Discussions

Dear DID Support Team,

We are currently integrating ElevenLabs voices into our application using the D-ID API for real-time Talk Streams. While everything else is working as expected, we are experiencing a significant issue with Tamil language output.

The voice generated by ElevenLabs for Tamil is unclear and often difficult to understand. The pronunciation lacks natural fluency and accuracy, which leads to poor user experience for our Tamil-speaking audience.

Issue Details:

API Used: D-ID Talk Stream (with ElevenLabs voice provider)

Language: Tamil (ta-IN)

Problem: Poor pronunciation and clarity; voice is not intelligible in normal conversational usage

Expected Behavior: Clear and fluent Tamil speech with accurate pronunciation

Impact: Tamil-speaking users are unable to effectively engage with the voice responses due to lack of clarity

We would like to know:

Whether Tamil is officially supported with high-quality voice synthesis on ElevenLabs

If any specific voice models are optimized for Tamil

If there are configuration tips or settings to improve the quality

When we can expect better Tamil language support if it is currently in beta or experimental stage

Please let us know how to proceed or if there's anything else required from our end to assist with diagnosing the issue.