🪙Link: https://archive.org/download/aquestalk1-f1-rvc-lc1.6/Yukkuri-AquesTalk1-Female1.zip
- Dataset: 19:39 (15:09 of normal pitch talk + 04:30 of high and low pitching talk for voice pitch variety)
- Saved by WAV export by AquesTalk Player (I used the built-in pitch changer, no audio app edit)
- Noise Reduced with UVR BSRoformer Denoise
🍎 Applio Kaggle
- Post-Normalized Dataset
- Pretrained with LegacyCore v1.6 (HifiGan & Contentvec)
- RVMPE/64 Hop-Length/32k Target Sample/Batch Size 8