Closed anthonio9 closed 6 months ago
This answer on stackoverflow look like a very helpful resource: https://stackoverflow.com/a/53688043/11287083
First part of the job is done, audio is augumented with noise where the SNR values are specified in the penn.GSET_AUGUMENT list. awgn_snr() function keeps the dtype of the original data.
Currently the dataset provided by GuitarSet is rather smalll, data augmentation by adding noise to the samples should help at least a bit.