Hacker News new | ask | show | jobs
by xyzhg 493 days ago
You can't lose word boundaries and phonemes don't tell you which part of the word is emphasized.
1 comments
popalchemist 493 days ago
Modern TTS engines use tokenizers to convert words to phonemes. See: https://github.com/FunAudioLLM/CosyVoice/issues/202
link