EUSIPCO'2002 banner

Paper data
Moving sound image representation method

Matsumoto Mitsuo, Kogakuin University
Yamanaka Susumu, Kogakuin University
Toyama Mikio, Kogakuin University
Nomura Hiroaki, University of East Asia

Page numbers in the proceedings:
Volume III pp 105-108

Multimedia Data Protection / Speech Analysis and Recognition

Paper abstract
This paper introduces an algorithm simulating a moving sound image as a time-variant acoustic event and reports on the results of an investigation of methods for interpolation of binaural responses. As moving sound images require numerous binaural responses, the use of interpolation methods is necessary. We investigated two interpolation methods; one is a linear waveform interpolation method considering of initial time delay difference and the other is a linear interpolation method using the minimum-phase component of a response and an inter-aural time difference model. The methods are evaluated by using a spectrogram of binaural signals simulated by the algorithm. The results showed that the linear interpolation considering the initial time-delay difference resulted in highly precise interpolation and that the inter-aural time difference was maintained in both the methods.

A PDF version is available here