使 Android 使用新的音频 HAL 需要什么

Question

我正在编写一个新的 Android 音频 HAL，让我的应用程序将音频馈送到其他应用程序，让我的手持远程麦克风输入到达 Google 应用程序。本质上，虚拟音频线。

这是一项正在进行的工作。我可能会覆盖 AUDIO_DEVICE_IN_BACK_MIC，但这是开放的建议。

我对如何确保 Android 使用此 HAL 进行输入有疑问。

我需要替换 audio.primary.default.so 还是应该保留 audio.vcable.default.so？

更具体地说：如果我不替换主要的，Android 如何知道使用我的 HAL 而不是主要的？

更新：

我真的可以在这项工作中得到任何帮助。任何指针都有帮助。

我写了一个Audio HAL模块。我在 audio_policy.conf 中添加了以下（粗体项）：

全球：

global_configuration {
  attached_output_devices AUDIO_DEVICE_OUT_SPEAKER|**AUDIO_DEVICE_OUT_LINE**
  default_output_device AUDIO_DEVICE_OUT_SPEAKER
  attached_input_devices AUDIO_DEVICE_IN_BUILTIN_MIC|AUDIO_DEVICE_IN_BACK_MIC|AUDIO_DEVICE_IN_REMOTE_SUBMIX|**AUDIO_DEVICE_IN_LINE**
}

及audio_hw_modules

以下

  vloop {
    inputs {
      vloop {
        sampling_rates 16000
        channel_masks AUDIO_CHANNEL_IN_MONO
        formats AUDIO_FORMAT_PCM_16_BIT
        devices AUDIO_DEVICE_IN_LINE
      }
    }
    outputs {
      vloop {
        sampling_rates 16000
        channel_masks AUDIO_CHANNEL_OUT_STEREO
        formats AUDIO_FORMAT_PCM_16_BIT
        devices AUDIO_DEVICE_OUT_LINE
        flags AUDIO_OUTPUT_FLAG_DIRECT
      }
    }
  }

我还在 AudioFlinger.cpp

中添加了以下内容（粗体）

static const char * const audio_interfaces[] = {
    AUDIO_HARDWARE_MODULE_ID_PRIMARY,
    AUDIO_HARDWARE_MODULE_ID_A2DP,
    AUDIO_HARDWARE_MODULE_ID_USB,
    **AUDIO_HARDWARE_MODULE_ID_VLOOP**
};

我可以看到在启动过程中，我的 HAL 被加载了，我得到了这些日志：

10-06 06:14:40.365 194-194/? I/AudioFlinger: Using default 3000 mSec as standby time.
10-06 06:14:46.664 194-194/? I/AudioPolicyService: AudioPolicyService CSTOR in new mode
10-06 06:14:46.673 194-194/? I/APM::ConfigParsingUtils: loadAudioPolicyConfig() loaded /system/etc/audio_policy.conf
10-06 06:14:46.681 194-194/? D/audio_hw_primary: adev_open: enter
10-06 06:14:46.797 194-194/? I/AudioFlinger: loadHwModule() Loaded primary audio interface from QCOM Audio HAL (audio) handle 1
10-06 06:14:46.797 194-194/? I/AudioFlinger: openOutput(), module 1 Device 2, SamplingRate 48000, Format 0x000001, Channels 3, flags 2
10-06 06:14:46.797 194-194/? I/AudioFlinger: AudioStreamOut::open(), mHalFormatIsLinearPcm = 1
10-06 06:14:46.798 194-194/? I/AudioFlinger: HAL output buffer size 240 frames, normal sink buffer size 960 frames
10-06 06:14:46.813 194-194/? I/AudioFlinger: Using module 1 has the primary audio interface
10-06 06:14:46.816 194-607/? I/AudioFlinger: AudioFlinger's thread 0xb4140000 ready to run
10-06 06:14:46.816 194-607/? D/audio_hw_primary: out_set_parameters: enter: usecase(1: low-latency-playback) kvpairs: routing=2
10-06 06:14:46.818 194-194/? I/AudioFlinger: openOutput(), module 1 Device 2, SamplingRate 48000, Format 0x000001, Channels 3, flags 8
10-06 06:14:46.818 194-194/? I/AudioFlinger: AudioStreamOut::open(), mHalFormatIsLinearPcm = 1
10-06 06:14:46.818 194-194/? I/AudioFlinger: HAL output buffer size 960 frames, normal sink buffer size 960 frames
10-06 06:14:46.818 194-608/? I/AudioFlinger: AudioFlinger's thread 0xb3dc0000 ready to run
10-06 06:14:46.818 194-607/? E/AudioFlinger: no wake lock to update!
10-06 06:14:46.818 194-608/? D/audio_hw_primary: out_set_parameters: enter: usecase(0: deep-buffer-playback) kvpairs: routing=2
10-06 06:14:46.818 194-608/? E/AudioFlinger: no wake lock to update!
10-06 06:14:46.820 194-609/? I/AudioFlinger: AudioFlinger's thread 0xb3c40000 ready to run
10-06 06:14:46.823 194-194/? I/AudioFlinger: loadHwModule() Loaded a2dp audio interface from A2DP Audio HW HAL (audio) handle 7
10-06 06:14:46.828 194-194/? I/AudioFlinger: loadHwModule() Loaded usb audio interface from USB audio HW HAL (audio) handle 8
10-06 06:14:46.832 194-194/? I/r_submix: adev_open(name=audio_hw_if)
10-06 06:14:46.832 194-194/? I/AudioFlinger: loadHwModule() Loaded r_submix audio interface from Wifi Display audio HAL (audio) handle 9
10-06 06:14:46.832 194-194/? D/r_submix: submix_audio_device_create_pipe_l(addr=0, idx=9)
10-06 06:14:46.833 194-610/? I/AudioFlinger: AudioFlinger's thread 0xb3bc0000 ready to run
10-06 06:14:46.833 194-194/? D/r_submix: submix_audio_device_release_pipe_l(idx=9) addr=0
10-06 06:14:46.833 194-194/? D/r_submix: submix_audio_device_destroy_pipe_l(): pipe destroyed
10-06 06:14:46.835 194-194/? D/audio_vloop: adev_open: audio_hw_if
10-06 06:14:46.835 194-194/? D/audio_vloop: adev_open(): 1678
10-06 06:14:46.835 194-194/? D/audio_vloop: adev_open(): 1685
10-06 06:14:46.835 194-194/? D/audio_vloop: adev_open(): 1688
10-06 06:14:46.835 194-194/? D/audio_vloop: adev_open(): 1722
10-06 06:14:46.835 194-194/? D/audio_vloop: adev_init_check(): 1252
10-06 06:14:46.835 194-194/? D/audio_vloop: adev_set_master_volume: 1.000000
10-06 06:14:46.835 194-194/? D/audio_vloop: adev_set_master_mute: 0
10-06 06:14:46.835 194-194/? I/AudioFlinger: loadHwModule() Loaded vloop audio interface from UI_audio_HW_HAL (audio) handle 11
10-06 06:14:46.835 194-194/? D/audio_vloop: adev_open_input_stream(): 1490
10-06 06:14:46.835 194-194/? D/audio_vloop: in_get_sample_rate(): 979
10-06 06:14:46.835 194-194/? D/audio_vloop: in_get_channels(): 1017
10-06 06:14:46.835 194-194/? D/audio_vloop: in_get_channels: 0x00000001
10-06 06:14:46.835 194-194/? D/audio_vloop: in_get_format(): 1029
10-06 06:14:46.836 194-194/? D/audio_vloop: in_get_format: 0x00000001
10-06 06:14:46.836 194-194/? D/audio_vloop: in_get_format(): 1029
10-06 06:14:46.836 194-194/? D/audio_vloop: in_get_format: 0x00000001
10-06 06:14:46.836 194-194/? D/audio_vloop: in_get_channels(): 1017
10-06 06:14:46.836 194-194/? D/audio_vloop: in_get_channels: 0x00000001
10-06 06:14:46.836 194-194/? D/audio_vloop: in_get_buffer_size(): 1005
10-06 06:14:46.836 194-194/? D/audio_vloop: in_get_buffer_size: 1600
10-06 06:14:46.836 194-194/? D/audio_vloop: in_get_buffer_size(): 1005
10-06 06:14:46.836 194-194/? D/audio_vloop: in_get_buffer_size: 1600
10-06 06:14:46.836 194-194/? D/audio_vloop: in_get_format(): 1029
10-06 06:14:46.836 194-194/? D/audio_vloop: in_get_format: 0x00000001
10-06 06:14:46.836 194-194/? D/audio_vloop: in_get_sample_rate(): 979
10-06 06:14:46.836 194-194/? D/audio_vloop: in_get_channels(): 1017
10-06 06:14:46.836 194-194/? D/audio_vloop: in_get_channels: 0x00000001
10-06 06:14:46.838 194-613/? I/AudioFlinger: AudioFlinger's thread 0xb3bc0000 ready to run
10-06 06:14:46.838 194-194/? D/audio_vloop: adev_close_input_stream(): 1570
10-06 06:14:46.839 194-194/? W/APM::AudioPolicyManager: Input device 00020000 unreachable
10-06 06:14:46.839 194-611/? D/audio_vloop: looper_thread(): 216: Entered
10-06 06:14:46.839 194-611/? D/audio_vloop: looper_thread(): 366: Exiting
10-06 06:15:07.137 616-616/? I/InputManager: Initializing input manager, mUseDevInputEventForAudioJack=false
10-06 06:15:10.155 616-616/? I/SystemServer: Audio Service
10-06 06:15:10.222 194-607/? E/AudioFlinger: no wake lock to update!
10-06 06:15:10.222 194-608/? E/AudioFlinger: no wake lock to update!
10-06 06:15:10.224 194-614/? D/audio_hw_primary: adev_set_mic_mute: state 0
10-06 06:15:10.224 194-614/? D/audio_vloop: adev_set_mic_mute: 0
10-06 06:15:14.061 194-614/? D/audio_hw_primary: adev_set_parameters: enter: A2dpSuspended=false
10-06 06:15:14.061 194-614/? D/audio_vloop: adev_set_parameters(): [A2dpSuspended=false]
10-06 06:15:14.084 194-194/? I/AudioFlinger: systemReady
10-06 06:15:16.308 194-194/? D/audio_hw_primary: adev_set_mic_mute: state 0
10-06 06:15:16.308 194-194/? D/audio_vloop: adev_set_mic_mute: 0
10-06 06:15:17.072 194-194/? D/audio_hw_primary: adev_set_parameters: enter: A2dpSuspended=false
10-06 06:15:17.072 194-194/? D/audio_vloop: adev_set_parameters(): [A2dpSuspended=false]
10-06 06:15:25.023 733-733/? W/AudioTrack: AUDIO_OUTPUT_FLAG_FAST denied by client; transfer 4, track 44100 Hz, output 48000 Hz
10-06 06:15:25.032 194-607/? D/audio_hw_primary: out_set_parameters: enter: usecase(1: low-latency-playback) kvpairs: routing=2
10-06 06:15:25.043 194-607/? D/audio_hw_primary: select_devices: out_snd_device(2: speaker) in_snd_device(0: none)
10-06 06:15:25.043 194-607/? D/audio_hw_primary: enable_snd_device: snd_device(2: speaker)
10-06 06:15:25.050 194-607/? D/audio_hw_primary: enable_audio_route: apply and update mixer path: low-latency-playback
10-06 06:15:26.431 1150-1298/? I/MicrophoneInputStream: mic_starting com.google.android.apps.gsa.speech.audio.ag@c6eb0e1
10-06 06:15:26.443 194-1585/? I/AudioFlinger: AudioFlinger's thread 0xb3bc0000 ready to run
10-06 06:15:26.447 1150-1298/? I/MicrophoneInputStream: mic_started com.google.android.apps.gsa.speech.audio.ag@c6eb0e1
10-06 06:15:26.457 194-1585/? D/audio_hw_primary: select_devices: out_snd_device(0: none) in_snd_device(38: voice-rec-mic)
10-06 06:15:26.457 194-1585/? D/audio_hw_primary: enable_snd_device: snd_device(38: voice-rec-mic)
10-06 06:15:26.460 194-1585/? D/audio_hw_primary: enable_audio_route: apply and update mixer path: audio-record
10-06 06:15:26.942 1150-1271/? I/AudioController: internalShutdown
10-06 06:15:26.943 1150-1271/? I/MicrophoneInputStream: mic_close com.google.android.apps.gsa.speech.audio.ag@c6eb0e1
10-06 06:15:26.943 1150-1298/? E/AudioRecord-JNI: Error -4 during AudioRecord native read
10-06 06:15:26.986 194-1585/? D/audio_hw_primary: disable_audio_route: reset and update mixer path: audio-record
10-06 06:15:26.987 194-1585/? D/audio_hw_primary: disable_snd_device: snd_device(38: voice-rec-mic)
10-06 06:15:27.066 194-607/? D/audio_hw_primary: out_set_parameters: enter: usecase(1: low-latency-playback) kvpairs: routing=2
10-06 06:15:27.100 194-607/? D/AudioFlinger: mixer(0xb4140000) throttle end: throttle time(7)
10-06 06:15:30.257 194-607/? D/audio_hw_primary: disable_audio_route: reset and update mixer path: low-latency-playback
10-06 06:15:30.257 194-607/? D/audio_hw_primary: disable_snd_device: snd_device(2: speaker)
10-06 06:15:30.262 194-607/? D/audio_hw_primary: out_set_parameters: enter: usecase(1: low-latency-playback) kvpairs: routing=2
10-06 06:15:30.272 194-607/? D/audio_hw_primary: select_devices: out_snd_device(2: speaker) in_snd_device(0: none)
10-06 06:15:30.273 194-607/? D/audio_hw_primary: enable_snd_device: snd_device(2: speaker)
10-06 06:15:30.280 194-607/? D/audio_hw_primary: enable_audio_route: apply and update mixer path: low-latency-playback
10-06 06:15:30.347 194-607/? D/AudioFlinger: mixer(0xb4140000) throttle end: throttle time(10)
10-06 06:15:31.517 194-607/? D/audio_hw_primary: out_set_parameters: enter: usecase(1: low-latency-playback) kvpairs: routing=2
10-06 06:15:31.751 1150-1298/? I/MicrophoneInputStream: mic_starting com.google.android.apps.gsa.speech.audio.ag@dd13203
10-06 06:15:31.762 194-1826/? I/AudioFlinger: AudioFlinger's thread 0xb3bc0000 ready to run
10-06 06:15:31.771 1150-1298/? I/MicrophoneInputStream: mic_started com.google.android.apps.gsa.speech.audio.ag@dd13203
10-06 06:15:31.780 194-1826/? D/audio_hw_primary: select_devices: out_snd_device(0: none) in_snd_device(38: voice-rec-mic)
10-06 06:15:31.780 194-1826/? D/audio_hw_primary: enable_snd_device: snd_device(38: voice-rec-mic)
10-06 06:15:31.783 194-1826/? D/audio_hw_primary: enable_audio_route: apply and update mixer path: audio-record
10-06 06:15:34.695 194-607/? D/audio_hw_primary: disable_audio_route: reset and update mixer path: low-latency-playback
10-06 06:15:34.695 194-607/? D/audio_hw_primary: disable_snd_device: snd_device(2: speaker)
10-06 06:15:34.850 1150-1271/? I/AudioController: internalShutdown
10-06 06:15:34.851 1150-1271/? I/MicrophoneInputStream: mic_close com.google.android.apps.gsa.speech.audio.ag@dd13203
10-06 06:15:34.851 1150-1298/? E/AudioRecord-JNI: Error -4 during AudioRecord native read
10-06 06:15:34.885 194-1826/? D/audio_hw_primary: disable_audio_route: reset and update mixer path: audio-record
10-06 06:15:34.885 194-1826/? D/audio_hw_primary: disable_snd_device: snd_device(38: voice-rec-mic)

我的是 audio_vloop。我可以看到 Android 打开我的设备，然后打开输入流，然后关闭输入流。它从不尝试打开输出流。 audio_vloop 实现了输入和输出流。此后，audio_vloop 中的任何内容都不会被 Android 调用。

我制作了一个播放音频的小应用程序（目前来自 pcm 文件）。我想将此输出重定向到我的 HAL。为了实现这一点，我相信我需要在我的音轨上做一个 AudioTrack.setPreferredDevice() 。我发现音频管理器应该有所有音频设备的列表。

所以我打电话给：

AudioDeviceInfo aDevInfo[] = am.getDevices(AudioManager.GET_DEVICES_OUTPUTS);

它只找到 1 个设备，有关此设备的更多信息：

10-06 06:37:01.962 3295-3663/? D/AudioPlayer: Have [1] devices
10-06 06:37:01.964 3295-3663/? D/AudioPlayer: devInfo[0]: [Landroid.media.AudioDeviceInfo;@90bd9da
10-06 06:37:01.965 3295-3663/? D/AudioPlayer: getProductName()AOSP on Flo
10-06 06:37:01.965 3295-3663/? D/AudioPlayer: getType()2
10-06 06:37:01.966 3295-3663/? D/AudioPlayer: isSink()true
10-06 06:37:01.966 3295-3663/? D/AudioPlayer: isSource()false

这貌似是我没有实现的audioPort。所以它不是来自我的 HAL。

在 Android 允许我的应用程序与我的设备对话之前，我显然错过了一个或多个步骤。

我需要能够将音频从我的应用程序发送到我的 HAL。稍后，我还需要能够从我的 HAL 接收音频（通过 AudioRecord 等）。

在将我的 HAL 集成到 Android 中时我遗漏了什么？我需要实现音频端口吗？还需要其他东西吗？

更新 2

我发现 AOSP 中有错别字，在 AudioPolicyManager.cpp@2922

而不是 Output，它打印 Input

我有这个日志 AudioPolicyManager: Input device 00020000 unreachable，假设它在谈论 BT/A2DP 输入设备，我忽略了它。

我修复了我设备的日志，原来是我们要使用的线路输出设备。我现在正在调试这个方向

Answer 1

找到答案。

对于输出流：

不支持 AUDIO_CHANNEL_OUT_STEREO 的输出未被 Android
提及标志 AUDIO_OUTPUT_FLAG_DIRECT 的输出不会被 Android 自动打开。这在 AudioPolicyManager.cpp
中的以下代码中很明显
```
    if ((outProfile->mFlags & AUDIO_OUTPUT_FLAG_DIRECT) != 0) {
        continue;
    }
```
来自 here

可能有一种以编程方式打开它们的方法，但我还没有找到答案。

这两个一旦修复就足以让 Android 开始使用我的输出流。

在流中：

流中已经是 AudioManager.getDevices()

结果的一部分

因此可以在 AudioTrack.setPreferredDevice() 之后从流中的 vloop 中读取。

为了确保其他应用程序能够从 vloop 读取麦克风输入，我不得不声明它来实现 AUDIO_DEVICE_IN_BUILTIN_MIC。为此，我还在 audio_policy.conf.

中从主 HAL 中删除了 AUDIO_DEVICE_IN_BUILTIN_MIC

此外，我在流立体声中制作只是为了保持与流缓冲区格式的兼容性。

经过这些改动，我看到vloop不断有读写调用。

更新：

我后来发现上述行为取决于音频策略管理器的实现。他们中的大多数行为相同，（例如，大多数打开 INBUILT_MIC 用于 VOICE_RECOGNITION 输入）但有些可能不是（Nexus Player）对于这些异常值，要么实施他们的 APM 打开的内容，要么修改 APM 以打开你的HAL 实现.

使 Android 使用新的音频 HAL 需要什么

What is required to make Android use a new Audio HAL

android

audio

hal