Spatial audio reproduction systems using loudspeakers produce coloration effects at high frequencies due to spatial interference between loudspeakers, in both those based on panning and those based on field synthesis. As a response to this problem and in order to reduce coloration, this paper studies the feasibility of an alternative approach where high-frequencies are reproduced from a single loudspeaker, with a different direction from that of its panned low-frequency counterpart. Listening tests are conducted to investigate the localization and quality of the source in the case that frequencies higher than 1.5 kHz are reproduced from a different direction than the low frequencies. In this context, the human ability to discriminate the spatial direction of low/high frequency bands and the error in the perceived direction of arrival for different separation angles is evaluated and quantified. The resulting data has been analyzed with ANOVA, providing significant results that allow us to establish a threshold in the angular separation of the high and low frequency parts where subjects do not perceive source location artifacts. The term just noticeable band splitting angle (JNBSA) is defined and introduced. It represents the minimum angle of separation between high and low frequencies from which the listener starts to perceive artifacts in the reproduction of a sound source using loudspeakers.
- Spatial acuity
- Spatial audio
- Vector Base Amplitude Panning (VBAP)
- Wave-field synthesis