Skip to content

feat: feed device-captured audio to senders through its track - #570

Merged
devopvoid merged 1 commit into
mainfrom
feat/device-audio-source
Nov 8, 2026
Merged

devopvoid merged 1 commit into
mainfrom
feat/device-audio-source

Conversation

@devopvoid

Copy link
Copy Markdown
Owner

createAudioSource() now returns a native DeviceAudioSource. The audio device module's proxy runs recorded audio through the factory's AudioProcessing and hands it to these sources' sinks, the senders among them, instead of to WebRTC's transport, which fed it into every send stream of the factory.

  • The factory builds its AudioProcessing itself, so recorded and playout audio reach the same instance and echo cancellation keeps working.
  • Sinks on a microphone track now receive the processed audio.
  • Media-source stats are unaffected: WebRTC computes the level in AudioSendStream::SendAudioData, which both paths reach.
  • The device/custom guard and WebRTC's control of when recording starts are unchanged; lifting the guard is the next step.

Adds a test-only ToneAudioDeviceModule that records a tone, and DeviceAudioSendTests.

createAudioSource() now returns a native DeviceAudioSource. The audio
device module's proxy runs the recorded audio through the factory's
AudioProcessing and hands it to these sources' sinks, the senders among
them, instead of to WebRTC's transport, which fed every send stream.
Local sinks on a microphone track now receive the processed audio.
@devopvoid
devopvoid added this pull request to stack #574 November 8, 2026 15:24
@devopvoid
devopvoid merged commit 41a2f94 into main Nov 8, 2026
41 of 42 checks passed
@devopvoid
devopvoid deleted the feat/device-audio-source branch November 8, 2026 15:26
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

1 participant