[x265] [PATCH 2 of 2] primitves: 8 bit : PredIntraAng4x4 function table implementation: code cleanup
mandar at multicorewareinc.com
mandar at multicorewareinc.com
Thu Jun 27 09:37:06 CEST 2013
# HG changeset patch
# User Mandar Gurav
# Date 1372318556 25200
# Node ID f71a4aff1b57e57f7d7ecdea006f6e6fbe340beb
# Parent e156dc24f05f4c2e6770fde1b46754cce640a96b
primitves: 8 bit : PredIntraAng4x4 function table implementation: code cleanup
diff -r e156dc24f05f -r f71a4aff1b57 source/common/vec/intrapred.inc
--- a/source/common/vec/intrapred.inc Wed Jun 26 02:59:47 2013 -0700
+++ b/source/common/vec/intrapred.inc Thu Jun 27 00:35:56 2013 -0700
@@ -34,6 +34,56 @@
using namespace x265;
+#define PRED_INTRA_ANG4_START \
+ Vec8s row11, row12, row21, row22, row31, row32, row41, row42; \
+ Vec16uc tmp16_1, tmp16_2; \
+ Vec2uq tmp2uq; \
+ Vec8s v_deltaFract, v_deltaPos(0), thirty2(32), thirty1(31), v_ipAngle(0); \
+ bool modeHor = (dirMode < 18);
+
+
+#define PRED_INTRA_ANG4_END \
+ v_deltaFract = v_deltaPos & thirty1; \
+ row11 = ((thirty2 - v_deltaFract) * row11 + (v_deltaFract * row12) + 16) >> 5; \
+ /*row2*/ \
+ v_deltaPos += v_ipAngle; \
+ v_deltaFract = v_deltaPos & thirty1; \
+ row21 = ((thirty2 - v_deltaFract) * row21 + (v_deltaFract * row22) + 16) >> 5; \
+ /*row3*/ \
+ v_deltaPos += v_ipAngle; \
+ v_deltaFract = v_deltaPos & thirty1; \
+ row31 = ((thirty2 - v_deltaFract) * row31 + (v_deltaFract * row32) + 16) >> 5; \
+ /*row4*/ \
+ v_deltaPos += v_ipAngle; \
+ v_deltaFract = v_deltaPos & thirty1; \
+ row41 = ((thirty2 - v_deltaFract) * row41 + (v_deltaFract * row42) + 16) >> 5; \
+ /* Flip the block */ \
+ if (modeHor) \
+ { \
+ Vec8s tmp1, tmp2, tmp3, tmp4; \
+ tmp1 = blend8s<0, 8, 1, 9, 2, 10, 3, 11>(row11, row31); \
+ tmp2 = blend8s<0, 8, 1, 9, 2, 10, 3, 11>(row21, row41); \
+ tmp3 = blend8s<0, 8, 1, 9, 2, 10, 3, 11>(tmp1, tmp2); \
+ tmp4 = blend8s<4, 12, 5, 13, 6, 14, 7, 15>(tmp1, tmp2); \
+ tmp16_1 = compress_unsafe(tmp3, tmp3); \
+ store_partial(const_int(4), pDst, tmp16_1); \
+ tmp2uq = reinterpret_i(tmp16_1); \
+ tmp2uq >>= 32; \
+ store_partial(const_int(4), pDst + dstStride, tmp2uq); \
+ tmp16_1 = compress_unsafe(tmp4, tmp4); \
+ store_partial(const_int(4), pDst + (2 * dstStride), tmp16_1); \
+ tmp2uq = reinterpret_i(tmp16_1); \
+ tmp2uq >>= 32; \
+ store_partial(const_int(4), pDst + (3 * dstStride), tmp2uq); \
+ } \
+ else \
+ { \
+ store_partial(const_int(4), pDst, compress_unsafe(row11, row11)); \
+ store_partial(const_int(4), pDst + (dstStride), compress_unsafe(row21, row21)); \
+ store_partial(const_int(4), pDst + (2 * dstStride), compress_unsafe(row31, row31)); \
+ store_partial(const_int(4), pDst + (3 * dstStride), compress_unsafe(row41, row41)); \
+ }
+
namespace {
const int angAP[17][64] =
{
@@ -1673,10 +1723,9 @@
#else /* if HIGH_BIT_DEPTH */
-void PredIntraAng4_32(pixel* pDst, int dstStride, pixel *refMain, int dirMode)
+void PredIntraAng4_32(pixel* pDst, int dstStride, pixel *refMain, int /*dirMode*/)
{
Vec16uc tmp16_1, tmp16_2;
- dirMode++;
tmp16_1 = (Vec16uc)load_partial(const_int(8), refMain + 2);
store_partial(const_int(4), pDst, tmp16_1);
tmp16_2 = (Vec16uc)load_partial(const_int(8), refMain + 3);
@@ -1686,16 +1735,13 @@
tmp16_2 = (Vec16uc)load_partial(const_int(8), refMain + 5);
store_partial(const_int(4), pDst + 3 * dstStride, tmp16_2);
}
+
void PredIntraAng4_26(pixel* pDst, int dstStride, pixel *refMain, int dirMode)
{
- Vec8s row11, row12, row21, row22, row31, row32, row41, row42;
- Vec16uc tmp16_1, tmp16_2;
- Vec2uq tmp2uq;
- Vec8s v_deltaFract, v_deltaPos(0), thirty2(32), thirty1(31), v_ipAngle(0);
- bool modeHor = (dirMode < 18);
- tmp16_1 = (Vec16uc)load_partial(const_int(8), refMain + 1);
-
+ PRED_INTRA_ANG4_START
+
+ tmp16_1 = (Vec16uc)load_partial(const_int(8), refMain + 1);
row11 = extend_low(tmp16_1); //offsets(0,1,2,3)
tmp2uq = reinterpret_i(tmp16_1);
@@ -1723,66 +1769,13 @@
row42 = extend_low(tmp16_2); //offsets(4,5,6,7)
v_deltaPos = v_ipAngle = 26;
-
- v_deltaFract = v_deltaPos & thirty1;
- row11 = ((thirty2 - v_deltaFract) * row11 + (v_deltaFract * row12) + 16) >> 5;
-
- //row2
- v_deltaPos += v_ipAngle;
- v_deltaFract = v_deltaPos & thirty1;
- row21 = ((thirty2 - v_deltaFract) * row21 + (v_deltaFract * row22) + 16) >> 5;
-
- //row3
- v_deltaPos += v_ipAngle;
- v_deltaFract = v_deltaPos & thirty1;
- row31 = ((thirty2 - v_deltaFract) * row31 + (v_deltaFract * row32) + 16) >> 5;
-
- //row4
- v_deltaPos += v_ipAngle;
- v_deltaFract = v_deltaPos & thirty1;
- row41 = ((thirty2 - v_deltaFract) * row41 + (v_deltaFract * row42) + 16) >> 5;
-
- // Flip the block
- if (modeHor)
- {
- Vec8s tmp1, tmp2, tmp3, tmp4;
-
- tmp1 = blend8s<0, 8, 1, 9, 2, 10, 3, 11>(row11, row31);
- tmp2 = blend8s<0, 8, 1, 9, 2, 10, 3, 11>(row21, row41);
-
- tmp3 = blend8s<0, 8, 1, 9, 2, 10, 3, 11>(tmp1, tmp2);
- tmp4 = blend8s<4, 12, 5, 13, 6, 14, 7, 15>(tmp1, tmp2);
-
- tmp16_1 = compress_unsafe(tmp3, tmp3);
- store_partial(const_int(4), pDst, tmp16_1);
-
- tmp2uq = reinterpret_i(tmp16_1);
- tmp2uq >>= 32;
- store_partial(const_int(4), pDst + dstStride, tmp2uq);
-
- tmp16_1 = compress_unsafe(tmp4, tmp4);
- store_partial(const_int(4), pDst + (2 * dstStride), tmp16_1);
-
- tmp2uq = reinterpret_i(tmp16_1);
- tmp2uq >>= 32;
- store_partial(const_int(4), pDst + (3 * dstStride), tmp2uq);
- }
- else
- {
- store_partial(const_int(4), pDst, compress_unsafe(row11, row11));
- store_partial(const_int(4), pDst + (dstStride), compress_unsafe(row21, row21));
- store_partial(const_int(4), pDst + (2 * dstStride), compress_unsafe(row31, row31));
- store_partial(const_int(4), pDst + (3 * dstStride), compress_unsafe(row41, row41));
- }
+
+ PRED_INTRA_ANG4_END
}
void PredIntraAng4_21(pixel* pDst, int dstStride, pixel *refMain, int dirMode)
{
- Vec8s row11, row12, row21, row22, row31, row32, row41, row42;
- Vec16uc tmp16_1, tmp16_2;
- Vec2uq tmp2uq;
- Vec8s v_deltaFract, v_deltaPos(0), thirty2(32), thirty1(31), v_ipAngle(0);
- bool modeHor = (dirMode < 18);
+ PRED_INTRA_ANG4_START
tmp16_1 = (Vec16uc)load_partial(const_int(8), refMain + 1);
@@ -1811,65 +1804,12 @@
v_deltaPos = v_ipAngle = 21;
- v_deltaFract = v_deltaPos & thirty1;
- row11 = ((thirty2 - v_deltaFract) * row11 + (v_deltaFract * row12) + 16) >> 5;
-
- //row2
- v_deltaPos += v_ipAngle;
- v_deltaFract = v_deltaPos & thirty1;
- row21 = ((thirty2 - v_deltaFract) * row21 + (v_deltaFract * row22) + 16) >> 5;
-
- //row3
- v_deltaPos += v_ipAngle;
- v_deltaFract = v_deltaPos & thirty1;
- row31 = ((thirty2 - v_deltaFract) * row31 + (v_deltaFract * row32) + 16) >> 5;
-
- //row4
- v_deltaPos += v_ipAngle;
- v_deltaFract = v_deltaPos & thirty1;
- row41 = ((thirty2 - v_deltaFract) * row41 + (v_deltaFract * row42) + 16) >> 5;
-
- // Flip the block
- if (modeHor)
- {
- Vec8s tmp1, tmp2, tmp3, tmp4;
-
- tmp1 = blend8s<0, 8, 1, 9, 2, 10, 3, 11>(row11, row31);
- tmp2 = blend8s<0, 8, 1, 9, 2, 10, 3, 11>(row21, row41);
-
- tmp3 = blend8s<0, 8, 1, 9, 2, 10, 3, 11>(tmp1, tmp2);
- tmp4 = blend8s<4, 12, 5, 13, 6, 14, 7, 15>(tmp1, tmp2);
-
- tmp16_1 = compress_unsafe(tmp3, tmp3);
- store_partial(const_int(4), pDst, tmp16_1);
-
- tmp2uq = reinterpret_i(tmp16_1);
- tmp2uq >>= 32;
- store_partial(const_int(4), pDst + dstStride, tmp2uq);
-
- tmp16_1 = compress_unsafe(tmp4, tmp4);
- store_partial(const_int(4), pDst + (2 * dstStride), tmp16_1);
-
- tmp2uq = reinterpret_i(tmp16_1);
- tmp2uq >>= 32;
- store_partial(const_int(4), pDst + (3 * dstStride), tmp2uq);
- }
- else
- {
- store_partial(const_int(4), pDst, compress_unsafe(row11, row11));
- store_partial(const_int(4), pDst + (dstStride), compress_unsafe(row21, row21));
- store_partial(const_int(4), pDst + (2 * dstStride), compress_unsafe(row31, row31));
- store_partial(const_int(4), pDst + (3 * dstStride), compress_unsafe(row41, row41));
- }
+ PRED_INTRA_ANG4_END
}
void PredIntraAng4_17(pixel* pDst, int dstStride, pixel *refMain, int dirMode)
{
- Vec8s row11, row12, row21, row22, row31, row32, row41, row42;
- Vec16uc tmp16_1, tmp16_2;
- Vec2uq tmp2uq;
- Vec8s v_deltaFract, v_deltaPos(0), thirty2(32), thirty1(31), v_ipAngle(0);
- bool modeHor = (dirMode < 18);
+ PRED_INTRA_ANG4_START
tmp16_1 = (Vec16uc)load_partial(const_int(8), refMain + 1);
@@ -1898,68 +1838,14 @@
v_deltaPos = v_ipAngle = 17;
- v_deltaFract = v_deltaPos & thirty1;
- row11 = ((thirty2 - v_deltaFract) * row11 + (v_deltaFract * row12) + 16) >> 5;
-
- //row2
- v_deltaPos += v_ipAngle;
- v_deltaFract = v_deltaPos & thirty1;
- row21 = ((thirty2 - v_deltaFract) * row21 + (v_deltaFract * row22) + 16) >> 5;
-
- //row3
- v_deltaPos += v_ipAngle;
- v_deltaFract = v_deltaPos & thirty1;
- row31 = ((thirty2 - v_deltaFract) * row31 + (v_deltaFract * row32) + 16) >> 5;
-
- //row4
- v_deltaPos += v_ipAngle;
- v_deltaFract = v_deltaPos & thirty1;
- row41 = ((thirty2 - v_deltaFract) * row41 + (v_deltaFract * row42) + 16) >> 5;
-
- // Flip the block
- if (modeHor)
- {
- Vec8s tmp1, tmp2, tmp3, tmp4;
-
- tmp1 = blend8s<0, 8, 1, 9, 2, 10, 3, 11>(row11, row31);
- tmp2 = blend8s<0, 8, 1, 9, 2, 10, 3, 11>(row21, row41);
-
- tmp3 = blend8s<0, 8, 1, 9, 2, 10, 3, 11>(tmp1, tmp2);
- tmp4 = blend8s<4, 12, 5, 13, 6, 14, 7, 15>(tmp1, tmp2);
-
- tmp16_1 = compress_unsafe(tmp3, tmp3);
- store_partial(const_int(4), pDst, tmp16_1);
-
- tmp2uq = reinterpret_i(tmp16_1);
- tmp2uq >>= 32;
- store_partial(const_int(4), pDst + dstStride, tmp2uq);
-
- tmp16_1 = compress_unsafe(tmp4, tmp4);
- store_partial(const_int(4), pDst + (2 * dstStride), tmp16_1);
-
- tmp2uq = reinterpret_i(tmp16_1);
- tmp2uq >>= 32;
- store_partial(const_int(4), pDst + (3 * dstStride), tmp2uq);
- }
- else
- {
- store_partial(const_int(4), pDst, compress_unsafe(row11, row11));
- store_partial(const_int(4), pDst + (dstStride), compress_unsafe(row21, row21));
- store_partial(const_int(4), pDst + (2 * dstStride), compress_unsafe(row31, row31));
- store_partial(const_int(4), pDst + (3 * dstStride), compress_unsafe(row41, row41));
- }
+ PRED_INTRA_ANG4_END
}
void PredIntraAng4_13(pixel* pDst, int dstStride, pixel *refMain, int dirMode)
{
- Vec8s row11, row12, row21, row22, row31, row32, row41, row42;
- Vec16uc tmp16_1, tmp16_2;
- Vec2uq tmp2uq;
- Vec8s v_deltaFract, v_deltaPos(0), thirty2(32), thirty1(31), v_ipAngle(0);
- bool modeHor = (dirMode < 18);
+ PRED_INTRA_ANG4_START
tmp16_1 = (Vec16uc)load_partial(const_int(8), refMain + 1);
-
row11 = extend_low(tmp16_1); //offsets(0,1,2,3)
tmp2uq = reinterpret_i(tmp16_1);
@@ -1981,65 +1867,12 @@
v_deltaPos = v_ipAngle = 13;
- v_deltaFract = v_deltaPos & thirty1;
- row11 = ((thirty2 - v_deltaFract) * row11 + (v_deltaFract * row12) + 16) >> 5;
-
- //row2
- v_deltaPos += v_ipAngle;
- v_deltaFract = v_deltaPos & thirty1;
- row21 = ((thirty2 - v_deltaFract) * row21 + (v_deltaFract * row22) + 16) >> 5;
-
- //row3
- v_deltaPos += v_ipAngle;
- v_deltaFract = v_deltaPos & thirty1;
- row31 = ((thirty2 - v_deltaFract) * row31 + (v_deltaFract * row32) + 16) >> 5;
-
- //row4
- v_deltaPos += v_ipAngle;
- v_deltaFract = v_deltaPos & thirty1;
- row41 = ((thirty2 - v_deltaFract) * row41 + (v_deltaFract * row42) + 16) >> 5;
-
- // Flip the block
- if (modeHor)
- {
- Vec8s tmp1, tmp2, tmp3, tmp4;
-
- tmp1 = blend8s<0, 8, 1, 9, 2, 10, 3, 11>(row11, row31);
- tmp2 = blend8s<0, 8, 1, 9, 2, 10, 3, 11>(row21, row41);
-
- tmp3 = blend8s<0, 8, 1, 9, 2, 10, 3, 11>(tmp1, tmp2);
- tmp4 = blend8s<4, 12, 5, 13, 6, 14, 7, 15>(tmp1, tmp2);
-
- tmp16_1 = compress_unsafe(tmp3, tmp3);
- store_partial(const_int(4), pDst, tmp16_1);
-
- tmp2uq = reinterpret_i(tmp16_1);
- tmp2uq >>= 32;
- store_partial(const_int(4), pDst + dstStride, tmp2uq);
-
- tmp16_1 = compress_unsafe(tmp4, tmp4);
- store_partial(const_int(4), pDst + (2 * dstStride), tmp16_1);
-
- tmp2uq = reinterpret_i(tmp16_1);
- tmp2uq >>= 32;
- store_partial(const_int(4), pDst + (3 * dstStride), tmp2uq);
- }
- else
- {
- store_partial(const_int(4), pDst, compress_unsafe(row11, row11));
- store_partial(const_int(4), pDst + (dstStride), compress_unsafe(row21, row21));
- store_partial(const_int(4), pDst + (2 * dstStride), compress_unsafe(row31, row31));
- store_partial(const_int(4), pDst + (3 * dstStride), compress_unsafe(row41, row41));
- }
+ PRED_INTRA_ANG4_END
}
void PredIntraAng4_9(pixel* pDst, int dstStride, pixel *refMain, int dirMode)
{
- Vec8s row11, row12, row21, row22, row31, row32, row41, row42;
- Vec16uc tmp16_1, tmp16_2;
- Vec2uq tmp2uq;
- Vec8s v_deltaFract, v_deltaPos(0), thirty2(32), thirty1(31), v_ipAngle(0);
- bool modeHor = (dirMode < 18);
+ PRED_INTRA_ANG4_START
tmp16_1 = (Vec16uc)load_partial(const_int(8), refMain + 1);
row11 = extend_low(tmp16_1); //offsets(0,1,2,3)
@@ -2059,65 +1892,12 @@
v_deltaPos = v_ipAngle = 9;
- v_deltaFract = v_deltaPos & thirty1;
- row11 = ((thirty2 - v_deltaFract) * row11 + (v_deltaFract * row12) + 16) >> 5;
-
- //row2
- v_deltaPos += v_ipAngle;
- v_deltaFract = v_deltaPos & thirty1;
- row21 = ((thirty2 - v_deltaFract) * row21 + (v_deltaFract * row22) + 16) >> 5;
-
- //row3
- v_deltaPos += v_ipAngle;
- v_deltaFract = v_deltaPos & thirty1;
- row31 = ((thirty2 - v_deltaFract) * row31 + (v_deltaFract * row32) + 16) >> 5;
-
- //row4
- v_deltaPos += v_ipAngle;
- v_deltaFract = v_deltaPos & thirty1;
- row41 = ((thirty2 - v_deltaFract) * row41 + (v_deltaFract * row42) + 16) >> 5;
-
- // Flip the block
- if (modeHor)
- {
- Vec8s tmp1, tmp2, tmp3, tmp4;
-
- tmp1 = blend8s<0, 8, 1, 9, 2, 10, 3, 11>(row11, row31);
- tmp2 = blend8s<0, 8, 1, 9, 2, 10, 3, 11>(row21, row41);
-
- tmp3 = blend8s<0, 8, 1, 9, 2, 10, 3, 11>(tmp1, tmp2);
- tmp4 = blend8s<4, 12, 5, 13, 6, 14, 7, 15>(tmp1, tmp2);
-
- tmp16_1 = compress_unsafe(tmp3, tmp3);
- store_partial(const_int(4), pDst, tmp16_1);
-
- tmp2uq = reinterpret_i(tmp16_1);
- tmp2uq >>= 32;
- store_partial(const_int(4), pDst + dstStride, tmp2uq);
-
- tmp16_1 = compress_unsafe(tmp4, tmp4);
- store_partial(const_int(4), pDst + (2 * dstStride), tmp16_1);
-
- tmp2uq = reinterpret_i(tmp16_1);
- tmp2uq >>= 32;
- store_partial(const_int(4), pDst + (3 * dstStride), tmp2uq);
- }
- else
- {
- store_partial(const_int(4), pDst, compress_unsafe(row11, row11));
- store_partial(const_int(4), pDst + (dstStride), compress_unsafe(row21, row21));
- store_partial(const_int(4), pDst + (2 * dstStride), compress_unsafe(row31, row31));
- store_partial(const_int(4), pDst + (3 * dstStride), compress_unsafe(row41, row41));
- }
+ PRED_INTRA_ANG4_END
}
void PredIntraAng4_5(pixel* pDst, int dstStride, pixel *refMain, int dirMode)
{
- Vec8s row11, row12, row21, row22, row31, row32, row41, row42;
- Vec16uc tmp16_1, tmp16_2;
- Vec2uq tmp2uq;
- Vec8s v_deltaFract, v_deltaPos(0), thirty2(32), thirty1(31), v_ipAngle(0);
- bool modeHor = (dirMode < 18);
+ PRED_INTRA_ANG4_START
tmp16_1 = (Vec16uc)load_partial(const_int(8), refMain + 1);
row11 = extend_low(tmp16_1); //offsets(0,1,2,3)
@@ -2134,65 +1914,12 @@
v_deltaPos = v_ipAngle = 5;
- v_deltaFract = v_deltaPos & thirty1;
- row11 = ((thirty2 - v_deltaFract) * row11 + (v_deltaFract * row12) + 16) >> 5;
-
- //row2
- v_deltaPos += v_ipAngle;
- v_deltaFract = v_deltaPos & thirty1;
- row21 = ((thirty2 - v_deltaFract) * row21 + (v_deltaFract * row22) + 16) >> 5;
-
- //row3
- v_deltaPos += v_ipAngle;
- v_deltaFract = v_deltaPos & thirty1;
- row31 = ((thirty2 - v_deltaFract) * row31 + (v_deltaFract * row32) + 16) >> 5;
-
- //row4
- v_deltaPos += v_ipAngle;
- v_deltaFract = v_deltaPos & thirty1;
- row41 = ((thirty2 - v_deltaFract) * row41 + (v_deltaFract * row42) + 16) >> 5;
-
- // Flip the block
- if (modeHor)
- {
- Vec8s tmp1, tmp2, tmp3, tmp4;
-
- tmp1 = blend8s<0, 8, 1, 9, 2, 10, 3, 11>(row11, row31);
- tmp2 = blend8s<0, 8, 1, 9, 2, 10, 3, 11>(row21, row41);
-
- tmp3 = blend8s<0, 8, 1, 9, 2, 10, 3, 11>(tmp1, tmp2);
- tmp4 = blend8s<4, 12, 5, 13, 6, 14, 7, 15>(tmp1, tmp2);
-
- tmp16_1 = compress_unsafe(tmp3, tmp3);
- store_partial(const_int(4), pDst, tmp16_1);
-
- tmp2uq = reinterpret_i(tmp16_1);
- tmp2uq >>= 32;
- store_partial(const_int(4), pDst + dstStride, tmp2uq);
-
- tmp16_1 = compress_unsafe(tmp4, tmp4);
- store_partial(const_int(4), pDst + (2 * dstStride), tmp16_1);
-
- tmp2uq = reinterpret_i(tmp16_1);
- tmp2uq >>= 32;
- store_partial(const_int(4), pDst + (3 * dstStride), tmp2uq);
- }
- else
- {
- store_partial(const_int(4), pDst, compress_unsafe(row11, row11));
- store_partial(const_int(4), pDst + (dstStride), compress_unsafe(row21, row21));
- store_partial(const_int(4), pDst + (2 * dstStride), compress_unsafe(row31, row31));
- store_partial(const_int(4), pDst + (3 * dstStride), compress_unsafe(row41, row41));
- }
+ PRED_INTRA_ANG4_END
}
void PredIntraAng4_2(pixel* pDst, int dstStride, pixel *refMain, int dirMode)
{
- Vec8s row11, row12, row21, row22, row31, row32, row41, row42;
- Vec16uc tmp16_1, tmp16_2;
- Vec2uq tmp2uq;
- Vec8s v_deltaFract, v_deltaPos(0), thirty2(32), thirty1(31), v_ipAngle(0);
- bool modeHor = (dirMode < 18);
+ PRED_INTRA_ANG4_START
tmp16_1 = (Vec16uc)load_partial(const_int(8), refMain + 1);
row11 = extend_low(tmp16_1); //offsets(0,1,2,3)
@@ -2209,65 +1936,12 @@
v_deltaPos = v_ipAngle = 2;
- v_deltaFract = v_deltaPos & thirty1;
- row11 = ((thirty2 - v_deltaFract) * row11 + (v_deltaFract * row12) + 16) >> 5;
-
- //row2
- v_deltaPos += v_ipAngle;
- v_deltaFract = v_deltaPos & thirty1;
- row21 = ((thirty2 - v_deltaFract) * row21 + (v_deltaFract * row22) + 16) >> 5;
-
- //row3
- v_deltaPos += v_ipAngle;
- v_deltaFract = v_deltaPos & thirty1;
- row31 = ((thirty2 - v_deltaFract) * row31 + (v_deltaFract * row32) + 16) >> 5;
-
- //row4
- v_deltaPos += v_ipAngle;
- v_deltaFract = v_deltaPos & thirty1;
- row41 = ((thirty2 - v_deltaFract) * row41 + (v_deltaFract * row42) + 16) >> 5;
-
- // Flip the block
- if (modeHor)
- {
- Vec8s tmp1, tmp2, tmp3, tmp4;
-
- tmp1 = blend8s<0, 8, 1, 9, 2, 10, 3, 11>(row11, row31);
- tmp2 = blend8s<0, 8, 1, 9, 2, 10, 3, 11>(row21, row41);
-
- tmp3 = blend8s<0, 8, 1, 9, 2, 10, 3, 11>(tmp1, tmp2);
- tmp4 = blend8s<4, 12, 5, 13, 6, 14, 7, 15>(tmp1, tmp2);
-
- tmp16_1 = compress_unsafe(tmp3, tmp3);
- store_partial(const_int(4), pDst, tmp16_1);
-
- tmp2uq = reinterpret_i(tmp16_1);
- tmp2uq >>= 32;
- store_partial(const_int(4), pDst + dstStride, tmp2uq);
-
- tmp16_1 = compress_unsafe(tmp4, tmp4);
- store_partial(const_int(4), pDst + (2 * dstStride), tmp16_1);
-
- tmp2uq = reinterpret_i(tmp16_1);
- tmp2uq >>= 32;
- store_partial(const_int(4), pDst + (3 * dstStride), tmp2uq);
- }
- else
- {
- store_partial(const_int(4), pDst, compress_unsafe(row11, row11));
- store_partial(const_int(4), pDst + (dstStride), compress_unsafe(row21, row21));
- store_partial(const_int(4), pDst + (2 * dstStride), compress_unsafe(row31, row31));
- store_partial(const_int(4), pDst + (3 * dstStride), compress_unsafe(row41, row41));
- }
+ PRED_INTRA_ANG4_END
}
void PredIntraAng4_m_2(pixel* pDst, int dstStride, pixel *refMain, int dirMode)
{
- Vec8s row11, row12, row21, row22, row31, row32, row41, row42;
- Vec16uc tmp16_1, tmp16_2;
- Vec2uq tmp2uq;
- Vec8s v_deltaFract, v_deltaPos(0), thirty2(32), thirty1(31), v_ipAngle(0);
- bool modeHor = (dirMode < 18);
+ PRED_INTRA_ANG4_START
tmp16_1 = (Vec16uc)load_partial(const_int(8), refMain);
row11 = extend_low(tmp16_1); //offsets(0,1,2,3)
@@ -2284,65 +1958,12 @@
v_deltaPos = v_ipAngle = -2;
- v_deltaFract = v_deltaPos & thirty1;
- row11 = ((thirty2 - v_deltaFract) * row11 + (v_deltaFract * row12) + 16) >> 5;
-
- //row2
- v_deltaPos += v_ipAngle;
- v_deltaFract = v_deltaPos & thirty1;
- row21 = ((thirty2 - v_deltaFract) * row21 + (v_deltaFract * row22) + 16) >> 5;
-
- //row3
- v_deltaPos += v_ipAngle;
- v_deltaFract = v_deltaPos & thirty1;
- row31 = ((thirty2 - v_deltaFract) * row31 + (v_deltaFract * row32) + 16) >> 5;
-
- //row4
- v_deltaPos += v_ipAngle;
- v_deltaFract = v_deltaPos & thirty1;
- row41 = ((thirty2 - v_deltaFract) * row41 + (v_deltaFract * row42) + 16) >> 5;
-
- // Flip the block
- if (modeHor)
- {
- Vec8s tmp1, tmp2, tmp3, tmp4;
-
- tmp1 = blend8s<0, 8, 1, 9, 2, 10, 3, 11>(row11, row31);
- tmp2 = blend8s<0, 8, 1, 9, 2, 10, 3, 11>(row21, row41);
-
- tmp3 = blend8s<0, 8, 1, 9, 2, 10, 3, 11>(tmp1, tmp2);
- tmp4 = blend8s<4, 12, 5, 13, 6, 14, 7, 15>(tmp1, tmp2);
-
- tmp16_1 = compress_unsafe(tmp3, tmp3);
- store_partial(const_int(4), pDst, tmp16_1);
-
- tmp2uq = reinterpret_i(tmp16_1);
- tmp2uq >>= 32;
- store_partial(const_int(4), pDst + dstStride, tmp2uq);
-
- tmp16_1 = compress_unsafe(tmp4, tmp4);
- store_partial(const_int(4), pDst + (2 * dstStride), tmp16_1);
-
- tmp2uq = reinterpret_i(tmp16_1);
- tmp2uq >>= 32;
- store_partial(const_int(4), pDst + (3 * dstStride), tmp2uq);
- }
- else
- {
- store_partial(const_int(4), pDst, compress_unsafe(row11, row11));
- store_partial(const_int(4), pDst + (dstStride), compress_unsafe(row21, row21));
- store_partial(const_int(4), pDst + (2 * dstStride), compress_unsafe(row31, row31));
- store_partial(const_int(4), pDst + (3 * dstStride), compress_unsafe(row41, row41));
- }
+ PRED_INTRA_ANG4_END
}
void PredIntraAng4_m_5(pixel* pDst, int dstStride, pixel *refMain, int dirMode)
{
- Vec8s row11, row12, row21, row22, row31, row32, row41, row42;
- Vec16uc tmp16_1, tmp16_2;
- Vec2uq tmp2uq;
- Vec8s v_deltaFract, v_deltaPos(0), thirty2(32), thirty1(31), v_ipAngle(0);
- bool modeHor = (dirMode < 18);
+ PRED_INTRA_ANG4_START
tmp16_1 = (Vec16uc)load_partial(const_int(8), refMain);
row11 = extend_low(tmp16_1); //offsets(0,1,2,3)
@@ -2359,65 +1980,12 @@
v_deltaPos = v_ipAngle = -5;
- v_deltaFract = v_deltaPos & thirty1;
- row11 = ((thirty2 - v_deltaFract) * row11 + (v_deltaFract * row12) + 16) >> 5;
-
- //row2
- v_deltaPos += v_ipAngle;
- v_deltaFract = v_deltaPos & thirty1;
- row21 = ((thirty2 - v_deltaFract) * row21 + (v_deltaFract * row22) + 16) >> 5;
-
- //row3
- v_deltaPos += v_ipAngle;
- v_deltaFract = v_deltaPos & thirty1;
- row31 = ((thirty2 - v_deltaFract) * row31 + (v_deltaFract * row32) + 16) >> 5;
-
- //row4
- v_deltaPos += v_ipAngle;
- v_deltaFract = v_deltaPos & thirty1;
- row41 = ((thirty2 - v_deltaFract) * row41 + (v_deltaFract * row42) + 16) >> 5;
-
- // Flip the block
- if (modeHor)
- {
- Vec8s tmp1, tmp2, tmp3, tmp4;
-
- tmp1 = blend8s<0, 8, 1, 9, 2, 10, 3, 11>(row11, row31);
- tmp2 = blend8s<0, 8, 1, 9, 2, 10, 3, 11>(row21, row41);
-
- tmp3 = blend8s<0, 8, 1, 9, 2, 10, 3, 11>(tmp1, tmp2);
- tmp4 = blend8s<4, 12, 5, 13, 6, 14, 7, 15>(tmp1, tmp2);
-
- tmp16_1 = compress_unsafe(tmp3, tmp3);
- store_partial(const_int(4), pDst, tmp16_1);
-
- tmp2uq = reinterpret_i(tmp16_1);
- tmp2uq >>= 32;
- store_partial(const_int(4), pDst + dstStride, tmp2uq);
-
- tmp16_1 = compress_unsafe(tmp4, tmp4);
- store_partial(const_int(4), pDst + (2 * dstStride), tmp16_1);
-
- tmp2uq = reinterpret_i(tmp16_1);
- tmp2uq >>= 32;
- store_partial(const_int(4), pDst + (3 * dstStride), tmp2uq);
- }
- else
- {
- store_partial(const_int(4), pDst, compress_unsafe(row11, row11));
- store_partial(const_int(4), pDst + (dstStride), compress_unsafe(row21, row21));
- store_partial(const_int(4), pDst + (2 * dstStride), compress_unsafe(row31, row31));
- store_partial(const_int(4), pDst + (3 * dstStride), compress_unsafe(row41, row41));
- }
+ PRED_INTRA_ANG4_END
}
void PredIntraAng4_m_9(pixel* pDst, int dstStride, pixel *refMain, int dirMode)
{
- Vec8s row11, row12, row21, row22, row31, row32, row41, row42;
- Vec16uc tmp16_1, tmp16_2;
- Vec2uq tmp2uq;
- Vec8s v_deltaFract, v_deltaPos(0), thirty2(32), thirty1(31), v_ipAngle(0);
- bool modeHor = (dirMode < 18);
+ PRED_INTRA_ANG4_START
tmp16_1 = (Vec16uc)load_partial(const_int(8), refMain - 1);
row41 = extend_low(tmp16_1); //offsets(-2,-1,0,1)
@@ -2439,65 +2007,12 @@
v_deltaPos = v_ipAngle = -9;
- v_deltaFract = v_deltaPos & thirty1;
- row11 = ((thirty2 - v_deltaFract) * row11 + (v_deltaFract * row12) + 16) >> 5;
-
- //row2
- v_deltaPos += v_ipAngle;
- v_deltaFract = v_deltaPos & thirty1;
- row21 = ((thirty2 - v_deltaFract) * row21 + (v_deltaFract * row22) + 16) >> 5;
-
- //row3
- v_deltaPos += v_ipAngle;
- v_deltaFract = v_deltaPos & thirty1;
- row31 = ((thirty2 - v_deltaFract) * row31 + (v_deltaFract * row32) + 16) >> 5;
-
- //row4
- v_deltaPos += v_ipAngle;
- v_deltaFract = v_deltaPos & thirty1;
- row41 = ((thirty2 - v_deltaFract) * row41 + (v_deltaFract * row42) + 16) >> 5;
-
- // Flip the block
- if (modeHor)
- {
- Vec8s tmp1, tmp2, tmp3, tmp4;
-
- tmp1 = blend8s<0, 8, 1, 9, 2, 10, 3, 11>(row11, row31);
- tmp2 = blend8s<0, 8, 1, 9, 2, 10, 3, 11>(row21, row41);
-
- tmp3 = blend8s<0, 8, 1, 9, 2, 10, 3, 11>(tmp1, tmp2);
- tmp4 = blend8s<4, 12, 5, 13, 6, 14, 7, 15>(tmp1, tmp2);
-
- tmp16_1 = compress_unsafe(tmp3, tmp3);
- store_partial(const_int(4), pDst, tmp16_1);
-
- tmp2uq = reinterpret_i(tmp16_1);
- tmp2uq >>= 32;
- store_partial(const_int(4), pDst + dstStride, tmp2uq);
-
- tmp16_1 = compress_unsafe(tmp4, tmp4);
- store_partial(const_int(4), pDst + (2 * dstStride), tmp16_1);
-
- tmp2uq = reinterpret_i(tmp16_1);
- tmp2uq >>= 32;
- store_partial(const_int(4), pDst + (3 * dstStride), tmp2uq);
- }
- else
- {
- store_partial(const_int(4), pDst, compress_unsafe(row11, row11));
- store_partial(const_int(4), pDst + (dstStride), compress_unsafe(row21, row21));
- store_partial(const_int(4), pDst + (2 * dstStride), compress_unsafe(row31, row31));
- store_partial(const_int(4), pDst + (3 * dstStride), compress_unsafe(row41, row41));
- }
+ PRED_INTRA_ANG4_END
}
void PredIntraAng4_m_13(pixel* pDst, int dstStride, pixel *refMain, int dirMode)
{
- Vec8s row11, row12, row21, row22, row31, row32, row41, row42;
- Vec16uc tmp16_1, tmp16_2;
- Vec2uq tmp2uq;
- Vec8s v_deltaFract, v_deltaPos(0), thirty2(32), thirty1(31), v_ipAngle(0);
- bool modeHor = (dirMode < 18);
+ PRED_INTRA_ANG4_START
tmp16_1 = (Vec16uc)load_partial(const_int(8), refMain - 1);
row41 = extend_low(tmp16_1); //offsets(-2,-1,0,1)
@@ -2519,65 +2034,12 @@
v_deltaPos = v_ipAngle = -13;
- v_deltaFract = v_deltaPos & thirty1;
- row11 = ((thirty2 - v_deltaFract) * row11 + (v_deltaFract * row12) + 16) >> 5;
-
- //row2
- v_deltaPos += v_ipAngle;
- v_deltaFract = v_deltaPos & thirty1;
- row21 = ((thirty2 - v_deltaFract) * row21 + (v_deltaFract * row22) + 16) >> 5;
-
- //row3
- v_deltaPos += v_ipAngle;
- v_deltaFract = v_deltaPos & thirty1;
- row31 = ((thirty2 - v_deltaFract) * row31 + (v_deltaFract * row32) + 16) >> 5;
-
- //row4
- v_deltaPos += v_ipAngle;
- v_deltaFract = v_deltaPos & thirty1;
- row41 = ((thirty2 - v_deltaFract) * row41 + (v_deltaFract * row42) + 16) >> 5;
-
- // Flip the block
- if (modeHor)
- {
- Vec8s tmp1, tmp2, tmp3, tmp4;
-
- tmp1 = blend8s<0, 8, 1, 9, 2, 10, 3, 11>(row11, row31);
- tmp2 = blend8s<0, 8, 1, 9, 2, 10, 3, 11>(row21, row41);
-
- tmp3 = blend8s<0, 8, 1, 9, 2, 10, 3, 11>(tmp1, tmp2);
- tmp4 = blend8s<4, 12, 5, 13, 6, 14, 7, 15>(tmp1, tmp2);
-
- tmp16_1 = compress_unsafe(tmp3, tmp3);
- store_partial(const_int(4), pDst, tmp16_1);
-
- tmp2uq = reinterpret_i(tmp16_1);
- tmp2uq >>= 32;
- store_partial(const_int(4), pDst + dstStride, tmp2uq);
-
- tmp16_1 = compress_unsafe(tmp4, tmp4);
- store_partial(const_int(4), pDst + (2 * dstStride), tmp16_1);
-
- tmp2uq = reinterpret_i(tmp16_1);
- tmp2uq >>= 32;
- store_partial(const_int(4), pDst + (3 * dstStride), tmp2uq);
- }
- else
- {
- store_partial(const_int(4), pDst, compress_unsafe(row11, row11));
- store_partial(const_int(4), pDst + (dstStride), compress_unsafe(row21, row21));
- store_partial(const_int(4), pDst + (2 * dstStride), compress_unsafe(row31, row31));
- store_partial(const_int(4), pDst + (3 * dstStride), compress_unsafe(row41, row41));
- }
+ PRED_INTRA_ANG4_END
}
void PredIntraAng4_m_17(pixel* pDst, int dstStride, pixel *refMain, int dirMode)
{
- Vec8s row11, row12, row21, row22, row31, row32, row41, row42;
- Vec16uc tmp16_1, tmp16_2;
- Vec2uq tmp2uq;
- Vec8s v_deltaFract, v_deltaPos(0), thirty2(32), thirty1(31), v_ipAngle(0);
- bool modeHor = (dirMode < 18);
+ PRED_INTRA_ANG4_START
tmp16_1 = (Vec16uc)load_partial(const_int(8), refMain - 2);
row41 = extend_low(tmp16_1); //offsets(-3,-2,-1,0)
@@ -2602,65 +2064,13 @@
row12 = extend_low(tmp16_2); //offsets(-1,0,1,2)
v_deltaPos = v_ipAngle = -17;
- v_deltaFract = v_deltaPos & thirty1;
- row11 = ((thirty2 - v_deltaFract) * row11 + (v_deltaFract * row12) + 16) >> 5;
-
- //row2
- v_deltaPos += v_ipAngle;
- v_deltaFract = v_deltaPos & thirty1;
- row21 = ((thirty2 - v_deltaFract) * row21 + (v_deltaFract * row22) + 16) >> 5;
-
- //row3
- v_deltaPos += v_ipAngle;
- v_deltaFract = v_deltaPos & thirty1;
- row31 = ((thirty2 - v_deltaFract) * row31 + (v_deltaFract * row32) + 16) >> 5;
-
- //row4
- v_deltaPos += v_ipAngle;
- v_deltaFract = v_deltaPos & thirty1;
- row41 = ((thirty2 - v_deltaFract) * row41 + (v_deltaFract * row42) + 16) >> 5;
-
- // Flip the block
- if (modeHor)
- {
- Vec8s tmp1, tmp2, tmp3, tmp4;
-
- tmp1 = blend8s<0, 8, 1, 9, 2, 10, 3, 11>(row11, row31);
- tmp2 = blend8s<0, 8, 1, 9, 2, 10, 3, 11>(row21, row41);
-
- tmp3 = blend8s<0, 8, 1, 9, 2, 10, 3, 11>(tmp1, tmp2);
- tmp4 = blend8s<4, 12, 5, 13, 6, 14, 7, 15>(tmp1, tmp2);
-
- tmp16_1 = compress_unsafe(tmp3, tmp3);
- store_partial(const_int(4), pDst, tmp16_1);
-
- tmp2uq = reinterpret_i(tmp16_1);
- tmp2uq >>= 32;
- store_partial(const_int(4), pDst + dstStride, tmp2uq);
-
- tmp16_1 = compress_unsafe(tmp4, tmp4);
- store_partial(const_int(4), pDst + (2 * dstStride), tmp16_1);
-
- tmp2uq = reinterpret_i(tmp16_1);
- tmp2uq >>= 32;
- store_partial(const_int(4), pDst + (3 * dstStride), tmp2uq);
- }
- else
- {
- store_partial(const_int(4), pDst, compress_unsafe(row11, row11));
- store_partial(const_int(4), pDst + (dstStride), compress_unsafe(row21, row21));
- store_partial(const_int(4), pDst + (2 * dstStride), compress_unsafe(row31, row31));
- store_partial(const_int(4), pDst + (3 * dstStride), compress_unsafe(row41, row41));
- }
+
+ PRED_INTRA_ANG4_END
}
void PredIntraAng4_m_21(pixel* pDst, int dstStride, pixel *refMain, int dirMode)
{
- Vec8s row11, row12, row21, row22, row31, row32, row41, row42;
- Vec16uc tmp16_1, tmp16_2;
- Vec2uq tmp2uq;
- Vec8s v_deltaFract, v_deltaPos(0), thirty2(32), thirty1(31), v_ipAngle(0);
- bool modeHor = (dirMode < 18);
+ PRED_INTRA_ANG4_START
tmp16_1 = (Vec16uc)load_partial(const_int(8), refMain - 2);
row41 = extend_low(tmp16_1); //offsets(-3,-2,-1,0)
@@ -2686,65 +2096,12 @@
v_deltaPos = v_ipAngle = -21;
- v_deltaFract = v_deltaPos & thirty1;
- row11 = ((thirty2 - v_deltaFract) * row11 + (v_deltaFract * row12) + 16) >> 5;
-
- //row2
- v_deltaPos += v_ipAngle;
- v_deltaFract = v_deltaPos & thirty1;
- row21 = ((thirty2 - v_deltaFract) * row21 + (v_deltaFract * row22) + 16) >> 5;
-
- //row3
- v_deltaPos += v_ipAngle;
- v_deltaFract = v_deltaPos & thirty1;
- row31 = ((thirty2 - v_deltaFract) * row31 + (v_deltaFract * row32) + 16) >> 5;
-
- //row4
- v_deltaPos += v_ipAngle;
- v_deltaFract = v_deltaPos & thirty1;
- row41 = ((thirty2 - v_deltaFract) * row41 + (v_deltaFract * row42) + 16) >> 5;
-
- // Flip the block
- if (modeHor)
- {
- Vec8s tmp1, tmp2, tmp3, tmp4;
-
- tmp1 = blend8s<0, 8, 1, 9, 2, 10, 3, 11>(row11, row31);
- tmp2 = blend8s<0, 8, 1, 9, 2, 10, 3, 11>(row21, row41);
-
- tmp3 = blend8s<0, 8, 1, 9, 2, 10, 3, 11>(tmp1, tmp2);
- tmp4 = blend8s<4, 12, 5, 13, 6, 14, 7, 15>(tmp1, tmp2);
-
- tmp16_1 = compress_unsafe(tmp3, tmp3);
- store_partial(const_int(4), pDst, tmp16_1);
-
- tmp2uq = reinterpret_i(tmp16_1);
- tmp2uq >>= 32;
- store_partial(const_int(4), pDst + dstStride, tmp2uq);
-
- tmp16_1 = compress_unsafe(tmp4, tmp4);
- store_partial(const_int(4), pDst + (2 * dstStride), tmp16_1);
-
- tmp2uq = reinterpret_i(tmp16_1);
- tmp2uq >>= 32;
- store_partial(const_int(4), pDst + (3 * dstStride), tmp2uq);
- }
- else
- {
- store_partial(const_int(4), pDst, compress_unsafe(row11, row11));
- store_partial(const_int(4), pDst + (dstStride), compress_unsafe(row21, row21));
- store_partial(const_int(4), pDst + (2 * dstStride), compress_unsafe(row31, row31));
- store_partial(const_int(4), pDst + (3 * dstStride), compress_unsafe(row41, row41));
- }
+ PRED_INTRA_ANG4_END
}
void PredIntraAng4_m_26(pixel* pDst, int dstStride, pixel *refMain, int dirMode)
{
- Vec8s row11, row12, row21, row22, row31, row32, row41, row42;
- Vec16uc tmp16_1, tmp16_2;
- Vec2uq tmp2uq;
- Vec8s v_deltaFract, v_deltaPos(0), thirty2(32), thirty1(31), v_ipAngle(0);
- bool modeHor = (dirMode < 18);
+ PRED_INTRA_ANG4_START
tmp16_1 = (Vec16uc)load_partial(const_int(8), refMain - 3);
row41 = extend_low(tmp16_1); //offsets(-4,-3,-2,-1)
@@ -2773,62 +2130,12 @@
v_deltaPos = v_ipAngle = -26;
- v_deltaFract = v_deltaPos & thirty1;
- row11 = ((thirty2 - v_deltaFract) * row11 + (v_deltaFract * row12) + 16) >> 5;
-
- //row2
- v_deltaPos += v_ipAngle;
- v_deltaFract = v_deltaPos & thirty1;
- row21 = ((thirty2 - v_deltaFract) * row21 + (v_deltaFract * row22) + 16) >> 5;
-
- //row3
- v_deltaPos += v_ipAngle;
- v_deltaFract = v_deltaPos & thirty1;
- row31 = ((thirty2 - v_deltaFract) * row31 + (v_deltaFract * row32) + 16) >> 5;
-
- //row4
- v_deltaPos += v_ipAngle;
- v_deltaFract = v_deltaPos & thirty1;
- row41 = ((thirty2 - v_deltaFract) * row41 + (v_deltaFract * row42) + 16) >> 5;
-
- // Flip the block
- if (modeHor)
- {
- Vec8s tmp1, tmp2, tmp3, tmp4;
-
- tmp1 = blend8s<0, 8, 1, 9, 2, 10, 3, 11>(row11, row31);
- tmp2 = blend8s<0, 8, 1, 9, 2, 10, 3, 11>(row21, row41);
-
- tmp3 = blend8s<0, 8, 1, 9, 2, 10, 3, 11>(tmp1, tmp2);
- tmp4 = blend8s<4, 12, 5, 13, 6, 14, 7, 15>(tmp1, tmp2);
-
- tmp16_1 = compress_unsafe(tmp3, tmp3);
- store_partial(const_int(4), pDst, tmp16_1);
-
- tmp2uq = reinterpret_i(tmp16_1);
- tmp2uq >>= 32;
- store_partial(const_int(4), pDst + dstStride, tmp2uq);
-
- tmp16_1 = compress_unsafe(tmp4, tmp4);
- store_partial(const_int(4), pDst + (2 * dstStride), tmp16_1);
-
- tmp2uq = reinterpret_i(tmp16_1);
- tmp2uq >>= 32;
- store_partial(const_int(4), pDst + (3 * dstStride), tmp2uq);
- }
- else
- {
- store_partial(const_int(4), pDst, compress_unsafe(row11, row11));
- store_partial(const_int(4), pDst + (dstStride), compress_unsafe(row21, row21));
- store_partial(const_int(4), pDst + (2 * dstStride), compress_unsafe(row31, row31));
- store_partial(const_int(4), pDst + (3 * dstStride), compress_unsafe(row41, row41));
- }
+ PRED_INTRA_ANG4_END
}
-void PredIntraAng4_m_32(pixel* pDst, int dstStride, pixel *refMain, int dirMode)
+void PredIntraAng4_m_32(pixel* pDst, int dstStride, pixel *refMain, int /*dirMode*/)
{
Vec16uc tmp16_1, tmp16_2;
- dirMode++;
tmp16_1 = (Vec16uc)load_partial(const_int(8), refMain); //-1,0,1,2
store_partial(const_int(4), pDst, tmp16_1);
tmp16_2 = (Vec16uc)load_partial(const_int(8), refMain - 1); //-2,-1,0,1
@@ -2879,8 +2186,8 @@
void xPredIntraAng4x4(int /*bitDepth*/, pixel* pDst, int dstStride, int width, int dirMode, pixel *refLeft, pixel *refAbove, bool bFilter = true)
{
assert(dirMode > 1); //no planar and dc
- int mode_to_angle_table[] = {32, 26, 21, 17, 13, 9, 5, 2, 0, -2, -5, -9, -13, -17, -21, -26, -32, -26, -21, -17, -13, -9, -5, -2, 0, 2, 5, 9, 13, 17, 21, 26, 32};
- int mode_to_invAng_table[] = {256, 315, 390, 482, 630, 910, 1638, 4096, 0, 4096, 1638, 910, 630, 482, 390, 315, 256, 315, 390, 482, 630, 910, 1638, 4096, 0, 4096, 1638, 910, 630, 482, 390, 315, 256};
+ static const int mode_to_angle_table[] = {32, 26, 21, 17, 13, 9, 5, 2, 0, -2, -5, -9, -13, -17, -21, -26, -32, -26, -21, -17, -13, -9, -5, -2, 0, 2, 5, 9, 13, 17, 21, 26, 32};
+ static const int mode_to_invAng_table[] = {256, 315, 390, 482, 630, 910, 1638, 4096, 0, 4096, 1638, 910, 630, 482, 390, 315, 256, 315, 390, 482, 630, 910, 1638, 4096, 0, 4096, 1638, 910, 630, 482, 390, 315, 256};
int intraPredAngle = mode_to_angle_table[dirMode-2];
int invAngle = mode_to_invAng_table[dirMode-2];
-------------- next part --------------
A non-text attachment was scrubbed...
Name: xhevc_workspace1-2.patch
Type: text/x-patch
Size: 39148 bytes
Desc: not available
URL: <http://mailman.videolan.org/pipermail/x265-devel/attachments/20130627/6ba3957e/attachment-0001.bin>
More information about the x265-devel
mailing list