John W. Mullennix
Dept. of Psychol., Wayne State Univ., Detroit, MI 48202
The perception of a synthetic male-female voice continuum was examined using the selective adaptation paradigm. In a series of experiments, voice categorization performance was assessed before and after adaptation with various adaptors. In the first experiment, the results showed that synthetic male and female endpoint adaptors and naturally produced male and female adaptors had similar effects on voice categorization, with voice boundary shifts toward the endpoints. In a second experiment, synthetic adaptors that overlapped with the male endpoint only in F0 or only in formant values were used. Neither of these adaptors had significant perceptual effects. Overall, the results indicate that auditory overlap of the adaptor with the voice continuum endpoint is not necessary or (in the case of isolated acoustic cues related to voice) is not sufficient to produce adaptation. These findings are discussed in terms of the levels of processing and representation used for voice encoding during speech perception.