Description

Before Start Remember, this is a model based on just 9 seconds of footage from a 2005 archive. It's understandable if the results are not perfect. - Additionally, the sample was thoroughly cleaned to eliminate as much noise as possible, including background conversations. - By the way, if you want to make voice previews with this model yourself, I think you'll have to use audio editing software to add ambient noise (because that's what makes the renderings so strange if you decide not to use it). Last Update : <t:1703614454:R> - Model URL Image URL Edited Version RVC V2.0 - Pitch Extraction Algorithm RMVPE - Epochs - Steps 1k- 3k - Dataset ~ 00:00:09 ( 2005 - Me at the Zoo - Full Background Noise Cleaned ) - Recommended Usage Speech - Search Feature Ratio 0.0 (0 is better to reduce the robotic effect, even if it's not entirely removed.) - Pitch Logic Pitch ( -4 ~ 0 = Man 12 = Women ) - You can adjust if you found an better result Previews Preview_TTS.wav *Contains External Effects* : No - *Pitch* : 0 - *Feature Ratio* : 0,75 - Preview_Cover.wav *Contains External Effects* : Yes - *Pitch* 4 - *Feature Ratio* : 0