ACTA ACUSTICA UNITED WITH ACUSTICA Vol. 103 (2017) 181 – 184 DOI 10.3813/AAA.919045 Subjective Evaluation of Dynamic Voice Directivity for Auralizations Barteld N. J. Postma, Hugo Demontis, Brian F. G. Katz Audio Acoustics Group, LIMSI, CNRS, Rue John von Neumann, Campus Universitaire d’Orsay, 91405 Orsay, France. bart.postma@limsi.fr Summary It is possible to include static directivity patterns into auralizations to better represent the way in which real sources radiate sound and subsequently excite the room. Sources such as the voice have directivity patterns which vary with frequency and also due to dynamic movements of the performer. This study presents an investigation of the inclusion of dynamic voice directivity into room acoustic auralizations based on geometrical acoustics (GA) software. Previous studies have performed this using multi-channel anechoic recordings. In contrast, this study employs single channel anechoic stimuli. Focus is on the incorporation of dynamic orientation presenting the means by which it is included into the GA software as well as the results of a subjective listening test. Results indicate that dynamic voice directivity orientation auralizations are perceived as more plausible, more enveloping, and exhibiting greater apparent source width than auralizations with a static voice directivity and omnidirectional sources. © 2017 The Author(s). Published by S. Hirzel Verlag · EAA. This is an open access article under the terms of the Creative Commons Attribution (CC BY 4.0) license (https://creativecommons.org/licenses/by/4.0/). 1. Introduction The use of room acoustic auralizations has increased due to the improving computing power available and the qual- ity of numerical modelling software. In such auralizations, it is often possible to prescribe the directivity of an acous- tic source in order to better represent the way in which it excites the room. Such directivities are static, being de- fined according to source excitation as a function of fre- quency for the numerical simulation. However, voice di- rectivity varies due to both phoneme dependent radiation patterns [1] and dynamic orientation of the talker/speaker. Studies by Rindel and Otondo [2, 3] proposed to achieve the inclusion of dynamic directivity through the use of multi-channel source directivity auralization 1 . This method employs anechoic multi-channel recordings. The source’s radiation sphere is divided into segments repre- senting each microphone position. The room impulse re- sponse (RIR) is then calculated for each segment and con- volved with the corresponding microphone channel of the anechoic recording. Convolutions of each channel are then down-mixed to create a multi-channel source directivity auralization. These studies compared multi-channel source directiv- ity auralizations to a static directivity source type. The ge- ometrical acoustics (GA) software ODEON was employed Received 28 November 2016, accepted 6 January 2017, published online 17 January 2017. Figure 1. Non-overlapping spatial wedge division approach for partial sources used for multi-channel source directivity aural- izations (from [2]). to create auralizations of an anechoic clarinet recording convolved with 2–, 5–, and 10–channels and a single chan- nel with a prescribed clarinet directivity. Figure 1 depicts the multi-channel sources which were combined, without overlap, to represent the spherical recording area around the source. A listening test judged the 10-channel version significantly less spacious than the three other source rep- resentations. Additionally, subjects significantly preferred the 10-channel auralization over the others in terms of per- ceived naturalness. 1 The original paper termed this application “multi-channel auralization”. In order to avoid confusion with distributed source multi-channel aural- izations, this article will employ the term multi-channel source directivity auralization. © 2017 The Author(s). Published by S. Hirzel Verlag · EAA. This is an open access article under the terms of the CC BY 4.0 license. 181