The script takes a set of .wav files and creates spectrogram images of each of them, then zips the images in files under 100mb
- matplotlib
- scipy.io
python main.py [path to file tree containing .wav files]
Alternatively, the script defaults to looking for an "assets" folder in the same directory as the script containing .wav files
- finish clean_up() in filehandler.py
- Spectrogram images of audio files less than 10 seconds should go flat after their duration show some unexpected artifacts
- Take 10 second clip of files greater than 10 seconds instead of first 10 seconds to give better context of the sound being classified
- add command line options to:
- Change clip length
- do clean_up() after job
- set destination of zip files
- set verbosity of script
- move sample audio files out of repo and populate a repo of non-copyrighted audio files