]> granicus.if.org Git - libvpx/commit
Bug fix in motion vector limit calculation
authorYunqing Wang <yunqingwang@google.com>
Thu, 30 Jun 2011 15:20:13 +0000 (11:20 -0400)
committerYunqing Wang <yunqingwang@google.com>
Thu, 30 Jun 2011 15:20:13 +0000 (11:20 -0400)
commitb7480454706a6b15bf091e659cd6227ab373c1a6
treebc695dcfaf24572130f10be7a0fb64ab8073491d
parent034cea5e726b851294baad4d77e81bb6ce45703c
Bug fix in motion vector limit calculation

Motion vector limits are calculated using right shifts, which
could give wrong results for negative numbers. James Berry's
test on one clip showed encoder produced some artifacts. This
change fixed that.

Change-Id: I035fc02280b10455b7f6eb388f7c2e33b796b018
vp8/encoder/mcomp.c
vp8/encoder/mcomp.h
vp8/encoder/pickinter.c
vp8/encoder/rdopt.c