We believe that individuals with reading problems can learn to self resolve them by interacting with a website (or standalone application) that allows them to compare their oral reading abilities with a synthesized or pre-recorded voice model and their own speaking style. The student would log on to the website and use mouse and a headset/microphone interface with the computer's sound card.
Proposed scenario: Student clicks on a green button
Step 1-Student hears a three word sentence by synthesized or pre-recorded voice (V) and is asked to repeat the sentence into the microphone. Student's vocal input is saved.
Step 2-Student is presented with the same sentence in text form along with V reading of it. V instructs student: "Please try to read the sentence silently. If after reading it silently, you would like to try reading it aloud, click the green button and begin. If you would like to hear the sentence again, click the red button."
Step 3-If student clicks green, this vocal input would be saved. The student then hears V say: "This is your voice saying the words" (Playback) "This is your voice reading the words." (Playback) "Do they sound the same? Were you comfortable reading the sentence?" If the answer is yes, click on the green button. If the answer is no, click on the red button."
Step 4a-If the student clicks red, the V repeats the sentence material and the instruction: "Please try to read the sentence silently. If after reading it silently, you would like to try reading it aloud, click the green button and begin. If you would like to hear the sentence again, click the red button."
Step 4b-If the student clicks green, the student hears another three word sentence as in Step 1, and Steps 1-4 follow.
When the student chooses Step 4b ten times without choosing step 4a, the sentence length is increased to four words, with a one word increase every time this criterion is met.
I believe this design could be coded in such a way to make it easy to localize to different languages. It might also have application in teaching languages other than the subject's first language.
The current version is written in PERL as a stand-alone application. This version is highly alpha and not considered ready for general use. I believe that further development would best be done in Java, but I am not a Java programmer.
It has been suggested that the speex codec would be well suited to this application, and there is a java implementation of this codec (http://jspeex.sourceforge.net/). But I would still need help in using this code.
It was also suggested that a browser based implementation with a java plugin would have trouble accessing the local hard drive to write the generated sound files, and that using a "web start" type application would avoid this problem.
Sourceforge Project Page/Download Source Code Here
Contact Webmaster/Project Administrator