Qualcomm 音频学习一

前言

最近在学习高通的音频驱动，在学习了高通音频 bring up 和 Audio overview 文档后，并在网上寻找到一篇比较重要的 blog进行学习后，将这部分学习笔记记录于此。

四个重要部分

高通音频框架大体分为以下四个部分：
音频前端（FE）
音频后端（BE）
DSP
音频设备（Audio Device）

(以下，音频前端使用 FE 代替；音频后端使用 BE 代替，音频设备使用 Device 代替)
其中，一个音频前端对应着一个 PCM 设备，一个音频后端对应着一个 DAI 口；DSP 处于音频前端和音频后端之间，起着连接 FE 和 BE 的作用；所有的 Device 都是挂在 DAI 上的。

1. PCM

PCM ：(Pulse-code modulation)脉冲编码调制，是将模拟信号转化为数字信号的一种方法（个人理解：PCM 数据流即是通过PCM方法转换后的二进制数据流）。大致包含以下这几种：
deep-buffer (//音乐、视频等对时延要求不高的播放音)
low-latency (//按键音、触摸音、游戏背景音等低延时的放音)
mutil-channel
compress-offload (//mp3、flac、aac等格式的音源播放)
audio-record (//普通录音)
usb-audio
a2dp-audio
voice-call (//语音通话)

Auido PCM 解释（ https://blog.csdn.net/shuyong1999/article/details/7165419 ）

2. DAI

DAI : Direct audio input （音频总线接口）
SLIM_BUS
Aux_PCM
Primary_MI2S
Secondary_MI2S
Tertiary_MI2S
Quatermary_MI2S

3. Audio Device

headset
handset
speaker
earpiece
mic
bt
modem
FM

具体其定义如下：

/* Playback devices */
SND_DEVICE_MIN,
SND_DEVICE_OUT_BEGIN = SND_DEVICE_MIN,
SND_DEVICE_OUT_HANDSET = SND_DEVICE_OUT_BEGIN,
SND_DEVICE_OUT_SPEAKER,
SND_DEVICE_OUT_SPEAKER_EXTERNAL_1,
SND_DEVICE_OUT_SPEAKER_EXTERNAL_2,
SND_DEVICE_OUT_SPEAKER_REVERSE,
SND_DEVICE_OUT_SPEAKER_WSA,
SND_DEVICE_OUT_SPEAKER_VBAT,
SND_DEVICE_OUT_LINE,
SND_DEVICE_OUT_HEADPHONES,
SND_DEVICE_OUT_HEADPHONES_DSD,
SND_DEVICE_OUT_HEADPHONES_44_1,
SND_DEVICE_OUT_SPEAKER_AND_HEADPHONES,
SND_DEVICE_OUT_SPEAKER_AND_LINE,
SND_DEVICE_OUT_SPEAKER_AND_HEADPHONES_EXTERNAL_1,
SND_DEVICE_OUT_SPEAKER_AND_HEADPHONES_EXTERNAL_2,
SND_DEVICE_OUT_VOICE_HANDSET,
SND_DEVICE_OUT_VOICE_SPEAKER,
SND_DEVICE_OUT_VOICE_SPEAKER_WSA,
SND_DEVICE_OUT_VOICE_SPEAKER_VBAT,
SND_DEVICE_OUT_VOICE_SPEAKER_2,
SND_DEVICE_OUT_VOICE_SPEAKER_2_WSA,
SND_DEVICE_OUT_VOICE_SPEAKER_2_VBAT,
SND_DEVICE_OUT_VOICE_HEADPHONES,
SND_DEVICE_OUT_VOICE_LINE,
SND_DEVICE_OUT_HDMI,
SND_DEVICE_OUT_SPEAKER_AND_HDMI,
SND_DEVICE_OUT_DISPLAY_PORT,
SND_DEVICE_OUT_SPEAKER_AND_DISPLAY_PORT,
SND_DEVICE_OUT_BT_SCO,
SND_DEVICE_OUT_BT_SCO_WB,
SND_DEVICE_OUT_BT_A2DP,
SND_DEVICE_OUT_SPEAKER_AND_BT_A2DP,
SND_DEVICE_OUT_VOICE_TTY_FULL_HEADPHONES,
SND_DEVICE_OUT_VOICE_TTY_VCO_HEADPHONES,
SND_DEVICE_OUT_VOICE_TTY_HCO_HANDSET,
SND_DEVICE_OUT_VOICE_TX,
SND_DEVICE_OUT_AFE_PROXY,
SND_DEVICE_OUT_USB_HEADSET,
SND_DEVICE_OUT_USB_HEADPHONES,
SND_DEVICE_OUT_SPEAKER_AND_USB_HEADSET,
SND_DEVICE_OUT_TRANSMISSION_FM,
SND_DEVICE_OUT_ANC_HEADSET,
SND_DEVICE_OUT_ANC_FB_HEADSET,
SND_DEVICE_OUT_VOICE_ANC_HEADSET,
SND_DEVICE_OUT_VOICE_ANC_FB_HEADSET,
SND_DEVICE_OUT_SPEAKER_AND_ANC_HEADSET,
SND_DEVICE_OUT_ANC_HANDSET,
SND_DEVICE_OUT_SPEAKER_PROTECTED,
SND_DEVICE_OUT_VOICE_SPEAKER_PROTECTED,
SND_DEVICE_OUT_VOICE_SPEAKER_2_PROTECTED,
SND_DEVICE_OUT_SPEAKER_PROTECTED_VBAT,
SND_DEVICE_OUT_VOICE_SPEAKER_PROTECTED_VBAT,
SND_DEVICE_OUT_VOICE_SPEAKER_2_PROTECTED_VBAT,
SND_DEVICE_OUT_SPEAKER_PROTECTED_RAS,
SND_DEVICE_OUT_SPEAKER_PROTECTED_VBAT_RAS,
#ifdef RECORD_PLAY_CONCURRENCY
SND_DEVICE_OUT_VOIP_HANDSET,
SND_DEVICE_OUT_VOIP_SPEAKER,
SND_DEVICE_OUT_VOIP_HEADPHONES,
#endif
SND_DEVICE_OUT_END,/** Note: IN_BEGIN should be same as OUT_END because total number of devices* SND_DEVICES_MAX should not exceed MAX_RX + MAX_TX devices.*/
/* Capture devices */
SND_DEVICE_IN_BEGIN = SND_DEVICE_OUT_END,
SND_DEVICE_IN_HANDSET_MIC  = SND_DEVICE_IN_BEGIN,
SND_DEVICE_IN_HANDSET_MIC_EXTERNAL,
SND_DEVICE_IN_HANDSET_MIC_AEC,
SND_DEVICE_IN_HANDSET_MIC_NS,
SND_DEVICE_IN_HANDSET_MIC_AEC_NS,
SND_DEVICE_IN_HANDSET_DMIC,
SND_DEVICE_IN_HANDSET_DMIC_AEC,
SND_DEVICE_IN_HANDSET_DMIC_NS,
SND_DEVICE_IN_HANDSET_DMIC_AEC_NS,
SND_DEVICE_IN_SPEAKER_MIC,
SND_DEVICE_IN_SPEAKER_MIC_AEC,
SND_DEVICE_IN_SPEAKER_MIC_NS,
SND_DEVICE_IN_SPEAKER_MIC_AEC_NS,
SND_DEVICE_IN_SPEAKER_DMIC,
SND_DEVICE_IN_SPEAKER_DMIC_AEC,
SND_DEVICE_IN_SPEAKER_DMIC_NS,
SND_DEVICE_IN_SPEAKER_DMIC_AEC_NS,
SND_DEVICE_IN_HEADSET_MIC,
SND_DEVICE_IN_HEADSET_MIC_FLUENCE,
SND_DEVICE_IN_VOICE_SPEAKER_MIC,
SND_DEVICE_IN_VOICE_HEADSET_MIC,
SND_DEVICE_IN_HDMI_MIC,
SND_DEVICE_IN_BT_SCO_MIC,
SND_DEVICE_IN_BT_SCO_MIC_NREC,
SND_DEVICE_IN_BT_SCO_MIC_WB,
SND_DEVICE_IN_BT_SCO_MIC_WB_NREC,
SND_DEVICE_IN_CAMCORDER_MIC,
SND_DEVICE_IN_VOICE_DMIC,
SND_DEVICE_IN_VOICE_SPEAKER_DMIC,
SND_DEVICE_IN_VOICE_SPEAKER_QMIC,
SND_DEVICE_IN_VOICE_TTY_FULL_HEADSET_MIC,
SND_DEVICE_IN_VOICE_TTY_VCO_HANDSET_MIC,
SND_DEVICE_IN_VOICE_TTY_HCO_HEADSET_MIC,
SND_DEVICE_IN_VOICE_REC_MIC,
SND_DEVICE_IN_VOICE_REC_MIC_NS,
SND_DEVICE_IN_VOICE_REC_DMIC_STEREO,
SND_DEVICE_IN_VOICE_REC_DMIC_FLUENCE,
SND_DEVICE_IN_VOICE_RX,
SND_DEVICE_IN_USB_HEADSET_MIC,
SND_DEVICE_IN_CAPTURE_FM,
SND_DEVICE_IN_AANC_HANDSET_MIC,
SND_DEVICE_IN_QUAD_MIC,
SND_DEVICE_IN_HANDSET_STEREO_DMIC,
SND_DEVICE_IN_SPEAKER_STEREO_DMIC,
SND_DEVICE_IN_CAPTURE_VI_FEEDBACK,
SND_DEVICE_IN_CAPTURE_VI_FEEDBACK_MONO_1,
SND_DEVICE_IN_CAPTURE_VI_FEEDBACK_MONO_2,
SND_DEVICE_IN_VOICE_SPEAKER_DMIC_BROADSIDE,
SND_DEVICE_IN_SPEAKER_DMIC_BROADSIDE,
SND_DEVICE_IN_SPEAKER_DMIC_AEC_BROADSIDE,
SND_DEVICE_IN_SPEAKER_DMIC_NS_BROADSIDE,
SND_DEVICE_IN_SPEAKER_DMIC_AEC_NS_BROADSIDE,
SND_DEVICE_IN_VOICE_FLUENCE_DMIC_AANC,
SND_DEVICE_IN_HANDSET_QMIC,
SND_DEVICE_IN_SPEAKER_QMIC_AEC,
SND_DEVICE_IN_SPEAKER_QMIC_NS,
SND_DEVICE_IN_SPEAKER_QMIC_AEC_NS,
SND_DEVICE_IN_THREE_MIC,
SND_DEVICE_IN_HANDSET_TMIC,
SND_DEVICE_IN_VOICE_REC_TMIC,
SND_DEVICE_IN_UNPROCESSED_MIC,
SND_DEVICE_IN_UNPROCESSED_STEREO_MIC,
SND_DEVICE_IN_UNPROCESSED_THREE_MIC,
SND_DEVICE_IN_UNPROCESSED_QUAD_MIC,
SND_DEVICE_IN_UNPROCESSED_HEADSET_MIC,

4. 音频 DSP

音频 DSP，大致可以等同于 codec。在高通MSM8953/MSM8937 平台上，codec分为两部分，一部分是数字 codec，其在 MSM 上；另一部分是模拟 codec,其在 PMIC 上。

二、音频场景(usecase)

usecase在逻辑上对应着音频前端，其定义如下：

1.playback:

Offload playback

A large-sized buffer is sent to the aDSP and the APSS goes to sleep. The aDSP
decodes, applies postprocessing effects, and outputs the PCM data to the
physical sound device. Before the aDSP decoder input runs out of data, it
interrupts the APSS to wake up and send the next set of buffers.
 Supported formats – MP3, AC3, EAC3, AAC, 24bit PCM, 16-bit PCM, FLAC
 Sampling rates in kHz – 8, 11.025, 16, 22.05, 32, 44.1, 48, 64, 88.2, 96, 176.4, 192
 Flags to be set in AudioTrack – AUDIO_OUTPUT_FLAG_DIRECT
|AUDIO_OUTPUT_FLAG_COMPRESS_OFFLOAD
|AUDIO_OUTPUT_FLAG_NON_BLOCKING
 Supported channels – 1, 2, 2.1, 4, 5, 5.1, 6, 7.1

Deep buffer playback

PCM data is sent to the aDSP, postprocessed, and rendered to an output sound
device. Audio effects can also be applied in the ARM or aDSP.
 Use cases – Ringtone, audio/video playback, audio streaming, YouTube streaming, and
so on.
 Supported format – PCM
 Sampling rates in kHz – 44.1 and 48
 Flag to be set in AudioTrack – AUDIO_OUTPUT_FLAG_PRIMARY
 Supported channel – Stereo

Low latency

Playback mode is similar to deep buffer, it uses a smaller buffer size and
minimal or no postprocessing in the aDSP so that the PCM stream is rendered
to the output sound device
 Use cases – Touchtone, gaming audio, and so on.
 Supported format – PCM
 Sampling rates in kHz – 44.1 and 48
 Flag to be set in AudioTrack – AUDIO_OUTPUT_FLAG_FAST
 Supported channel – Stereo

Multichannel

Playback mode where the PCM output of the multichannel decoder is sent to
the aDSP, postprocessed, and rendered at the output device
 Examples – AAC 5.1 channel, Dolby AC3/eAC3 playback
 Supported format – PCM
 Sampling rates in kHz – 44.1 and 48
 Flag to be set in AudioTrack – AUDIO_OUTPUT_FLAG_DIRECT
 Channels supported – 6 (default); changes dynamically

Playback over A2DP and USB does not go through aDSP
FM playback

In FM playback, PCM data from the FM chip is routed to the output sound
device via the ADSP.
 When a headset is connected and an FM app is launched, the FM app sets the
device to AudioSystem.DEVICE_OUT_FM.
 fm.c in hardware/qcom/audio/hal/audio_extn implements the code to open and
starts two PCM hostless sessions.
 One session for PCM capture
 One session for PCM playback
 A hostless session is required to keep the AFE ports in the aDSP active when
the APSS is asleep.
 The audio HAL sends the routing command to enable loopback from internal FM
Tx port to the primary MI2S Rx port using the mixer control settings in the
mixer_paths.xml file.
 Example of mixer settings
path name=“play-fm”
ctl name=“Internal FM RX Volume” value=“1”
ctl name=“PRI_MI2S_RX Port Mixer INTERNAL_FM_TX” value=“1”
ctl name=“MI2S_DL_HL Switch” value=“1”
######2.Recording
Compress mode
Mode of recording where encoded packets are
received by the APSS directly from the ADSP; it is supported for AMR WB
format only

Nontunnel mode

Mode of recording where PCM data from the mic is
preprocessed in DSP and received by the APSS, which then encodes the
PCM to the required encoding format by using the DSP-based or
software-based encoder
 Examples include camcorder recording and in-call recording

Multichannel mode

Used for capturing more than 2 channels of the PCM
stream and encoding them into a multichannel codec format like AC3
 Examples include surround sound camcorder recording, 4 to 6 channel
upsampling for 5.1 channel encoding

FM recording

 PCM data from the FM chip is routed to the APSS via the aDSP for recording in
the intended format.
 Recording occurs as normal use case from the internal FM Tx port.
 Example of mixer settings
name="audio-record capture-fm
ctl name=“MultiMedia1 Mixer INTERNAL_FM_TX” value=“1”

/* Playback usecases */
USECASE_AUDIO_PLAYBACK_DEEP_BUFFER = 0,
USECASE_AUDIO_PLAYBACK_LOW_LATENCY,
USECASE_AUDIO_PLAYBACK_MULTI_CH,
USECASE_AUDIO_PLAYBACK_OFFLOAD,
USECASE_AUDIO_PLAYBACK_OFFLOAD2,
USECASE_AUDIO_PLAYBACK_OFFLOAD3,
USECASE_AUDIO_PLAYBACK_OFFLOAD4,
USECASE_AUDIO_PLAYBACK_OFFLOAD5,
USECASE_AUDIO_PLAYBACK_OFFLOAD6,
USECASE_AUDIO_PLAYBACK_OFFLOAD7,
USECASE_AUDIO_PLAYBACK_OFFLOAD8,
USECASE_AUDIO_PLAYBACK_OFFLOAD9,
USECASE_AUDIO_PLAYBACK_ULL,/* FM usecase */
USECASE_AUDIO_PLAYBACK_FM,/* HFP Use case*/
USECASE_AUDIO_HFP_SCO,
USECASE_AUDIO_HFP_SCO_WB,/* Capture usecases */
USECASE_AUDIO_RECORD,
USECASE_AUDIO_RECORD_COMPRESS,
USECASE_AUDIO_RECORD_COMPRESS2,
USECASE_AUDIO_RECORD_COMPRESS3,
USECASE_AUDIO_RECORD_COMPRESS4,
USECASE_AUDIO_RECORD_LOW_LATENCY,
USECASE_AUDIO_RECORD_FM_VIRTUAL,/* Voice usecase */
USECASE_VOICE_CALL,/* Voice extension usecases */
USECASE_VOICE2_CALL,
USECASE_VOLTE_CALL,
USECASE_QCHAT_CALL,
USECASE_VOWLAN_CALL,
USECASE_VOICEMMODE1_CALL,
USECASE_VOICEMMODE2_CALL,
USECASE_COMPRESS_VOIP_CALL,USECASE_INCALL_REC_UPLINK,
USECASE_INCALL_REC_DOWNLINK,
USECASE_INCALL_REC_UPLINK_AND_DOWNLINK,
USECASE_INCALL_REC_UPLINK_COMPRESS,
USECASE_INCALL_REC_DOWNLINK_COMPRESS,
USECASE_INCALL_REC_UPLINK_AND_DOWNLINK_COMPRESS,USECASE_INCALL_MUSIC_UPLINK,
USECASE_INCALL_MUSIC_UPLINK2,USECASE_AUDIO_SPKR_CALIB_RX,
USECASE_AUDIO_SPKR_CALIB_TX,USECASE_AUDIO_PLAYBACK_AFE_PROXY,
USECASE_AUDIO_RECORD_AFE_PROXY,USECASE_AUDIO_PLAYBACK_EXT_DISP_SILENCE,    /* Playback usecases */
USECASE_AUDIO_PLAYBACK_DEEP_BUFFER = 0,
USECASE_AUDIO_PLAYBACK_LOW_LATENCY,
USECASE_AUDIO_PLAYBACK_MULTI_CH,
USECASE_AUDIO_PLAYBACK_OFFLOAD,
USECASE_AUDIO_PLAYBACK_OFFLOAD2,
USECASE_AUDIO_PLAYBACK_OFFLOAD3,
USECASE_AUDIO_PLAYBACK_OFFLOAD4,
USECASE_AUDIO_PLAYBACK_OFFLOAD5,
USECASE_AUDIO_PLAYBACK_OFFLOAD6,
USECASE_AUDIO_PLAYBACK_OFFLOAD7,
USECASE_AUDIO_PLAYBACK_OFFLOAD8,
USECASE_AUDIO_PLAYBACK_OFFLOAD9,
USECASE_AUDIO_PLAYBACK_ULL,/* FM usecase */
USECASE_AUDIO_PLAYBACK_FM,/* HFP Use case*/
USECASE_AUDIO_HFP_SCO,
USECASE_AUDIO_HFP_SCO_WB,/* Capture usecases */
USECASE_AUDIO_RECORD,
USECASE_AUDIO_RECORD_COMPRESS,
USECASE_AUDIO_RECORD_COMPRESS2,
USECASE_AUDIO_RECORD_COMPRESS3,
USECASE_AUDIO_RECORD_COMPRESS4,
USECASE_AUDIO_RECORD_LOW_LATENCY,
USECASE_AUDIO_RECORD_FM_VIRTUAL,/* Voice usecase */
USECASE_VOICE_CALL,/* Voice extension usecases */
USECASE_VOICE2_CALL,
USECASE_VOLTE_CALL,
USECASE_QCHAT_CALL,
USECASE_VOWLAN_CALL,
USECASE_VOICEMMODE1_CALL,
USECASE_VOICEMMODE2_CALL,
USECASE_COMPRESS_VOIP_CALL,USECASE_INCALL_REC_UPLINK,
USECASE_INCALL_REC_DOWNLINK,
USECASE_INCALL_REC_UPLINK_AND_DOWNLINK,
USECASE_INCALL_REC_UPLINK_COMPRESS,
USECASE_INCALL_REC_DOWNLINK_COMPRESS,
USECASE_INCALL_REC_UPLINK_AND_DOWNLINK_COMPRESS,USECASE_INCALL_MUSIC_UPLINK,
USECASE_INCALL_MUSIC_UPLINK2,USECASE_AUDIO_SPKR_CALIB_RX,
USECASE_AUDIO_SPKR_CALIB_TX,USECASE_AUDIO_PLAYBACK_AFE_PROXY,
USECASE_AUDIO_RECORD_AFE_PROXY,USECASE_AUDIO_PLAYBACK_EXT_DISP_SILENCE

三、音频通路配置(音频控件配置)

简单的来说，就是将音频前端(FE)经过音频后端(BE)，与音频设备(Audio Device)连接起来。
(FE <=> BE <==> Devices)
usecase 通过路由与音频设备相关联。其配置一般放在某个 xml 文件，以 MSM8953 为例，其放在 mixer_paths_qrd_sku3.xml 文件中。该文件中的配置，又称为音频控件配置。
音频控件配置。
在高通 bring up 音频的时候，一般不会先去修改 mixer-path.xml 文件，而是先使用以下几个工具，进行通路配置验证：

tinymix: 配置音频路由；
tinycap：录音；
tinyplay：播放

1. 外部 SPK 调试举例

比如在调试外部 SPK，切其音源为右声道(HPHR)，可以使用 tinymix + tinyplay 进行音频通路验证：

tinymix "PRI_MI2S_RX Audio Mixer MultiMedia1" "1"
tinymix "MI2S_RX Channels" "One"
tinymix "RX2 MIX1 INP1" "RX1"
tinymix "RDAC2 MUX" "RX2"
tinymix "HPHR" "Switch"
tinymix "Ext Spk Switch" "On"
tinyplay /sdcard/test.wav

如何理解 tinymix “MI2S_RX Channels” “One”
MI2S RX 线路上通道的个数：One or Two

如何理解 tinymix “RX2 MIX1 INP1” “RX1”
表示 SPK 以内部的 Rx Mix2 作为 mixer 作为输入，在 mixer 端又以 mixer的 Rx1 作为输入

如何理解 tinymix “RDAC2 MUX” “RX2”
表示音源右声道的前端数据来自 DAC2 模数转换器，而 DAC2 的又以 RDAC2 的 Rx2 作为输入

如何理解 tinymix “HPHR” “Switch”
表示右声道是打开还是关闭，其值为 Switch or Zero

如何理解 tinymix “Ext Spk Switch” “On”
表示打开外部 SPK，一般的，外部SPK都带有 PA 使能引脚，需要特别的打开，其值有 On or Off

则整个外部 SPK 的音频链路可以简化为：

Rx Mix2 --> DAC2 --> HPHR --> SPK

上面对 SPK 的配置，整个路径如图所示：

2.单 Mic 调试举例(主麦)

tinymix "MultiMedia1 Mixer TERT_MI2S_TX" "1"
tinymix "MI2S_TX Channels" "One"
tinymix "ADC1 Volume" "6"
tinymix "DEC1 MUX" "ADC1"
tinymix "IIR1 INP1 MUX" "DEC1"
tinycap /data/input1.wav –C 1 –R 44100 –T 20

字段意思，同 SPK 的差不多，只是需要注意的是，MIC 和 SPK 的音频流方向是不同的，并且，不同的 Audio Device 是挂在不同的 DAI 口上，这就体现在了以下两条命令：

tinymix “PRI_MI2S_RX Audio Mixer MultiMedia1” “1” // SPK ： BE DAI —> DSP —> FE PCM

tinymix “MultiMedia1 Mixer TERT_MI2S_TX” “1” // MIC: FE PCM —> DSP —> BE DAI
【
经网友提示，现将上面内容调整修改为：
tinymix “PRI_MI2S_RX Audio Mixer MultiMedia1” “1” // SPK ： FE PCM —> DSP —> BE DAI
tinymix “MultiMedia1 Mixer TERT_MI2S_TX” “1” // MIC: BE DAI —> DSP —> FE PCM
】
整个单 MIC的路径如图所示：

在确定通路没有问题的情况下，再将 tinymix 的参数导入到 mixer-path.xml 对应的音频控件中。并将该 mixer-path.xml push 到 /etc/ 目录下替换原来的文件，然后重启设备，通过系统 apk 进行测试。

引用参考

1. 高通文档

相关的 overview & bring up & audio debug 文档，此处不方便透漏文档编号和名字

2. csdn blog 引用

本篇笔记，在下面这篇博客内容上进行了修改和扩充。

Qualcomm Audio HAL 音频通路设置

感谢原作者的分享！