Description

Usual Info: ================================= Dataset is 45 minutes. It's voice lines and dialogue from throughout the series. Speech was tested using Elevenlabs. Shouldn't matter what model you use. Transpose range for speech will probably be -12. Test with 0 first. Songs will depend. Model can sound too deep for original instrumentals, transposing up the vocals might help in that case. RVC Setting: ================================= * Search Feature Ratio: 0-0.5 (For singing) & 0.5 - 1 (For speech) * Volume Envelope: 0.5 * Protect Voiceless Consonants: 0.5 (Disable) * Resample Audio - 0 or 48000 (Only if you need to, but it isn't necessary.) Model