Include cartesia's voice controls on docs + update synthesizer

[sauhardjain]

• Fix end of sentence issues
  • Send ctx.no_more_inputs() after a 1s timeout on the last ctx.send()
  • Pad out the last chunk to make it uniform
• Fix the emotion and speed settings
• Add docs for the Cartesia synthesizer

[ajar98]

let me know when lint passes as well!

[sauhardjain]

also getting lint errors from vocode/streaming/vector_db/pinecone.py

[cyrilS-dev]

This pull request is causing a problem; I have no audio output from my synthesizer, even though my configuration remains unchanged :

synthesizer=CartesiaSynthesizer(
            synthesizer_config=CartesiaSynthesizerConfig.from_output_device(
                output_device=output_device,
                model_id = "sonic-multilingual",
                voice_id = "a249eaff-1e96-4d2c-b23b-12efa4f66f41"
            )
        )

[sauhardjain]

@cyrilS-dev happy to get this resolved! What cartesia version are you using and do you have any logs associated with the issue here?

[cyrilS-dev]

thanks for your reply @sauhardjain

pip show cartesia returns Version: 1.0.11

I don't see any errors in the logs, and everything appears normal, but there is no audio output :

2024-08-13 14:19:11.602 | DEBUG    | vocode.streaming.streaming_conversation:process:481 - Synthesizing speech for message
2024-08-13 14:19:11.604 | WARNING  | vocode.streaming.synthesizer.audio_cache:safe_create:23 - Redis ping failed on startup, disabling audio cache
2024-08-13 14:19:11.748 | DEBUG    | vocode.streaming.streaming_conversation:send_speech_to_output:929 - Start sending speech Bonjour comment allez vous? to output
2024-08-13 14:19:11.749 | DEBUG    | vocode.streaming.streaming_conversation:process:429 - Sending end of turn
2024-08-13 14:19:11.880 | INFO     | vocode.streaming.livekit.livekit_conversation:_on_track_subscribed:39 - track subscribed
2024-08-13 14:19:12.497 | DEBUG    | vocode.streaming.streaming_conversation:send_speech_to_output:963 - Finished sending chunks to the output device

As additional information, i'm using
output_device = LiveKitOutputDevice()

Sample rate is 48000 and audio encoding is LINEAR16

[sauhardjain]

@cyrilS-dev my speaker output also uses sample rate 48000 and audio encoding LINEAR16 so not sure what's happening here! let's connect on the vocode discord to debug! my user is sauahrd