预测出来的快歌 感觉软绵绵的,没有力度,响度还原了也不行
NeilGY opened this issue · comments
Can you tell me more details? I still can't figure it out.
If you want to talk about the wrong voice energy, I think you should review your dataset, because if your dataset only has 1 voice timbre or voice color, it may be difficult for so-vits-svc 5.0 to have it. You can have as much energy as you want