<div dir="ltr">Min, Now we have done the first improvement in SAO. The second one I believe we can write only two functions, <div>one for BO statistics and one for EO statistics using some data manipulation. So that we have<div>only two primitives for BO and EO. No need to write separate functions for each EO. Though it may not give</div><div>significant performance but the code will be in very good shape.</div></div></div><div class="gmail_extra"><br><div class="gmail_quote">On Wed, Oct 7, 2015 at 4:25 AM, Min Chen <span dir="ltr"><<a href="mailto:chenm003@163.com" target="_blank">chenm003@163.com</a>></span> wrote:<br><blockquote class="gmail_quote" style="margin:0 0 0 .8ex;border-left:1px #ccc solid;padding-left:1ex"># HG changeset patch<br>
# User Min Chen <<a href="mailto:chenm003@163.com">chenm003@163.com</a>><br>
# Date 1444167726 18000<br>
# Node ID 36a54b2cf7c4c96067bafb67077651d30d83e8e9<br>
# Parent 8fdd1b8fd4529b6966ab787f6c624f6056f77593<br>
improve sao diff[] by primivites.sub_ps<br>
---<br>
source/encoder/sao.cpp | 23 +++++++++++++++++------<br>
source/encoder/sao.h | 1 +<br>
2 files changed, 18 insertions(+), 6 deletions(-)<br>
<br>
diff -r 8fdd1b8fd452 -r 36a54b2cf7c4 source/encoder/sao.cpp<br>
--- a/source/encoder/sao.cpp Tue Oct 06 16:42:02 2015 -0500<br>
+++ b/source/encoder/sao.cpp Tue Oct 06 16:42:06 2015 -0500<br>
@@ -106,6 +106,7 @@<br>
bool SAO::create(x265_param* param)<br>
{<br>
m_param = param;<br>
+ m_chromaFormat = param->internalCsp;<br>
m_hChromaShift = CHROMA_H_SHIFT(param->internalCsp);<br>
m_vChromaShift = CHROMA_V_SHIFT(param->internalCsp);<br>
<br>
@@ -715,14 +716,24 @@<br>
ALIGN_VAR_32(int16_t, diff[MAX_CU_SIZE * MAX_CU_SIZE]);<br>
<br>
// Calculate (fenc - frec) and put into diff[]<br>
- // WARNING: *) May read beyond bound on video than width or height is NOT multiple of cuSize<br>
- // *) MUST BE handle ColorSpace other than 420 yourself!<br>
- //<a href="http://primitives.cu" rel="noreferrer" target="_blank">primitives.cu</a>[g_maxLog2CUSize - 2 - (plane != 0)].sub_ps(diff, MAX_CU_SIZE, fenc0, rec0, stride, stride);<br>
- for(int y = 0; y < ctuHeight; y++)<br>
+ if ((lpelx + ctuWidth < picWidth) & (tpely + ctuHeight < picHeight))<br>
{<br>
- for(int x = 0; x < ctuWidth; x++)<br>
+ // WARNING: *) May read beyond bound on video than ctuWidth or ctuHeight is NOT multiple of cuSize<br>
+ X265_CHECK((ctuWidth == ctuHeight) || (m_chromaFormat != X265_CSP_I420), "video size check failure\n");<br>
+ if (plane)<br>
+ primitives.chroma[m_chromaFormat].cu[g_maxLog2CUSize - 2].sub_ps(diff, MAX_CU_SIZE, fenc0, rec0, stride, stride);<br>
+ else<br>
+ <a href="http://primitives.cu" rel="noreferrer" target="_blank">primitives.cu</a>[g_maxLog2CUSize - 2].sub_ps(diff, MAX_CU_SIZE, fenc0, rec0, stride, stride);<br>
+ }<br>
+ else<br>
+ {<br>
+ // path for non-square area (most in edge)<br>
+ for(int y = 0; y < ctuHeight; y++)<br>
{<br>
- diff[y * MAX_CU_SIZE + x] = (fenc0[y * stride + x] - rec0[y * stride + x]);<br>
+ for(int x = 0; x < ctuWidth; x++)<br>
+ {<br>
+ diff[y * MAX_CU_SIZE + x] = (fenc0[y * stride + x] - rec0[y * stride + x]);<br>
+ }<br>
}<br>
}<br>
<br>
diff -r 8fdd1b8fd452 -r 36a54b2cf7c4 source/encoder/sao.h<br>
--- a/source/encoder/sao.h Tue Oct 06 16:42:02 2015 -0500<br>
+++ b/source/encoder/sao.h Tue Oct 06 16:42:06 2015 -0500<br>
@@ -83,6 +83,7 @@<br>
int8_t m_offsetBo[SAO_NUM_BO_CLASSES];<br>
int8_t m_offsetEo[NUM_EDGETYPE];<br>
<br>
+ int m_chromaFormat;<br>
int m_numCuInWidth;<br>
int m_numCuInHeight;<br>
int m_hChromaShift;<br>
<br>
_______________________________________________<br>
x265-devel mailing list<br>
<a href="mailto:x265-devel@videolan.org">x265-devel@videolan.org</a><br>
<a href="https://mailman.videolan.org/listinfo/x265-devel" rel="noreferrer" target="_blank">https://mailman.videolan.org/listinfo/x265-devel</a><br>
</blockquote></div><br></div>