Finding the genre of a song with Deep Learning

Re:Finding the genre of a song with Deep Learning by dkarapetyan on Thursday December 01, 2016 @11:33PM (#13086042)

Hmm, convolution is perfectly good operation to run on wave forms as well. In fact the wikipedia article (https://en.wikipedia.org/wiki/Convolution) shows the operation on functions which would correspond to time-domain wave forms. What is the point of converting everything to pictures and then using 2D convolutions when that step could have been skipped entirely?

Converting to pictures is unnecessary. It makes the processing harder. The pooling should just happen on segments of the wave form instead of the fourier transform (frequency-domain) picture spectrograms.

[ Parent | on HN ]

Re:Finding the genre of a song with Deep Learning by highd on Thursday December 01, 2016 @11:46PM (#13086105)

The idea is that the vertical axis of the spectrogram is basically already an hierarchical set of features (in scale/frequency). Then convolutions on that is a lot like how DenseNets combine hierarchical features.

I agree it seems a little jank, but the features are pretty good - and a lot of network architectures / training techniques are most practiced in an image processing context.

[ Parent | on HN ]

Re:Finding the genre of a song with Deep Learning by Despoisj on Friday December 02, 2016 @04:35AM (#13087302)

Thanks for your inputs, it's true that we can use convolutions on raw waveform, however the main reason I've used a spectrogram was to work on precomputed relevant features as highd pointed out, instead of running the convolution on lots of data.

[ Parent | on HN ]