Hi, all I am studying the ogg-vorbis algorithm but I have less knowledge about the codebook structure of vorbis. After tracing the release source code, I found it doesn't use huffman tree to decode the huffman code instead of using a special mapping and pre-look. The memory requirement seems similar with huffman tree decoding. Could anyone tell me how the vorbis huffman decoding works and the difference or advantage. Or where can I get the information. Thanks... trimp _________________________________________________________________ ¥ß§Y¥Ó½Ð MSN Mobile ªA°È¡G¥Î¤â¾÷©M MSN Messenger ºô¤ÍÀH®É¥æ½Í http://msn.com.tw/msnmobile --- >8 ---- List archives: http://www.xiph.org/archives/ Ogg project homepage: http://www.xiph.org/ogg/ To unsubscribe from this list, send a message to 'vorbis-request@xiph.org' containing only the word 'unsubscribe' in the body. No subject is needed. Unsubscribe messages sent to the list will be ignored/filtered.
Áô©~ªº ¦Ñ¤H wrote:> Hi, all > > I am studying the ogg-vorbis algorithm but I have less knowledge about > the codebook structure of vorbis. After tracing the release source code, > I found it doesn't use huffman tree to decode the huffman code instead > of using a special mapping and pre-look.This is just an optimized decoding algorithm; the Huffman tree is still there, just in a different data structure. [Also, there's a shortcut for very short Huffman words]. > The memory requirement seems similar with huffman tree decoding. Yes, the memory requirements are the same as with some other Huffman decoding algo's, but only half that of the "obvious" algorithm. > Could anyone tell me how the vorbis> huffman decoding works and the difference or advantage. Or where can I > get the information. Thanks...a) This is just how libvorbis works, there's no requirement for a different Vorbis implementation to do it in the same way. b) Read the code, again and again and again... c) If you have more specific questions, ask at vorbis-dev@. But I don't think anyone there will take you by the hand and give you a walkthrough, either; you'll just have to read the code yourself. Sorry... <p>Segher --- >8 ---- List archives: http://www.xiph.org/archives/ Ogg project homepage: http://www.xiph.org/ogg/ To unsubscribe from this list, send a message to 'vorbis-request@xiph.org' containing only the word 'unsubscribe' in the body. No subject is needed. Unsubscribe messages sent to the list will be ignored/filtered.
Áô©~ªº ¦Ñ¤H wrote:> Hi, all > > I am studying the ogg-vorbis algorithm but I have less knowledge about > the codebook structure of vorbis. After tracing the release source code, > I found it doesn't use huffman tree to decode the huffman code instead > of using a special mapping and pre-look. The memory requirement seems > similar with huffman tree decoding. Could anyone tell me how the vorbis > huffman decoding works and the difference or advantage. Or where can I > get the information. Thanks...Oh just a note -- if you are familiar with arithmetic coding: consider the Huffman code as an arithmetic code, with the left subtree of every node equal probability to the right subtree; then everything false into place. <p>Segher --- >8 ---- List archives: http://www.xiph.org/archives/ Ogg project homepage: http://www.xiph.org/ogg/ To unsubscribe from this list, send a message to 'vorbis-request@xiph.org' containing only the word 'unsubscribe' in the body. No subject is needed. Unsubscribe messages sent to the list will be ignored/filtered.