Data compression

Abstract

A study of data compression techniques involving linear interpolation and linear prediction showed that redundancy is a problem that can be significantly reduced by various polynomial approximations. A more recent compressor, the continuous secant compressor which determines the optimum sampling interval prior to sampling, was found to be the most efficient compressor examined. The continuous secant compressor bases its reduction technique on a straight-line approximation. Data compression results when the system in question does not occupy its entire bandwidth. The addition of white noise over the entire bandwidth was found to reduce the efficiency of the continuous secant compressor by only a small amount. The probability distribution of the straight-line approximation in the presence of noise had a gaussian distribution and a relatively small standard deviation.