Abstract: A video codec that allows for improved performance of embedded devices using vector quantization and the high correlation between frames in a scene to reduce the overhead associated with displaying video data. Video coding includes determining scene changes in a movie, producing one codebook for each scene, and coding each frame within a scene with the same codebook. The quantized video data is then losslessly compressed and formed into a video stream. Video decoding includes decoding the codebook for a frame and supplying the codebook and frame indices for frame rendering. Alternatively, video coding also includes dividing CPU time between codebook processing, indices processing, and frame rendering to ensure smooth video output.
Abstract: An improved Huffman code deflating method including look-up tables that provide improved performance on limited resource devices, such as hand held gaming devices. The tables are generated from Huffman binary tree data into a primary table and/or one or more secondary tables. For codes having secondary tables, the primary table contains an offset to the secondary table, which contains the information of the secondary table with a minimum number of bits of storage.