By Topic

Adaptive predictive run-length encoding for analogue sources

Sign In

Cookies must be enabled to login.After enabling cookies , please use refresh or reload or ctrl+f5 on the browser for the login options.

Formats Non-Member Member
$31 $31
Learn how you can qualify for the best price for this item!
Become an IEEE Member or Subscribe to
IEEE Xplore for exclusive pricing!
close button

puzzle piece

IEEE membership options for an individual and IEEE Xplore subscriptions for an organization offer the most affordable access to essential journal articles, conference papers, standards, eBooks, and eLearning courses.

Learn more about:

IEEE membership

IEEE Xplore subscriptions

1 Author(s)
Mark, J.W. ; University of Waterloo, Department of Electrical Engineering & Computer Communications Networks Group, Waterloo, Canada

An adaptive predictive run-length encoding scheme in which the predictor, the quantiser and the run-length encoder are coupled together in their adaptive operation is described. The static quantiser parameters are chosen to produce a prescribed probability of occurrence of the quantiser output levels. Run-length encoding results in the production of a codeword at a variable rate. To maintain a constant transmission rate, a fixed-size buffer is employed for data storage. Dynamic sensing of the buffer content to alter the actual quantiser output levels provides an effective means of avoiding buffer overflow and underflow. The zero output level can severely limit the system operation, particularly when the number of quantising levels is only three or five. A process, which has been termed dithering, is introduced to overcome this shortcoming by making the actual (2L + 1)-level quantising system operate at an effective (2L + 2)-level whenever the first dynamic threshold t0(n) is greater than a design acceptable tolerance t¿. Dithering improves system convergence, particularly for the 3- or 5-level quantiser cases. The transmission rate for a 3-bit codeword run-length encoding and a quantiser output distribution to give a design probability pD = pr(vn = 0)=0.85 is approximately 0.7 [log2 (2L+1)] bits per source sample, where [.] denotes the integer greater than or equal to the argument.

Published in:

Electrical Engineers, Proceedings of the Institution of  (Volume:123 ,  Issue: 11 )