The psychophysics of concurrent sound segregation

To perceptually separate concurrent complex sounds, normally hearing listeners simultaneously combine information across a wide range of frequency components. Three psychoacoustical experiments are described which investigate different forms of this across-frequency processing. The first two experiments investigate the role of coherence of frequency modulation (FM) between widely separated frequency components of a complex sound. The first experiment bolsters existing evidence that, for harmonic sounds, listeners can discriminate coherent from incoherent FM, but only by detecting the mistuning that arises from incoherent FM. The second demonstrates that, for inharmonic sounds, coherence of FM has no effect on the phenomenon of modulation detection interference (see Moore & Shailer, this symposium) once within-channel cues (combination tones and beating) are masked by background noise. It is concluded that there is not an across-frequency mechanism specific to the detection of FM incoherence. The third experiment investigates the extent to which the detection of mistuning of one component of a harmonic complex is impaired by an interfering sound (the 'interferer') with a frequency spectrum similar to that of the mistuned component. When the interferer is gated on and off with the harmonic complex, it has only a small effect provided that its level is more than 3 dB below that of the target. However, when the interferer starts before and ends after the complex, thresholds are elevated more, and this elevation occurs even for low-level interferers. Explanations of this effect in terms of adaptation and of auditory streaming are discussed.