Chimera Audio Capture
Capture audio from microphone with echo cancellation
Chimera Audio Capture
The ChimeraAudioCaptureComponent captures audio from the microphone and provides WebRTC-based audio processing including echo cancellation, noise suppression, and automatic gain control.
Properties
| Property | Type | Description |
|---|---|---|
bEnableEchoCancellation | bool | Enable WebRTC AEC3 echo cancellation |
bEnableNoiseSuppression | bool | Enable noise suppression |
bEnableGainControl | bool | Enable automatic gain control |
AecWarmUpTimeMs | int32 | Warm-up delay before transmitting (100ms default) |
SampleRate | int32 | Audio sample rate (default: 48000) |
Echo Cancellation
The component uses WebRTC's AEC3 algorithm to remove speaker echo from the microphone signal. This enables full-duplex conversations where the AI can speak while the user is also speaking.
How AEC Works
- Render Audio: Audio played through speakers (AI voice) is fed to the AEC
- Capture Audio: Microphone audio is captured
- Processing: AEC removes the speaker audio from the mic signal
- Clean Output: Echo-free audio is sent to the server
Feeding Render Audio
For AEC to work, you must feed the received audio (AI voice) to the audio capture component:
Methods
StartCapture
Begin capturing audio from the microphone.
StopCapture
Stop audio capture.
FeedRenderAudioFloat
Feed speaker audio to the AEC for echo cancellation.
Events
OnAudioCaptured
Fired when audio is captured and processed.
Usage with Publisher
The AudioCaptureComponent automatically connects to the Publisher if bAutoConnectMicCapture is enabled on the Publisher:
Blueprint Example
- Add both
Chimera PublisherandChimera Audio Captureto your Actor - In Audio Capture details, enable Echo Cancellation
- Bind
OnAudioDataReceivedfrom Publisher toFeedRenderAudioFloaton Audio Capture - Capture starts automatically when Publisher connects