asteroid-team / torch-audiomentations

Fast audio data augmentation in PyTorch. Inspired by audiomentations. Useful for deep learning.
MIT License
969 stars 88 forks source link

Get rid of librosa dependency in favor of torchaudio #134

Open hbredin opened 2 years ago

hbredin commented 2 years ago

torchaudio now does a good job loading and resampling audio.

Would be nice to get rid of (sometimes tricky to install) librosa in favor of torchaudio.

Related: #9

iver56 commented 2 years ago

Good! Yeah, I agree, let's get rid of the librosa dependency. I haven't followed the progress in torchaudio closely. Did they speed up their resampling since https://github.com/asteroid-team/torch-audiomentations/issues/37#issuecomment-724691882 ?

Bomme commented 2 years ago

Since version 0.8.0 the resampling in torchaudio is based on julius.

Leveraged julius resampling implementation to make resampling faster. (https://github.com/pytorch/audio/pull/1087)

hbredin commented 2 years ago

Related: https://pytorch.org/audio/stable/tutorials/audio_resampling_tutorial.html#comparison-against-librosa