My project is to take the sound input and find the actual present on my computer.
I'm not sure what you mean. Do you want to record a block of sound and see which file matches this block most closely?
That's extremely difficult. How do you define "match"? A simple cross-correlation will give an answer, but only if the sound is the same length, starting at the same point, the same amplitude, AND the same rate, will it be meaningful.
The Gabor Spectrogram (in the JTFA toolkit) will give an amplitude vs. time. vs frequency pattern. Comparing them visually (on the screen) is easy, but comparing them programmatically would still be difficult.
What is it you need to do?
