TO DISPLAY HUMAN VOICE
1. Connect the microphone to the PowerMac, launch
WildSpectra2, and then . . .
2. Say something, like "ee . . . ay . . .ah . . . oh . . . oo".
The emphasized frequencies (or formants) in vowels result from
resonance in the pharyngeal cavity (as modified by the tongue and jaw
positions).
Compare a series of vowels, like those above, with a selection of
consonants. For instance, say the words "Wild
Spectra" (emphasizing the consonants)!
3. Compare an individual's voice at Transform Sizes 256 and 128
Pay close attention to the fine structure of the spectrogram near the
baseline.
Transform size changes the temporal and frequency resolution of the
spectral analysis. With a larger transform size, we have lower
temporal resolution but higher frequency resolution.
With higher temporal resolution (Transform Size = 128), you might find
(especially with male voices) that the opening and closing of the vocal
chords appear as distinct impulses (vertical lines) on the spectrogram.
With lower temporal resolution (Transform Size = 256 or 512), these
impulses instead appear in the frequency domain as a series of narrowly
spaced harmonics (horizontal lines).
Be sure you understand the explanations of spectral analysis in
See Fundamentals of Spectral Analysis.
If you do, you can then explain why measurements of the periodicity of the
vocal chord impulses are the same in both the time domain (1 / time between impulses in secs) and the frequency domain
(distance between harmonics in Hz).