Acoustic characteristics of less-masculine-sounding male speech

This study compared samples of less-masculine-sounding (LMS) and more-masculine-sounding (MMS) male speech to identify acoustic characteristics, other than fundamental frequency, that might contribute to the perception of these categories. In the first phase, audiorecorded speech samples provided by 35 males were presented to 35 female listeners in a paired-comparison perceptual experiment. Nineteen of the male speech samples were judged reliably to fall within the LMS or MMS categories. Within those 19 samples, 8 speakers (4 LMS and 4 MMS) exhibited similar distributions of habitual fundamental frequency values in connected speech and in sustained phonation. In the second phase of the experiment, various acoustic measures of these eight connected speech samples were conducted. Significant differences between measures of fundamental frequency contours, vowel formant midpoint values, and in the first, third and fourth spectral moments of two fricatives were revealed. These findings may be useful in creating stylized synthetic speech that varies on the dimension of masculinity, and they may have clinical relevance for patients wishing to modify the perception of masculinity invoked by their speech.