Spectral analysis of sound has been used in speech therapy and vocology for quite a bit now. It helps understand more about the acoustics of voice, circumstances of voice production and perception of speech and singing. However, all they can normally do is show the distribution of energy in given frequencies. Nowadyas Spectrograms and Sonagrams are no longer tools for experts or expensive studios, but available in many shapes, often in freeware audio studios.
However, there are spectrograms that can be manipulated visually, muting single frequencies, for instance.
"Cockos Spectro" (stillwellaudio) is a good example for that. It COULD BE an instructive tool for voice teachers and therapists. The program is too big though, to be attractive to the mentioned group of users.
I would like to issue a one-track studio, analysing the audio signal in a Songram-way offering a little less of the possibilities Spectro does. Isolating frequency areas and replay them. File or incoming signals through mike.