Letter to the Editor In response to the question, “Why are male voices easier to understand on the telephone than female voices?” the main answer Stevens offered, in brief, is that the spectra of certain consonants have less high-frequency rise for female talkers than those for male talkers. This property certainly could contribute to ... Letter to the Editor
Letter to the Editor  |   November 01, 1993
Letter to the Editor
 
Author Affiliations & Notes
  • Lawrence J. Revit
    Etymonic Design Dorchester, Ontario
Article Information
Speech, Voice & Prosody / Letters to the Editor
Letter to the Editor   |   November 01, 1993
Letter to the Editor
American Journal of Audiology, November 1993, Vol. 2, 68-a. doi:10.1044/1059-0889.0203.68b
 
American Journal of Audiology, November 1993, Vol. 2, 68-a. doi:10.1044/1059-0889.0203.68b
In response to the question, “Why are male voices easier to understand on the telephone than female voices?” the main answer Stevens offered, in brief, is that the spectra of certain consonants have less high-frequency rise for female talkers than those for male talkers. This property certainly could contribute to poorer intelligibility of high-frequency consonants with female talkers, but there is another property, seemingly ignored in the literature, that could be important for vowel identification. That property could be called “spectral density.” (One could think of the same property as “spectral resolution” or “spectral redundancy.”)
Spectral density is related inversely to the pitch, or fundamental frequency, of a voice. In a voice having a fundamental frequency of 100 Hz (typical for males), spectral energy is present at every integer-multiple of 100 Hz. But in a voice having a fundamental frequency of 200 Hz (typical for females), spectral energy is present only at every even integer-multiple of 100 Hz. Thus the spectral density of a male voice is typically twice that of a female voice. This alone could account for a 3 dB lower overall level in female versus male voices. But I hypothesize that, rather than overall level, spectral resolution and/or redundancy might be more the problem. My hypothesis is illustrated in the accompanying figure.
First Page Preview
First page PDF preview
First page PDF preview ×
View Large
Order a Subscription
Pay Per View
Entire American Journal of Audiology content & archive
24-hour access
This Article
24-hour access