An efficient narrowband active sound profiling system with online secondary path modeling

Abstract

Abstract Active noise equalization or active sound profiling techniques for sound quality enhancement tend to balance the sound pressure levels (SPLs) of certain frequency points or frequency bands of target noise to meet human perception requirements rather than simply minimizing the SPLs like active noise control (ANC) technique. Therefore, when the secondary path is time-varying, online secondary path modeling (OSPM) appears particularly important in the active sound quality enhancement systems. This paper presents a novel parallel-structure narrowband active sound profiling (NASP) system with OSPM. In the system, the OSPM subsystem uses the control signals to estimate low-order local secondary path models at the controlled frequencies instead of the usual high-order global model. This way may ensure the convergence speed and accuracy of modeling. Meanwhile, the main sound profiling subsystem employs the estimated local secondary path models to profile the targeted multi-tonal narrowband noise based on the internal model FXLMS algorithm. By using bandpass filter banks, each tonal component may be tuned relatively independently. In addition, the proposed system enjoys good profiling accuracy and stability since no additional random noise is introduced. Several simulations are provided to demonstrate the superiority of the proposed NASP system with OSPM in cancellation mode over the conventional NANC system with OSPM and evaluate the behavior of the proposed system in different time-varying scenarios. Finally, real-time experiments are also conducted to confirm the system’s ability to deal with sudden changes in the secondary path and its sound profiling capability.