The error "Ref Audio Path Required" indicates that you are trying to use a voice cloning feature (like GPT-SoVITS, F5-TTS, or CosyVoice) but have not provided the required reference audio file.
To fix this, follow these steps:
- Select a Clone Voice Role: In the main interface, under the "Voiceover (TTS)" section, choose a voice role that is marked as "Clone" (e.g., "GPT-SoVITS-Clone", "F5-TTS-Clone").
- Provide Reference Audio: After selecting a clone role, a new button or field will appear (usually labeled "Select Reference Audio" or similar). Click it and select a clean audio file (e.g., WAV, MP3) containing the voice you want to clone. This file should be a clear recording of a single person speaking.
- Process Again: After setting the reference audio, start the translation process again.
Important Notes:
- If you are not trying to use voice cloning, ensure you have selected a standard TTS role (like "Edge-TTS", "VITS-cnen", or "Azure-TTS") instead of a clone role.
- The reference audio file path must not contain special characters or non-English letters. Use a simple path like
D:/ref_audio.wav.
Please refer to the documentation for more details: