Azimuth and elevation errors in binaural reproduction of ambisonics sound
Abstract:
Binaural decoding of an ambisonic sound is reproducing the information about a soundfield over headphones. It is done based on the spherical harmonics representation of the spatial sound and on the use of Head Related Transfer Function (HRTF). Inaccuracies in the decoding process, which can be caused for example by using non-personalized HRTF, may lead to difficulties in localizing the sound source by the listener. Especially in the elevation plane, localization errors can be significant. In this study, listening tests were conducted for naive listeners in order compare azimuth and elevation errors for first and third-order ambisonics recordings of pink noise bursts recorded in anechoic chamber. The tests involved 16 participants who used headphones to listen to ambisonic sound recordings, specifically bursts of pink noise, captured using Sennheiser Ambeo (1st order) and Zylia (3rd order) microphones. These recordings were then converted to B-format and decoded into binaural format for the listening tests. In the azimuth plane, the highest errors occurred at 0°, indicative of front-back confusions. In contrast, the elevation plane exhibited generally larger errors, with third-order ambisonics resulting in notably higher errors compared to first-order.