Dear developer, I just discovered a deficiency/defect in the software algorithm when testing the new version. Chinese traditional opera cannot be correctly identified. As the greatest separation software on earth, I think that by calculating and analyzing this kind of music, the separation purity can be greatly improved.
If you need more, you can give me feedback and I will continue to provide it through the network disk, or you can search by keywords (京剧,豫剧,戏曲唱段).
[XiQuCN202408.zip]
1drv.ms/u/s!AlKMNMAC8GizikB97SFyGqMux0_Z?e=d1cfd1
1 Like
I just downloaded your file and all the elements/components can easily be unmixed. The reason why Spectralayers might struggle with this is because the power(Amplitude) of the frequency (of your files) is mostly in the mid-tone range, however you can try different algorithms (outside of Spectralayers) to get different results.
I could easily unmix this(all instruments and vocals), however I’m not going to do it for free. If you want this unmix then directly private message me and we can discuss details further.
1 Like