Divide Spectrograms into Chunks
Closed this issue · 1 comments
johnmartinsson commented
Implement a method which divides the spectrograms into equal chunks.
Implementation details:
- Split spectrogram into chunks of equal size (length 512).
Motivation:
- Need fixed sized input for the neural network architecture.
- Allow to pad only the last part, and keep step size constant
- Each chunk can be used as a unique sample for training (since "empty" parts have been removed)
- Network can make multiple predictions per sound file, and average them to generate a final prediction.
johnmartinsson commented
This issue has been resolved.