[x265] [PATCH] bug fix in scalefactor 0

Kavitha Sampath kavitha at multicorewareinc.com
Tue May 8 11:05:59 CEST 2018


I tested locally in my machine and looks like it fixes load/save issues of
tests. We can move to stable branch if the test on default branch run
successfully.


Regards,
Kavitha

On Tue, May 8, 2018 at 1:36 PM, Ashok Kumar Mishra <
ashok at multicorewareinc.com> wrote:

> Is this patch fixed all load/save issues in smoke/regression test and
> should be pushed into stable tip?
>
> On Mon, May 7, 2018 at 5:55 PM, <kavitha at multicorewareinc.com> wrote:
>
>> # HG changeset patch
>> # User Kavitha Sampath <kavitha at multicorewareinc.com>
>> # Date 1525425970 -19800
>> #      Fri May 04 14:56:10 2018 +0530
>> # Node ID b97a65d1227fbe8e80ed15a426ec4bc44537c247
>> # Parent  147b7dcee675a7cd394125788e12909cdce0c373
>> bug fix in scalefactor 0
>>
>> diff -r 147b7dcee675 -r b97a65d1227f source/encoder/analysis.cpp
>> --- a/source/encoder/analysis.cpp       Wed May 02 15:14:30 2018 +0530
>> +++ b/source/encoder/analysis.cpp       Fri May 04 14:56:10 2018 +0530
>> @@ -2500,6 +2500,15 @@
>>                          uint8_t candDir[MRG_MAX_NUM_CANDS];
>>                          mode.cu.getInterMergeCandidates(pu.puAbsPartIdx,
>> part, candMvField, candDir);
>>                          uint8_t mvpIdx = mode.cu.m_mvpIdx[0][pu.puAbsPa
>> rtIdx];
>> +                        if (mode.cu.isBipredRestriction())
>> +                        {
>> +                            /* do not allow bidir merge candidates if PU
>> is smaller than 8x8, drop L1 reference */
>> +                            if (candDir[mvpIdx] == 3)
>> +                            {
>> +                                candDir[mvpIdx] = 1;
>> +                                candMvField[mvpIdx][1].refIdx =
>> REF_NOT_VALID;
>> +                            }
>> +                        }
>>                          mode.cu.setPUInterDir(candDir[mvpIdx],
>> pu.puAbsPartIdx, part);
>>                          mode.cu.setPUMv(0, candMvField[mvpIdx][0].mv,
>> pu.puAbsPartIdx, part);
>>                          mode.cu.setPUMv(1, candMvField[mvpIdx][1].mv,
>> pu.puAbsPartIdx, part);
>>
>> _______________________________________________
>> x265-devel mailing list
>> x265-devel at videolan.org
>> https://mailman.videolan.org/listinfo/x265-devel
>>
>>
>
> _______________________________________________
> x265-devel mailing list
> x265-devel at videolan.org
> https://mailman.videolan.org/listinfo/x265-devel
>
>


-- 
Regards,
Kavitha
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mailman.videolan.org/pipermail/x265-devel/attachments/20180508/1a63a4d6/attachment.html>


More information about the x265-devel mailing list