Abstract: Disclosed is a method of compressing data contained in variable length or universal variable length code words to be carried in a digital bitstream. The method includes formatting a first set of code words for data and then constructing a second set of code words containing code words from the first set as well as concatenations of code words from the first set. Code words from the second set are to carry data in compressed form in the digital bitstream. One way in which the concatenation is accomplished includes selecting a code word from said first set of code words and applying it as a prefix to itself and to all of the other words in the first set, thereby constructing the second set. The code word selected from the first set for application as a prefix may be the shortest code word of the first set.

Abstract: A method of video data compression is provided which is comparatively efficient in compressing data, and at the same time enables transmission of compressed data with sufficient robustness to withstand losses in the transmission channel. A frame of data is divided into coded macroblocks and skipped macroblocks which are grouped into respective slices in a sequence of slices. Respective slices are allowed to contain arbitrary numbers of consecutive macroblocks, and can begin and end at arbitrary positions within the frame. A run-length codeword, comprising a small number of bits, is inserted at the end of a slice to indicate the total number of skipped macroblocks between the final coded macroblock and the end of the slice. To further enhance error resilience an additional codeword, which also indicates the total number of skipped macroblocks, may be inserted at the beginning of the next-following slice.