My article on the spectrogram pre-processing for audio classification. This discuss how several pre processing methods for spectrograms would affect the classification accuracy of an audio emotion classification model.

Read the Medium article