Piece of voice input has not been voiced
# supportm
Moody Reaper
07/04/2024, 7:50 AMCall log: https://dashboard.vapi.ai/calls/e8948b68-bd73-43a9-87f3-1ee9a4bd3923
Expected Result:
The whole phrase has been voiced:
> Please wait a moment while I check if the 12 PM slot is available for tomorrow. It might take 1-2 minutes.
Actual Result:
Only part of the phrase has been voiced:
> Please wait a moment while I check
https://cdn.discordapp.com/attachments/1258328950652796939/1258328950896332912/image.png?ex=6687a5a8&is=66865428&hm=0450fe96bd24f2c4df2cb083364862da28b3e767c6f0b2c9fa8b7f9a254bb598&
v
Vapi
07/04/2024, 7:51 AMmessage has been deleted
Vapi
07/04/2024, 7:51 AM@Moody Reaper
Thank you for bringing this to our attention. It seems there might have been an issue with the speech synthesis process. To assist you further, could you please provide more details? Specifically, I need to know if you would like to attempt to synthesize the phrase again or if there's another action you'd prefer to take regarding this call log.
Vapi
07/04/2024, 7:51 AMYou can continue this conversation by mentioning me in the message.
Vapi
07/04/2024, 7:51 AMGot it. Tagging @User for additional help.
s
Shubham Bajaj
07/04/2024, 11:04 AMTTS was not ready because of slow llm response and only voiced out the partial transcript.
m
Moody Reaper
07/04/2024, 3:39 PM@Shubham Bajaj, what exactly does it mean?
We've sent the whole piece as a single item ("model outpit" >> "voice input").
Which "slow LLM response" do you mean?
s
Shubham Bajaj
07/04/2024, 9:27 PMBy slow llm response: AI LLM is slow to respond, and the Avg latency between LLM output tokens is 5107 ms
Shubham Bajaj
07/04/2024, 9:29 PMTry changing voice provider, this issue has been reported repeatedly.
Shubham Bajaj
07/04/2024, 9:30 PMJTC: By this you didn't meant sending audio buffer?
> "We've sent the whole piece as a single item ("model outpit" >> "voice input")."