On Every Note a Griff: Looking for a Useful Representation of Basso Continuo Performance Style

TL;DR

This work addresses how to represent and analyze contemporary basso continuo realizations by proposing griff, a transposition-invariant, historically informed feature encoding built from aligned performance-to-score data in ACoRD. Griff constructs two representations—Ordered and Pooled—by grouping performance notes aligned to each score note and encoding them as chromatic-interval tokens, enabling a loss-minimal view of pitch content and structure. The authors provide basic griff statistics on 17,152 extracted griffs and perform experiments on performance profiles to assess individuality of players via cross-entropy in griff spaces, contrasted with interval-based controls. Results suggest that griff-based representations capture intra-player similarity more effectively than inter-player similarity, supporting griff as a meaningful feature space for robust empirical study and larger-scale analysis of continuo practice. The work highlights the value of preserving structural improvisational content for historically informed performance analysis and paves the way for clustering and classification studies.

Abstract

Basso continuo is a baroque improvisatory accompaniment style which involves improvising multiple parts above a given bass line in a musical score on a harpsichord or organ. Basso continuo is not merely a matter of history; moreover, it is a historically inspired living practice, and The Aligned Continuo Dataset (ACoRD) records the first sample of modern-day basso continuo playing in the symbolic domain. This dataset, containing 175 MIDI recordings of 5 basso continuo scores performed by 7 players, allows us to start observing and analyzing the variety that basso continuo improvisation brings. A recently proposed basso continuo performance-to-score alignment system provides a way of mapping improvised performance notes to score notes. In order to study aligned basso continuo performances, we need an appropriate feature representation. We propose griff, a representation inspired by historical basso continuo treatises. It enables us to encode both pitch content and structure of a basso continuo realization in a transposition-invariant way. Griffs are directly extracted from aligned basso continuo performances by grouping together performance notes aligned to the same score note in a onset-time ordered way, and they provide meaningful tokens that form a feature space in which we can analyze basso continuo performance styles. We statistically describe griffs extracted from the ACoRD dataset recordings, and show in two experiments how griffs can be used for statistical analysis of individuality of different players' basso continuo performance styles. We finally present an argument why it is desirable to preserve the structure of a basso continuo improvisation in order to conduct a refined analysis of personal performance styles of individual basso continuo practitioners, and why griffs can provide a meaningful historically informed feature space worthy of a more robust empirical validation.

Figures (5)

• Figure 1: Griffs are built using performances aligned to scores, and each score note forms its own griff from performance notes aligned to it (a). MIDI pitches of performance notes are grouped to vectors based on the order and proximity of their onset times (b). A griff is a sequence of these vectors. Griffs converted to chromatic intervals from the score note pitch (c) are encoded as strings (d).
• Figure 2: Distribution of different griff types in both representations.
• Figure 3: Three most occurring ordered griff types without and with vertical lines belonging to the three most frequent chord types. The referential score note and a root position of the chords are shown. The figure contains string encoding of each griff, frequency of occurrence and respective rank in the whole dataset.
• Figure 4: Cumulative coverage of ordered griff types ranked from the most occurring to the least occurring ones for each player. The rightmost segment of a player’s line corresponds to griffs used only once.
• Figure 5: Similarity matrices of different players’ mean cross-entropies in three different representations for each score.