Room impulse response estimation using perfect sequences for Legendre nonlinear filters

Abstract

The paper proposes a novel method for room impulse response estimation that is robust towards nonlinearities affecting the power amplifier or the loudspeaker of the measurement system. The method is based on measurements of the first order kernel of the Legendre nonlinear filter modeling the acoustic path. In the proposed approach, the first order kernel is efficiently estimated with the cross-correlation method using perfect periodic sequences for Legendre filters. Perfect sequences with period suitable for room impulse response identification are also developed within the paper. Simulation results in a realistic scenario illustrate the effectiveness and robustness towards nonlinearities of the proposed approach.