Skip to content

Latest commit

 

History

History

Advanced-Audio-Driver-Kotlin

README.md

Advanced Audio Driver

This app shows how to use the audio driver for publisher and subscriber audio. Just like the custom video driver, most applications will be fine using the default. If you want to add custom audio manipulation, look here.

AdvancedAudioDevice instance is passed into the AudioDeviceManager.setAudioDevice method:

AdvancedAudioDevice AdvancedAudioDevice = new AdvancedAudioDevice(this);
AudioDeviceManager.setAudioDevice(AdvancedAudioDevice);

The AdvancedAudioDevice class extends the BaseAudioDevice class, defined in the OpenTok Android SDK. This class includes methods for setting up and using a custom audio driver. The audio driver contains an audio capturer -- used to get audio samples from the audio source -- and an audio renderer -- used to playback audio samples from the OpenTok streams the client has subscribed to.

Note: that you must call the method AudioDeviceManager.setAudioDevice() before you instantiate a Session object (and connect to the session).

The constructor for the AdvancedAudioDevice class instantiates two instances of the BaseAudioDevice.AudioSettings class, defined in the OpenTok Android SDK. These are settings for audio capturing and audio rendering:

captureSettings = new AudioSettings(captureSamplingRate, NUM_CHANNELS_CAPTURING);
rendererSettings = new AudioSettings(outputSamplingRate, NUM_CHANNELS_RENDERING);

The AdvancedAudioDevice class overrides the initCapturer method, defined in the BaseAudioDevice class. This method initializes the app's audio capturer, instantiating a an andriod.media.AudioRecord instance to be used to capture audio from the device's audio input hardware:

int channelConfig = NUM_CHANNELS_CAPTURING == 1 ? AudioFormat.CHANNEL_IN_MONO : AudioFormat.CHANNEL_IN_STEREO;

audioRecord = new AudioRecord(
        AudioSource.VOICE_COMMUNICATION,
        captureSettings.getSampleRate(),
        channelConfig,
        AudioFormat.ENCODING_PCM_16BIT, recBufSize);

The initCapturer method also sets up a thread to capture audio from the device:

new Thread(captureThread).start();

The AdvancedAudioDevice overrides the startCapturer method, which is called when the app starts sampling audio to be sent to the publisher's stream. The audio capture thread reads audio samples from the AudioRecord object into a buffer, m_recbuffer:

int lengthInBytes = (samplesToRec << 1) * NUM_CHANNELS_CAPTURING;
int readBytes = audioRecord.read(tempBufRec, 0, lengthInBytes);

recBuffer.rewind();
recBuffer.put(tempBufRec);

samplesRead = (readBytes >> 1) / NUM_CHANNELS_CAPTURING;

The getAudioBus method, defined in the BaseAudioDevice class, returns a BaseAudioDevice.AudioBus object, also defined in the OpenTok Android SDK. This audio bus object includes a writeCaptureData method, which you call to send audio samples to be used as audio data for the publisher's stream:

getAudioBus().writeCaptureData(recBuffer, samplesRead);

The AdvancedAudioDevice class overrides the initRenderer method, defined in the BaseAudioDevice class. This method initializes the app's audio renderer, instantiating an andriod.media.AudioTrack instance. This object will be used to playback audio to the device's audio output hardware:

int channelConfig = (NUM_CHANNELS_RENDERING == 1) ? AudioFormat.CHANNEL_OUT_MONO : AudioFormat.CHANNEL_OUT_STEREO;

audioTrack = new AudioTrack(
        AudioManager.STREAM_VOICE_CALL,
        rendererSettings.getSampleRate(),
        channelConfig,
        AudioFormat.ENCODING_PCM_16BIT,
        minPlayBufSize >= 6000 ? minPlayBufSize : minPlayBufSize * 2,
        AudioTrack.MODE_STREAM
);

The initRenderer method also sets up a thread to playback audio to the device's audio output hardware:

new Thread(m_renderThread).start();

The AdvancedAudioDevice overrides the startRenderer method, which is called when the app starts receiving audio from subscribed streams.

The AudioBus object includes a readRenderData method, which the audio render thread calls to read audio samples from the subscribed streams into a playback buffer:

int samplesRead = getAudioBus().readRenderData(m_playBuffer, samplesToPlay);

Sample data is written from the playback buffer to the audio track:

int bytesRead = (samplesRead << 1) * NUM_CHANNELS_RENDERING;
playBuffer.get(tempBufPlay, 0, bytesRead);

int bytesWritten = audioTrack.write(tempBufPlay, 0, bytesRead);

// increase by number of written samples
bufferedPlaySamples += (bytesWritten >> 1) / NUM_CHANNELS_RENDERING;

// decrease by number of played samples
int pos = audioTrack.getPlaybackHeadPosition();
if (pos < playPosition) {
        // wrap or reset by driver
        playPosition = 0;
}

bufferedPlaySamples -= (pos - playPosition);
playPosition = pos;

Further Reading