Opens in a new tab

AES E-Library

← Back to search

Convention Paper

Investigations on Nonlinearity in a Gammatone Filter Bank Based Perceptual Model

Authors: Hou, Yaqiong; Schaller, Fabian; Schäfer, Paula; Meier, Paul Emil; Meier, Bernd

Convention Paper · Paper 10282 · May 2026

Abstract

Perceptual models are playing an important role in effectively balancing the data compression and fidelity in audio encoders by leveraging the masking effects in human auditory perception. For deriving well suitable masking thresholds, considering tonality is important. In this study, a novel filter bank is proposed, which uses narrow complex-valued all-pole gammatone filters followed by a non-linear spectral spreading processing. With an appropriate non-linear mapping before spreading, and inverse non-linear mapping afterwards, differences between masking strengths of tonal and noise-like maskers can be directly obtained without explicit tonality estimation. By applying a suitable non-linearity, level-dependency of spectral spreading in the human auditory system can also be modeled. The performance of the proposed approach is evaluated through subjective listening tests, which include comparisons with results obtained using partial spectral flatness measures.

Details

AES Convention
160
Paper number
10282
Publication date
May 28, 2026
Session subject
Audio Processing
Type
Convention Paper