Abstract: A waveform classification system uses a detector 10 for detecting an analog input x(t), as well as sampler 14 and quantizer 16 to convert this analog input into a series of digital values {J.sub.k-1, J.sub.k, J.sub.k+1 }. This series of digital values {J.sub.k-1, J.sub.k, J.sub.k+1 } is represented by a fixed number of quantized digital values. A collection of predetermined feasible values is held in a feasible set table 20. A waveform classifier 18 compares the digital series from the quantizer 16 to the predetermined series held in the feasible set table 20. The waveform classifier 18 has an output, f(Z.sub.k), that marks the data in the digital series as either high or low interest. Low interest data can either be eliminated or reduced in priority to improve the overall system performance.