rhasspy / piper

A fast, local neural text to speech system
https://rhasspy.github.io/piper-samples/
MIT License
6.32k stars 465 forks source link

incorrect output for simple Chinese phrase? #243

Open MarkoSchuetz opened 1 year ago

MarkoSchuetz commented 1 year ago

I'm running

echo "一点儿" | piper-tts --model ~/src/piper/medium/zh_CN-huayan-medium.onnx --output_file 一点儿.wav

and the output is incorrect.

https://github.com/rhasspy/piper/assets/944309/16be765c-0835-45f1-b6c4-9a078c4ad13a

This is what it should sound like. https://zh.forvo.com/search/%E4%B8%80%E7%82%B9%E5%84%BF/

MarkoSchuetz commented 1 year ago

I have now tried different ways of installing piper:

in all cases I get the same result.

Could it have something to do with phonemization? There are more syllables than there should be.