The following projects and abstracts represent some of the work published or currently in progress, and further information will be added as soon as it is known.


Pablo Padilla, Francis Knights and Dan Tidhar, Approaches to the Computational Analysis of Classical Music: Methodology and Applications (forthcoming 2019)


1: 'Identification and evolution of musical style I - Hierarchical transition networks and their modular structure
In this paper we propose a methodology to assist with the problem of identifying musical styles using mathematical tools that is related to the transition network approach developed by David Cope in his Experiments on Musical Intelligence. This extension allows for the possibility of defining stylistic cells at different scales as motifs and moduli of networks at the corresponding scale. We also outline how this methodology can be used to systematically study stylistic changes in different contexts by incorporating probabilistic and statistical tools and connections with other approaches. Published in Agustín-Aquino, O A, Lluis-Puebla, E and Montiel, M (eds.), Mathematics and Computation in Music (2017), 258-79,

3: 'Louis Couperin's viol fantasies - a question of attribution'
As well as the several hundred harpsichord and organ works attributed to Louis Couperin in modern editions, there are a number of chamber works, including two viol fantaisies and three 'symphonies'. These comprise his surviving non-keyboard works; and are found in the late-17th century Bauyn manuscript from which the harpsichord works come. Couperin was both organist and treble viol player, so one might expect such repertoire to exist. The question of the secure attribution of the viol works can now be discussed by stylistic comparison with the 'Monsieur Couperin' harpsichord (possibly by Charles or Louis) and organ (definitely by Louis) works, using new computational and mathematical models developed for a whole-corpus study of this important repertoire. [Paper in final draft]

4: 'Chambonnières versus Louis Couperin: attributing the F major Chaconne'
Chambonnières left four extant chaconnes in the Bauyn manuscript. Of these, the fine Chaconne in F has been assigned to Louis Couperin by a number of editors and writers on grounds of style (see for example, Curtis [1970] vol.ii, pp.ix and 117). The aim of this study is to use computational and mathematical models developed for a whole-corpus study of the organ works of Louis Couperin and the harpsichord works now proposed to be by his brother Charles Couperin (Wilson, 2013), in order to compare analytically the style of this Chaconne with other works by Louis, by 'Charles' and Chambonnières himself, and arrive at a more secure attribution. In addition, the two extra couplets that appear in the Parville but not the Bauyn source, which have sometimes been regarded as a later addition, are compared in order to judge their stylistic similarities. Published in Harpsichord & fortepiano, xx/i (Autumn 2017), 28-32

5: The attribution of Taverner and Tye's O Splendor gloriae

The five-part motet O splendor gloriae is attributed to Taverner in two important sources, but to both Taverner and Tye in the other two sixteenth-century sources. Joint works of this kind are extremely rare, and common authorship difficult to assess. Scoring and many other musical features here are consistent, but the style of each half seems different. Davison (1987, xvii-xviii) attributes the second part as likely by Tye on grounds of stylistic characteristics. It certainly seems more modern in its imitative treatment, which may be consistent with the fact that Tye was a little younger than Taverner. In this paper we employ a new mathematical modelling methodology in order to provide an objective criterion for analysis, as additional evidence to standard musicological methods. The results support the claim attributing each half to a different composer.  [Paper ready for submission]

7: 'Analytical methods for the attribution of early Bach keyboard works'
A surprising number of early Bach keyboard works have historical attributions that are contested by modern scholars on grounds of style, and therefore remain on the margins of the canon. Using a variety of mathematical modelling tools developed for a whole-corpus analysis of the keyboard works attributed to members of the Couperin couperin, we compare the stylistic components of one such work, the Suite in Bb BWV821. The dance movements are compared with equivalent works from later in Bach's career, using Principal Component Analysis and Clustering techniques, to search for patterns that can help identify if there are convincing Bachian fingerprints in this and other early works. [Paper in final revision]

10: 'The mathematical modelling of aspects of stylistic development in Haydn's string quartets'
Haydn's string quartets contain some of his finest music, and were written over half a century of his compositional career. The standard forms he employed mean that there is sufficient consistency of content to use mathematical modelling tools and pattern recognition processes, including a graph-theoretical approach, to study stylistic change during Haydn's career, observing both consistencies and differences in, for example, melodic and harmonic patterns, during a period which transitioned all the way from the late Baroque to the early Romantic. [Paper submitted]

23: 'Natural mathematics and the derivation of new formal structural parameters for composers'
Contemporary composers face an interesting challenge: to generate new musical structures that depart from traditional forms. The use of mathematical procedures and algorithms in music (which date back to at least Guido d'Arezzo in the 11th century) can provide one interesting starting point as a source of ideas. We present a series of examples using mathematical ideas accessible to anyone familiar with basic arithmetic. They should be taken not as instructions on their own, but as material to be modified and further elaborated. In other words, we do not intend to automatically generate music by a given procedure, but rather to give composers additional tools and materials.  For a practical outcome see [Paper submitted]

50 'Completing the Art of Fugue: matching Bach's style'
There are numerous expert completions of the unfinished final Contrapunctus from Die Kunst Der Fugue, and comparative evaluations of these have tended to judge them according to claimed parameters of fugal and structural correctness. An information-theoretical approach allows statistical assessment of the actual voice-leading of various completions, in order to see which conforms most closely to the extant Bach portion. [Paper in final revision]

69: 'A proposal for notation-related modifications to midi data for the statistical analysis of Renaissance polyphony'
In this paper we study the structural importance of rests in the melodic content of a classical music work and its implications for stylistic classification. We apply a methodology previously developed using statistical methods based on the pitch transitions of the musical material. Specific examples from 16th century motets by John Sheppard are analyzed to show the importance of this component. [Paper submitted]

71: 'New unmeasured prelude attributions to 'Monsieur Couperin'"
The Parville Manuscript is one of the principal sources of harpsichord music by 'Monsieur Couperin' (formerly believed to be Louis Couperin). It also contains a number of anonymous works which are very similar in style, and technical analysis shows two of these unmeasured preludes to be convincingly similar to the attributed 'Monsieur Couperin' works. [Paper in final draft]