The Use of Asymmetric Numeral Systems Entropy Encoding in Video Compression

Abstract

Modern image and video compression technologies includes many compression methods both lossless and lossy. Entropy encoding has a special place among lossless compression methods - it performs final data compression before its conversion to output bitstream. Huffman coding was used for these purposes, but arithmetic coding is widely used nowadays. This algorithm allows to achieve higher compression rate than Huffman coding, however, it is inferior in speed. New method called ANS (Asymmetric Numeral Systems) was proposed a few years ago. This statistical coding algorithm is comparable in terms of compression ratio with arithmetic coder and in speed with Huffman coder. ANS is currently used in such data compressors as Zstandard (Google), LZFSE (Apple), DivANS (Dropbox), etc. In this paper, it is proposed to use ANS in wavelet video codecs. The research and development of an effective compression method of wavelet transform coefficients with usage of ANS entropy encoding was performed.