[en] Numerous audio systems are nonlinear. It is thus of great importance to study them and understand how they work. Volterra series model and its subclass (cascade Hammerstein-Wiener model) are usual ways to modelize nonlinear systems. However the identification methods of these models are still considered as an open topic. Therefore we have developed a new optimized identification tool ready for use and presented as a Matlab toolbox. This toolbox provides the parameters of the optimized sine sweep needed for the identification method, it is able to calculate the parameters of the Hammerstein model and to emulate the output signal of a nonlinear device for a given input signal. To evaluate the toolbox, we modelize a guitar distortion effect (the Tubescreamer) having a total harmonic distortion (THD) comprised in the range 10-23\%. We report a mean error of less than 0.7\% between the emulated signal and the signal coming from the distortion effect.
Disciplines :
Electrical & electronics engineering
Author, co-author :
Schmitz, Thomas ; Université de Liège > Dép. d'électric., électron. et informat. (Inst.Montefiore) > Dép. d'électric., électron. et informat. (Inst.Montefiore)
Embrechts, Jean-Jacques ; Université de Liège > Dép. d'électric., électron. et informat. (Inst.Montefiore) > Techniques du son et de l'image
Language :
English
Title :
A new toolbox for the identification of diagonal Volterra kernels allowing the emulation of nonlinear audio devices
Publication date :
10 September 2016
Event name :
The 22nd International Congress on Acoustics
Event place :
Argentina
Event date :
du 5 au 9 septempbre 2016
Audience :
International
Main work title :
International Congress on Acoustics : ICA, Argentine 5-9 Septembre 2016
Author, co-author :
Schmitz, Thomas ; Université de Liège - ULiège > Montefiore Institute of Electrical Engineering and Computer Science