Table of Contents
Fetching ...

Perpetual Dialogues: A Computational Analysis of Voice-Guitar Interaction in Carlos Paredes's Discography

Gilberto Bernardes, Nádia Moura, António Sá Pinto

Abstract

Computational musicology enables systematic analysis of performative and structural traits in recorded music, yet existing approaches remain largely tailored to notated, score-based repertoires. This study advances a methodology for analyzing voice-guitar interaction in Carlos Paredes's vocal collaborations - an oral-tradition context where compositional and performative layers co-emerge. Using source-separated stems, physics-informed harmonic modelling, and beat-level audio descriptors, we examine melodic, harmonic, and rhythmic relationships across eight recordings with four singers. Our commonality-diversity framework, combining multi-scale correlation analysis with residual-based detection of structural deviations, reveals that expressive coordination is predominantly piece-specific rather than corpus-wide. Diversity events systematically align with formal boundaries and textural shifts, demonstrating that the proposed approach can identify musically salient reorganizations with minimal human annotation. The framework further offers a generalizable computational strategy for repertoires without notated blueprints, extending Music Performance Analysis into oral-tradition and improvisation-inflected practices.

Perpetual Dialogues: A Computational Analysis of Voice-Guitar Interaction in Carlos Paredes's Discography

Abstract

Computational musicology enables systematic analysis of performative and structural traits in recorded music, yet existing approaches remain largely tailored to notated, score-based repertoires. This study advances a methodology for analyzing voice-guitar interaction in Carlos Paredes's vocal collaborations - an oral-tradition context where compositional and performative layers co-emerge. Using source-separated stems, physics-informed harmonic modelling, and beat-level audio descriptors, we examine melodic, harmonic, and rhythmic relationships across eight recordings with four singers. Our commonality-diversity framework, combining multi-scale correlation analysis with residual-based detection of structural deviations, reveals that expressive coordination is predominantly piece-specific rather than corpus-wide. Diversity events systematically align with formal boundaries and textural shifts, demonstrating that the proposed approach can identify musically salient reorganizations with minimal human annotation. The framework further offers a generalizable computational strategy for repertoires without notated blueprints, extending Music Performance Analysis into oral-tradition and improvisation-inflected practices.
Paper Structure (10 sections, 1 equation, 7 figures)

This paper contains 10 sections, 1 equation, 7 figures.

Figures (7)

  • Figure 1: Computational pipeline for the extraction of musical descriptors. Dashed boxes indicate expert-driven manual annotations used to compute specific descriptors and to impose hierarchical segmentations.
  • Figure 2: Empirical Mode Decomposition of the guitar loudness descriptor in 'Balada do Mar' by Paredes and Goes, showing the raw signal alongside intrinsic mode functions IMF 1, IMF 2, and IMF 3 and higher.
  • Figure 3: Correlation matrix for the song ‘Balada do Mar’ (Paredes and Goes), aggregated using Fisher’s z-transformation of Pearson correlations.
  • Figure 4: Time series of the Empirical Mode Decomposition (IMF 3+) for guitar loudness and voice rhythmic density in 'Balada do Mar' by Paredes and Goes. Masked segments correspond to periods of vocal inactivity below -40 dB. Vertical black lines denote section boundaries, while vertical grey lines denote phrase boundaries. Section labels are displayed at the bottom of the plot, and phrase labels are displayed at the top.
  • Figure 5: Residual analysis of the guitar loudness and voice rhythmic density interaction in 'Balada do Mar' by Paredes and Goes. The highlighted point identifies a positive outlier with a residual z-score exceeding $+2.5$, indicating a breakdown of the typical contrapuntal relationship at the beginning of section B. Vertical black lines denote section boundaries, while vertical grey lines denote phrase boundaries. Section labels are displayed at the bottom of the plot, and phrase labels are displayed at the top.
  • ...and 2 more figures