A Hybrid Symbolic–Acoustic Computational Framework for the Analysis of Choros: The Case of Garoto

Resumo

This article proposes a hybrid symbolic–acoustic computational framework for the analysis of Brazilian choro repertoire, integrating methods from computational musicology and musicological interpretation. The study focuses on works by Aníbal Augusto Sardinha (Garoto), a central figure in the development of Brazilian instrumental music, combining symbolic feature extraction from digitized scores using the Python library music21 with acoustic descriptors derived from historical recordings through librosa. The symbolic corpus is based on the Songbook Choros de Garoto (SESC/IMS), while the acoustic corpus consists of selected historical recordings performed by the composer. Extracted features — including tonal organization, harmonic progressions, melodic patterns, MFCCs, tempo estimation, and energy curves — are integrated into a unified analytical structure using pandas, enabling cross-domain comparison between compositional and performative dimensions. The application of the model to the piece Jorge do Fusa reveals correspondences between formal structure, harmonic mobility, timbral behavior, and dynamic shaping, suggesting a close interaction between compositional design and expressive performance practices. The study argues that computational descriptors gain analytical relevance when interpreted within musicological frameworks, contributing both to methodological discussions in computational musicology and to the analytical understanding of Brazilian popular music. The proposed framework offers a replicable approach for future stylistic investigations of choro and other repertoires that combine written and oral traditions.