop point specification can be in effect in a set of waveform data. If there are two or more loop point blocks, the block closest to the endpoint becomes the loop point when sound is actually generated. If an address is set in the LSAXH/L register after sound generation has started, the loop point specification in the waveform data is disregarded until the next time the voice is keyed on, and the set address becomes the loop point. For waveform data with a loop point set, the LOOP bit of the header must be set to 1 in all blocks. [SSAH / L] LOOP=1 Start Block [LSAXH / L] LOOP=1, LOOP/START=1 Loop Point Block LOOP=1, LOOP/END=1 Endpoint Block Figure 1-5 Loop Processing © SCEI -16- SCE CONFIDENTIAL SPU2 Overview Version 6.0 1.4. Reset The sound library provides a resetting feature. In resetting, neither register values nor data in the local memory is guaranteed. © SCEI -17- SCE CONFIDENTIAL SPU2 Overview Version 6.0 (This page is left blank intentionally) © SCEI -18- SCE CONFIDENTIAL SPU2 Overview Version 6.0 2. Sound Generation This chapter describes sound generation in each core. © SCEI -19- SCE CONFIDENTIAL SPU2 Overview Version 6.0 2.1. Voice Processing Voice processing generates sound primarily by decoding waveform data and varying the decoded sound data by time. The generated sound is transferable to the host as data, and can also be processed on the host. The entire flow of voice processing is shown as follows: OUTX(i-1) Local Memory [PITCH] Pitch Modulation [SSAH/L] [LSAXH/L] VOL(L) Control [ADSR1] [ADSR2] [VMIXL] Envelope Control Local Memory Waveform data Sound Data Output [VOLL] [VOLXL] [PMON] ADPCM Decode Pitch Control Dry L ƒ° Wet L ƒ° Dry R [VMIXEL] [VMIXR] [NON] Noise Generator ƒ° Wet R ƒ° [VMIXER] [MMIX] VOL(R) Control [VOLR] [VOLXR] OUTX(i) Figure 2-1 Voice Processing 2.1.1. Sound Sources Waveform data stored in the local memory or the noise generator (each core has 1 unit) can be used as a sound source for each voice. This selection is specified by the NON0/1 register. [PMON0/1] [PITCH] [SSAH / L] Waveform Data ADPCM Decoding Pitch Processing Local Memory [NON0/1] Noise Generator Figure 2-2 Switching between Waveform Data and Noise Generator © SCEI -20- SCE CONFIDENTIAL SPU2 Overview Version 6.0 When waveform data is the sound source, the address in the local memory where the waveform data is located is specified by the SSAH/L register of the voice attribute. Specify the location of the header in the waveform data block to the SSAH/L register. Since each voice generates a single tone, it is necessary to vary the pitch of the same sound in two or more voices to generate a chord. In this case, however, the same waveform data address is specified to the SSAH/L register in each voice. The waveform data is decoded by hardware. The user can know the decoding progress from each of the following registers: • Address of the waveform data to be read next (NAXH/L register) • Header address in the loop point block (LSAXH/L register: after passing the loop point) • Endpoint block passing flag (ENDX0/1 register) The NAXH/L register is incremented as decoding advances and shows up to which sample data the sound generation has been completed. That is, sound processing has been completed up to the address immediately preceding the one indicated by the NAXH/L register. Since the waveform data includes the header, however, the above does not mean all the addresses before the one indicated by the NAXH/L register have been processed. When replacing the sound-generated waveform data, replacements can be made in block units up to the block immediately before the one having the address specified by the NAXH/L register. (However, if a replacement is made at a place between the loop point and endpoint in waveform data including loop processing, decoding becomes discontinuous and might cause a noise.) When decoding advances to the loop point block, the header address in the block is written to the LSAXH/L register. The loop point can be changed by rewriting the value of this register while sound is being generated 4 Ts after the key-on. (Rewriting within 4 Ts is disregarded.) When rewriting, specify the location of the header in the waveform data block in the LSAXH/L register. After the change, the address in this LSAXH/L register is used as a loop point, and the loop point information in the header of the waveform data is disregarded until the next time the voice is keyed on. When decoding of the endpoint block is finished, regardless of the presence of the loop, the bit corresponding to the voice is set to 1 in the ENDX register and kept until the next key-on. © SCEI -21- SCE CONFIDENTIAL SPU2 Overview Version 6.0 SSAH/L LSAXH/L LOOP/START=1 NAXH/L ENDX0/1 1 LOOP/END=1 Figure 2-3 Sound Generation Status 2.1.2. Pitch Transformation Sound can be generated from waveform data by varying the pitch within the range of -12 to +2 octaves. The pitch transformation is specified by the PITCH register of voice attribute. Assuming the original pitch of the sound source is f0, the value of the PITCH register is [PITCH], and the finally generated sound is f, the following expression is met: f= [PITCH] f0 212 That is, sound is generated to the pitch of the original sound by specifying 0x1000(=212) in the PITCH register. The above relationship is met only when the waveform data is sampled at 48 kHz. The sound from waveform data sampled at a rate other than 48 kHz (24 kHz, for example) is generated one octave higher than the original sound when specifying 0x1000 in the PITCH register to generate sound. Assuming the sampling rate to be s kHz, the above-mentioned expression is expanded as follows. 48 [PITCH ] f0 s 212 If an appropriate value is specified to the PITCH register according to this expression, sound can be generated to the pitch of the original sound even from the waveform data whose sampling rate is not 48 kHz. However, the acoustic characteristic varies along with pitch transformation processing. f= © SCEI -22- SCE CONFIDENTIAL SPU2 Overview Version 6.0 -2 octaves -1 octave Pitch of Original Sound +1 octave +2 octaves 0x400 0x800 0x1000 0x2000 0x3fff [PITCH] Figure 2-4 Pitch Transformation of Waveform Data Sampled at 48 kHz -1 octave Pitch of Original Sound +1 octave 0x400 0x800 0x1000 [PITCH] Figure 2-5 Pitch Transformation of Waveform Data Sampled at 24 kHz When the sound source is the noise generator, the acoustic pitch can be specified in each core by using the sound library. When there are two or more voices allocated to the noise generator in each core, their sound will be all generated at the same acoustic pitch. The speed of sound generation advance changes with the specification of the pitch. The lower the pitch is set, the slower the advance in the transition of the waveform data address shown by the NAXH/L register becomes. 2.1.3. Pitch Modulation In two voices with consecutive voice numbers, voice n can be modulated by using the output value from voice n-1. The value used for modulation is the product of the crest value immediately after decoding and the envelope value in the waveform data for voice n-1, which is 1Ts before in terms of time. This is called the OUTX of voice n-1. OUTX is not reflected in the register. [PITCH] Voice n-1 Waveform Data Decoding Processing 1Ts before Voice n Waveform Data Decoding Pitch Transformation Envelope Processing (OUTX) [PITCH] [PMON0/1.Vn] Pitch Transformation Envelope Processing Figure 2-6 Pitch Modulation © SCEI -23- SCE CONFIDENTIAL SPU2 Overview Version 6.0 Assuming the crest value of voice n-1 used for modulation of voice n to be OUTX and the pitch of voice n to be P, then P', the value to be used for pitch transformation of voice n, is decided by the following expression. P' = P (1 + OUTX) Whether pitch modulation is performed or not can be specified with the PMON0/1 register. When not performed, the result becomes the same as the case for OUTX=0. 2.1.4. Envelope The envelope specifies the volume variation by time from key-on to key-off according to the following five parameters. Parameter Attack Rate Decay Rate Sustain Level Sustain Rate Code AR DR SL SR Release Rate RR Description Rising immediately after key-on Attenuation from the maximum value Transition point from Decay to Sustain Attenuation (or increment) from Sustain Level to key-off Attenuation after key-off For Attack Rate, Sustain Rate and Release Rate, curves of variation by time can be selected. Attack Rate and Release Rate can use two kinds and Sustain Rate can use four kinds of curves as shown in the table below. Decay Rate is fixed to an exponential decrement curve. Parameter Attack Rate Sustain Rate Release Rate Selection Linear increment (+lin) Pseudo exponential increment (+exp) Linear increment (+lin) Linear decrement(-lin) Pseudo exponential increment (+exp) Exponential decrement (-exp) Linear decrement (-lin) Exponential decrement (-exp) Set Value ADSR1.X=0 ADSR1.X=1 ADSR2.Y=000 ADSR2.Y=010 ADSR2.Y=100 ADSR2.Y=110 ADSR2.Z=0 ADSR2.Z=1 The "pseudo exponential increment" is a variation in line, in which linear volume increment is lowered in increment rate when 75% of the maximum value (0x6000) is exceeded. © SCEI -24- SCE CONFIDENTIAL SPU2 Overview Version 6.0 1 SR(+exp) DR(-exp) 0.75 SR(+lin) RR(-lin) SL AR(+exp) SR(-lin) RR(-exp) AR(+lin) SR(-exp) RR(-lin) RR(-exp) 0 key-off key-on t Figure 2-7 Parameters for Envelope and Their Curves The value of the envelope varies successively, but the value is reflected in the ENVX register and can be referred to. When sound is generated from loop-less waveform data, ENVX is set to 0 regardless of the envelope status, at the moment the ENDX register bit corresponding to the voice is set to 1. 2.1.5. Volume In each voice, the volume for the L channel and R channel can be set independently. By setting different values, the panpot can be configured. The volume settings are made in the VOLL and VOLR registers. [VOLL] Volume(L) Control Envelope Control Volume(R) Control [VOLR] Figure 2-8 Volume Processing © SCEI -25- SCE CONFIDENTIAL SPU2 Overview Version 6.0 The mode specification, by which the volume varies over time, can be set for the L channel and R channel independently, as in the case of the Sustain Rate of the envelope. Mode Direct +lin -lin +exp -exp VOLL/R Register Setting Bit 15=0 Bit 15:13=100 Bit 15:13=101 Bit 15:13=110 Bit 15:13=111 Volume Variation Constant (No variation) Linear increment Linear decrement Pseudo exponential increment Exponential decrement First specify the standard volume in direct mode and start sound generation, even when specifying a mode other than Direct mode. If a mode other than Direct mode is specified again later with a variation rate, variation of the volume starts instantly. The phase can be reversed by setting a negative value in Direct mode. 1 +exp +lin 0.75 Direct -exp 0 Key on -lin VOLL/R Register Setting t Figure 2-9 Volume Variation by Time The volume varies successively in modes other than Direct mode, but the value is reflected in the VOLX register and can be referred to. 2.1.6. Key-On/Key-Off Key-on (starting sound generation) and key-off (stopping sound generation) can be controlled by the KON0/1 and KOF0/1 registers, respectively, for each voice. At key-on, sound generation of the waveform data indicated by the SSAH/L register is started according to the parameters of pitch, envelope, volume, etc. At key-off, the envelope enters the Release phase and sound generation stops according to the Release Rate. If there is no loop specification in the waveform data, sound generation of the voice ends when reaching the endpoint block of waveform data even before key-off, and muting is applied by hardware. © SCEI -26- SCE CONFIDENTIAL SPU2 Overview Version 6.0 2.1.7. Mixing Switch The output from each voice is mixed into four channels of Dry L, Wet L, Dry R and Wet R in each core through the control of the mixing switch. For the mixing switch, on/off of each voice’s output to the corresponding channels is specified in the VMIXL0/1, VMIXR0/1, VMIXEL0/1 and VMIXER0/1 registers. When a voice’s output to all the channels is off, it is equivalent to a voice to which muting is applied. The mixing results are successively stored in the local memory sound data output area, and become the final output from the core through switching by the MMIX register at the same time. For the final output, on/off can be specified for Dry L, Wet L, Dry R and Wet R independently. Turning all the MMIX switches off is equivalent to applying muting to all the voice outputs. Channel Dry L (Voice direct output) Dry R (Voice direct output) Wet L (Voice effect output) Wet R (Voice effect output) Mixing Switch VMIXL0/1 VMIXR0/1 VMIXEL0/1 VMIXER0/1 Output Switch MMIX.MSNDL MMIX.MSNDR MMIX.MSNDEL MMIX.MSNDER Volume(L) Control Dry L Wet L Dry R Wet R Volume(R) Control Mixing Mixing of Switches for All Voices Each voice to 4 Channels [VMIXL] [VMIXEL] [VMIXR] [VMIXER] Switches for To Sound Data Mixing Results Output Area in Local Memory [MMIX] Figure 2-10 Voice Mixing © SCEI -27- SCE CONFIDENTIAL SPU2 Overview Version 6.0 2.2. Sound Data Input Processing Successive 16-bit data (little endian) transfer to the local memory sound data input area by the host enables the SPU2 to apply volume processing to the data as sound data, mix it with the output from voice processing, and then apply digital effects to it. It is also possible to output the sound data to the output block directly by bypassing the internal processing in the SPU2. 2.2.1. Sound Data Input Area Each core is provided with one stereo unit as the sound data input. The addresses in the input area are as follows. These areas are reserved areas, and other data cannot be placed there. Address 2000-21FF 2200-23FF 2400-25FF 2600-27FF Area CORE0 MEMIN(L) CORE0 MEMIN(R) CORE1 MEMIN(L) CORE1 MEMIN(R) Description CORE0 L channel sound data input area CORE0 R channel sound data input area CORE1 L channel sound data input area CORE1 R channel sound data input area The sound data input area is 512 short words (1024 bytes) in size and is composed of double buffers of 256 short words. Data transfer to the sound data input area can be realized easily by using the Auto DMA write transfer. For details, refer to the appropriate sound library document. Local Memory Host [MINL/R] [MINEL/ER] Volume [BVOL] Voice Processing Figure 2-11 Sound Data Input 2.2.2. Volume Processing The volume can be set to the sound obtained from the sound data input with the BVOLL and BVOLR registers. However, the mode of variation by time cannot be specified and the constant mode is always set. © SCEI -28- SCE CONFIDENTIAL SPU2 Overview Version 6.0 2.2.3. Mixing with Voice Output The sound obtained from the sound data input can be mixed with the Dry L/Wet L/Dry R/Wet R channel of the voice output. This is controlled by the flags of the MMIX register. Sound Data Input MEMIN(L) MEMIN(R) MEMIN(EL) MEMIN(ER) Voice Output Dry L Dry R Wet L Wet R Switch MMIX.MINL MMIX.MINR MMIX.MINEL MMIX.MINER 2.2.4. Bypass Processing (CORE0) It is possible to specify a mode that connects directly to the S/PDIF digital output of the output block by bypassing internal processing, for sound data input to CORE0. Since volume processing is not performed, the encoded data, which is not ordinary 16-bit digital data, can be directly output from the host. In this mode, however, CORE1 output, the final output from the core, is not connected to the S/PDIF digital output. Moreover, this switching is performed only for the S/PDIF digital output and sound data input to CORE0 cannot be directly connected to the D/A converter output. For details, refer to the appropriate sound library document. 2.2.5. 32-bit Sound Data Input (CORE1) It is possible to specify a mode that connects directly to the output block by bypassing the internal processing, for sound data input to CORE1. In this mode, the sound data input is processed in 32-bit units (24 bits enabled and lower 8 bits disabled). It can mix the 32-bit data (little endian) transferred from the host as higher quality digital data with the D/A converter output or the S/PDIF digital output. The sound data input area is processed in 32-bit units in this mode. Since the unit of data volume doubles, the speed of data reading also doubles. For details, refer to the appropriate sound library document. 2.2.6. Monaural Output The SPU2 does not have the ability to generate monaural sound from the sound obtained from the sound data input. If monaural output is required, prepare monaural sound data at the authoring level. © SCEI -29- SCE CONFIDENTIAL SPU2 Overview Version 6.0 2.3. Sound Data Output Processing The SPU2 can write the generated 16-bit sound data to the local memory sound data output area at any time. Various processes can be applied to the sound data being generated by the SPU2 by reading the data successively on the host. The contents of the sound data to be output and the output area are fixed as shown in the table. This area is a reserved area, and cannot be used for other purposes. Core CORE0 Channel Voice1 Voice3 CORE1 MEMOUT(L) MEMOUT(R) MEMOUT(EL) MEMOUT(ER) SIN(L) SIN(R) Voice1 Voice3 MEMOUT(L) MEMOUT(R) MEMOUT(EL) MEMOUT(ER) Contents Crest value after multiplication of envelope in voice1 Crest value after multiplication of envelope in voice3 Dry L after mixing 24 voices Dry R after mixing 24 voices Wet L after mixing 24 voices Wet R after mixing 24 voices CORE0 output L CORE0 output R Crest value after multiplication of envelope in voice1 Crest value after multiplication of envelope in voice3 Dry L after mixing 24 voices Dry R after mixing 24 voices Wet L after mixing 24 voices Wet R after mixing 24 voices © SCEI -30- Output Area 000400 – 0005FF 000600 – 0007FF 001000 – 0011FF 001200 – 0013FF 001400 – 0015FF 001600 – 0017FF 000800 – 0009FF 000A00 – 000BFF 000C00 – 000DFF 000E00 – 000FFF 001800 – 0019FF 001A00 – 001BFF 001C00 – 001DFF 001E00 – 001FFF SCE CONFIDENTIAL SPU2 Overview Version 6.0 Local Memory Host 00 0000 00 0400 00 1FFF (Reserved) Voice Processing CORE0 Voice1 CORE0 Voice3 CORE1 SIN(L) CORE1 SIN(R) CORE1 Voice1 CORE1 Voice3 CORE0 MEMOUT(L) CORE0 MEMOUT(R) CORE0 MEMOUT(EL) CORE0 MEMOUT(ER) CORE1 MEMOUT(L) CORE1 MEMOUT(R) CORE1 MEMOUT(EL) CORE1 MEMOUT(ER) Mixing Voice Processing Mixing Output Block Figure 2-12 Sound Data Output The output area is 512 short words (1024 bytes) in size for each channel, and is composed of double buffers of 256 short words each. Data can be read easily from the sound data output area by using the Auto DMA read transfer. For details, refer to the appropriate sound library document. © SCEI -31- SCE CONFIDENTIAL SPU2 Overview Version 6.0 2.4. Mixing Mixing performs volume processing and switching to voice output, sound data input and external input, and distributes the data to direct output and digital effects. Each core has the following four sources: Source Voice Direct Output (Dry L/R) Voice Effect Output (Wet L/R) Sound Data Input External Input (Final Output from CORE0) Volume - Switch MMIX.MSNDL/R Destination Direct Output - MMIX.MSNDEL/ER Digital Effect BVOLL/R MMIX.MINL/R MMIX.MINEL/ER MMIX.SINL/R MMIX.SINEL/ER Direct Output Digital Effect Direct Output Digital Effect AVOLL/R * External input is performed only to CORE1. Dry L/R Voice Processing [MSNDL/R] Wet L/R [MSNDEL/ER] To Direct Output [MINL/R] Sound Data Input [MINEL/ER] Volume [BVOL] [SINL/R] External Input [SINEL/ER] Volume To Digital Effect [AVOL] Figure 2-13 Mixing For sound data input and external input, only the constant mode is set for volume variation with time. © SCEI -32- SCE CONFIDENTIAL SPU2 Overview Version 6.0 2.5. Digital Effect Processing Various sound effects such as reverb, echo and delay can be added by performing digital effect processing on the sounds generated by each core and input from the sound data input. For other digital effect procedures, refer to the sound library document. 2.5.1. Signal Flow for Digital Effect Processing The input to digital effect processing is the mixture of the following sound sources as described in "2.4. Mixing". • Voice Effect Output (Wet L/R) • Sound Data Input • External Input *CORE1 only However, various connection forms can be taken by passing through the host with switch on/off setting and sound data output function. CORE0 CORE1 Voice Processing Voice Processing Mixing Effect Volume Sound Data Input Sound Data Input Mixing Effect Volume [EVOL] Digital Effect To Output Block [EVOL] External Input Digital Effect Local Memory Figure 2-14 Signal Flow of Digital Effects 2.5.2. Work Area for Digital Effect Processing When performing digital effect processing, it is necessary to reserve a work area in the local memory. The start address of the work area is specified with the ESAH/L register and the end address is specified with the EEAH register. Decide the start address by totaling the size required by each delay block for digital effects. The end address is specified in the upper 6 bits only, and it is assumed that the lower 16 bits are 0xFFFF. That is, the end address in the work area is always on a 64K short-word (128 KB) boundary in the local memory. © SCEI -33- SCE CONFIDENTIAL SPU2 Overview Version 6.0 CORE0 [ESAH / L] CORE1 [ESAH / L] CORE0 [EEAH] Digital Effect Work Area for CORE0 0? FFFF CORE1 [EEAH] Digital Effect Work Area for CORE1 0? FFFF Figure 2-15 Work Area for Digital Effects 2.5.3. Effect Volume When the output from a digital effect is mixed with the direct output, volume control can be performed to the output from the digital effect. The effect volume is set with the EVOLL/R register. The panpot of the effect can be decided by setting the L channel and R channel independently. Only the constant mode is set for the variation with time. The volume set by the EVOLL/R register corresponds to the return volume of the effect. The following volume settings of each source correspond to the send volume respectively. Source Voice Output Sound Data Input External Input Send Volume VOLL/R BVOLL/R AVOLL/R © SCEI -34- Remarks CORE1 only SCE CONFIDENTIAL SPU2 Overview Version 6.0 2.6. Master Volume The final volume is determined by adding the master volume to the mixing result of the output from the digital effect and direct output (Dry L/R). The master volume of the L channel and R channel can be set independently by the MVOLL and MVOLR register. By setting the values, the whole panpot is decided. To Output Mixing Effect Volume [EVOLL/R] Master Volume [MVOLL/R] Digital Effect Figure 2-16 Master Volume Processing The mode of varying the master volume with time can be specified as follows, as well as the voice volume: Mode Direct +lin -lin +exp -exp MVOLL/R Register Setting Bit 15=0 Bit 15:13=100 Bit 15:13=101 Bit 15:13=110 Bit 15:13=111 Volume Variation Constant (no variation with time) Linear increment Linear decrement Pseudo exponential increment Exponential decrement The L channel and R channel can be specified independently for the volume mode as well. When specifying a mode other than "Direct", first specify the standard volume value in Direct mode. Then, the volume variation starts when other volume mode variation rates are re-specified. Moreover, the phase can be reversed by specifying a negative value as a variation rate. The current value of the master volume can be read from the MVOLXL/MVOLXR register. © SCEI -35- SCE CONFIDENTIAL SPU2 Overview Version 6.0 2.7. Interrupt Processing In most SPU2 processing, when each core accesses a specific address in the local memory, an interrupt can be generated to the host. The address where an interrupt is generated is set by the IRQAH and IRQAL registers. CORE 0 [IRQAH / L] Local Memory Interrupt to Host Arbitrary Processing CORE 1 [IRQAH / L] Figure 2-17 Interrupt Processing No matter which core accesses this address, an interrupt is generated. Interrupts generated by both cores are detected at the same time on the host. An interface provided by the sound library represents which core has generated the interrupt. In the following situations, there are limitations and warnings applied to the relationship between each core's access and interrupts. Initial state and waveform data without a loop Voices without key-on after resetting the SPU2, and voices generated after decoding the LOOP/END block of loop-less waveform data, access the local memory unnecessarily (free-run the entire local memory area) as an internal operation of the SPU2. This may cause an unexpected interrupt. Therefore, when applying an interrupt to loop-less waveform data, suppress unnecessary accesses from a voice not in process, by adding a soundless block whose LOOP/START, LOOP, and LOOP/END bits are set to 1 to the end of the data. (For waveform data format, refer to "1.3. Waveform Data Format".) Actions for the voices to which key-on has not been applied after resetting are handled via the sound library. Position of waveform data which sets interrupt generation address If an address between the starting and end addresses in waveform data is specified to the SSAH/L or LSAXH/L register by mistake and the IRQA/H register is specified to the same address, an interrupt does not occur. Local memory data transfer When local memory data transfer is performed, the internal pointer stays at the following addresses at the end of the transfer: Transfer from host to local memory: TSAH/L + Data Size + 1 Transfer from local memory to host: TSAH/L + Data Size + 0 x 20 (Both are in short-word units) © SCEI -36- SCE CONFIDENTIAL SPU2 Overview Version 6.0 Therefore, if the IRQAH/L register has been set at these positions, an interrupt occurs after the transfer has been ended. Digital effect processing When digital effect processing is disabled, interrupts by digital effect processing are not generated even if the interrupt generation address is specified to the digital effect work area. © SCEI -37- SCE CONFIDENTIAL SPU2 Overview Version 6.0 (This page is left blank intentionally) © SCEI -38- SCE CONFIDENTIAL SPU2 Overview Version 6.0 3. Register List © SCEI -39- SCE CONFIDENTIAL SPU2 Overview Version 6.0 3.1. Classification of Registers The registers of the SPU2 are classified as follows: Voice basic parameter registers These registers show the basic parameters of each voice. Each voice in each core has a set of registers. Voice control parameter registers These registers control on/off of each function in voice processing. Each core has a set of registers. Addressing registers These registers perform address specification in the local memory. The registers, which show the upper 6 bits and the lower 16 bits in the address, are paired. Each core has a set of registers. Digital effect addressing registers These registers perform addressing related to digital effects. The registers, which show the upper 6 bits and the lower 16 bits in the address, are paired. Each core has a set of registers. Volume registers These registers specify the mixing volume of each voice. Each core has a set of registers. © SCEI -40- SCE CONFIDENTIAL SPU2 Overview Version 6.0 3.2. Registers in Pairs The SPU2 registers include registers to be used in pairs. Addresses in the local memory are specified by a pair of registers showing the upper 6 bits and the lower 16 bits of one address. For example, the starting address of the waveform data for each voice is maintained in the SSAH register (the upper 6 bits) and the SSAL register (the lower 16 bits). Therefore, SSAH and SSAL are treated as a pair and written as the SSAH/L register. The EEAH register is an exception to the addressing registers, and does not include a register which shows the lower 16 bits. Registers that specify the switching for each voice are composed of a pair of registers corresponding to Voice 015 and Voice 16-23. As for the specification of the voice output to the Dry L channel, for example, Voice0-15 and Voice16-23 are specified by the VMIXL0 register and VMIXL1 register respectively. These two registers are treated as a pair and written as the VMIXL0/1 register. © SCEI -41- SCE CONFIDENTIAL SPU2 Overview Version 6.0 VOLL / VOLR : Voice volume Constant specification mode (direct mode) 15 12 8 4 0 VALUE 0 Name VALUE Pos. 14:0 Format int 1:0:14 Contents Constant volume value The phase reverses for a negative value. Linear increment mode (+lin mode) 15 1 12 0 Name VALUE X 0 X Pos. 6:0 12 8 4 (reserved) Format int 0:7:0 int 0:1:0 0 VALUE Contents Addition constant per Ts Polarity specification 0 Normal phase (specifiable when the current value is positive.) Linear increment to +1.0 1 Reverse phase (specifiable when the current value is negative.) Linear decrement to –1.0 Linear decrement mode (-lin mode) 15 1 12 0 Name VALUE X 1 X Pos. 6:0 12 8 (reserved) Format int 0:7:0 int 0:1:0 4 0 0 VALUE Contents Subtraction constant per Ts Polarity specification 0 Normal phase (specifiable when the current value is positive.) Linear decrement to 0 1 Reverse phase (specifiable when the current value is negative.) Linear increment to 0 © SCEI -42- SCE CONFIDENTIAL SPU2 Overview Version 6.0 Pseudo inverse-exponential increment mode (+exp mode) 15 1 12 1 Name VALUE X 0 X Pos. 6:0 12 8 (reserved) Format int 0:7:0 int 0:1:0 4 0 0 VALUE Contents Addition constant per Ts Polarity specification 0 Normal phase (specifiable when the current value is positive.) Increment to +1.0 in a line 1 Reverse phase (specifiable when the current value is negative.) Decrement to –1.0 in a line Exponential decrement mode (-exp mode) 15 1 12 1 Name VALUE 1 Pos. 6:0 8 (reserved) Format int 0:7:0 4 0 0 VALUE Contents Multiplication constant per Ts Description These registers specify the volume for each voice. The values of the upper three bits specify the pattern of the volume variation with time. When specifying a mode other than constant mode, the volume value varies with time, because the value corresponding to the value of the VALUE field is added to, subtracted from or multiplied by the volume value per Ts. For the relationship between the VALUE field value and actual volume duration, refer to "4.1. Rate Parameter Table". © SCEI -43- SCE CONFIDENTIAL SPU2 Overview Version 6.0 PITCH : Pitch when sound is generated 15 0 12 8 0 Name VALUE 4 0 VALUE Pos. 13:0 Format int 0:14:0 Contents Pitch specification value Description This register specifies the pitch (degree of highness or lowness of sound) of each voice. Assuming the pitch of the original sound (waveform data) to be f0, the relationship between the VALUE, or the pitch specification, and the pitch f, from which sound is generated, is as follows: VALUE f0 4096 When the sound source is a noise generator, there is no acoustic variation even though the pitch specification is changed. The pitch of the noise can be specified for each core by using the sound library. f = Notes Pitch specification affects the progressing speed of sound generation. The lower the pitch is specified, the slower the sound generation proceeds. © SCEI -44- SCE CONFIDENTIAL SPU2 Overview Version 6.0 ADSR1 / ADSR2 : Envelope ADSR1 15 12 X 8 4 Name SL DR AR X Pos. 3:0 7:4 14:8 15 SL DR AR Format int 0:4:0 int 0:4:0 int 0:7:0 int 0:1:0 0 Contents Sustain Level Decay Rate Attack Rate Mode specification for Attack Rate 0 Linear increment mode (+lin mode) 1 Pseudo exponential increment mode (+exp mode) ADSR2 15 12 8 4 Z SR Y Name RR Z Pos. 4:0 5 Format int 0:5:0 int 0:1:0 SR Y 12:6 15:13 int 0:7:0 int 0:3:0 0 RR Contents Release Rate Mode specification for Release Rate 0 Linear decrement mode (-lin mode) 1 Exponential decrement mode (-exp mode) Sustain Rate Mode specification for Sustain Rate 000 Linear increment mode (+lin mode) 010 Linear decrement mode (-lin mode) 100 Pseudo exponential increment mode (+exp mode) 110 Exponential decrement mode (-exp mode) Description These registers specify each parameter for the envelope. For the relationship between the Rate/Level field values and actual envelope duration, refer to "4.1. Rate Parameter Table". © SCEI -45- SCE CONFIDENTIAL SPU2 Overview Version 6.0 ENVX : Current value of envelope 15 12 8 4 0 VALUE Name VALUE Pos. 15:0 Format int 1:0:15 Contents Current value of envelope Description This register indicates the current value of the envelope. When SR and RR for the envelope specify linear decrement, a negative value is set only for 1 Ts. When sound is generated from loop-less waveform data, ENVX is set to 0 regardless of the envelope status, at the moment the ENDX register bit corresponding to the voice is set to 1. © SCEI -46- SCE CONFIDENTIAL SPU2 Overview Version 6.0 VOLXL / VOLXR : Current value of volume 15 12 8 4 0 VALUE Name VALUE Pos. 15:0 Format int 1:0:15 Contents Current value of voice volume Description These registers indicate the current volume of each voice. When VOL is in a mode other than constant specification mode, the value varies per Ts according to the volume variation. © SCEI -47- SCE CONFIDENTIAL SPU2 Overview Version 6.0 PMON0 / PMON1 : Pitch modulation specification PMON0 15 V 15 12 V 14 Name V1 V 13 8 V 12 Pos. 1 (Omitted) V15 15 V 11 V 10 V 9 V 8 4 V 7 V 6 V 5 V 4 0 V 3 V 2 V 1 - V 17 V 16 Format int 0:1:0 Contents Pitch modulation specification for Voice1 0 Pitch modulation off 1 Pitch modulation by Voice0 output int 0:1:0 Pitch modulation specification for Voice15 0 Pitch modulation off 1 Pitch modulation by Voice14 output PMON1 15 12 (reserved) Name V16 Pos. 0 (Omitted) V23 7 8 4 V 23 V 22 V 21 V 20 0 V 19 V 18 Format int 0:1:0 Contents Pitch modulation specification for Voice16 0 Pitch modulation off 1 Pitch modulation by Voice15 output int 0:1:0 Pitch modulation specification for Voice23 0 Pitch modulation off 1 Pitch modulation by Voice22 output Description These registers specify whether to apply the pitch modulation to each voice by using the crest value of the voice of a number lower. Voice0 is disabled. © SCEI -48- SCE CONFIDENTIAL SPU2 Overview Version 6.0 NON0 / NON1 : Voice allocation to noise generator NON0 15 V 15 12 V 14 Name V0 V 13 V 12 Pos. 0 (Omitted) V15 15 8 V 11 V 10 V 9 V 8 4 V 7 V 6 V 5 V 4 0 V 3 V 2 Format int 0:1:0 Contents Sound source specification for Voice0 0 Waveform data 1 Noise generator int 0:1:0 Sound source specification for Voice15 0 Waveform data 1 Noise generator V 1 V 0 V 17 V 16 NON1 15 12 8 (reserved) Name V16 Pos. 0 (Omitted) V23 7 4 V 23 V 22 V 21 V 20 0 V 19 V 18 Format int 0:1:0 Contents Sound source specification for Voice16 0 Waveform data 1 Noise generator int 0:1:0 Sound source specification for Voice23 0 Waveform data 1 Noise generator Description These registers specify whether to allocate each voice to the noise generator as a sound source. © SCEI -49- SCE CONFIDENTIAL SPU2 Overview Version 6.0 VMIX* : Mixing specification of voice output VMIXL0, VMIXEL0, VMIXR0, VMIXER0 15 V 15 12 V 14 Name V0 V 13 V 12 Pos. 0 (Omitted) V15 15 8 V 11 V 10 V 9 V 8 4 V 7 V 6 V 5 V 4 0 V 3 V 2 Format int 0:1:0 Contents Output switch for Voice0 0 No output to applicable channel. 1 Output to applicable channel. int 0:1:0 Output switch for Voice15 0 No output to applicable channel. 1 Output to applicable channel. V 1 V 0 V 17 V 16 VMIXL1, VMIXEL1, VMIXR1, VMIXER1 15 12 8 (reserved) Name V16 Pos. 0 (Omitted) V23 7 4 V 23 V 22 V 21 V 20 0 V 19 V 18 Format int 0:1:0 Contents Output switch for Voice16 0 No output to applicable channel. 1 Output to applicable channel. int 0:1:0 Output switch for Voice23 0 No output to applicable channel. 1 Output to applicable channel. Description These registers specify whether to output the output from each voice to each channel of Dry L/Wet L/Dry R/Wet R. Each register corresponds to each channel as shown below. Register VMIXL0 VMIXL1 VMIXEL0 VMIXEL1 VMIXR0 VMIXR1 VMIXER0 VMIXER1 Channel Dry L (Direct output (L)) Dry L (Direct output (L)) Wet L (Effect output (L)) Wet L (Effect output (L)) Dry R (Direct output (R)) Dry R (Direct output (R)) Wet R (Effect output (R)) Wet R (Effect output (R)) © SCEI -50- SCE CONFIDENTIAL SPU2 Overview Version 6.0 MMIX : Output specification after voice mixing 15 12 (reserved) Name SINER SINEL SINR SINL MINER MINEL MINR MINL MSNDER MSNDEL MSNDR MSNDL Pos. 0 1 2 3 4 5 6 7 8 9 10 11 M S N D L M S N D R Format int 0:1:0 int 0:1:0 int 0:1:0 int 0:1:0 int 0:1:0 int 0:1:0 int 0:1:0 int 0:1:0 int 0:1:0 int 0:1:0 int 0:1:0 int 0:1:0 M S N D E L 8 M S N D E R M I N L M I N R M I N E L 4 M I N E R S I N L S I N R S I N E L 0 S I N E R Contents External input (R) ! Effect output External input (L) ! Effect output External input (R) ! Direct output External input (L) ! Direct output Sound data input (R) ! Effect output Sound data input (L) ! Effect output Sound data input (R) ! Direct output Sound data input (L) ! Direct output Voice output Wet R ! Effect output Voice output Wet L ! Effect output Voice output Dry R ! Direct output Voice output Dry L ! Direct output Description This is a switching specification register, which divides the voice, sound data input and external input into direct output and digital effect. When each bit is 0, the output is off. When 1, the output is on. For SINL/R and SINEL/ER, specify 0 in CORE0 at all times. © SCEI -51- SCE CONFIDENTIAL SPU2 Overview Version 6.0 IRQAH / IRQAL : Interrupt address specification IRQAH 15 12 8 4 ADDRH (reserved) Name ADDRH Pos. 5:0 Format int 0:6:0 0 Contents Local memory address where an interrupt is generated (upper 6 bits) IRQAL 15 12 8 4 0 ADDRL Name ADDRL Pos. 15:0 Format int 0:16:0 Contents Local memory address where an interrupt is generated (lower 16 bits) Description When each core accesses a specific address in the local memory, an interrupt can be generated for the host. The above registers specify the address. For details, refer to "2.7. Interrupt Processing". © SCEI -52- SCE CONFIDENTIAL SPU2 Overview Version 6.0 KON0 / KON1 : Key-on specification KON0 15 V 15 12 V 14 V 13 V 12 Name Pos. V0 0 (Omitted) V15 15 8 V 11 V 10 V 9 V 8 4 V 7 V 6 V 5 V 4 Format int 0:1:0 Contents Key-on switch of Voice0 int 0:1:0 Key-on switch of Voice15 0 V 3 V 2 V 1 V 0 V 19 V 18 V 17 V 16 KON1 15 12 8 (reserved) Name Pos. V16 0 (Omitted) V23 7 4 V 23 V 22 V 21 V 20 Format int 0:1:0 Contents Key-on switch of Voice16 int 0:1:0 Key-on switch of Voice23 0 Description These registers specify key-on (start of sound generation) of each voice. When writing these registers, the sound generation of the voice, which corresponds to the bit set to 1 among the written values, is started. Notes The value read from this register does not reflect the voice that has actually been generated. Do not write to any bits of the same register (KON0 or KON1) twice within 2 Ts. The period of time between commands may not be sufficient for the commands to execute properly. Do not specify key-on and key-off of the same voice within 2 Ts. If specified, the voice which actually starts / ends sound generation is indeterminate. Key-on can be specified for the voice in the process of sound generation, by writing 1 to the bit again without specifying key-off. © SCEI -53- SCE CONFIDENTIAL SPU2 Overview Version 6.0 KOF0 / KOF1 : Key-off specification KOF0 15 V 15 12 V 14 V 13 V 12 Name Pos. V0 0 (Omitted) V15 15 8 V 11 V 10 V 9 V 8 4 V 7 V 6 V 5 V 4 Format int 0:1:0 Contents Key-off switch of Voice0 int 0:1:0 Key-off switch of Voice15 0 V 3 V 2 V 1 V 0 V 19 V 18 V 17 V 16 KOF1 15 12 8 (reserved) Name Pos. V16 0 (Omitted) V23 7 4 V 23 V 22 V 21 V 20 Format int 0:1:0 Contents Key-off switch of Voice16 int 0:1:0 Key-off switch of Voice23 0 Description These registers specify key-off (end of sound generation) of each voice. When writing these registers, the envelope of the voice, which corresponds to the bit set to 1 among the written values, goes to the release phase. Notes Do not write to any bits of the same register (KOF0 or KOF1) twice within 2 Ts. The period of time between commands may not be sufficient for the commands to execute properly. Do not specify key-on and key-off of the same voice within 2 Ts. If specified, the voice which actually starts / ends sound generation is indeterminate. © SCEI -54- SCE CONFIDENTIAL SPU2 Overview Version 6.0 TSAH / TSAL : Transfer start address TSAH 15 12 8 4 (reserved) Name ADDRH Pos. 5:0 Format int 0:6:0 0 ADDRH Contents The starting address of the transfer area in the local memory (upper 6 bits) TSAL 15 12 8 4 0 ADDRL Name ADDRL Pos. 15:0 Format int 0:16:0 Contents The starting address of the transfer area in the local memory (lower 16 bits) Description These registers specify the starting address in the local memory area, which becomes the source/destination in DMA read transfer, DMA write transfer, Auto DMA read transfer and single write transfer. In the SSAL, which specifies the starting address of the waveform data (lower 16 bits) to the voice, the lower 3 bits must be specified to 0 when transferring the waveform data to the local memory. Therefore, in the case of the TSAL, it is also necessary to transfer data to the address whose lower 3 bits are set to 0. Notes The value is invariant regardless of the transfer execution status. If the value is changed during data transfer, operation and transferred data become indeterminate. © SCEI -55- SCE CONFIDENTIAL SPU2 Overview Version 6.0 SSAH / SSAL : Starting address of waveform data SSAH 15 12 8 4 (reserved) Name ADDRH Pos. 5:0 0 ADDRH Format int 0:6:0 Contents Starting address of waveform data (upper 6 bits) SSAL 15 12 8 4 0 ADDRL Name ADDRL Pos. 15:0 Format int 0:16:0 0 0 0 Contents Starting address of waveform data (lower 16 bits; 0 is specified to lower 3 bits) Description These registers specify the starting address of the waveform data, which becomes the sound source of each voice. © SCEI -56- SCE CONFIDENTIAL SPU2 Overview Version 6.0 LSAXH / LSAXL : Address of loop point LSAXH 15 12 8 4 ADDRH (reserved) Name ADDRH Pos. 5:0 Format int 0:6:0 0 Contents Loop point address (upper 6 bits) LSAXL 15 12 8 4 0 ADDRL Name ADDRL Pos. 15:0 Format int 0:16:0 0 0 0 Contents Loop point address (lower 16 bits; when writing, 0 is specified to lower 3 bits) Description These registers show the starting address of the block specified as a loop point (the block whose LOOP/START bit of the header is set to 1) in the waveform data. These are set when the loop point block is passed through with the advance of sound generation. The loop point can be set or changed by writing the address of an appropriate block header to this register during sound generation (4 Ts after the key-on). (Rewriting within 4 Ts is disregarded.) In this case, the loop point specification of the waveform data is disregarded temporarily until the next time the voice is keyed on. © SCEI -57- SCE CONFIDENTIAL SPU2 Overview Version 6.0 NAXH / NAXL : Address of waveform data to be read next NAXH 15 12 8 4 ADDRH (reserved) Name ADDRH Pos. 5:0 Format int 0:6:0 0 Contents Address of the waveform data to be read next (upper 6 bits) NAXL 15 12 8 4 0 ADDRL Name ADDRL Pos. 15:0 Format int 0:16:0 Contents Address of the waveform data to be read next (lower 16 bits) Description These registers show the address of the waveform data to be read next. They are updated automatically with the advance of sound generation. © SCEI -58- SCE CONFIDENTIAL SPU2 Overview Version 6.0 ESAH / ESAL : Starting address in the work area for effect processing ESAH 15 12 8 4 ADDRH (reserved) Name ADDRH Pos. 5:0 Format int 0:6:0 0 Contents Starting address in the work area for effect (upper 6 bits) ESAL 15 12 8 4 0 ADDRL Name ADDRL Pos. 15:0 Format int 0:16:0 Contents Starting address in the work area for effect (lower 16 bits) Description These registers specify the starting address in the work area to be used for digital effect processing. © SCEI -59- SCE CONFIDENTIAL SPU2 Overview Version 6.0 EEAH : End address in the work area for effect processing EEAH 15 12 8 4 (reserved) Name ADDRH Pos. 5:0 Format int 0:6:0 0 ADDRH Contents End address in the work area for effect processing (upper 6 bits) Description This register specifies the end address in the work area to be used for digital effect processing. Unlike other addressing registers, this register, which specifies the upper 6 bits only, is not paired with a register, which specifies the lower 16 bits. Because of this, the end address in the work area can only be specified on the 64K short-word (128 KB) boundary. © SCEI -60- SCE CONFIDENTIAL SPU2 Overview Version 6.0 ENDX0 / ENDX1 : Endpoint passing flag ENDX0 15 V 15 12 V 14 Name V0 V 13 V 12 Pos. 0 (Omitted) V15 15 8 V 11 V 10 V 9 V 8 4 V 7 V 6 V 5 V 4 0 V 3 Format int 0:1:0 Contents Endpoint passing flag for Voice0 0 Has not been passed. 1 Has been passed. int 0:1:0 Endpoint passing flag for Voice15 0 Has not been passed. 1 Has been passed. V 2 V 1 V 0 V 18 V 17 V 16 ENDX1 15 12 8 (reserved) Name V16 Pos. 0 (Omitted) V23 7 4 V 23 V 22 V 21 V 20 0 V 19 Format int 0:1:0 Contents Endpoint passing flag for Voice16 0 Has not been passed. 1 Has been passsed. int 0:1:0 Endpoint passing flag for Voice23 0 Has not been passed. 1 Has been passed. Description These registers show whether or not the endpoint block has been reached with the advance of sound generation of each voice. The bit corresponding to the voice is set to 0 by specifying key-on. All the bits are cleared to 0 by writing an arbitrary value (including a value other than 0) to these registers. © SCEI -61- SCE CONFIDENTIAL SPU2 Overview Version 6.0 MVOLL / MVOLR : Master volume Constant specification mode (direct mode) 15 12 8 0 4 0 VALUE Name VALUE Pos. 14:0 Format int 1:0:14 Contents Volume value The phase reverses for a negative value. Linear increment mode (+lin mode) 15 1 12 0 0 Name VALUE X X Pos. 6:0 12 8 0 (reserved) Format int 0:7:0 int 0:1:0 4 0 VALUE Contents Addition constant per Ts Polarity specification 0 Normal phase (specifiable when the current value is positive.) Linear increment to +1.0 1 Reverse phase (specifiable when the current value is negative.) Linear decrement to –1.0 Linear decrement mode (-lin mode) 15 1 12 0 Name VALUE X 1 X Pos. 6:0 12 8 (reserved) Format int 0:7:0 int 0:1:0 4 0 0 VALUE Contents Subtraction constant per Ts Polarity specification 0 Normal phase (specifiable when the current value is positive.) Linear decrement to 0 1 Reverse phase (specifiable when the current value is negative.) Linear increment to 0 © SCEI -62- SCE CONFIDENTIAL SPU2 Overview Version 6.0 Pseudo inverse-exponential increment mode (+exp mode) 15 1 12 1 0 Name VALUE X X Pos. 6:0 12 8 0 (reserved) Format int 0:7:0 int 0:1:0 4 0 VALUE Contents Addition constant per Ts Polarity specification 0 Normal phase (specifiable when the current value is positive.) Increment to +1.0 in a line. 1 Reverse phase (specifiable when the current value is negative.) Decrement to –1.0 in a line. Exponential decrement mode (-exp mode) 15 1 12 1 Name VALUE 1 8 (reserved) Pos. 6:0 Format int 0:7:0 4 0 0 VALUE Contents Multiplication constant Description These registers specify the master volume for each core. The values of the upper three bits specify the pattern of the volume variation with time. When specifying a mode excluding the constant mode, the master volume value varies with time, because the value corresponding to the value of the VALUE field is added to, subtracted from or multiplied by the master volume value per Ts. For the relationship between the VALUE field value and actual volume duration, refer to "4.1. Rate Parameter Table". © SCEI -63- SCE CONFIDENTIAL SPU2 Overview Version 6.0 EVOLL / EVOLR : Return volume of effect 15 12 8 4 0 VALUE Name VALUE Pos. 15:0 Format int 1:0:15 Contents Return volume of effect The phase reverses for a negative value. Description These registers specify the volume when mixing the output from digital effect processing with the Dry L/Dry R channel. © SCEI -64- SCE CONFIDENTIAL SPU2 Overview Version 6.0 AVOLL / AVOLR : Volume for external input 15 12 8 4 0 VALUE Name VALUE Pos. 15:0 Format int 1:0:15 Contents Volume for external input The phase reverses for a negative value. Description These registers specify the volume for the external input. They are disabled in CORE0. © SCEI -65- SCE CONFIDENTIAL SPU2 Overview Version 6.0 BVOLL / BVOLR : Volume for sound data input 15 12 8 4 VALUE Name VALUE Pos. 15:0 Format int 1:0:15 Contents Volume for sound data input The phase reverses for a negative value. Description These registers specify the volume for sound data input. © SCEI -66- 0 SCE CONFIDENTIAL SPU2 Overview Version 6.0 MVOLXL / MVOLXR : Current value of master volume 15 12 8 4 0 VALUE Name VALUE Pos. 15:0 Format int 1:0:15 Contents Current value of master volume Description These registers indicate the current value of the master volume. When MVOL is in a mode other than constant specification mode, the value varies per Ts according to the volume variation. © SCEI -67- SCE CONFIDENTIAL SPU2 Overview Version 6.0 (This page is left blank intentionally) © SCEI -68- SCE CONFIDENTIAL SPU2 Overview Version 6.0 4. Appendix © SCEI -69- SCE CONFIDENTIAL SPU2 Overview Version 6.0 4.1. Rate Parameter Table The following tables show the volume variation rate for values of the envelope rate/level parameters (AR, DR, SL, SR and RR) and the parameters to specify volume variation with time. The variation rate is shown with the time required for the volume to vary from 0 to 1 (1 to 0, or 1 to 0.1). For SL, the variation rate is shown with level. © SCEI -70- SCE CONFIDENTIAL SPU2 Overview Version 6.0 4.1.1. +Lin Mode The variation is shown in linear increment for normal phase and in linear decrement for reverse phase. Set Value 000 0000 000 0001 000 0010 000 0011 000 0100 000 0101 000 0110 000 0111 000 1000 000 1001 000 1010 000 1011 000 1100 000 1101 000 1110 000 1111 001 0000 001 0001 001 0010 001 0011 001 0100 001 0101 001 0110 001 0111 001 1000 001 1001 001 1010 001 1011 001 1100 001 1101 001 1110 001 1111 010 0000 010 0001 010 0010 010 0011 010 0100 010 0101 010 0110 0 -> 1 Time 0.05 (msec) 0.06 0.07 0.09 0.10 0.12 0.15 0.18 0.21 0.24 0.29 0.36 0.41 0.48 0.58 0.73 0.83 0.97 1.2 1.5 1.7 1.9 2.3 2.9 3.3 3.9 4.6 5.8 6.6 7.7 9.3 12 13 15 19 23 27 31 37 Set Value 010 0111 010 1000 010 1001 010 1010 010 1011 010 1100 010 1101 010 1110 010 1111 011 0000 011 0001 011 0010 011 0011 011 0100 011 0101 011 0110 011 0111 011 1000 011 1001 011 1010 011 1011 011 1100 011 1101 011 1110 011 1111 100 0000 100 0001 100 0010 100 0011 100 0100 100 0101 100 0110 100 0111 100 1000 100 1001 100 1010 100 1011 100 1100 100 1101 0 -> 1 Time 46 (msec) 53 62 74 93 0.11 (sec) 0.12 0.15 0.19 0.21 0.25 0.30 0.37 0.42 0.50 0.59 0.74 0.85 0.99 1.2 1.5 1.7 2.0 2.4 3.0 3.4 4.0 4.8 5.9 6.8 7.9 9.5 12 14 16 19 24 27 32 Set Value 100 1110 100 1111 101 0000 101 0001 101 0010 101 0011 101 0100 101 0101 101 0110 101 0111 101 1000 101 1001 101 1010 101 1011 101 1100 101 1101 101 1110 101 1111 110 0000 110 0001 110 0010 110 0011 110 0100 110 0101 110 0110 110 0111 110 1000 110 1001 110 1010 110 1011 110 1100 : 111 1110 0 -> 1 Time 38 (sec) 48 54 63 76 95 109 127 152 190 218 254 304 380 436 508 608 760 872 1016 1216 1520 1744 2032 2432 3040 3488 4064 4864 6080 (reserved) 111 1111 infinity © SCEI -71- SCE CONFIDENTIAL SPU2 Overview Version 6.0 4.1.2. –Lin Mode The variation is shown in linear decrement for normal phase and in linear increment for reverse phase. -Lin Value 000 0000 000 0001 000 0010 000 0011 000 0100 000 0101 000 0110 000 0111 000 1000 000 1001 000 1010 000 1011 000 1100 000 1101 000 1110 000 1111 001 0000 001 0001 001 0010 001 0011 001 0100 001 0101 001 0110 001 0111 001 1000 001 1001 001 1010 001 1011 001 1100 001 1101 001 1110 001 1111 010 0000 010 0001 010 0010 010 0011 010 0100 010 0101 010 0110 1 –> 0 Time 0.04 (msec) 0.05 0.06 0.07 0.09 0.10 0.12 0.15 0.18 0.21 0.24 0.29 0.36 0.41 0.48 0.58 0.73 0.83 0.97 1.2 1.5 1.7 1.9 2.3 2.9 3.3 3.9 4.6 5.8 6.6 7.7 9.3 12 13 15 19 23 27 31 -Lin Value 010 0111 010 1000 010 1001 010 1010 010 1011 010 1100 010 1101 010 1110 010 1111 011 0000 011 0001 011 0010 011 0011 011 0100 011 0101 011 0110 011 0111 011 1000 011 1001 011 1010 011 1011 011 1100 011 1101 011 1110 011 1111 100 0000 100 0001 100 0010 100 0011 100 0100 100 0101 100 0110 100 0111 100 1000 100 1001 100 1010 100 1011 100 1100 100 1101 1 –> 0 Time 37 (msec) 46 53 62 74 93 0.11 (sec) 0.12 0.15 0.19 0.21 0.25 0.30 0.37 0.42 0.50 0.59 0.74 0.85 0.99 1.2 1.5 1.7 2.0 2.4 3.0 3.4 4.0 4.8 5.9 6.8 7.9 9.5 12 14 16 19 24 27 © SCEI -72- -Lin Value 100 1110 100 1111 101 0000 101 0001 101 0010 101 0011 101 0100 101 0101 101 0110 101 0111 101 1000 101 1001 101 1010 101 1011 101 1100 101 1101 101 1110 101 1111 110 0000 110 0001 110 0010 110 0011 110 0100 110 0101 110 0110 110 0111 110 1000 110 1001 110 1010 110 1011 110 1100 : 111 1110 1 –> 0 Time 32 (sec) 38 48 54 63 76 95 109 127 152 190 218 254 304 380 436 508 608 760 872 1016 1216 1520 1744 2032 2432 3040 3488 4064 4864 (reserved) 111 1111 infinity SCE CONFIDENTIAL SPU2 Overview Version 6.0 4.1.3. +Exp Mode (Normal phase) Pseudo Inverse-Exponential Increment Mode +Exp Set Value 000 0000 000 0001 000 0010 000 0011 000 0100 000 0101 000 0110 000 0111 000 1000 000 1001 000 1010 000 1011 000 1100 000 1101 000 1110 000 1111 001 0000 001 0001 001 0010 001 0011 001 0100 001 0101 001 0110 001 0111 001 1000 001 1001 001 1010 001 1011 001 1100 001 1101 001 1110 001 1111 010 0000 010 0001 010 0010 010 0011 1 –> 0 Time 0.09 (msec) 0.11 0.13 0.16 0.18 0.21 0.25 0.32 0.36 0.42 0.51 0.64 0.73 0.85 1.0 1.3 1.5 1.7 2.0 2.5 2.9 3.4 4.1 5.1 5.8 6.8 8.1 10 12 14 16 20 23 27 33 41 +Exp Set Value 010 0100 010 0101 010 0110 010 0111 010 1000 010 1001 010 1010 010 1011 010 1100 010 1101 010 1110 010 1111 011 0000 011 0001 011 0010 011 0011 011 0100 011 0101 011 0110 011 0111 011 1000 011 1001 011 1010 011 1011 011 1100 011 1101 011 1110 011 1111 100 0000 100 0001 100 0010 100 0011 100 0100 100 0101 100 0110 100 0111 1 –> 0 Time 46 (msec) 54 65 81 93 0.11 (sec) 0.13 0.16 0.19 0.22 0.26 0.33 0.37 0.43 0.52 0.65 0.74 0.87 1.0 1.3 1.5 1.7 2.1 2.6 3.0 3.5 4.2 5.2 5.9 6.9 8.3 10 12 14 17 21 +Exp Set Value 100 1000 100 1001 100 1010 100 1011 100 1100 100 1101 100 1110 100 1111 101 0000 101 0001 101 0010 101 0011 101 0100 101 0101 101 0110 101 0111 101 1000 101 1001 101 1010 101 1011 101 1100 101 1101 101 1110 101 1111 110 0000 110 0001 110 0010 110 0011 110 1100 : 111 1110 111 1111 1 –> 0 Time 24 (sec) 28 33 42 48 55 67 83 95 111 133 166 190 222 266 333 380 444 532 666 760 888 1064 1332 1520 1776 2128 2664 (reserved) infinity © SCEI -73- SCE CONFIDENTIAL SPU2 Overview Version 6.0 4.1.4. +Exp Mode (Reverse phase) Pseudo Inverse-Exponential Mode with Negative Volume Value +Exp Set Value 000 0000 000 0001 000 0010 000 0011 000 0100 000 0101 000 0110 000 0111 000 1000 000 1001 000 1010 000 1011 000 1100 000 1101 000 1110 000 1111 001 0000 001 0001 001 0010 001 0011 001 0100 001 0101 001 0110 001 0111 001 1000 001 1001 001 1010 001 1011 001 1100 001 1101 001 1110 001 1111 010 0000 010 0001 010 0010 010 0011 1 -> 0 Time 0.08 (msec) 0.09 0.11 0.13 0.16 0.18 0.21 0.25 0.32 0.36 0.42 0.51 0.64 0.73 0.85 1.0 1.3 1.5 1.7 2.0 2.5 2.9 3.4 4.1 5.1 5.8 6.8 8.1 10 12 14 16 20 23 27 33 +Exp Set Value 010 0100 010 0101 010 0110 010 0111 010 1000 010 1001 010 1010 010 1011 010 1100 010 1101 010 1110 010 1111 011 0000 011 0001 011 0010 011 0011 011 0100 011 0101 011 0110 011 0111 011 1000 011 1001 011 1010 011 1011 011 1100 011 1101 011 1110 011 1111 100 0000 100 0001 100 0010 100 0011 100 0100 100 0101 100 0110 100 0111 1 -> 0 Time 41 (msec) 46 54 64 81 93 0.11 (sec) 0.13 0.16 0.19 0.22 0.26 0.33 0.37 0.43 0.52 0.65 0.74 0.87 1.0 1.3 1.5 1.7 2.1 2.6 3.0 3.5 4.2 5.2 5.9 6.9 8.3 10 12 14 17 © SCEI -74- +Exp Set Value 100 1000 100 1001 100 1010 100 1011 100 1100 100 1101 100 1110 100 1111 101 0000 101 0001 101 0010 101 0011 101 0100 101 0101 101 0110 101 0111 101 1000 101 1001 101 1010 101 1011 101 1100 101 1101 101 1110 101 1111 110 0000 110 0001 110 0010 110 0011 110 1100 : 111 1110 1 -> 0 Time 111 1111 infinity 21 (sec) 24 28 33 42 48 55 67 83 95 111 133 166 190 222 266 333 380 444 532 666 760 888 1064 1332 1520 1776 2128 (reserved) SCE CONFIDENTIAL SPU2 Overview Version 6.0 4.1.5. –Exp Mode Exponential Decrement Mode -Exp Set Value 000 0000 000 0001 000 0010 000 0011 000 0100 000 0101 000 0110 000 0111 000 1000 000 1001 000 1010 000 1011 000 1100 000 1101 000 1110 000 1111 001 0000 001 0001 001 0010 001 0011 001 0100 001 0101 001 0110 001 0111 001 1000 001 1001 001 1010 001 1011 001 1100 001 1101 001 1110 001 1111 010 0000 010 0001 010 0010 010 0011 010 0100 010 0101 010 0110 1 –> 0.1 Time 0.07 (msec) 0.09 0.11 0.14 0.18 0.21 0.25 0.31 0.39 0.45 0.53 0.64 0.81 0.93 1.1 1.3 1.6 1.9 2.2 2.6 3.3 3.8 4.4 5.3 6.7 7.6 8.9 11 13 15 18 21 27 31 36 43 53 61 71 -Exp Set Value 010 0111 010 1000 010 1001 010 1010 010 1011 010 1100 010 1101 010 1110 010 1111 011 0000 011 0001 011 0010 011 0011 011 0100 011 0101 011 0110 011 0111 011 1000 011 1001 011 1010 011 1011 011 1100 011 1101 011 1110 011 1111 100 0000 100 0001 100 0010 100 0011 100 0100 100 0101 100 0110 100 0111 100 1000 100 1001 100 1010 100 1011 100 1100 100 1101 1 –> 0.1 Time 86 (msec) 0.11 (sec) 0.12 0.14 0.17 0.21 0.24 0.29 0.34 0.43 0.49 0.57 0.68 0.86 0.98 1.1 1.4 1.7 2.0 2.3 2.7 3.4 3.9 4.6 5.5 6.8 7.8 9.1 11 14 16 18 22 27 31 36 44 55 63 -Exp Set Value 100 1110 100 1111 101 0000 101 0001 101 0010 101 0011 101 0100 101 0101 101 0110 101 0111 101 1000 101 1001 101 1010 101 1011 101 1100 101 1101 101 1110 101 1111 110 0000 110 0001 110 0010 110 0011 110 0100 110 0101 110 0110 110 0111 110 1000 110 1001 110 1010 110 1011 110 1100 : 111 1110 1 –> 0.1 Time 73 (sec) 88 109 125 146 175 219 250 292 350 438 500 584 700 876 1000 1168 1400 1752 2000 2336 2800 3504 4000 4672 5600 7008 8000 9344 11200 (reserved) 111 1111 infinity © SCEI -75- SCE CONFIDENTIAL SPU2 Overview Version 6.0 4.1.6. Decay Rate (DR) Exponential Decrement Mode -Exp Set Value 0000 0001 0010 0011 0100 0101 0110 0111 1000 1001 1010 1011 1100 1101 1110 1111 1 –> 0.10 Time 0.07 (msec) 0.18 0.39 0.81 1.6 3.3 6.7 13 27 53 0.11 (sec) 0.21 0.43 0.86 1.7 3.4 © SCEI -76- SCE CONFIDENTIAL SPU2 Overview Version 6.0 4.1.7. Sustain Level (SL) Set Value Envelope Level 0000 1/16 0001 2/16 0010 3/16 0011 4/16 0100 5/16 0101 6/16 0110 7/16 0111 8/16 1000 9/16 1001 10/16 1010 11/16 1011 12/16 1100 13/16 1101 14/16 1110 15/16 1111 1 Note: when the set value is 1111, the Decay section immediately before is processed only for 1 Ts. © SCEI -77- SCE CONFIDENTIAL SPU2 Overview Version 6.0 4.1.8. –Lin Mode for Release Rate (RR) Linear Decrement Mode -Lin Set Value 0 0000 0 0001 0 0010 0 0011 0 0100 0 0101 0 0110 0 0111 0 1000 0 1001 0 1010 0 1011 0 1100 0 1101 0 1110 0 1111 1 0000 1 0001 1 0010 1 0011 1 0100 1 0101 1 0110 1 0111 1 1000 1 1001 1 1010 1 1011 : 1 1110 1 1111 1 –> 0 Time 0.04 (msec) 0.09 0.18 0.36 0.73 1.5 2.9 5.8 12 23 46 93 0.19 (sec) 0.37 0.74 1.5 3.0 5.9 12 24 48 95 190 380 760 1520 3040 (reserved) infinity © SCEI -78- SCE CONFIDENTIAL SPU2 Overview Version 6.0 4.1.9. –Exp Mode for Release Rate (RR) Exponential Decrement Mode -Exp Set Value 0 0000 0 0001 0 0010 0 0011 0 0100 0 0101 0 0110 0 0111 0 1000 0 1001 0 1010 0 1011 0 1100 0 1101 0 1110 0 1111 1 0000 1 0001 1 0010 1 0011 1 0100 1 0101 1 0110 1 0111 1 1000 1 1001 1 1010 1 1011 : 1 1110 1 1111 1 –> 0.1 Time 0.07 (msec) 0.18 0.39 0.81 1.6 3.3 6.7 13 27 53 0.11 (sec) 0.21 0.43 0.86 1.7 3.4 6.8 14 27 55 109 219 438 876 1752 3504 7008 (reserved) infinity © SCEI -79- SCE CONFIDENTIAL SPU2 Overview Version 6.0 (This page is left blank intentionally) © SCEI -80-
Source Exif Data:
File Type : PDF File Type Extension : pdf MIME Type : application/pdf PDF Version : 1.4 Linearized : Yes Encryption : Standard V1.2 (40-bit) User Access : Print, Fill forms, Extract, Assemble, Print high-res Modify Date : 2002:06:10 14:26:20-07:00 Create Date : 2002:05:31 20:33:14Z Page Count : 80 Creation Date : 2002:05:31 20:33:14Z Mod Date : 2002:06:10 14:26:20-07:00 Producer : Acrobat Distiller 5.0.5 (Windows) Author : Sony Computer Entertainment Metadata Date : 2002:06:10 14:26:20-07:00 Creator : Sony Computer Entertainment Title : SPU2 Overview Page Mode : UseOutlines Page Layout : SinglePageEXIF Metadata provided by EXIF.tools