Subjective perception analysis of active noise control algorithms in an encapsulated structure: An experimental study

This study investigates the subjective perception of active noise control (ANC) performance, focusing on how individuals evaluate the noise reduction provided by different ANC algorithms. While the performance of the ANC algorithms has already been evaluated using objective metrics, this study aims to assess their effectiveness from a subjective perspective. In a simulated vehicle interior created using a noise box, two ANC algorithms were tested: the normalized least-mean-square (NLMS) algorithm and the hybrid selective fixed-filter active noise control normalized least-mean-square (SFANC-NLMS) algorithm. Participants were exposed to 27 stimuli, which combined three types of noise (motorcycle, street, and train), three sound pressure levels (55, 65, and 72 dB(A)), and three ANC conditions (no control, NLMS, and SFANC-NLMS). Subjective evaluations were collected using three indicators: perceived annoyance (PAY), perceived affective quality (PAQ), and perceived loudness (PLN). These metrics captured participants' impressions of the noise environment and the impact of noise control. The study is structured around three research questions (RQ1, RQ2, and RQ3), each addressing different aspects of ANC performance evaluation. In response to RQ1, the results demonstrated that the SFANC-NLMS algorithm outperformed NLMS in reducing perceived annoyance and loudness. Regarding RQ2, higher sound levels (72 dB) led to greater perceived annoyance, but sound level did not significantly alter the relationship between ANC algorithm type and perceived annoyance. Finally, in addressing RQ3, noise type influenced ANC effectiveness, with SFANC-NLMS showing more significant reductions in perceived annoyance compared to NLMS. Overall, the findings confirm that the SFANC-NLMS algorithm provides better noise reduction in encapsulated structures.