SVM to extract features

Hi, everyone.
Could you please tell me how to use SVM to extract audio features?

Thanks
best

It all depends on all you formalize you problem. But why dont you use svm from sklearn.