DocumentCode
3354512
Title
Efficient data flow variable length decoding implementation for the MPEG reconfigurable video coding framework
Author
Li, Jianjun ; Ding, Dandan ; Lucarz, Christophe ; Keller, Samuel ; Mattavelli, Marco
Author_Institution
Microelectron. Syst. Lab. (GR-LSM), Ecole Polytech. Fed. de Lausanne, Lausanne
fYear
2008
fDate
8-10 Oct. 2008
Firstpage
188
Lastpage
193
Abstract
In 2004, ISO/IEC SC29 better known as MPEG started a new standard initiative aiming at facilitating the deployment of multi-format video codec design and to enable the possibility of reconfiguring video codecs using a library of standard components. The new standard under development is called MPEG Reconfigurable Video Coding (RVC) framework. Whereas video coding tools are specified in the RVC library, when a new decoder is reconfigured choosing in principle any (sub)-set of tools, the corresponding bitstream syntax, described using MPEG-21 BSDL schema, and the associated parser need to be respectively derived and instantiated reconfiguration by reconfiguration. Therefore, the development of an efficient systematic procedure able to instantiate efficient bitstream parsing and particularly variable length decoding is an important component in RVC. This paper introduces an efficient data flow based implementation of the variable length decoding (VLD) process particularly adapted for the instantiation and synthesis of CAL parsers in the MPEG RVC framework.
Keywords
variable length codes; video codecs; video coding; ISO/IEC SC29; MPEG; bitstream parsing; data flow variable length decoding; multiformat video codec design; reconfigurable video coding; Code standards; Decoding; IEC standards; ISO standards; Libraries; Standardization; Standards development; Video codecs; Video coding; Video sharing; Bitstream Syntax Description Language; CAL language; Huffman coding; Reconfigurable Video Coding Variable Length Decoding;
fLanguage
English
Publisher
ieee
Conference_Titel
Signal Processing Systems, 2008. SiPS 2008. IEEE Workshop on
Conference_Location
Washington, DC
ISSN
1520-6130
Print_ISBN
978-1-4244-2923-3
Electronic_ISBN
1520-6130
Type
conf
DOI
10.1109/SIPS.2008.4671760
Filename
4671760
Link To Document