Based on kernel version 4.9. Page generated on 2016-12-21 14:37 EST.
1 compress_offload.txt 2 ===================== 3 Pierre-Louis.Bossart <firstname.lastname@example.org> 4 Vinod Koul <email@example.com> 5 6 Overview 7 8 Since its early days, the ALSA API was defined with PCM support or 9 constant bitrates payloads such as IEC61937 in mind. Arguments and 10 returned values in frames are the norm, making it a challenge to 11 extend the existing API to compressed data streams. 12 13 In recent years, audio digital signal processors (DSP) were integrated 14 in system-on-chip designs, and DSPs are also integrated in audio 15 codecs. Processing compressed data on such DSPs results in a dramatic 16 reduction of power consumption compared to host-based 17 processing. Support for such hardware has not been very good in Linux, 18 mostly because of a lack of a generic API available in the mainline 19 kernel. 20 21 Rather than requiring a compatibility break with an API change of the 22 ALSA PCM interface, a new 'Compressed Data' API is introduced to 23 provide a control and data-streaming interface for audio DSPs. 24 25 The design of this API was inspired by the 2-year experience with the 26 Intel Moorestown SOC, with many corrections required to upstream the 27 API in the mainline kernel instead of the staging tree and make it 28 usable by others. 29 30 Requirements 31 32 The main requirements are: 33 34 - separation between byte counts and time. Compressed formats may have 35 a header per file, per frame, or no header at all. The payload size 36 may vary from frame-to-frame. As a result, it is not possible to 37 estimate reliably the duration of audio buffers when handling 38 compressed data. Dedicated mechanisms are required to allow for 39 reliable audio-video synchronization, which requires precise 40 reporting of the number of samples rendered at any given time. 41 42 - Handling of multiple formats. PCM data only requires a specification 43 of the sampling rate, number of channels and bits per sample. In 44 contrast, compressed data comes in a variety of formats. Audio DSPs 45 may also provide support for a limited number of audio encoders and 46 decoders embedded in firmware, or may support more choices through 47 dynamic download of libraries. 48 49 - Focus on main formats. This API provides support for the most 50 popular formats used for audio and video capture and playback. It is 51 likely that as audio compression technology advances, new formats 52 will be added. 53 54 - Handling of multiple configurations. Even for a given format like 55 AAC, some implementations may support AAC multichannel but HE-AAC 56 stereo. Likewise WMA10 level M3 may require too much memory and cpu 57 cycles. The new API needs to provide a generic way of listing these 58 formats. 59 60 - Rendering/Grabbing only. This API does not provide any means of 61 hardware acceleration, where PCM samples are provided back to 62 user-space for additional processing. This API focuses instead on 63 streaming compressed data to a DSP, with the assumption that the 64 decoded samples are routed to a physical output or logical back-end. 65 66 - Complexity hiding. Existing user-space multimedia frameworks all 67 have existing enums/structures for each compressed format. This new 68 API assumes the existence of a platform-specific compatibility layer 69 to expose, translate and make use of the capabilities of the audio 70 DSP, eg. Android HAL or PulseAudio sinks. By construction, regular 71 applications are not supposed to make use of this API. 72 73 74 Design 75 76 The new API shares a number of concepts with the PCM API for flow 77 control. Start, pause, resume, drain and stop commands have the same 78 semantics no matter what the content is. 79 80 The concept of memory ring buffer divided in a set of fragments is 81 borrowed from the ALSA PCM API. However, only sizes in bytes can be 82 specified. 83 84 Seeks/trick modes are assumed to be handled by the host. 85 86 The notion of rewinds/forwards is not supported. Data committed to the 87 ring buffer cannot be invalidated, except when dropping all buffers. 88 89 The Compressed Data API does not make any assumptions on how the data 90 is transmitted to the audio DSP. DMA transfers from main memory to an 91 embedded audio cluster or to a SPI interface for external DSPs are 92 possible. As in the ALSA PCM case, a core set of routines is exposed; 93 each driver implementer will have to write support for a set of 94 mandatory routines and possibly make use of optional ones. 95 96 The main additions are 97 98 - get_caps 99 This routine returns the list of audio formats supported. Querying the 100 codecs on a capture stream will return encoders, decoders will be 101 listed for playback streams. 102 103 - get_codec_caps For each codec, this routine returns a list of 104 capabilities. The intent is to make sure all the capabilities 105 correspond to valid settings, and to minimize the risks of 106 configuration failures. For example, for a complex codec such as AAC, 107 the number of channels supported may depend on a specific profile. If 108 the capabilities were exposed with a single descriptor, it may happen 109 that a specific combination of profiles/channels/formats may not be 110 supported. Likewise, embedded DSPs have limited memory and cpu cycles, 111 it is likely that some implementations make the list of capabilities 112 dynamic and dependent on existing workloads. In addition to codec 113 settings, this routine returns the minimum buffer size handled by the 114 implementation. This information can be a function of the DMA buffer 115 sizes, the number of bytes required to synchronize, etc, and can be 116 used by userspace to define how much needs to be written in the ring 117 buffer before playback can start. 118 119 - set_params 120 This routine sets the configuration chosen for a specific codec. The 121 most important field in the parameters is the codec type; in most 122 cases decoders will ignore other fields, while encoders will strictly 123 comply to the settings 124 125 - get_params 126 This routines returns the actual settings used by the DSP. Changes to 127 the settings should remain the exception. 128 129 - get_timestamp 130 The timestamp becomes a multiple field structure. It lists the number 131 of bytes transferred, the number of samples processed and the number 132 of samples rendered/grabbed. All these values can be used to determine 133 the average bitrate, figure out if the ring buffer needs to be 134 refilled or the delay due to decoding/encoding/io on the DSP. 135 136 Note that the list of codecs/profiles/modes was derived from the 137 OpenMAX AL specification instead of reinventing the wheel. 138 Modifications include: 139 - Addition of FLAC and IEC formats 140 - Merge of encoder/decoder capabilities 141 - Profiles/modes listed as bitmasks to make descriptors more compact 142 - Addition of set_params for decoders (missing in OpenMAX AL) 143 - Addition of AMR/AMR-WB encoding modes (missing in OpenMAX AL) 144 - Addition of format information for WMA 145 - Addition of encoding options when required (derived from OpenMAX IL) 146 - Addition of rateControlSupported (missing in OpenMAX AL) 147 148 Gapless Playback 149 ================ 150 When playing thru an album, the decoders have the ability to skip the encoder 151 delay and padding and directly move from one track content to another. The end 152 user can perceive this as gapless playback as we don't have silence while 153 switching from one track to another 154 155 Also, there might be low-intensity noises due to encoding. Perfect gapless is 156 difficult to reach with all types of compressed data, but works fine with most 157 music content. The decoder needs to know the encoder delay and encoder padding. 158 So we need to pass this to DSP. This metadata is extracted from ID3/MP4 headers 159 and are not present by default in the bitstream, hence the need for a new 160 interface to pass this information to the DSP. Also DSP and userspace needs to 161 switch from one track to another and start using data for second track. 162 163 The main additions are: 164 165 - set_metadata 166 This routine sets the encoder delay and encoder padding. This can be used by 167 decoder to strip the silence. This needs to be set before the data in the track 168 is written. 169 170 - set_next_track 171 This routine tells DSP that metadata and write operation sent after this would 172 correspond to subsequent track 173 174 - partial drain 175 This is called when end of file is reached. The userspace can inform DSP that 176 EOF is reached and now DSP can start skipping padding delay. Also next write 177 data would belong to next track 178 179 Sequence flow for gapless would be: 180 - Open 181 - Get caps / codec caps 182 - Set params 183 - Set metadata of the first track 184 - Fill data of the first track 185 - Trigger start 186 - User-space finished sending all, 187 - Indicate next track data by sending set_next_track 188 - Set metadata of the next track 189 - then call partial_drain to flush most of buffer in DSP 190 - Fill data of the next track 191 - DSP switches to second track 192 (note: order for partial_drain and write for next track can be reversed as well) 193 194 Not supported: 195 196 - Support for VoIP/circuit-switched calls is not the target of this 197 API. Support for dynamic bit-rate changes would require a tight 198 coupling between the DSP and the host stack, limiting power savings. 199 200 - Packet-loss concealment is not supported. This would require an 201 additional interface to let the decoder synthesize data when frames 202 are lost during transmission. This may be added in the future. 203 204 - Volume control/routing is not handled by this API. Devices exposing a 205 compressed data interface will be considered as regular ALSA devices; 206 volume changes and routing information will be provided with regular 207 ALSA kcontrols. 208 209 - Embedded audio effects. Such effects should be enabled in the same 210 manner, no matter if the input was PCM or compressed. 211 212 - multichannel IEC encoding. Unclear if this is required. 213 214 - Encoding/decoding acceleration is not supported as mentioned 215 above. It is possible to route the output of a decoder to a capture 216 stream, or even implement transcoding capabilities. This routing 217 would be enabled with ALSA kcontrols. 218 219 - Audio policy/resource management. This API does not provide any 220 hooks to query the utilization of the audio DSP, nor any preemption 221 mechanisms. 222 223 - No notion of underrun/overrun. Since the bytes written are compressed 224 in nature and data written/read doesn't translate directly to 225 rendered output in time, this does not deal with underrun/overrun and 226 maybe dealt in user-library 227 228 Credits: 229 - Mark Brown and Liam Girdwood for discussions on the need for this API 230 - Harsha Priya for her work on intel_sst compressed API 231 - Rakesh Ughreja for valuable feedback 232 - Sing Nallasellan, Sikkandar Madar and Prasanna Samaga for 233 demonstrating and quantifying the benefits of audio offload on a 234 real platform.