Data Compression in Microprocessor-Based Data Acquisition Systems

A first-order predictor routine and an encoding routine were implemented on an Intel-8080-based microprocessor system to determine their ability to compress seismic data. The predictor routine obtained a maximum compression ratio of 2.33. It appears that higher order predictors offer no significant advantage over the first-order predictor in compressing seismic data. The encoder routine obtained a larger compression ratio of 3.94. The accuracy of reconstructed seismic traces is bounded by a maximum average error per point. For the predictor, the maximum average error per point is shown to be equal to the specified user tolerance. For the encoder, it is shown to be proportional to the maximum quantization error associated with the ranges used in the encoding algorithm. In general, the average error per point is found to be equal to the square root of the mean-squared error regardless of the compression technique used. A design for a basic data acquisition system utilizing data compression to reduce memory requirements is proposed. Examination of the design indicates that such a system can process no more than three separate data channels at a 1-ms sampling rate and that such a system is cost effective only when small numbers of data channels are to be processed.