Virtual Bass Enhancement for Distributed Mode Loudspeaker

• Main Author: Lintervo, Arvi
• Other Authors: Tossavainen, Antero, Sähkötekniikan korkeakoulu, Välimäki, Vesa, Aalto-yliopisto, Aalto University
• Format: Master Thesis
• Language: English
• Published: 2022
• Subjects: Audio algorithms; audio signal processing; loudspeakers; psychoacoustics; Sarja; DML
• Online Access: https://aaltodoc.aalto.fi/handle/123456789/116402

A Distributed Mode Loudspeaker (DML) is a sound reproduction system, which is based on the modal waves of a thin, stiff panel. A typical issue with flat panel speakers is weak bass reproduction capability. In this thesis, the feasibility of Virtual Bass Enhancement (VBE) algorithms for improving the bass performance of DML systems is studied. VBE algorithms utilize the feature of human hearing, where a series of harmonic frequency components of a certain fundamental frequency, is perceived to produce the same pitch as the fundamental frequency alone. Two ways to generate the harmonic frequencies are time-domain methods and frequency-domain methods. In this thesis, an application was developed to implement both methods. A listening test was conducted to evaluate the performance of different VBE algorithms for DML. The algorithms under the test were the Arc-Tangent Square Root Nonlinear Device (ATSR-NLD), the Phase Vocoder (PV) and the hybrid method which combines aforesaid methods. The results of 18 participants were analyzed using Analysis of Variance (ANOVA) and Tukey’s Honestly Significant Difference tests. It was found that each processing scheme increased the perceived bass significantly. The increase of distortion was negligible with the ATSR-NLD, but significant with the PV and hybrid methods. Furthermore, the overall subjective audio quality increased when the ATSR-NLD and hybrid methods were used. The PV processing decreased the overall audio quality, probably due to the increased amount of distortion. Hajautettujen moodien kaiutin on äänentoistojärjestelmä, jonka äänen muodostus perustuu ääniaaltojen moodeihin ohuessa ja jäykässä paneelissa. Tyypillisesti ohuiden paneelikaiuttimien heikkous on rajoittunut kyky toistaa matalia taajuuksia. Tässä opinnäytetyössä tutkitaan virtuaalisten basson parannusalgoritmien käytettävyyttä matalien taajuuksien toiston parantamiseen. Basson parannusalgoritmit perustuvat ihmisen kuulon ominaisuuteen, jossa harmonisten taajuuksien sarja aistitaan matalampana äänenä. ...