Modeling perceptual categories of parametric musical systems

Abstract

In computer music fields, such as algorithmic composition and live coding, the aural exploration of parameter combinations is the process through which systems’ capabilities are learned and the material for different musical tasks is selected and classified. Despite its importance, few models of this process have been proposed. Here, a rule extraction algorithm is presented. It works with data obtained during a user auditory exploration of parameters, in which specific perceptual categories are searched. The extracted rules express complex, but general relationships, among parameter values and categories. Its formation is controlled by functions that govern the data grouping. These are given by the user through heuristic considerations. The rules are used to build two more general models: a set of “extended or Inference Rules” and a fuzzy classifier which allow the user to infer unheard combinations of parameters consistent with the preselected categories from the extended rules and between the limits of the explored parameter space, respectively. To evaluate the models, user tests were performed. The constructed models allow to reduce complexity in operating the systems, by providing a set of “presets” for different categories, and extend compositional capacities through the inferred combinations, alongside a structured representation of the information.Peer ReviewedPostprint (author's final draft

On-The-Fly Syntheziser Programming with Fuzzy Rule Learning

“…This complexity considers the dissimilarity and create_rule functions described. This complexity is better than a previous version of the algorithm presented in [ 11 ].…”

“…To test how the algorithm models the feature space of a synthesis algorithm, we used the data set described in [ 11 ]. This dataset was generated by user tests, in which different configurations of a Band Limited Impulse Oscillator [ 14 ] were programmed by users and tagged either as rhythmic , rough or pure tone .…”

“…Data sets Wine, Wine-quality-red, Glass and Ionosphere, selected from the UCI repository [15]. The Blip data set was obtained from [11]. The accuracy was calculated using 10-fold cross validation.…”

“…This, together with the positive criticism that the performances and recordings received, suggests that rule learning is a promising approach, able to build models from few observations of complex systems. In addition, to the best of the author’s knowledge, inductive rule learning has not been explored beyond our work [ 3 , 11 ] neither for automatic synthesizer programming nor within live coding.…”

Charting Perceptual Spaces with Fuzzy Rules

“…As mentioned in section I, there are no benchmarks available in this field, since the parameter combinations that codify what is beautiful, good, or suitable depend on the specific context. Therefore, to evaluate the extracted models the data generated during the user test described in [14], was used. It is available at [15].…”