]> granicus.if.org Git - libvpx/commit
projects / libvpx / commit
? search:
summary | shortlog | log | commit | commitdiff | tree
(parent: a1b7261) | patch
Fixed idct16x16_10 highbitdepth transform
authorPeter de Rivaz <peter.derivaz@gmail.com>
Mon, 10 Nov 2014 16:17:49 +0000 (16:17 +0000)
committerPeter de Rivaz <peter.derivaz@gmail.com>
Mon, 10 Nov 2014 16:17:49 +0000 (16:17 +0000)
commit08d2f548007fd8d6fd41da8ef7fdb488b6485af3
treed054f8ba630979b38185d9321ca6b96c655e00a6tree | snapshot
parenta1b726117f5470f227bc90cd030b7d25045dc510commit | diff
Fixed idct16x16_10 highbitdepth transform

In the case when there are only non-zero coefficients
in the first 4x4 block a special routine is called.
The highbitdepth optimized version of this routine
examined the wrong positions when deciding whether
to call an assembler or C inverse transform.

Change-Id: I62da663ca11775dadb66e402e42f4a1cb1927893
vp9/common/x86/vp9_idct_intrin_sse2.c diff | blob | history
Unnamed repository; edit this file 'description' to name the repository.