Delay translating DCT tokens into coefficients until immediately before IDCT
authorDavid Conrad <lessen42@gmail.com>
Wed, 3 Mar 2010 23:27:43 +0000 (23:27 +0000)
committerDavid Conrad <lessen42@gmail.com>
Wed, 3 Mar 2010 23:27:43 +0000 (23:27 +0000)
commitc72625f29902bfd7f01184c3eb1c29c94ec2a648
tree5b37448d3683813aa90e8d3a6e8a318dd6816fd2
parent7a095ea6fbf0bd26b02a783a0b2f2afd46f58578
Delay translating DCT tokens into coefficients until immediately before IDCT

This is generally around 12% faster than the prior method of creating a
linked list for each block as tokens are read, but can be anywhere from
8% to 28% faster depending on file and CPU.

Originally committed as revision 22190 to svn://svn.ffmpeg.org/ffmpeg/trunk
libavcodec/vp3.c