Melody recognition using three types of dichotic-pitch stimulus

The recognition of 10 different 16-note melodies, constructed using either dichotic-pitch stimuli or diotic pure-tone stimuli. was measured. The dicho tic pitches were created by placing a frequency-dependent transition in the interaural phase of a noise burst. Three different configurations for the transition were used in order to give Huggins pitch, binaural-edge pitch, a nd binaural-coherence-edge pitch. Forty-nine inexperienced listeners partic ipated. The melodies evoked by the dichotic stimuli were consistently ident ified well in the first block of trials, indicating that the sensation of d ichotic pitch was relatively immediate and did not require prolonged listen ing experience. There were only small improvements across blocks of trials. The mean scores were 97% ( I pure tones), 93% (Huggins pitch), 89% (binaur al-edge pitch), and 77% (binaural-coherence-edge pitch). All pairwise diffe rences were statistically significant, indicating that Huggins pitch was th e most salient of the dichotic pitches and binaural-coherence-edge pitch wa s weakest. To account for these differences in salience, a simulation of la teral inhibition was applied to the recovered spectrum generated by the mod ified equalization cancellation model [J. F. Culling, A. Q. Summerfield, an d D. H. Marshall. J. Acoust. Soc. Am. 103. 3509-3526 (1998)]. The height of the peak in the resulting "edge-enhanced" recovered spectrum reflected the relative strength of the different dichotic pitches. (C) 2001 Acoustical S ociety of America.