[x265] uhdbd: fix sourceHeight check

Deepthi Nandakumar deepthi at multicorewareinc.com
Fri Jun 3 08:36:53 CEST 2016


# HG changeset patch
# User Mahesh Pittala <mahesh at multicorewareinc.com>
# Date 1461582611 -19800
#      Mon Apr 25 16:40:11 2016 +0530
# Node ID f90f359eb770652e822e3a813c0b9eaf47036122
# Parent  91de20bfe2affa6e225681be129208e654c2ba48
uhdbd: fix sourceHeight check

when sourceHeight is 1080, it is padded by some value to become multiple of
the
minimum CU size(for ex- 1088) so just moved down those calculations after
uhd-bd.

diff -r 91de20bfe2af -r f90f359eb770 source/encoder/encoder.cpp
--- a/source/encoder/encoder.cpp Fri Jun 03 12:02:05 2016 +0530
+++ b/source/encoder/encoder.cpp Mon Apr 25 16:40:11 2016 +0530
@@ -1845,16 +1845,6 @@
         m_conformanceWindow.rightOffset = padsize;
     }

-    /* set pad size if height is not multiple of the minimum CU size */
-    if (p->sourceHeight & (p->minCUSize - 1))
-    {
-        uint32_t rem = p->sourceHeight & (p->minCUSize - 1);
-        uint32_t padsize = p->minCUSize - rem;
-        p->sourceHeight += padsize;
-
-        m_conformanceWindow.bEnabled = true;
-        m_conformanceWindow.bottomOffset = padsize;
-    }
     if (p->bEnableRdRefine && (p->rdLevel < 5 || !p->rc.aqMode))
     {
         p->bEnableRdRefine = false;
@@ -1974,6 +1964,15 @@
         }
     }

+    /* set pad size if height is not multiple of the minimum CU size */
+    if (p->sourceHeight & (p->minCUSize - 1))
+    {
+        uint32_t rem = p->sourceHeight & (p->minCUSize - 1);
+        uint32_t padsize = p->minCUSize - rem;
+        p->sourceHeight += padsize;
+        m_conformanceWindow.bEnabled = true;
+        m_conformanceWindow.bottomOffset = padsize;
+    }

     if (p->bLogCuStats)
         x265_log(p, X265_LOG_WARNING, "--cu-stats option is now
deprecated\n");


-- 
Deepthi Nandakumar
Engineering Manager, x265
Multicoreware, Inc
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mailman.videolan.org/pipermail/x265-devel/attachments/20160603/321a9b4f/attachment-0001.html>


More information about the x265-devel mailing list