Project developed for the Music Information Retrieval course, at UPF.
Several experiments were conducted with music classification in one of the Valence-Arousal quadrants.
We used 3 datasets with annotated music clips from various languages:
- English: MER_audio_taffc_dataset
- Chinese: CH-818
- Turkish: tr-music-dataset
We only kept here the .csv files with the extracted features for each.
For individual experiments with each dataset, see EnglishMER.ipynb, ChineseMER.ipynb, TurkishMER.ipynb.
For cross-cultural experiments, see CompareMER.ipynb. This can be run independently from the others, as it only loads data from the .csv files.
For more details, see Cross_CulturalMER.pdf, the associated paper for this project. This paper is made on the original template of the ISMIR conference, but was NOT submitted or accepted to this event, it is just for informative purposes here.
Some parts of the code in this project were adapted from https://github.com/MTG/MIRCourse/blob/master/notebooks/Lecture1_IntroWithACaseStudy.ipynb.