Environmental Sound Classification Using Local Binary Pattern and Audio Features Collaboration. This code is referenced from https://github.com/ohinitoffa/esc/tree/main.

Dataset is from:

@data{DVN/YDEPUT_2015,
author = {Karol J. Piczak},
publisher = {Harvard Dataverse},
title = {{ESC: Dataset for Environmental Sound Classification}},
year = {2015},
version = {V2},
doi = {10.7910/DVN/YDEPUT},
url = {<https://doi.org/10.7910/DVN/YDEPUT}>
}

1. Model training and evaluating
2. Offline processing
3. Inline processing

1. cry/laughter
2. fan/motor/noise (low freq)
3. TV, conversation (high freq)
4. phone ring, door-bell
5. traffic (police, firefighter, emergency, trash, normal)

Multiprocessing is so bugged so how.

|--- dataset folder |--- category A |--- file A |--- category B |--- file B