]> granicus.if.org Git - libvpx/commit
Adjust the forward 16x16 DCT computation steps
authorJingning Han <jingning@google.com>
Mon, 19 May 2014 19:33:40 +0000 (12:33 -0700)
committerJingning Han <jingning@google.com>
Mon, 19 May 2014 19:39:26 +0000 (12:39 -0700)
commit7f547336b787af764c90e01375aa3298c1d82bb7
tree18e059f7871bddbfdca0cb2b9169598e77b0ec89
parent3f96081a15093604f9eca116c9e7a9493db37253
Adjust the forward 16x16 DCT computation steps

This commit adjusts the forward 16x16 DCT computation steps to
simplify the register level operations. It fixes the corresponding
sse2 version accordingly.

Change-Id: I72a9c25b8ca9442fc5e113f47cd701ae55aa7f08
test/dct16x16_test.cc
vp9/encoder/vp9_dct.c
vp9/encoder/x86/vp9_dct_sse2.c