project did not utilize the power of the supercomputer. We were interested in pursuing a project in the biometric

        identification field, and finally narrowed the possibilities down to voice-recognition.

                Becoming acquainted with the basics of voice-recognition required that we learn more about digital signal

        processing in PCs and the mathematical basis of Fourier Transforms. We originally planned to construct a program which

        took audio input, graphed the input in spectrogram format, and printed a list of tens of thousands of points in the audio
 
        file. Several weeks into the project we came across a program that performed all of these functions, and in the interest of

        time and achieving our basic goal of identifying voices, we opted to use this program.

                The process of identifying a human voice involves running approximately half a million numbers with precision in

        the hundred-thousandths through a complex mathematical equation. Performing such a task on any normal computer

        could take anywhere from four to eight minutes to complete, provided the machine didn’t crash in the process. The

        supercomputer allowed us to perform the enumerated calculations in a fraction of a second.