[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

pitch in speech

Dear List,
I am studying the quasiperiodic nature of speech and am looking for some
kind comments on whether anyone could shed some light as to the strange
counter intutive results that were indicated. Basically, I am measuring the
wavelength manually by visually identifying the periodicity of the waveform
for some short words e.g. no, go, me etc, using to their horror, family
members for the samples. The variations from the begining of the word to the
end in the worst case was about 25hz based on the wavelength but was much
smaller for words that ended in the vowel 'o', where I would have expected
it to be much larger. To the extent that the variation was this small,
tended to reinforce my belief that the length of the vocal tract does not
change much and therefore it is the shape of the waveform that provides the
preceived variation in pitch for spoken words. But this also led to a
problem in that, the shape of the waveform varies widely between speakers,
though the transition from consonant to vowel seems to be identifiable.
Hence even if the modalities of vision and hearing are assumed to be the
same, then I would have thought that the shape of the input waveforms should
also be the same, unless one reason the shape changes could be due to the
differing basic wavelengths between the speakers. I am using *.wav files
with the inherent PCM coding and am wondering whether this could also
contribute to the variation in the waveforms that I am seeing.
Thanking everybody,
Yours Truly
Randy Randhawa