https://vapi.ai/ logo
Join Discord
Powered by
11Labs Slurring Voice (Urgent)
# support
n
Call ID: 
4be01233-b2d4-498f-81e6-60ddffc90a96
At the end of this call, 11labs starts slurring speech it sounds like. Not sure how we can adjust this.
v
Thank you for reporting the slurring issue with 11Labs. To help diagnose the problem, could you let me know if you have recently adjusted the voice speed, formatting settings, or if the slurring only happens after long responses or at specific points in the call? Sources: - [Vapi ElevenLabs Integration Documentation](https://docs.vapi.ai/providers/voice/elevenlabs) - [Vapi Voice Formatting Plan](https://docs.vapi.ai/assistants/voice-formatting-plan) - [Vapi Changelog: ElevenLabs Speed Control](https://docs.vapi.ai/changelog/2025/3/2)
To address the slurring issue with ElevenLabs voices, try fine-tuning the TTS parameters—especially speed and buffer settings—to improve clarity. For example, you can adjust the speech parameters in your integration:
Copy code
json
{
  "voice": "elevenlabs",
  "parameters": {
    "speed": 1.0,
    "pitch": 1.0,
    "emphasis": 1.0
  }
}
Start by lowering the speed slightly (e.g., 
speed: 0.95
) and test if the slurring improves. If the issue persists, review your buffer optimization and network latency settings as these can also impact real-time audio quality ([source](https://docs.vapi.ai/providers/voice/elevenlabs)). Source: - [How Vapi Integrates Text-to-Speech Platforms: ElevenLabs](https://docs.vapi.ai/providers/voice/elevenlabs)
n
I don't think those properties actually exist
s
@Noah Johnson are you able to identify the elevenlabs setting you have assigned to the agent?
n
@SlaviSavanovic
Copy code
{
  "provider": "11labs",
  "voiceId": "OYTbf65OHHFELVut7v2H",
  "model": "eleven_flash_v2_5",
  "speed": 0.94,
  "stability": 0.6,
  "similarityBoost": 0.78
}
s
Remove the speed setting. It doesn’t work well and screws up playback Use turbo instead of flash, same cost but much better quality
@Noah Johnson
n
@SlaviSavanovic Turbo is a little slower than flash, right?
s
You will not notice a difference in real time generation. Time to first byte is maybe 20-40 ms slower. Audio improvement is significantly better
n
@SlaviSavanovic Thank you. We made that switch.
s
Sweet LMK if that helps
c
Hello, We sincerely appreciate your understanding and apologize for the inconvenience this has caused. We’d like to check in to see if you need any further assistance. We’re sorry for the inconvenience and are more than happy to help if there’s anything else you need.
2 Views
PreviousNext
Powered by