Differences in the timbre of sounds in a sequence can affect their perceptu
al organization. Using a performance measure, Hartmann and Johnson (1991) c
oncluded that streaming could be predicted primarily by the extent to which
sounds were passed by different peripheral channels. However, results from
a rating task by Dannenbring and Bregman (1976) suggested that sounds in t
he same spectral region (passed by the same peripheral channels) can be all
ocated to different streams. In Experiment 1, it was found, using an interl
eaved melody task, that target sounds could be selected from distracters in
the same spectral region more easily when they differed in timbre. This fi
nding might result from primitive stream segregation or schema-driven selec
tion, but not from peripheral channeling. In Experiment 2, a rhythm discrim
ination task was used, requiring the sounds to be integrated for good perfo
rmance. Differences in timbre impaired performance, indicating the occurren
ce of primitive stream segregation.