I want to use the librosa package to convert music in wav format into a 34-dimensional npy matrix based on 34 features for model training, but I don’t know how to restore the trained 34-dimensional npy matrix to wav format. Please tell me. Is there a way to do this?
I used librosa to extract thirty-four features, and used this thirty-four-dimensional matrix as a data set in the GAN network to transfer music styles. But I don’t know if we restore the thirty-four-dimensional feature matrix generated by the model to wav format. Our purpose is to use the generated music to do something.