Sarvam TTS - error decoding WAV audio

Agents
1

I’m using Sarvam for STT and TTS and I got an error that says "livekit.agents error decoding WAV audio ", seems like Sarvam TTS is giving outputs in mp3. How to fix this. I’m using v1.5.1.

I saw the livekit codebase and see its supporting mp3 too but not sure whay I’m facing this issue:
ALLOWED_OUTPUT_AUDIO_CODECS: set[str] = {

"mp3",

“opus”,

“flac”,

“aac”,

“wav”,

}
**My code:
**
stt=sarvam.STT(model=config.stt_model, language=config.stt_language, mode=“transcribe”, flush_signal=True ),

    llm=inference.LLM(

        model=config.gemini_llm_model,

    ),

    tts=sarvam.TTS(

            target_language_code=config.tts_language,

            model=config.tts_model,

            speaker=config.tts_voice,

            pace=1.0,

            temperature=0.61,

            output_audio_codec='mp3'

    ),
2

I believe this was fixed in fix: allow codec format specification via the user for Sarvam TTS by pUrGe12 · Pull Request #5209 · livekit/agents · GitHub which I would anticipate being in the next release (though it’s not present in 1.5.1)

3

Hey @darryncampbell ,

Can you please share when can we expect the latest release?

4

Hey @darryncampbell , Would you happen to have an estimated timeline for when this fix might be released, and which version it’s expected to be included in?

5

This fix will be in the next release, looks like it just narrowly missed inclusion in 1.5.1. I don’t have a specific date I can share I’m afraid as it looks like the team are trying to align with a partner release, but I hope soon.

6

Thanks a lot for the response @darryncampbell . Hopefully we will release this with in a week or so as many people are waiting for this change.

1 Like
7

Hey @darryncampbell Could you please share any updates on the issue above? I would appreciate your update. Thanks

8

I don’t have an exact date when the next release will be but it should be soon. I’ll put in a request.

9

Apologies for the delay, but Agents 1.5.2 was just released: