Hello all, I need to know what is the best way to do a data-augmentation using Matlab on a small database (62 speakers and 1116 .wav files ) for training and testing an i-vector speaker recognition system
Thank you.
Abbas Cheddad
Hello,
I am not sure on 1D signal data augmentation, however, images (2D signals) are augmented using some image manipulation/transformation/deformation techniques such as: rotation, translation, rescaling, affine transformation, etc.
You can probably get, from the above examples, an idea on how to do that on audio files. For instance, could regenerating audio files by varying the sampling rate in Hz and/or passing the audio signal through a filter be valid choices?
/Abbas
0 votes
0 thanks
- Badges
- Science topic
Similar questions and discussions