Seems there is shift between deepfake and real singing #4

Thanks for your great work ! However, it appears that the same audio was used to construct a deepfake audio using the SVC method. The style predictions generated by Stars for this audio differ significantly from the style of the original audio.