About

CloneVoiceOnline uses viXTTS model. viXTTS is a text-to-speech voice generation tool that offers voice cloning voices in Vietnamese and other languages. This model is a fine-tuned version based on the XTTS-v2.0.3 model, utilizing the viVoice dataset. This repository is primarily intended for inference purposes.

The model can be accessed at: viXTTS on Hugging Face

Features

• Multi-lingual: Supports spech generation in 17 languages.
• Cross-language voice cloning: Can use a voice in one language to generate speech in another language.

Limitations

• Not Perfect: Works pretty well but not perfect. Generated audio may have some artifacts.
• Needs good input audio quality: Requires a good reference audio for voice cloning. The better the reference audio, the better the generated audio.
• Subpar performance for input sentences under 10 words in Vietnamese language (yielding inconsistent output and odd trailing sounds).
• This model is only fine-tuned in Vietnamese. The model's effectiveness with languages other than Vietnamese hasn't been tested, potentially reducing quality.

Usage Tips

• Only 1 Voice in Reference Speaker File: The reference speaker file should only contain 1 voice. If it contains multiple voices, the generated audio will sound bad.
• Clear Speech: The reference speaker should talk clearly.