Abstract:
Recent studies failed to find effects of frequency modulation (FM) on the segregation of concurrent harmonic sounds, beyond those imputed to induced differences in instantaneous fundamental frequency (F0). This study renews the attempt using a paradigm designed to be sensitive to segregation schemes involving FM. Subjects were presented with stimuli consisting of single vowels (a, e, i, o, u) or pairs of concurrent vowels. Vowels within a pair were different. Mean F0s were either the same or differed by 6%. F0s were either constant, or modulated by an ``n''- or ``u''-shaped cosine waveform. Levels differed by 15 dB to reduce ceiling effects for the weaker vowel and enhance sensitivity. Stimulus duration was 250 ms. The number of vowels reported per stimulus and the identification rate of the weaker vowel were recorded for each condition. Both were greater for conditions that differed in F0 than at unison, but they were not affected by the modulation state of the target (weaker vowel) or the background (stronger vowel). Additional conditions were tested in which both vowels were inharmonic, on the assumption that FM might enhance the fusion of their components. Again no effect of FM was found.