I am trying to develop an application which will record voice and plot a graph accordingly.
Following is the output which I am expecting.
When User1 records a sound say "Hello" system should generate a graph say "Graph A". Now when User2 records same sound as "Hello" system should generate "Graph B" i.e (Graph A != Graph B). Now again User1 records same sound as "Hello" system should generate same graph "Graph A".
I am using java sampling algorithm for this, but sound recorded second time always vary than recorded previously (may be because of background noise, natural variations in voice or other factors).
My questions are 1. Is it really doable? 2. Is there any built in library in JAVA to achieve the output.
I was playing a little bit with simple sound compression/recognition libraries and I have come across this
http://code.google.com/p/musicg-sound-api/downloads/list
You can record sound to wav format and then compare two (or more) wav files getting similatity level (%)