LAMA - LatinAmerica, Asia, MiddleEastern, Africa Genre Dataset
This Dataset consists of the .wav files, training dataset (MFCC), and graph plots (FFTs, MFCCs, STFTs, Waveforms) of YouTube videos classified into four categories: LatinAmerica, Asia, MiddleEastern, and Africa.
I hope that this work can help in several Deep Learning, Machine Learning projects in Music Genre Classification.
Anyone is free to use/change/contribute to this Dataset. Please cite if you use this Dataset in your research/projects.
Some data couldn't be uploaded to GitHub because the file size was too large. Instead, I attached a Harvard Dataverse link below to retrieve the data.
The data contained in LAMA can be classified into three categories:
- .wav files (
Data_Original,Data_Cropped) -> Uploaded in Harvard Dataverse. Link below. - Waveform, FFT, STFT, MFCC graph plots (
Graphs,Sample_Graphs) -> Uploaded in Harvard Dataverse. Link below. - mfcc training data (
training_data.json)
Data_Original: 113 (Africa), 83 (Asia), 90 (LatinAmerica), 80 (MiddleEast)Data_Cropped(1 min clips of original) : 108 (Africa), 77 (Asia), 87 (LatinAmerica), 77 (MiddleEast)Graphs-> FFT : 108 (Africa), 77 (Asia), 87 (LatinAmerica), 77 (MiddleEast)
from Graphs
from Sample_Graphs
The classification of genre in this Dataset is mostly from the "AudioSet" project by the Sound and Video Understanding teams at Google Research. I chose the best examples from their website and preprocessed them to create this Dataset.
In addition, I obtained much of the knowledge needed to create this Dataset from the YouTube Channel, "Valerio Velardo - The Sound of AI. Mr. Velardo makes amazing videos, and I believe that he deserves more recognition.
@data{DVN/13BPFB_2020,
author = {Lee, Bruce W},
publisher = {Harvard Dataverse},
title = {{LAMA World Music Genre Dataset}},
year = {2020},
version = {DRAFT VERSION},
doi = {10.7910/DVN/13BPFB},
url = {https://doi.org/10.7910/DVN/13BPFB}
}