Challenge Team Interim Report
Team Number: 050
School Name: Lovington High School
Area of Science: Digital Signal Processing/Audio Database Creation
Project Title: Using Unique Human Voice-Traits For Identification Systems
Our project goal is to construct a system which analyzes a human voice and compares it with a data- base of other human voices. Using digital technology, we intend to write a program which accurately matches the voice input to one of the pre-recorded voices in the database. Constructing the data- base of voices will also be one part of our project. Ultimately our finished project would provide for an effective high-security system which would reliably identify a person by using uniquenesses in their voice-pattern.
Gather human voice input from an analog source and using a Fast Fourier Transform (mathematical matrix) convert the input into useful digital code. A database of pre-recorded voices will be stored on a UNIX based machine, not necessarily the SuperComputer itself. We intend to write a program in C or C++ which allow for input by a user, uses a FFT to convert the data into digital form, and then analyzes and compares unique patterns in the input to the voices in the database. The super computer will be needed for the latter tasksdue to the fact that comparing and analyzing the uniquenesses in very large voice files is extensive, complicated work.
We are receiving technical support from one of the nation's leading voice analysts, Norman Perle. Mr. Perle has testified as an expert witness in a number of important trials, including O.J. Simpson and Atlanta Olympic-Park Bombing. He is helping us understand the basics of voice patterns, etc... After getting a grasp of the technical aspects of our project, it will not be extremelly difficult to code the programs which do the work.
Research to this point has led us to understand the basics digital signal processing. Although we are not familiar with the fine technical details of Fast Fourier Transforms (FFTs), we understand their basic structure and how they work, and that they will be the "back-bone" of our project. As mentioned in the Computational Plan, we are receiving help from Mr. Norman Perle. Mr. Perle has aided us in understanding how a project such as ours would work. We feel that we have made much valuable project so far. We feel, however, that we are on the verge of a major breakthrough, and will gain an even greater understanding of Digital Signal Processing and Fast Fourier Trans- forms.
After we complete our project, we optimisticly believe that it will accurately identify any given human, provided that voice is in our database. If we put in quality work on the project, we expect it to show that not only are biometric identification systems effective, but they are fairly efficient and simple. With further work put into our project after the New Mexico Super Computing Challenge, we should be able to advance our project to a professional level which could be used by a variety of customers.