Modeling auditory grouping based on a temporal coherence analysis

Current models of auditory streaming rely primarily on frequency separation for sound segregation. However, spectral components that are well separated in frequency are no longer heard as separate streams if presented synchronously rather than consecutively. Elhilali et al. [1] suggested a conceptual model to account for grouping based on synchrony, but the model was not evaluated with experimental data.

In this study, it was experimentally tested how the temporal overlap between spectral components affected the perception of one or two streams. This was investigated for a range of tone-repetition rates. The data suggested that the perceptual organization depends on the absolute asynchrony of the tones.

When the asynchrony of the tones was less than 20-44 ms, the tones were organized into the same perceptual stream, and when the asynchrony was larger, two separate streams were perceived. The conceptual model performed well in the tested conditions, however, some issues associated with the peripheral processing stages in the model were observed, caused by the frequency dependent delay of the filters.

An alternative peripheral model is thus required for the conceptual model to function properly.