Simple displays of talker location improve voice identification performance in multitalker, spatialized audio...

The aim of this study was to assess the voice identification benefits of visual depictions of the relative locations of spatialized talkers in a serial listening task.

Although spatialized audio is known to improve speech intelligibility and voice identification accuracy within multitalker environments, prior studies have not found any additional benefit for augmenting spatialized audio with visual depictions of relative voice locations. These studies, however, were restricted to small audio environments (four voices), potentially limiting the ability of simple talker location displays to provide additional identification benefit.

In the first experiment, 18 participants performed a voice identification task for four- and eight-voice environments under three display conditions: (a) nonspatialized voices with an audio-only display, (b) spatialized voices with an audio-only display, and (c) spatialized voices augmented by a visual display of relative talker locations. In the second experiment, 32 participants performed the same voice identification task within a spatialized eight-voice environment but with audio and visual displays of differing angular scale.

Visually depicting relative talker locations improved voice identification performance in terms of both accuracy and response time, particularly for more populous auditory spaces. Both auditory and visual display scale affected these benefits, with large-angle displays performing the best for both modalities.

Results indicate that simple visual representations of spatialized audio environments help listeners identify voices and that these representations are more effective when the angular spacing (auditory and visual) between talker locations is increased.

These results have important implications for the design and implementation of collaborative audio environments for shared, desktop, and portable communication devices.