Back to list

Untersuchung von Phantomschallquellen vertikaler Lautsprecheranordnungen

Authors Wendt, F.
Year 2013
Thesis Type Master's thesis
Topic Spatial Audio
Keywords audio recording, audio reproduction
Abstract Localization is well studied for single sound sources in the horizontal plane, and inter-aural level and time differences (ILD and ITD) are proven as localization cues. For localization in vertical planes, and especially in the median plane, such cues are weak or absent. Nevertheless, the localization of elevated single sound sources is possible and other cues such as spectral properties play an important role complementing the inter-aural differences. These direction-dependent spectral cues are characterized by the head-related transfer functions (HRTFs). Previous studies showed that the human auditory localization mechanism is mainly sensitive to spectral cues above 4kHz, which are caused by the pinna. A single phantom source is typically created by two loudspeakers playing the same signal. For the horizontal plane, well-known studies explain phantom source localization in terms of ILD and ITD. This is because the superposition of the HRTFs for the left and right loudspeaker yields inter-aural differences resembling those of a single sound source. Phantom sources created by a vertical pair of loudspeakers arranged in the median plane largely deliver correct but indifferent ILD and ITD cues. However, the superimposed HRTFs of the loudspeaker pair do not necessarily yield a spectral shape that is sufficiently similar to one of a single sound source. Still a few studies exist which imply that localization of vertical phantom sources is possible. Therefore, this thesis deals with the localization of vertical phantom sources. To begin with, existing localization models are applied to vertical phantom sources and their functional principle is analyzed and validated. As a next step, a listening test is performed to evaluate the localizability of phantom sources of different vertical loudspeaker setups. Furthermore, the positioning of the phantom source within the loudspeaker array by use of inter-channel level differences is examined in the listening test. In addition to this quantitative approach, a qualitative evaluation of vertical phantom sources is performed. Besides coloration, subjects were asked to evaluate the apparent source width and height of phantom sources created by different loudspeaker arrays. The results are meant to inspire the development of new and suitable loudspeaker arrays for a stable creation of vertical phantom sources.
Supervisors Frank, M., Zotter, F.