Explaining two-tone suppression and forward masking data using a compressive gammachirp auditory filterbank

Abstract

The gammatone filter was imported from auditory physiology to provide a time-domain version of the roex auditory filter and enable the development of a realistic auditory filterbank for models of auditory perception. The gammachirp auditory filter was developed to extend the domain of the gammatone auditory filter and simulate the changes in filter shape that occur with changes in stimulus level. Recently, the gammachirp was extended to explain the level-independent frequency glide of the impulse response and a large volume of simultaneous masking data quantitatively. Although this could be implemented with a static filter, we used a time-varying filter whose active component is an IIR asymmetric compensation filter. In this case, it is necessary to estimate the signal level that controls the level dependency, and explain how the level measurement produces the time course of threshold observed in forward masking experiments. In this talk, we propose a new implementation of the time-varying, compressive gammachirp auditory filterbank to explain two-tone suppression and forward masking data qualitatively. We will also show that the system can resynthesize compressed speech sounds and so provide the basis for a gammachirp analysis/synthesis filterbank. [Work supported by GASR(B)(2) No. 15300061, JSPS.]