A lossless video coding method is provided. The method comprises: accepting a current input macroblock (MB) of pixel information; subtracting a current prediction MB from the current input MB, creating a current difference MB; entropy encoding the current difference MB with current predictive information; summing the current difference MB with the current prediction MB, creating a current summed MB; buffering the current summed MB, creating a current buffered MB; and, using buffered MBs to generate a subsequent prediction MB and subsequent predictive information, for subsequently accepted input MBs. Generating a subsequent prediction MB and subsequent predictive information includes using either inter-mode or intra-mode coding. The method may further comprise: transmitting the entropy encoded MB; sending a predetermined quantization parameter (QP), signaling that the entropy encoded MB is losslessly encoded; and, losslessly decoding.

 
Web www.patentalert.com

> Quantization and compression of information in a direct acyclic graph

~ 00381