I am working on a website that will be used to teach kids English as a 2nd language. One of the applications we want to develop is the ability to record a students voice, generate a sign wave, and then using the output from that FFT/sign wave determine the accuracy of the speaker's spoken English. We believe we can do this by creating a database of native English words and phrases. The student would then be asked to speak a word or phrase and the recording would be compared with the correctly spoken English in the database.
Our website is under development, we have a mobile application built and are working on the simple database integration piece now.
What we need is a scientist - you - who can figure out how to use the data generated by recording the speaker and comparing it with the recorded voice to provide a real time grade of the spoken english.
I would love to talk to you about this if you are interested.
I put MATLAB as a skill because there is an opensource software called SoundRuler
[url removed, login to view] that may be a good starting point.
If you are up for a challenge let me know and we can begin to figure out how to approach this.