Hi everyone,
I am reporting a consistent issue with the Speech Generation feature in Google AI Studio. Whenever I input text in Bangla (Bengali), the generation fails with the error message: “No audio generated.”
Details:
-
Location: Bangladesh
-
Language: Bangla (bn-BD)
-
Models Tested: Gemini 2.5-pro-preview-tts.
-
Voices Tested: I have tried all available voices (Puck, Charon, etc.), but none seem to work for this specific language.
The Issue: I am a digital content creator working on educational materials for children and language learners. While the models are excellent at writing Bangla text, the audio modality seems to fail immediately upon clicking generate. This happens regardless of the length of the text—even simple words like “নমস্কার” (Hello) or “ধন্যবাদ” (Thank you) result in the same error.
Questions:
-
Is Bangla speech generation currently supported in the AI Studio preview, or is this a regional restriction for users in Bangladesh?
-
Is there a specific model or configuration I should be using to enable Bangla TTS?
I’ve seen others mention similar issues with non-English languages, but I wanted to provide specific feedback for Bangla. Any guidance from the team or the community would be greatly appreciated!