Black-Box System Identification of a Multilayered Wiener-Hammerstein Model for Distortion Pedals.
System identification; Non-Linear systems; Audio; Signal proccessing.
The advencement of electronics allowed the development of several types of audio effects, mostly applied to musical instruments. One of these effects is distortion, which shaped the sound of electric guitar as we know it today. Equipments that provided such effects became more robust and compact, combined with a greater diversity of tones. The digital electronics allowed the creation of equipments that have several of these effects, called multi-effect pedals, and unities that can model effects. One of the challenges in the development of these equipments is to truly replicate famous distortion tones. For that purpose, it is possible to model the desired effect through full knowledge of its electronic circuit or through modeling based on the identification of the system in question. This work proposes the development, application and evaluation of a parametric multilayered Wiener-Hammerstein model for digital modeling of distortion pedals based on existing pedals. The model parameters will be optimized by a genetic algorithm. The results will be evaluated by frequency-error, mel scale error calculus and through a hearing test.