dear author, thanks for your contribution,it is a useful tool, there are several questions ,could you help ? I will appreciate it if you can reply
1、which boundaries_id and labels_id have best effect on Chinese songs and English songs?
2、when use different labels_id , different label ids are given, could you help to reply what label id represent?(verse, chorus and etc)
I haven't done an experiment with Chinese songs, but my guess is that the best combination of algorithms is still sf (structural features) for boundaries and fmc2d (fourier magnitude coefficients 2D) for labels.
Unfortunately, MSAF does not support classification of label identifiers into semantic labels such as verse, chorus, etc.
dear author, thanks for your contribution,it is a useful tool, there are several questions ,could you help ? I will appreciate it if you can reply 1、which boundaries_id and labels_id have best effect on Chinese songs and English songs? 2、when use different labels_id , different label ids are given, could you help to reply what label id represent?(verse, chorus and etc)