Comment by wczekalski

have you tested longer utterances with both ElevenLabs and with StyleTTS? Short audio synthesis is a ~solved problem in the TTS world but things start falling apart once you want to do something like create an audiobook with text to speech.