Title :
A hybrid virtual bass system with improved phase vocoder and high efficiency
Author :
Shaofei Zhang ; Lei Xie ; Zhong-Hua Fu ; Yougen Yuan
Author_Institution :
Sch. of Comput. Sci., Northwestern Polytech. Univ., Xi´an, China
Abstract :
Virtual bass has drawn wide attention from consumer electronics manufacturers who demand good bass quality from small loundspeakers that have no low-frequency capability. Based on the missing fundamental phenomena, virtual bass aims to extend the low frequency bandwidth psychoacoustically by generating a series of harmonics in the mid-frequency range where loudspeakers can reproduce well. State-of-the-art virtual bass systems usually combine the phase vocoder (PV) approach and the non-linear device (NLD) approach, which aims to overcome their individual drawbacks and achieve good bass performance through a hybrid approach. However, these bass enhancement approaches still lack in audio quality and computation efficiency. In this paper, we propose a new hybrid system with an improved phase vocoder and high efficiency. Previous studies show that artificially-added harmonics may change the timbre of the sound tracks and damage audio quality. We solve this problem by a subjective preference based harmonic magnitude control (HMC) in PV, which is motivated by a recent study on the relationship between the audio timbre and the relative intensity of harmonics. We also significantly improve the computation efficiency through a fast transient content detector based on constant-Q transform (CQT), a direct FFT/IFFT implementation of PV and a more efficient mechanism of HMC. Experiments show that the proposed approach is five times faster than a recent hybrid approach [1] and achieves the best bass effect and audio quality.
Keywords :
audio signal processing; fast Fourier transforms; loudspeakers; vocoders; CQT; audio quality; audio timbre; bass enhancement approach; constant-Q transform; direct FFT-IFFT; harmonic magnitude control; harmonics; hybrid virtual bass system; loundspeaker; low frequency bandwidth; nonlinear device approach; phase vocoder; relative intensity; transient content detector; Abstracts; Educational institutions; Entertainment industry; HTML; Transforms; Virtual bass; audio effects; bass enhancement; digital signal processing; nonlinear acoustics; psychoacoustics;
Conference_Titel :
Chinese Spoken Language Processing (ISCSLP), 2014 9th International Symposium on
Conference_Location :
Singapore
DOI :
10.1109/ISCSLP.2014.6936703