Publications

The research addresses swing as a complex performance phenomenon, studying it through musicology, sound analysis, statistics and sociolinguistics. The work starts from a critical review of the main computational studies dedicated to jazz, from Schuller, Ellis and Keil to the models of Friberg, Abel, Busse, Collier and Ramirez. Many of these researches focus attention on the swing ratio, on the positioning of the notes with respect to the beat or on the synthetic reproduction of expressiveness. However, the strong dispersions found in the data show the insufficiency of a general rhythmic proportion: swing simultaneously involves microtiming, articulation, dynamics, timbre, pulse progression, melodic structure, harmony, form and interaction between the performers.

On this basis, a method based on the audiotactile principle and on the distinction, formulated by Vincenzo Caporaletti, between Swing-structure and Swing-idiolect is developed. Swing is interpreted as a performative language of which each musician develops an individual pronunciation. The analogy with sociolinguistics allows us to observe the historical, geographical and cultural variations, the changes produced by the role assumed in the ensemble, the responses to the musical context and the differences linked to the expressive possibilities of the individual instruments.

The investigation directly uses phonographic recordings, analyzed with Sonic Visualiser and subsequently processed through statistical tools. The different case studies show the breadth of the model: the comparison between 1920s jazz and Hard Bop highlights a historical transformation of pronunciation; Duke Ellington and Don Byron's versions of Cotton Club Stomp reveal the persistence of era-related microtemporal behaviors; Scott LaFaro changes dynamics and rhythmic rhythm in the transition from walking bass to solo role; Charlie Parker also organizes swing through recurring forms of timbral envelope; the performance of Well You Needn't by the Kronos Quartet allows us to define a "pidgin swing", based on the superficial assimilation of a few rhythmic traits.

The experimental nucleus consists of the analysis of Wynton Kelly's solo in Freddie Freeloader. Each note is classified according to position in the blues form, phrase, duration, intensity, articulation and deviation from the pulse expressed by Paul Chambers' double bass. The results show a temporal trend shaped by the form and interplay, specific recurrences in the management of eighth notes, triplets and semiquavers, blue notes placed further back on the beat and accents constructed in relation to the melodic and harmonic profile of the phrase. The subsequent comparison with Herbie Hancock reveals affinities in the management of triplets, blue notes and the relationship between dynamics and melodic design, together with different strategies in the management of semiquavers.

The work therefore proposes a use of computation aimed at stylistic analysis. Measurements become meaningful when interpreted in light of performance, musical culture, and formal context. Swing emerges as a mobile and individual expressive practice, in which gesture, sound, structure and interaction contribute to the construction of each musician's personal pronunciation.