generic-library/vpx
1
0
Fork 0
Code Issues Pull Requests Releases Wiki Activity

Change to use correct check for halfpel

Browse Source 
In motion estimation stage for subpel motion, subpel variance is
computed use bilinear interpolation. The motion vector precision
used is at 1/8 pel and three bits are used to represent the x and y
subpel offsets. Based on this, the half pel check should be against
4, not 8.

Change-Id: I1f56fa1fa3f2f5e19a20d27983efe628557f170e
This commit is contained in:
Yaowu Xu
2016-05-11 13:52:59 -07:00
parent a45596cff7
commit 17fae3ad0a
1 changed files with 4 additions and 4 deletions
Download Patch File Download Diff File Expand all files Collapse all files

8
vpx_dsp/x86/subpel_variance_sse2.asm
View File

@@ -274,7 +274,7 @@ SECTION .text
STORE_AND_RET
.x_zero_y_nonzero:
cmp y_offsetd, 8
cmp y_offsetd, 4
jne .x_zero_y_nonhalf
; x_offset == 0 && y_offset == 0.5
@@ -478,7 +478,7 @@ SECTION .text
STORE_AND_RET
.x_nonzero:
cmp x_offsetd, 8
cmp x_offsetd, 4
jne .x_nonhalf
; x_offset == 0.5
test y_offsetd, y_offsetd
@@ -546,7 +546,7 @@ SECTION .text
STORE_AND_RET
.x_half_y_nonzero:
cmp y_offsetd, 8
cmp y_offsetd, 4
jne .x_half_y_nonhalf
; x_offset == 0.5 && y_offset == 0.5
@@ -921,7 +921,7 @@ SECTION .text
STORE_AND_RET
.x_nonhalf_y_nonzero:
cmp y_offsetd, 8
cmp y_offsetd, 4
jne .x_nonhalf_y_nonhalf
; x_offset == bilin interpolation && y_offset == 0.5