android audio arch

最新推荐文章于 2023-02-28 19:28:16 发布

qianxuedegushi

最新推荐文章于 2023-02-28 19:28:16 发布

阅读量1.5k

点赞数

分类专栏： arm android audio

android audio 同时被 2 个专栏收录

36 篇文章 24 订阅

订阅专栏

arm

6 篇文章 0 订阅

订阅专栏

ALSA System on Chip(ASoC)驱动将一个audio子系统分成四个部分：
Machine driver， Platform driver， CPU driver以及Codec driver。
Machine 驱动 :
将平台，CPU以及codec驱动绑定在一块
实现在kernel/sound/soc/msm/<chipset.c>
定义前端FE和后端BE，DAI（digital audio interface） links
platform 驱动:
包括了平台相关的音频数据流的传输和路由的控制
被区分成FE和BE两个平台驱动
FE :

Audio -例化PCM播放和采集，将播放的pcm数据从用户空间传递给DSP，采集是逆过程，高通使用ASM接口，实现于kernel/sound/soc/msm-pcm-q6.c

提供了ALSA SoC framework需要的FE PCM设备信息，由平台驱动提供的ASoC framework和音频路由表可以指示PCM 播放/采集。
FE CPU DAI实现于kernel/sound/soc/msm/msm-dai-fe.c

Voice-初始化和销毁语音电话建立，实现与kernel/sound/soc/msm-pcm-voice-v2.c
VoIP-初始化和销毁MVS接口在DSP和用户空间互传数据。kernel/sound/soc/msm-pcm-voip-v2.c
Compress offload-Compress offload 播放数据传输到DSP。kernel/sound/soc/msm-compress-q6-v2.c

BE :

根据初始化的PCM播放和采集配置DSP AFE模块对应的audio 音频端口

定义了BE CPU　ＤＡＩ实现于kernel/sound/soc/msm/qdsp6v2/msm-dai-q6-v2.c

Routing-audio 通路选择，kernel/sound/soc/msm-pcm-routing-v2.c

CPU driver :

分为FE和BE CPU驱动

Codec 驱动 User Spcace Moudles :
Audio Hardware Abstraction Layer(AHAL)使用tinyalsa将Audioflinger调用映射到ASoC驱动
tinyalsa －kernel Asoc驱动接口，ＡＨＡＬ使用，提供了stream和设备管理需要的基本PCM和Mixer控制API
Audio Route－该模块从ＸＭＬ文件读取ALSA mixer控制器，并根据ＡＨＡＬ设置mixer控制器。
Multimedia framework－stagefright 1) 使用标准音频格式的播放和采集；2)和编解码库以及OpenMAX IL组件通信，实现音频编解码
Audio service 1) 运行时服务，由system server启动，server manager管理； 2) 注册intents，当接收到来自不同应用（HDMI，Bluetooth）的信息后通知Audio系统
Audio Flinger 1) 通过libaudio，蓝牙A2DP管理所有音频输入／输出设备，2) 将多个audio stream合成单一PCM，混合的output输出源被路由给输出设备　3)　music stream播放的音量控制
Audio Policy Manager（APM） 1) 定义了多个音频源并发的管理策略,管理各种输入／输出设备接口mic,speaker,headphone,headset,a2dp,bluetooth sco 2) 设置场景（电话，音乐，系统音，通知） 3) 基于stream模式选择和定义合适的路由策略.定义了音频类型（语音，播放，ring）在何种设备上播放（Bluetooth ,speaker，headset）.　４）管理每一个stream的音量.

Audio Output Stream and Volume Control :

缺省的音频输出（primary output）使用AudioFlinger中的Mixer线程，音量和音效将在Flinger中使用
Lowlatency输出使用Audioflinger中的快速mixer，音量控制在Flinger中
Compress offload播放使用audio flinger中的compress offload线程，音量和音效在DSP中。

Voip播放在Audioflinger中的mixerthread， volume 和echo cancelation 由DSP完成。

设备和流管理 :

Frontend DAI :

对用户空间作为PCM设备可见
可以将Audio路由到多个BE DAI
Routing 由用户空间mixer 控制
PCM由ALSA直接控制
定义于 msm-dai-fe.c

static struct snd_soc_dai_driver msm_fe_dais[]={
.playback = {
.stream_name = "Multimedia1 Playback",
.aif_name = "MM_DL1",
.rates
...
}

}

当front DAI使能时，AIF将会被使能。

用户空间和audio相关的代码用于调试和定制化：

sm8974 – Contains the audio Hardware Abstraction Layer (HAL)-related code

<APSS_BUILD>/external/tinyalsa/ – Contains the code related to tinymix, tinyplay, and tinycap

<APSS_BUILD>/hardware/qcom/audio/mm-audio – Contains the implementation of QTI OMX components for the audio encoder and decoders

<APSS_BUILD>/frameworks/av/media/libstagefright/ – Contains the source code for Google’s Stagefright implementation

<APSS_BUILD>/frameworks/av/media/libmediaplayerservice/nuplayer - Contains the source code for Google’s nuplayer implementation

<APSS_BUILD>/frameworks/av/services/audioflinger/ – Contains the source code for AudioFlinger that manages audio streams from the user space

<APSS_BUILD>/vendor/qcom/proprietary/mm-audio/ – Contains the code related to the Audio Calibration Database (ACDB) driver, parsers for DTS and AC3, surround sound, SVA, etc.

<APSS_BUILD>/external/bluetooth/bluedroid/ – Contains the code related to Bluetooth® (BT) A2DP used in a QTI platform

<APSS_BUILD>/external/bluetooth/bluedroid/audio_a2dp_hw/ – Contains the A2DP audio HAL implementation

<APSS_BUILD>/hardware/libhardware/modules/usbaudio/ – Contains the USB HAL implementation for a USB dock use case

<APSS_BUILD>/hardware/qcom/audio/hal/audio_extn/:

audio_extn.c – Implements the wrapper function for audio extension features, such as FM, Dolby, Compress capture, HFP, SVA (listen), speaker protection, SSR, USB audio over headset

usb.c – Contains the HAL implementation for USB playback and record over the headset

compress_capture.c – Contains the HAL implementation for Compress capture

dolby.c – Contains the HAL implementation for the Dolby postprocessing feature

fm.c – Contains the HAL implementation for the FM playback and recording feature

hfp.c – Contains the HAL implementation for the hands-free profile feature where the MSM™ chipset can be used as as a BT headset device

listen.c – Contains the HAL implementation for the Snapdragon™ Voice Activation (SVA) feature

spkr_protection.c – Contains the HAL implementation for the speaker protection feature

ssr.c – Contains the HAL implementation for the surround sound recording feature

<APSS_BUILD>/vendor/qcom/proprietary/wfd/mm/source/framework/src/ – Contains the Wi-Fi Display (WFD) frameworks-related code; WFDMMSourceAudioSource.cpp configures the RT Proxy port via ALSA APIs and gets the PCM data from the audio layer

<APSS_BUILD>/system/core/include/system/ – Contains audio.h and audio_policy.h that contain enum definitions and inline functions used all over the code for audio in the user space

<APSS_BUILD>/frameworks/base/media/java/android/media/ – Contains .java files for audio that expose APIs that can be called by Android™ applications written in Java