Room Acoustic Response Modeling and Equalization with Linear Predictive Coding and Parametric Filters for Speech and Audio Enhancement

Room acoustic response modeling is a challenging problem. Typical applications include speech dereverberation and loudspeaker correction. Traditionally, infinite-duration impulse response (IIR) or finite-duration impulse response (FIR) filters have been used for acoustic response modeling and equalization. The IIR filter, also called a parametric filter, has a bell-shaped magnitude response and is characterized by its center frequency, the gain at the center frequency, and a Q factor (which is inversely related to the bandwidth of the filter) and is easily implemented as a cascade for purposes of room response modeling and equalization. In this paper we present a technique for determining the coefficients of a second order IIR using a linear predictive coding (LPC) model, where the poles or roots of a high-order LPC dictate the parameters of the parametric filter. Due to the band interactions between the IIR filters, forming the cascade to model the room response, we also present a technique to optimize the Q values so as to better characterize the room response. An accurate model allows for better equalization, for correcting the loudspeaker and room acoustics for speech/audio enhancement, particularly at low frequencies. Alternatively, this technique can be utilized for speech dereverberation applications where the room responses have been estimated a priori. The advantages of the proposed method is the fast computation of the IIR filter parameters, from to the LPC model, since (i) the LPC model is efficient to compute since it uses the Levinson-Durbin recursion to solve the normal equations that arise from the least squares formulation, and (ii) a reasonably high-order LPC model is able to accurately model the low-frequency room response modes.