Description

Before Start I didn't extract everything from the vod, so there may be an extended version, with more datasets. - If this "new" version is created, I'd really like to clean up the dataset even more, because even with 3-4 hours of editing for 18 minutes, I still find the result imperfect. - It's important to note that Nyatasha's voice is very different when she sings, and that, for example, unlike Gawr Gura, it's harder to distinguish between the two (although I think someone who's a fan or used to listening to her voice [ which I'm not ] would find it easier). Last Update : <t:1703252981:R> - Model URL Version RVC V2.0 - Pitch Extraction Algorithm RMVPE - Epochs - Steps 400 - 17.6 K - Dataset ~ 00:18:40 ( Nyanners Sings Her Heart Out For Charity ) - Mastering, remove breath, suppress the moments when she's in fits of giggles ... - Recommended Usage Cover - Search Feature Ratio 0.75 - Pitch Logic Pitch ( 12 = Man / 0 = Women ) - You can adjust if you found an better result - I'm not sure if I used the right pitch for the previews, but try watching the vod available on Youtube (using the title of the video I entered in the dataset section) and compare your result with this one. Previews Preview_Cover_1.wav *Contains External Effects* : Yes - *Pitch* : +0 - *Feature Ratio* : 0.75 - Preview_Cover_2.wav *Contains External Effects* : Yes - *Pitch* : +6 - *Feature Ratio* : 0.75