Sequential stream segregation in the absence of spectral cues

This paper investigates the cues used by the auditory system in the percept ual organization of sequential sounds. In particular, the ability to organi ze sounds in the absence of spectral cues is studied. in the first experime nt listeners were presented with a tone sequence ABA ABA..., where the fund amental frequency (f0) of tone A was fixed at 100 Hz and the f0 difference between tones A and B varied across trials between 1 and 11 semitones. Thre e spectral conditions were tested: pure tones, harmonic complexes filtered with a bandpass region between 500 and 2000 Hz, and harmonic complexes filt ered with a bandpass region chosen so that only harmonics above the tenth w ould be passed by the filter, thus severely limiting spectral information. Listeners generally reported that they could segregate tones A and B into t wo separate perceptual streams when the f0 interval exceeded about four sem itones. This was true for all conditions. The second experiment showed that most listeners were better able to recognize a short atonal melody interle aved with random distracting tones when the distracting tones were in an f0 region 11 semitones higher than the melody than when the distracting tones were in the same f0 region. The results were similar for both pure tones a nd complex tones comprising only high, unresolved harmonics'. The results f rom both experiments show that spectral separation is not a necessary condi tion for perceptual stream segregation. This suggests that models of stream segregation that are based solely on spectral properties may require some revision. (C) 1999 Acoustical Society of America. [S0001-4966(99)03501-8].