silk field
This is an experimental model of KLM VC, capable of inferring speech, singing, coughing, high-pitched tones, and whispering.
This model does not include a feature index file.
Test Note – As can be seen from the spectrogram, this model does not produce perfectly clean silence during muted sections. However, this type of noise pattern does not pose any issues during actual use. The model is highly robust against ambient and environmental noise, and is specifically trained to infer real environmental sounds instead of artifacts when such noise enters through the microphone above a certain decibel level.
That said, if you intend to use this model for covers or other applications that require post-processing to achieve a noise reduction effect, a proper cleanup process is necessary. The primary purpose of this model is to enable voice changer functionality in harsh acoustic environments, such as spaces filled with loud white noise or disruptive background sounds.
Model Link - https://huggingface.co/SeoulStreamingStation/RVC_Voice_Models/resolve/main/Voice_Jane_Exp.zip?download=true
sacred moth
thorn turtle