Subjective evaluation of spatial analysis and synthesis methods using different microphone arrays

Paper: https://ieeexplore.ieee.org/document/9610905

Abstract: Multiple spatial analysis and synthesis methods were proposed up to this date, however, there is still a lack of scientific evidence to conclude how proposed methods and measurement setups compare to each other. This paper presents a subjective study that investigates how spatial analysis and synthesis method affects spatial and timbral fidelities of the binaural renderings for three different stimuli and three source positions. Evaluated positions: (a) +30°, (b) +90°, (c) +135° were motivated by the 4+7+0 loudspeaker layout from the ITU-R BS.2051-2 recommendation. Spatial room impulse responses were measured in the ITU-R BS.1116-compliant listening room using two different microphone arrays. Measurements were used to create auralisations from Spatial Decomposition Method (SDM) and Higher-Order Spatial Impulse Response Rendering (HO-SIRR). The test conditions were selected to investigate: (i) rendering method (ii) direction of arrival (DOA) estimation method, (iii) microphone array, (iv) use of a dedicated centre reference microphone with SDM and (v) band-pass filtering of the spatial room impulse responses prior to the TDOA-based DOA estimation and DOA enforcement for the direct sound. The listening experiment followed MUSHRA-like methodology with a hidden reference, employing a five-grade similarity scale. Subjects’ task was to evaluate how similar the test conditions are to the reference (measured BRIR) with regard to the assessed attribute. This article includes a short review of the existing methods, describes the experimental design and presents the results of the formal listening tests followed by the discussion.