Listener Discrimination Between Common Speaker-Based 3D Audio Reproduction Formats

Perceptual Comparison of 3D Audio Reproduction With and Without Bottom Channels

“…Differences related to envelopment/immersion/spaciousness were also common across all three groups, although reported comparatively less frequently. These results are somewhat comparable with those of a previous study by a subset of the current authors [22] examining listener discrimination between 9+10+3, 4+7+0, 3+7+0, and 4+5+0 audio reproduction formats. In that study, differences related to vertical imaging and apparent source width, a similar concept to "horizontal image spread," were mentioned by many listeners as being important factors in discriminating between the various formats, with the vertical imaging being a primary cue for when the bottom channels (9+10+3) were active.…”

“…When comparing reproduction systems of varying numbers or layers of loudspeakers, one typically begins by recording/mixing a given sound scene for the largest reproduction condition under test and then either actively remixing for all other formats under test, as seen in studies by Francombe et al [11] and Howie et al [22], or using a downmixing scheme or algorithm to create additional stimuli, as carried out in studies by Sugimoto et al [23] and Ando [24].…”

“…Global gain adjustments were then applied to each multichannel audio file until playback of all stimuli was level-matched to within 0.1 LUFS of each other, within each sound scene. This method has been used effectively in several previous studies comparing 3D audio stimuli [22,[26][27][28]. The levels of each sound scene group were then adjusted relative to each other to make for a comfortable and balanced listening experience throughout the test.…”

“…These comments were then searched for terms or synonyms of terms common to subjective spatial audio evaluation. Terms with similar or identical meanings, such as "low end," and "low frequencies," were pooled together, referencing lists of attributes from previous work [36][37][38] and a thesaurus, a method adopted from [22]. The counts for identical or similar terms (e.g., "envelopment," "enveloping" ) were then summed, per group.…”