[FFmpeg-cvslog] avfilter/vf_deshake: fix loss of precission with odd resolutions

Michael Niedermayer git at videolan.org
Mon Jun 23 18:52:53 CEST 2014


ffmpeg | branch: release/1.2 | Michael Niedermayer <michaelni at gmx.at> | Wed Apr 16 02:06:37 2014 +0200| [8d60d8bef228321a86cc62a9a2e0de975dbe78d7] | committer: Michael Niedermayer

avfilter/vf_deshake: fix loss of precission with odd resolutions

Fixes part of Ticket3466
Found-by: Andrey_Karpov / PVS-Studio
Signed-off-by: Michael Niedermayer <michaelni at gmx.at>
(cherry picked from commit 73734282e0e4df92269984ee1671424e39249481)

Signed-off-by: Michael Niedermayer <michaelni at gmx.at>

> http://git.videolan.org/gitweb.cgi/ffmpeg.git/?a=commit;h=8d60d8bef228321a86cc62a9a2e0de975dbe78d7
---

 libavfilter/vf_deshake.c |    4 ++--
 1 file changed, 2 insertions(+), 2 deletions(-)

diff --git a/libavfilter/vf_deshake.c b/libavfilter/vf_deshake.c
index 26f92f4..f9e42ac 100644
--- a/libavfilter/vf_deshake.c
+++ b/libavfilter/vf_deshake.c
@@ -346,8 +346,8 @@ static void find_motion(DeshakeContext *deshake, uint8_t *src1, uint8_t *src2,
         //av_log(NULL, AV_LOG_ERROR, "\n");
     }
 
-    p_x = (center_x - width / 2);
-    p_y = (center_y - height / 2);
+    p_x = (center_x - width / 2.0);
+    p_y = (center_y - height / 2.0);
     t->vector.x += (cos(t->angle)-1)*p_x  - sin(t->angle)*p_y;
     t->vector.y += sin(t->angle)*p_x  + (cos(t->angle)-1)*p_y;
 



More information about the ffmpeg-cvslog mailing list