A sonogram is a
visualization of a
sound generated by filtering
the sound through a windowed
FFT. If the sound is generated by an
animal, be a bird, human or other, the sonogram looks like a series
of
dark bands layered one on top of one another with varying distances
between them. Another way of describing a sonogram is several similar
functions chopped into little pieces that do not cross each other
drawn on the same
chalkboard in different places (
times) that have
then been drawn more thickly with the edge of a piece of chalk.
Each phoneme a person makes roughly maps onto a
single series of these bands (much like a UPC going vertically),
and when a person changes the phomeme to make a dipthong sound
consisting of two sounds blended together whether it be a
vowel-consonant, vowel-vowel or other, the two different series
of bands blend together as well.
The shape of a sonogram is affected mostly by vowels;
that is /tu/ and /ku/ are "closer" in shape on a sonogram than
/ku/ is to /ki/ or /tu/ to /ti/.
Sonograms are often used in studying and determining the species of birds
(http://research.amnh.org/ornithology/crossbills/diagnosis.html)
and in training people to simulate foreign accents and in speech
pathology. Musicians such as Aphex Twin have also been known to
embed images in sonograms to produce
strange and novel sound effects.
Nodeshell rescue