• Bounded Error Data Compression at Constant Rate

      Gonsalves, Robert A.; Evans, Norman E.; Shea, Alicia; EIKONIX Corporation (International Foundation for Telemetering, 1978-11)
      We present a technique to compress M-bit data to N-bit data with a bounded error between the original and de-compressed data. The bounded error is established by quantizing the error signal of DPCM and by entropy encoding the quantized error. The fixed data rate is maintained by non-linear adaptation of the quantization interval on an intermittent basis. Examples of image data compression are given.