CMLS Homework 1 , Group 2, Assignment 3 - Audio Effect Classification by students:
- 10577360
- 10650559
- 10574134
- 10532582
A system consisted of 3 python programs that combined are able to predict the audio effect used in recordings of electric guitar and bass.
The 3 classes to discriminated are :
- distortion
- tremolo
- no effect
The dataset is a set of around 11k monophonic guitar and bass sample sounds plus 4k polyphonic sounds.
Description of the python programs that can be executed independently:
-
preprocessing.py: this script reads the audio files and produces as output the files_information.csv table, which contains the index of the audio sample in which the attack phase ends.
-
featureExtraction.py: based on the files_information.csv this program computes the low-level features for every audio file of the data folder and produces the results folder csv tables for every effect (Tremolo, Distortion, NoFX): – X_Tremolo.csv – X_Distortion.csv – X_NoFX.csv
-
classify.py: this script implements the estimator and produces the metrics and the confusion matrix result.