Synthesis and implementation of virtual bass system with a phase-vocoder approach

Abstract

A bass enhancement technique based oil a phase-vocoder approach is presented. Instead of direct bass boosting, the proposed method creates a bass impression by exploiting the psychoacoustic properties of humans. This technique is most useful in audio reproduction using small loudspeakers that have no low-frequency capability, where direct boosting will likely result in nonlinear distortions. In light of psychoacoustics, the bass effect is synthesized by augmenting the original signals with high-frequency harmonics. Unlike conventional methods that rely on nonlinear processing, the proposed method performs the required frequency transformation by using a phase-vocoder approach. Apart from frequency transformation, another key element of the proposed technique is the magnitude adjustment of the generated harmonics. The underlying principle for magnitude adjustment is based on a polynomial model of equal-loudness contours. The method is implemented on a digital signal processor with the aid of multirate signal processing. To validate the proposed technique, objective and subjective experiments are conducted for PC multimedia loudspeakers and handset loudspeakers. The subjective listening experiment followed the procedure of multistimuli with the hidden reference and anchor (MUSHRA), and the data were analyzed by using the multianalysis of variance (MANOVA) method. As indicated by the results, the proposed technique proved effective in rendering bass impression with acceptable audio quality.