engineerchuan

I would like to work on this. Would we prefer doing this in one PR or okay to split across a 2-3 PR?

Couple of comments: From this [test run](https://app.circleci.com/pipelines/github/pytorch/audio/2710/workflows/724b2b85-4fba-4d01-accf-aa743b54cc3f/jobs/73107/steps), I learned: 1. `torchaudio/datasets/gtzan.py` probably uses `genres/noise/noise.0000.wav` so we shouldn't remove that. I can't find a reference in the code though but the...

Are we sure we can remove `steam-train-whistle-daniel_simon.mp3 `? In file `torchaudio/test/test_io.py`, it seems to use this file to test MP3 reading IO.

I did a grep for the mp3 file ``` $ grep -R steam-train-whistle-daniel_simon.mp3 test test/common_utils/backend_utils.py: test_filepath = data_utils.get_asset_path('steam-train-whistle-daniel_simon.mp3') test/test_dataloader.py: sound_files = ["sinewave.wav", "steam-train-whistle-daniel_simon.mp3"] test/test_io.py: "steam-train-whistle-daniel_simon.mp3") test/test_sox_effects.py: test_filepath = common_utils.get_asset_path("steam-train-whistle-daniel_simon.mp3") ```...

I would like to work on this.

Not off top of my head. Let me study it first for a day and come up with a proposal.

Hi @mthrok, I would like to follow this approach with some questions: 1. For testing `compute-fbank-feats` and `compute-mfcc-feats`, first extract the default argument values. 2. As suggested in https://github.com/pytorch/audio/issues/679, use...