Addition to Prefix Padding in Audio Capture

[Zakonildo]

Hello, I'm testing your solution to build a voice agent. However, I noticed that in LiveKit, there are two parameters for SileroVAD: prefix_padding_duration and max_buffered_speech. These parameters help improve transcription accuracy, as VAD may cut off parts of words and even cause model hallucinations. So, I was wondering if there are any plans to implement these parameters in your solution, as they are available in LiveKit.

LiveKit Silero VAD Documentation for reference: https://docs.livekit.io/agents/build/turns/vad/

[freddyaboulton]

Hi @Zakonildo - I will look into this!