]> granicus.if.org Git - libvpx/commit
projects / libvpx / commit
? search:
summary | shortlog | log | commit | commitdiff | tree
(parent: 0faa8a0) | patch
enable trellis quantization for 2nd order blocks
authorYaowu Xu <yaowu@google.com>
Fri, 1 Oct 2010 03:41:37 +0000 (20:41 -0700)
committerYaowu Xu <yaowu@google.com>
Sat, 2 Oct 2010 13:20:33 +0000 (06:20 -0700)
commitdcd29e369f35055fd39208a230bf64397217dfa6
treed69299ebc977b5a9b12c9e74555617d24412837dtree | snapshot
parent0faa8a0861a827bdaff6d4839bebf7531a891cdacommit | diff
enable trellis quantization for 2nd order blocks

Experimented with different value for Y2_RD_MULT ranging f[1, 32],
without adapting the value to MB coding mode/frame type/Q value,
4 works out best among all values, providing overall 0.1% coding
gain on the test set.

Change-Id: I6b2583a8aa5db5e7e5c65c646301909c0c58f876
vp8/encoder/encodemb.c diff | blob | history
Unnamed repository; edit this file 'description' to name the repository.