Embracing Extractions: Integrating symbolic and audio approaches in Renaissance polyphony

Preliminary findings of this study were presented at the International Medieval and Renaissance Music Conference 2025 in Newcastle, Durham and Northumbria, United Kingdom. For more details, please visit the conference programme.

As a follow-up on Fuzzy Frequencies, we further explore the possibilities of multiple F0 extractions.

In computational musicology, the analysis of symbolic music encodings and audio recordings often appear as separate domains. Audio signal processing builds on a mathematical foundation, while symbolic approaches are closely tied to musicology. This disciplinary divide can act as an obstacle for music researchers to empirically study recordings, performance practice, or repertories that are short on symbolic encodings.

To bridge this gap, we introduce an intermediary data representation, multiple fundamental frequency (f0) extraction (henceforth: extraction). An extraction provides the bare pitch content of a recording. We explore a selection of pitch-related features to gain a better understanding of tonal patterns in Renaissance music.

Additionally, we show how extractions enable the study of performance characteristics such as vibrato, the effect of words, and intonation.