When theere is noise in the background, the tts model waits for input instead of closing

Thanks for reaching out, Apratim. I'm an AI assistant and I'm here to help. Give me a moment while I research this topic.

Can you explain a bit more about you noise issue? Have you tried noise cancelling options, including ai-coustics with our dedicated VAD?

Thanks — I’m following up about the TTS session hanging when there’s background noise and your question about noise-cancelling and ai-coustics/VAD. Yes — LiveKit provides noise cancellation and a dedicated VAD you can enable to reduce background-noise-triggered waiting: • Use ai-coustics audio enhancement (example:

ai_coustics.audio_enhancement(model=ai_coustics.EnhancerModel.QUAIL_VF_L)

) to filter voice from background noise. • Enable noise cancellation in the agent/room options when starting the session; pick

BVC

for typical WebRTC audio or

BVCTelephony

for SIP/narrow-band audio. • LiveKit’s default VAD is Silero VAD; enabling VAD improves end-of-turn detection so the session can close when speech stops rather than waiting through noisy input. If you want, tell me: • Which TTS model/version and SDK (and version) you’re using, and whether participants connect via WebRTC or SIP. • Your current room/session options or code that configures noise cancellation, VAD, or timeouts. With those details I’ll suggest the exact room options and code changes (or a timeout fallback) to prevent the TTS session from waiting indefinitely. Sources: Turns overview | LiveKit Documentation | Using Noise Cancellation for SIP and WebRTC Participants | Google AI and LiveKit | LiveKit Documentation | Turn Detection for Voice Agents: VAD, Endpointing, and Model-Based Detectio... | LiveKit Platform | Build, run, and observe voice AI agents | LiveKit, with interactive elements