Android Audio 分析

1、架構

 

2、MediaServer初始化

全部的media服務都在進程mediaserver裏。其代碼在framework/base/media/mediaserver/main_mediaserver.cpp中。初始化過程以下。

1. 先初始化AudioFlinger

其初始化經過AudioFlinger的父類BindService建立惟一的AudioFlinger實例。

2. 而後初始化MediaPlayerService和CameraService

3. 最後初始化AudioPolicyService

 

3、 AudioFlinger分析

AudioFlinger負責管理每一個音軌AudioTrack及RecordTrack，主音量控制，每種聲音流的屬性設置，設備控制，音效控制。

 

4、 AudioPolicyService分析

AudioPolicyService用於聲音策略管理

客戶端代碼在framework/base/media/libmedia/AudioSystem.cpp中，其實該客戶端只是監測AudioPolicyService是否死亡（binderDied)。真正的客戶端爲BpAudioPolicyService類，在framework/base/media/libmedia/IAudioPolicyService.cpp中。

接口

 1 adjustVolume
 2 adjustSuggestedStreamVolume
 3 adjustStreamVolume
 4 setStreamVolume
 5 setStreamSolo
 6 setStreamMute
 7 getStreamVolume
 8 getStreamMaxVolume
 9 setRingerMode
10 getRingerMode
11 setVibrateSetting
12 getVibrateSetting
13 shouldVibrate
14 setMode
15 getMode
16 playSoundEffect
17 playSoundEffectVolume
18 loadSoundEffects
19 unloadSoundEffects
20 reloadAudioSettings
21 setSpeakerphoneOn
22 isSpeakerphoneOn
23 setBluetoothScoOn
24 isBluetoothScoOn
25 requestAudioFocus
26 abandonAudioFocus
27 unregisterAudioFocusClient
28 registerMediaButtonEventReceiver
29 unregisterMediaButtonEventReceiver
30 startBluetoothSco
31 stopBluetoothSco

 

服務端代碼在framework/base/services/audioflinger/AudioPolicyService.cpp中，實現了AudioPolicyService本地代碼及服務內部使用的AudioPolicyClientInterface接口，該接口被AudioPolicyManager調用。

 1 openOutput
 2 openDuplicateOutput
 3 closeOutput
 4 suspendOutput
 5 restoreOutput
 6 openInput
 7 loseInput
 8 setStreamVolume
 9 setStreamOutput
10 setParameters
11 getParameters
12 startTone
13 stopTone
14 setVoiceVolume
15 moveEffects

 

AudioPolicyService會啓動兩個AudioCommandThread類的線程mTonePlaybackThread和mAudioCommandThread，而後構造AudioPolicyManager的實例mpPolicyManager。基本上全部的服務接口都會在檢查權限後調用mpPolicyManager的相應實現。

mTonePlaybackThread用戶播放通話中的通知音

mAudioCommandThread用於設置AudioFlinger的一些設置(爲何使用AudioCommandThread，而後調用AudioSystem中接口控制AudioFlinger?)

 

5、 AudioManager分析

1. 音頻焦點轉

1 requestAudioFocus         請求焦點
2 abandonAudioFocus         放棄焦點

requestAudioFocus的第一參數是Listener,能夠是下面枚舉的後三個值

第二個參數是當前流類型

最後一個參數能夠是如下的前三個值

1 AUDIOFOCUS_GAIN 1 
2 AUDIOFOCUS_GAIN_TRANSIENT 2
3 AUDIOFOCUS_GAIN_TRANSIENT_MAY_DUCK = 3;
4 AUDIOFOCUS_LOSS = -1 * AUDIOFOCUS_GAIN;
5 AUDIOFOCUS_LOSS_TRANSIENT = -1 * AUDIOFOCUS_GAIN_TRANSIENT;
6 AUDIOFOCUS_LOSS_TRANSIENT_CAN_DUCK =
7  -1 * AUDIOFOCUS_GAIN_TRANSIENT_MAY_DUCK;

 

6、 AudioPolicyManager分析

流類型stream_type(在 AudioSystem中定義)

TYPE                          VALUE                                STRATEGY
DEFAULT(->MUSIC)                -1                             STRATEGY_MEDIA(0)
VOICE_CALL                       0                             STRATEGY_PHONE(1)
SYSTEM                           1                             STRATEGY_MEDIA(0)
RING                             2                             STRATEGY_SONIFICATION(2)
MUSIC                            3                             STRATEGY_MEDIA(0)
ALARM                            4                             STRATEGY_SONIFICATION(2)
NOTIFICATION                     5                             STRATEGY_SONIFICATION(2)
BLUETOOTH_SCO                    6                             STRATEGY_PHONE(1)
// Sounds that cannot be muted by user and must be routed to speaker
ENFORCED_AUDIBLE                 7                             STRATEGY_SONIFICATION(2)
DTMF                             8                             STRATEGY_DTMF(3)
TTS                              9                             STRATEGY_MEDIA(0)
HDMI                             10  MV的自定義類型              STRATEGY_MEDIA(0)
FM                               11  MV的自定義類型              STRATEGY_MEDIA(0)

 

音頻狀態audio_mode(在 AudioSystem中定義)

1  enum audio_mode {
2  MODE_INVALID = -2,
3  MODE_CURRENT = -1,
4  MODE_NORMAL = 0,
5  MODE_RINGTONE,
6  MODE_IN_CALL,
7  MODE_IN_COMMUNICATION,
8  NUM_MODES  // not a valid entry, denotes end-of-list
9  };

 

策略類型routing_strategy:

STRATEGY_MEDIA             0
STRATEGY_PHONE             1
STRATEGY_SONIFICATION      2
STRATEGY_DTMF              3

 

設備類型audio_devices：

 1 enum audio_devices {
 2  // output devices
 3  DEVICE_OUT_EARPIECE = 0x1,
 4  DEVICE_OUT_SPEAKER = 0x2,
 5  DEVICE_OUT_WIRED_HEADSET = 0x4,
 6  DEVICE_OUT_WIRED_HEADPHONE = 0x8,
 7  DEVICE_OUT_BLUETOOTH_SCO = 0x10,
 8  DEVICE_OUT_BLUETOOTH_SCO_HEADSET = 0x20,
 9  DEVICE_OUT_BLUETOOTH_SCO_CARKIT = 0x40,
10  DEVICE_OUT_BLUETOOTH_A2DP = 0x80,
11  DEVICE_OUT_BLUETOOTH_A2DP_HEADPHONES = 0x100,
12  DEVICE_OUT_BLUETOOTH_A2DP_SPEAKER = 0x200,
13  DEVICE_OUT_AUX_DIGITAL = 0x400,
14  DEVICE_OUT_DEFAULT = 0x8000,
15  DEVICE_OUT_ALL = (DEVICE_OUT_EARPIECE | DEVICE_OUT_SPEAKER | DEVICE_OUT_WIRED_HEADSET |
16  DEVICE_OUT_WIRED_HEADPHONE | DEVICE_OUT_BLUETOOTH_SCO | DEVICE_OUT_BLUETOOTH_SCO_HEADSET |
17  DEVICE_OUT_BLUETOOTH_SCO_CARKIT | DEVICE_OUT_BLUETOOTH_A2DP | DEVICE_OUT_BLUETOOTH_A2DP_HEADPHONES |
18  DEVICE_OUT_BLUETOOTH_A2DP_SPEAKER | DEVICE_OUT_AUX_DIGITAL | DEVICE_OUT_DEFAULT),
19  DEVICE_OUT_ALL_A2DP = (DEVICE_OUT_BLUETOOTH_A2DP | DEVICE_OUT_BLUETOOTH_A2DP_HEADPHONES |
20  DEVICE_OUT_BLUETOOTH_A2DP_SPEAKER),
21 
22  // input devices
23  DEVICE_IN_COMMUNICATION = 0x10000,
24  DEVICE_IN_AMBIENT = 0x20000,
25  DEVICE_IN_BUILTIN_MIC = 0x40000,
26  DEVICE_IN_BLUETOOTH_SCO_HEADSET = 0x80000,
27  DEVICE_IN_WIRED_HEADSET = 0x100000,
28  DEVICE_IN_AUX_DIGITAL = 0x200000,
29  DEVICE_IN_VOICE_CALL = 0x400000,
30  DEVICE_IN_BACK_MIC = 0x800000,
31  DEVICE_IN_DEFAULT = 0x80000000,
32 
33  DEVICE_IN_ALL = (DEVICE_IN_COMMUNICATION | DEVICE_IN_AMBIENT | DEVICE_IN_BUILTIN_MIC |
34  DEVICE_IN_BLUETOOTH_SCO_HEADSET | DEVICE_IN_WIRED_HEADSET | DEVICE_IN_AUX_DIGITAL |
35  DEVICE_IN_VOICE_CALL | DEVICE_IN_BACK_MIC | DEVICE_IN_DEFAULT)
36  };

 

流程分析：

1.音量控制

initStreamVolume
setStreamVolumeIndex
setPhoneState

2.播放控制

入口函數

 1 startOutput
 2 
 3 getStrategy
 4 
 5 getDeviceForStrategy
 6 
 7 setOutputDevice
 8 
 9 applyStreamVolume
10 
11 checkAndSetVolume