Title :
Quantized predictive coding
Author :
Lee, Don H. ; Neuhoff, David L.
Author_Institution :
Dept. of Electr. Eng. & Comput. Sci., Michigan Univ., Ann Arbor, MI, USA
Abstract :
A new form of predictive quantization, dubbed quantized predictive coding (QPC), is presented. One version uses variable-length coding and another uses fixed-length coding. For autoregressive sources, it is shown both analytically and experimentally that the performance of QPC with (without) entropy-coding is, essentially, the same as that of DPCM (differential pulse code modulation) with uniform quantization with (without) entropy-coding. The principal advantage of QPC is that, unlike DPCM, it is a priori designed for digital implementation. As a result, its digital implementation is simpler than DPCM and suffers no loss in performance due to round-off errors
Keywords :
encoding; filtering and prediction theory; roundoff errors; DPCM; QPC; autoregressive sources; digital implementation; entropy-coding; fixed-length coding; predictive quantization; quantized predictive coding; round-off errors; variable-length coding; Costs; Decoding; Fixed-point arithmetic; Performance analysis; Performance loss; Predictive coding; Pulse modulation; Quantization; Redundancy; Roundoff errors;
Conference_Titel :
Circuits and Systems, 1990., Proceedings of the 33rd Midwest Symposium on
Conference_Location :
Calgary, Alta.
Print_ISBN :
0-7803-0081-5
DOI :
10.1109/MWSCAS.1990.140777