This message was deleted.

It seems likely that when using a custom TTS model with LiveKit, the audio encoding and decoding format should match the expected format of the LiveKit audio processing components. According to a Slack discussion, an issue was resolved by ensuring that the server sends audio data in a format compatible with

rtc.AudioFrame

, which expects data to be unpacked as

int16

. This suggests that using

int16

PCM format for audio data might be necessary for compatibility with LiveKit's audio processing. Would you like more details on how to implement this or any other aspect of using custom TTS with LiveKit? Sources: - Slack discussion on audio format 👍 0 button 👎 0 button 💡 Hint: Mention @runllm in the thread for followups.