From: Yunqing Wang <yunqingwang@google.com>
Date: Tue, 4 Aug 2015 19:16:47 +0000 (-0700)
Subject: Minor adjustment in diagonal sub-pixel point checking
X-Git-Tag: v1.5.0~327^2
X-Git-Url: https://granicus.if.org/sourcecode?a=commitdiff_plain;h=726d1b841b8101fbfa75eb9832e9baa9e477c1d5;p=libvpx

Minor adjustment in diagonal sub-pixel point checking

Choose a different diagonal point to check when the two costs are
the same, making it consistent with the way we choose the best mv.
This slightly changes the encoding result, and the derflr set borg
test at speed 0 shows 0.027% Overall PSNR gain, 0.024% Avg PSNR
gain, and 0.043% SSIM gain.

Change-Id: Ic8ee3a6767394866d159e4f9e1c777604dd73c17
---

diff --git a/vp9/encoder/vp9_mcomp.c b/vp9/encoder/vp9_mcomp.c
index 6e0c561c3..63a35526b 100644
--- a/vp9/encoder/vp9_mcomp.c
+++ b/vp9/encoder/vp9_mcomp.c
@@ -703,8 +703,8 @@ int vp9_find_best_sub_pixel_tree(const MACROBLOCK *x,
     }
 
     // Check diagonal sub-pixel position
-    tc = bc + (cost_array[0] < cost_array[1] ? -hstep : hstep);
-    tr = br + (cost_array[2] < cost_array[3] ? -hstep : hstep);
+    tc = bc + (cost_array[0] <= cost_array[1] ? -hstep : hstep);
+    tr = br + (cost_array[2] <= cost_array[3] ? -hstep : hstep);
     if (tc >= minc && tc <= maxc && tr >= minr && tr <= maxr) {
       const uint8_t *const pre_address = y + (tr >> 3) * y_stride + (tc >> 3);
       MV this_mv = {tr, tc};