SIMULATION AND CALIBRATION OF HEADPHONES WITH ADAPTIVE ACTIVE NOISE CANCELLATION FOR COMPRISING MULTIFREQUENCY COMPOSITE AND REAL NOISE

In the noise environment of the modern society, everybody may be influenced by noise. Noise exerts physiology, psychology, and hearing effects, and there are many situations in which it would be desirable to be able to reduce low-frequency noise with active noise cancellation (ANC). An adaptive cancellation algorithm can be used when wearing headphones to reduce environmental noise as well as accommodate variations of the sound field in the cancellation. However, most past analyses of noise have focused on noise with high stability or a single noise source, with there being insufficient discussions and comparisons of algorithms for cancelling noise comprising multifrequency composite waves. This study investigated four common adaptive algorithms that can be implemented in hardware to achieve adaptive ANC in headphones. Furthermore, different noises comprising multifrequency composite waves and real noise in three industrial environments were used for signal estimation. The results show that the algorithms could reduce the level of pure-tone-based noise by more than 20 dB. FuRLMS (filtered-u recursive least mean square) and lattice-based ANC methods exhibited superior cancellation effects on real environmental noise, achieving reductions of 14 dB. The outcomes could represent effective reference data for the design of a system-on-chip for use in headphone noise cancellation.