Difference between revisions of "YLC"

From MultimediaWiki
Jump to navigation Jump to search
 
Line 1: Line 1:
* Website: http://spring-fragrance.mints.ne.jp/aviutl/ (Japanese)
* Website: http://spring-fragrance.mints.ne.jp/aviutl/ (Japanese)
* FourCC: YLC0
* FourCC: YLC0
* Sample: http://samples.libav.org/V-codecs/ylc0.avi


The website names it as "YUY2 Lossless Codec".
The website names it as "YUY2 Lossless Codec".
=== Format ===
bytes  0- 3 'YLC0'
bytes  4- 7 zero
bytes  8-11 offset to luma plane (should be 16)
bytes 12-15 offset to chroma plane
Bitstream is read as 32-bit little-endian words MSB.
Luma plane contains residues stored as exp-Golomb codes (unary-coded exponent + N bits of mantiss).
Chroma plane contains chroma values coded in groups and uses static Huffman trees:
  if (get_bit()) {
      val = decode_sym(tree1);
      if (val < 0xE1) {
          output chroma_table[val];
      } else {
          for (i = 0; i < (val - 0xDF); i++)
              output {0, 0, 0, 0}
      }
  } else {
      u0 = decode_sym(tree2);
      v0 = decode_sym(tree3);
      u1 = decode_sym(tree2);
      v1 = decode_sym(tree4);
      output { u0, v0, u0 + u1, v1 }
  }
Looks like for luma plane median prediction is used and for chroma only left prediction is used.


[[Category:Video Codecs]]
[[Category:Video Codecs]]
[[Category:Lossless Video Codecs]]
[[Category:Lossless Video Codecs]]
[[Category:Undiscovered Video Codecs]]
[[Category:Incomplete Video Codecs]]

Latest revision as of 04:14, 24 June 2012

The website names it as "YUY2 Lossless Codec".

Format

bytes  0- 3 'YLC0'
bytes  4- 7 zero
bytes  8-11 offset to luma plane (should be 16)
bytes 12-15 offset to chroma plane

Bitstream is read as 32-bit little-endian words MSB.

Luma plane contains residues stored as exp-Golomb codes (unary-coded exponent + N bits of mantiss).

Chroma plane contains chroma values coded in groups and uses static Huffman trees:

 if (get_bit()) {
     val = decode_sym(tree1);
     if (val < 0xE1) {
         output chroma_table[val];
     } else {
         for (i = 0; i < (val - 0xDF); i++)
             output {0, 0, 0, 0}
     }
 } else {
     u0 = decode_sym(tree2);
     v0 = decode_sym(tree3);
     u1 = decode_sym(tree2);
     v1 = decode_sym(tree4);

     output { u0, v0, u0 + u1, v1 }
 }

Looks like for luma plane median prediction is used and for chroma only left prediction is used.