issue with voice being cut off even though prompt ...
# support
k
@User I am having a problem with 11labs voices cutting off our sentence and going to silence. Here is a call for reference: d1d990d5-cd1c-45a2-97ae-c4b13a11a18e
v
@kbmaine I'm sorry to hear that you're experiencing issues with the 11labs voices cutting off sentences. To assist you better, could you please provide more details about the problem? For instance, are there specific conditions under which this occurs, such as a particular word or phrase, or is it happening randomly? Additionally, could you let me know if there have been any recent changes to your system or the way you're using the VAPI platform that might correlate with the onset of this issue?
You can continue this conversation by mentioning me in the message.
k
@Vapi we have a sentnce that says The word team keeps getting cut off and the bot is saying TEE and then going silent. Here are some calls logs where it is happening: 22fa0e1a-6e9b-4792-ac60-baf9be5ad107
v
@kbmaine Thank you for providing the call log reference. It seems like there might be an issue with the speech synthesis process for the word "team." To address this, I would recommend checking the following: 1. **Transcript Review**: Ensure that the transcript being sent to the voice API is correct and includes the full word "team." 2. **Audio Configuration**: Verify that the audio settings, such as bitrate and encoding, are correctly configured to support clear voice synthesis. 3. **API Parameters**: Check if there are any specific parameters or settings in the API call that might be causing the voice to cut off prematurely. If everything seems to be in order on your end, it might be necessary to investigate this issue further on the server side. I can report this to our technical team for a more in-depth analysis. Would you like me to proceed with that?
k
yes please
I am not an engineer so I need further explanation... we have 15 different assistants with different voices from eachother and they are all doing it. Can I do an audio configuration in bulk across all assistants? @User
s
@kbmaine could you also share: - What response you expected to get - What response you actually got instead This would really help me figure out what went wrong!
k
@Shubham Bajaj The bot is supposed to be saying "the information has been sent to the TEAM" and then it is supposed to ask the caller if there is anything else they can help with. But instead it is saying "the information has been sent to the TEE" and going silent. There is no follow up question happening unless the caller speaks again.
@Shubham Bajaj in the background the callers details are being emailed with a 'sendEmail' function/tool. Not sure if that is helpful to what may be happening
s
🔵 16:44:23:229 Got Tool Messages "request-complete". Messages: [ { "type": "request-complete", "content": "The emails was sent to the tea", "contents": [], "conditions": [], "endCallAfterSpokenEnabled": false, "role": "assistant" } ] The bot said "the information has been sent to the TEE" because of your request-complete message and then it goes silent because of it's turn was done and bot was waiting for user to speak. I suggest giving a try openAI Model, and then you can observer the difference by yourself.
Because of no response closing this ticket.
4 Views