Abstract: A decoder (100) is disclosed, which selectively reconstructs at least one predetermined data unit out of a coded bit stream (c). The decoder (100) includes a parsing unit (120) to search and parse unambiguous information (up) characterizing the at least one predetermined data unit uniquely. Next, the parsing unit (120) provides this parsed information (up) either directly to a decoding unit (130) or to a look-up table (140) such that the decoding unit (130) could reconstruct the predetermined data unit directly out of the coded bit stream (c) based on the parsed information (up) without decoding and reconstructing the coded bit stream (c) completely. Optionally, a specification unit (110) provides an start position (us) of the coded bit stream (c) to the parsing unit (120) used for starting the search. Therewith, data traffic, computation time and memory resources could be saved for decoding and displaying e.g. compressed JPEG pictures.

Abstract: A decoder (100) is disclosed, which selectively reconstructs at least one predetermined data unit out of a coded bit stream (c). The decoder (100) includes a parsing unit (120) to search and parse unambiguous information (up) characterizing the at least one predetermined data unit uniquely. Next, the parsing unit (120) provides this parsed information (up) either directly to a decoding unit (130) or to a look-up table (140) such that the decoding unit (130) could reconstruct the predetermined data unit directly out of the coded bit stream (c) based on the parsed information (up) without decoding and reconstructing the coded bit stream (c) completely. Optionally, a specification unit (110) provides an start position (us) of the coded bit stream (c) to the parsing unit (120) used for starting the search. Therewith, data traffic, computation time and memory resources could be saved for decoding and displaying e.g. compressed JPEG pictures.

Abstract: In an arrangement for transforming picture data while using a method of two-dimensional hierarchic sub-band segmentation, particularly a two-dimensional Discrete Wavelet Transformation (2-D DWT), in which pictures, possibly in a plurality of planes, are segmented into sub-bands by means of filtering, a minimal storage space, a minimal bus load and an optimal 2-D DWT transformation ensure that a picture memory (41) is provided in which the data of a picture are stored prior to the start of the first transformation plane and in which the data of a further sub-band to be transformed are stored after the first transformation plane, in which process the data of the picture are partially overwritten, in that the data of a possible further sub-band to be segmented are stored in the picture memory (41) after every further transformation plane, in that a sub-band memory (44) is provided in which, after a transformation process of a plane, the data gained during this transformation of those sub-bands which are no longe

Abstract: A memory-efficient trellis coded quantizer divides consecutive trellis columns into groups or windows each containing the maximum number of columns for which each pair of columns in a window still retains statistical dependence with respect to the states of the column and in relation to any trellis path connecting the two columns. Back searching for code words operates on each individual window, rather than on the overall trellis. Windows are sequentially processed in pairs to derive the code words for the input that has been trellis coded. Required storage is reduced, since trellis and code word storage is required only for the windows currently being processed.