Fraunhofer-Gesellschaft

Publica

Hier finden Sie wissenschaftliche Publikationen aus den Fraunhofer-Instituten.

Reproduction of moving sound sources by wave field synthesis - An analysis of artifacts

Wiedergabe sich bewegender Schallquellen mit der Wellenfeldsynthese - Analyse von Artefakten
 
: Franck, A.; Gräfe, A.; Korn, T.; Strauß, M.

Pedersen, J.A. ; Audio Engineering Society -AES-:
DSP for loudspeakers : The proceedings of the AES 32nd international conference. 2007 September 21 - 23, Hillerød, Copenhagen, Denmark
New York, NY: AES, 2007
ISBN: 978-0-937803-60-8
pp.188-196
Audio Engineering Society (International Conference) <32, 2007, Copenhagen>
English
Conference Paper
Fraunhofer IDMT ()
Audiosignal; Digitalsignalverarbeitung; Doppler-Effekt; Schallquelle; Tonwiedergabetechnik; Tonfarbe; Bewegungsartefakt; Regelalgorithmus; Fehlerkorrektur

Abstract
Wave field synthesis is a spatial audio reproduction concept that is based on the synthesis of a wave field over an extended listening area. The reproduction of moving sound sources is an important feature for many areas of application, for instance theatres, cinemas or virtual reality applications. Although several current WFS systems are able to reproduce moving sound sources, the synthesis causes a number of distinct, audible artifacts. In this article the authors classify and describe these artifacts, explain their causes and discuss means to reduce audible deviations. The key question is whether the correct reproduction is possible using the theory underlying the current WFS technology. Time-variant coloration is caused by the spatial aliasing of static WFS sources and therefore not directly caused by the dynamic motion. This effect often degrades audio reproduction quality, but can be handled by either choosing a closer secondary speaker distance or employing one of the methods for the reduction of spatial aliasing proposed in the literature. The digital signal processing for the synthesis of moving sources, namely the fractional delay interpolation, may cause different audible errors. Therefore, much care has to be taken to choose or develop suitable, efficient algorithms. A third, systematic error of the synthesized doppler shift is caused by neglecting the so-called retarded time. It is possible to correct this error algorithmically at the expense of some computational load. But, due to the fact that a quantitative correct doppler shift is not required in most cases, this error may be neglected in many applications. An artifact that results in a spectral broadening of a moving sound source was described and analyzed. This artifact is perceivable as a spectral broadening of the signal and is, depending on the audio material, often disturbing. Supported by a number of simulations, the analysis suggests that this artifact directly follows from the interpretation of the Kirchhoff-Helmholtz integral which is used as the foundation of the WFS theory.
Entnommen aus TEMA

: http://publica.fraunhofer.de/documents/N-82194.html