I have developed over more than 10 years a Java applet that delivers language learning exercises with intelligent feedback, and what I badly need now is an audio-fingerprinting module that would work as follows:
- A student is asked to read a list of words or phrases, one by one into the microphone.
- The student then does an exercise in which the correct answer to a question will be one of the words or phrases previously recorded. The student must say the answer into the microphone.
- The module must audio-record the original list of words and phrases, then the answers spoken by the student.
- It must compare the sound file produced in the exercise with the sound files originally recorded by the student, and determine which of the words or phrases from the initial list the student has used in the answer. If the answer is correct, or incorrect, appropriate action will be taken, just as if the answer had been typed in and compared as a character string.
Do you know of anything like this already done in Java and available for incorporation into an applet such as mine?