The project demonstrates how to use Twilio's Programmable Voice SDK with audio playback and recording functionality provided by a custom TVOAudioDevice.
The example demonstrates the custom audio device ExampleAVAudioEngineDevice, which uses CoreAudio's VoiceProcessingIO audio unit to playback and record audio at up to 48KHz with built-in echo cancellation. The example uses two AVAudioEngine in manual rendering mode:
- The upstream AVAudioEngine is used for mixing the Remote Participant's audio with audio from a file. The AudioUnit receives mixed audio samples from AVAudioEngine's output node.
- The downstream AVAudioEngine is used for mixing the Local Participant's microphone audio with audio from a file. The Voice SDK receives the mixed audio samples from the AVAudioEngine's output node.
This diagram describes how ExampleAVAudioEngineDevice uses TwilioVoice, AVAudioEngine, and CoreAudio -
Please note, ExampleAVAudioEngineDevice requires iOS 11.0 or above.
See the master README for instructions on how to generate access tokens and make an outbound Call.
This example requires Xcode 11.0 and the iOS 12.0 SDK, as well as a device running iOS 11.0 or above.
Once you have configured your access token, build and run the example. Use the text field to make an outgoing call:
Once the Call is connected, tap the "Play Music" button and let ExampleAVAudioEngineDevice mix some tunes into your track.
The AVAudioSession is configured and activated at device initialization time. Ideally, it would be better to activate the AVAudioSession only when audio playback or recording is needed.