MPB Corpus v1.0

Read full text Summarise with VIRAT
Authors: Carlos de Lemos Almada, Hugo Tremonte de Carvalho, Felipe Defensor Martins
Cited by: 0

Summary

No summary yet. VIRAT can read this paper and write a plain-language summary of its aim, methods, key findings and limitations.

Summarise with VIRAT Free to try with your Science Catalogs account.

Abstract

This release corresponds to version 1.1 of the MPB Corpus. It contains analytical data derived from a collection of 500 songs by Tom Jobim, Ivan Lins, Chico Buarque, Edu Lobo, Caetano Veloso, Djavan, João Bosco, Milton Nascimento, Gilberto Gil, and Rita Lee. The corpus includes structured representations of melodic contour, melodic rhythm, harmonic content, and the relationship between harmony and melody. These materials are intended to support research in computational musicology of Brazilian popular music. No original copyrighted scores or audio files are distributed. This release provides derived data only. Changes from v1.0 No annotation changed. Every value is identical to v1.0 — 8,426 contour and rhythm segments, 17,053 chords, and 23,447 notes, in the same order and under the same column names. Analyses run against v1.0 do not need to be recomputed. Added dataset_by_corpus/ — the same three tables split into one folder per composer, e.g. dataset_by_corpus/JOBIM/harmony.csv. Useful for working with a single composer without loading the whole corpus. Changed dataset/ is now dataset_aggregated/. Code that loads the tables by path will need updating. Line endings are now LF on every platform, where v1.0 used CRLF. This is the only byte-level difference in the annotation files, and the reason the checksums differ between versions while the data does not. About this release This release contains the dataset only. Documentation, metadata, analysis code, and how to cite this work are on the main branch. Full details in CHANGELOG.md.

Fields of Study & Subjects

Music and Audio Processing Musicology and Musical Analysis Music Technology and Sound Studies

Publication details

Journal / source
Zenodo (CERN European Organization for Nuclear Research)
Publication date
25 March 2026
Language
EN
Document type
Dataset