经常要搞音频调试，很麻烦，现在记录下。

常用数据结构

/system/media/audio/include/system/audio.h

定义了常用的stream类型,例如3就是音乐的stream。

/* Audio stream types */

typedef enum {

/* These values must kept in sync with

* frameworks/base/media/java/android/media/AudioSystem.java

*/

AUDIO_STREAM_DEFAULT = -1,

AUDIO_STREAM_MIN = 0,

AUDIO_STREAM_VOICE_CALL = 0,

AUDIO_STREAM_SYSTEM = 1,

AUDIO_STREAM_RING = 2,

AUDIO_STREAM_MUSIC = 3,

AUDIO_STREAM_ALARM = 4,

AUDIO_STREAM_NOTIFICATION = 5,

AUDIO_STREAM_BLUETOOTH_SCO = 6,

AUDIO_STREAM_ENFORCED_AUDIBLE = 7, /* Sounds that cannot be muted by user

* and must be routed to speaker

*/

AUDIO_STREAM_DTMF = 8,

AUDIO_STREAM_TTS = 9, /* Transmitted Through Speaker.

* Plays over speaker only, silent on other devices.

*/

AUDIO_STREAM_ACCESSIBILITY = 10, /* For accessibility talk back prompts */

AUDIO_STREAM_REROUTING = 11, /* For dynamic policy output mixes */

AUDIO_STREAM_PATCH = 12, /* For internal audio flinger tracks. Fixed volume */

AUDIO_STREAM_PUBLIC_CNT = AUDIO_STREAM_TTS + 1,

AUDIO_STREAM_CNT = AUDIO_STREAM_PATCH + 1,

} audio_stream_type_t;

定义了输出设备id,例如0x4对应设备就是有线耳机

/* output devices */

AUDIO_DEVICE_OUT_EARPIECE = 0x1,

AUDIO_DEVICE_OUT_SPEAKER = 0x2,

AUDIO_DEVICE_OUT_WIRED_HEADSET = 0x4,

AUDIO_DEVICE_OUT_WIRED_HEADPHONE = 0x8,

2.常用调试方法

1.查看音量的log

关键tag是AudioMTKGainController

插入耳机，调小音量，出现下面log

01-01 00:29:51.752251 499 906 D AudioMTKGainController: setNormalVolume(), stream 8, devices 0x8, index 8, mode 0x0

上面的log就是把设备耳机的stream 8(DTMF声)的音量设置为8。

拔出耳机，调小音量，出现下面log

AudioMTKGainController: setNormalVolume(), stream 1, devices 0x2, index 13, mode 0x0

AudioMTKGainController: getGainDevice(), input devices = 0x2, return gainDevice = 2

AudioMTKGainController: setSpeakerGain(), gain = 10, spkAnaType = 2, spkLMixerName = Audio_Speaker_PGA_gain, spkRMixerName = Audio_Speaker_PGA_gain

上面的log就是把设备扬声器的stream 1(系统声)的音量设置为13。这里的13只是上层表示的一个音量index，要转换成底层的gain，例如index=13时，gain就是10.

下面是设置speaker音量的流程

status_t AudioMTKGainController::setNormalVolume(int stream, int index, int devices, audio_mode_t mode)

{

ALOGD("setNormalVolume(), stream %d, devices 0x%x, index %d, mode 0x%x", stream, devices, index, mode);

// get gain device

GAIN_DEVICE gainDevice = getGainDevice(devices); //获取要设置增益的设备，2是speaker

if (isSpeakerCategory(gainDevice))

{

if (mSpec.spkAnaType >= 0 && mSpec.spkAnaType < NUM_GAIN_ANA_TYPE)

{

unsigned char gain = mGainTable.streamGain[stream][gainDevice][index].analog[mSpec.spkAnaType]; //把index转换成底层的gain增益

setSpeakerGain(gain); //设置speaker的增益

}

}

2.查看设置Parameter的log

在音频控制中，常常有一种比较粗暴的设置方法，就是直接设置参数Parameter，这个可以在java层直接设置，然后在hal层进行处理。例如在fm中,在java层设置参数

/packages/apps/FMRadio/src/com/android/fmradio/FmService.java

mAudioManager.setParameters("AudioFmPreStop=1");

status_t AudioSystem::setParameters(audio_io_handle_t ioHandle, const String8& keyValuePairs)

{

const sp& af = AudioSystem::get_audio_flinger();

if (af == 0) return PERMISSION_DENIED;

return af->setParameters(ioHandle, keyValuePairs);

}

String8 AudioSystem::getParameters(audio_io_handle_t ioHandle, const String8& keys)

{

const sp& af = AudioSystem::get_audio_flinger();

String8 result = String8("");

if (af == 0) return result;

result = af->getParameters(ioHandle, keys);

return result;

}

上面是原生的代码，是在AudioFlinger中处理，而在mtk中，往往是有AudioPolicy处理

aps->SetPolicyManagerParameters (POLICY_SET_FM_PRESTOP,value,0,0);

3.了解路由规则

下面的log表示路由到耳机

01-03 06:44:36.621514 570 1265 D AudioALSAStreamOut: +setParameters(): routing=8

/* Routing bits for the former setRouting/getRouting API */

/** @deprecated */

@Deprecated public static final int ROUTE_EARPIECE = (1 << 0);

/** @deprecated */

@Deprecated public static final int ROUTE_SPEAKER = (1 << 1);

/** @deprecated use {@link #ROUTE_BLUETOOTH_SCO} */

@Deprecated public static final int ROUTE_BLUETOOTH = (1 << 2);

/** @deprecated */

@Deprecated public static final int ROUTE_BLUETOOTH_SCO = (1 << 2);

/** @deprecated */

@Deprecated public static final int ROUTE_HEADSET = (1 << 3);

/** @deprecated */

@Deprecated public static final int ROUTE_BLUETOOTH_A2DP = (1 << 4);

/** @deprecated */

@Deprecated public static final int ROUTE_ALL = 0xFFFFFFFF;

3.了解Audio Patch

Audio Patch是一个通路，包含源source和目标sink。例如FM就是源，耳机就是目标sink。

4.了解Audio Mode

/* modes for setPhoneState, must match AudioSystem.h audio_mode */

public static final int MODE_INVALID = -2;

public static final int MODE_CURRENT = -1;

public static final int MODE_NORMAL = 0;

public static final int MODE_RINGTONE = 1;

public static final int MODE_IN_CALL = 2;

public static final int MODE_IN_COMMUNICATION = 3;

public static final int NUM_MODES = 4;

5.了解Audio Trace

看看下面的log

01-03 08:26:37.051925 5353 5353 V ToneGenerator: AudioTrack(0xf2f36700) created

01-03 08:26:37.051999 5353 5353 D AudioTrack: set(): 0xf2f36700, streamType 8, sampleRate 44100, format 0x1, channelMask 0x1, frameCount 882, flags #4, notificationFrames 0, sessionId 0, transferType 1, uid -1, pid -1

上面的log是点击拨号盘，发出声音打出来的。0xf2f36700是AudioTrack对象的ID。streamType 是按键音。

6.关于SessionId

抄自网上:

一个Session就是一个会话。每个会话都有一个独一无二的Id来标识。该Id的最终管理在AudioFlinger中。

一个会话可以被多个AudioTrack对象和MediaPlayer共用。

共用一个Session的AudioTrack和MediaPlayer共享相同的AudioEffect。