OK I will bite!
Just shooting from the hip, how about a microphone, amplifier to give a volt or so output, and then feed the amplifier outputs into a few 567 tone decoders, each set to a different frequency.
When the tone decoder detects the frequency, it lights an LED.
Second idea, microphone, amplifier, frequency to voltage converter, and use the output of the F to V converter to drive a bargraph display using something like an LM3914.
Third idea. I cant help but think that you may find these solutions very limited in practical applications for "voice therapy".
Consider using some kind of audio spectrum analyser such as "Spectrogram"
This may be taking you in a direction which you dont want to go, but I think it has advantages.
JimB