This library is created because of the discrepancy between jLibrosa and Librosa. As of right now, our AI uses a .csv file to train the AI, and the .csv created using features generated with librosa differs from the ones generated with jLibrosa, sometimes by 10%, and other times by more than 50% This library creates a .csv file with the features generated with jLibrosa.
an input and output directories are needed where the program will run
the input directory should have subdirectories which are the genres of the songs inside of them
eg.
+input/
+++blues/
+++rock/
+++classical/
+++...
All subdirectories (genres) and will get copied to the output/ directory. All .wav files will get copied to their respective subdirectories
and all .mp3 files will get converted and then saved as .wav in the output directory
Afterwards the program will extract the features from the songs in the output directory and write them into the .csv