diff options
author | John Koleszar <jkoleszar@google.com> | 2013-04-15 13:18:24 -0700 |
---|---|---|
committer | John Koleszar <jkoleszar@google.com> | 2013-04-18 17:47:17 -0700 |
commit | 2987fa1dc130256562e661bafb0136ac9b704abf (patch) | |
tree | 310eae4de869a0ec216b66d1e70ab82e387b217c /vp9/common/x86/vp9_mask_sse3.asm | |
parent | 4924934d2bf860a13ebd8b3f28fe78ea7d9ccb25 (diff) | |
download | libvpx-2987fa1dc130256562e661bafb0136ac9b704abf.tar libvpx-2987fa1dc130256562e661bafb0136ac9b704abf.tar.gz libvpx-2987fa1dc130256562e661bafb0136ac9b704abf.tar.bz2 libvpx-2987fa1dc130256562e661bafb0136ac9b704abf.zip |
Removing rounding from UV MV calculation
Consider the previous behavior for the MV 1 3/8 (11/8 pel). In the
existing code, the fractional part of the MV is considered separately,
and rounded is applied, giving a result of 6/8. Rounding is not required
in this case, as we're increasing the precision from a q3 to a q4, and
the correct value 11/16 can be represented exactly.
Slight gain observed (+.033 average on derf)
Change-Id: I320e160e8b12f1dd66aa0ce7966b5088870fe9f8
Diffstat (limited to 'vp9/common/x86/vp9_mask_sse3.asm')
0 files changed, 0 insertions, 0 deletions