Here we explore the use of the filter class called
A formant is a resonance frequency associtated with a
humans vocal track. (You may think of the idealized human as a
as simply a tube out of which sound emits)
The idea is to filter out of the speech signal
those spectral components which
do not lie in a given range.
In the above example
we are filtering on the first (lowest) formant of the word
"three". In the below we filter out on second formant.
The range selected is defined by the region between
the blue and red horizontal lines with boxes attached.
(The FormantFilter class uses a frequency and a
width, which we have translated to a range) One final
note, you can a normal voice and transform it into
some awful sounding screeches with this application.