Pretty printed the C++ code with astyle

Ethan Hugg committed 12 years ago

Commit ff6b6691 parent 063709c225 files · +48457 −51954Browse files

Showing the changes to codec/decoder/core/src/deblocking.cpp (1 of 225 files). Show all

Unified · Split

codec/decoder/core/src/deblocking.cpp+795−881

@@ -34,7 +34,7 @@ · 33 unmodified lines
34	34	* \brief Interfaces introduced in frame deblocking filtering
35	35	*
36	36	* \date 08/02/2010
37		− *
	37	+ *
38	38	*************************************************************************************
39	39	*/
40	40
@@ -86,51 +86,54 @@ namespace WelsDec { · 45 unmodified lines
86	86	iBeta = g_kiBetaTable((iQp + iBetaOffset));\
87	87	}
88	88
89		−static const uint8_t g_kuiAlphaTable[52+24] = { //this table refers to Table 8-16 in H.264/AVC standard
90		− 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0,
91		− 0, 0, 0, 0, 0, 0, 0, 0, 0, 0,
92		− 0, 0, 0, 0, 0, 0, 4, 4, 5, 6,
93		− 7, 8, 9, 10, 12, 13, 15, 17, 20, 22,
94		− 25, 28, 32, 36, 40, 45, 50, 56, 63, 71,
95		− 80, 90,101,113,127,144,162,182,203,226,
96		− 255, 255
97		− ,255, 255,255, 255,255, 255,255, 255,255, 255,255, 255
	89	+static const uint8_t g_kuiAlphaTable[52 + 24] = { //this table refers to Table 8-16 in H.264/AVC standard
	90	+ 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0,
	91	+ 0, 0, 0, 0, 0, 0, 0, 0, 0, 0,
	92	+ 0, 0, 0, 0, 0, 0, 4, 4, 5, 6,
	93	+ 7, 8, 9, 10, 12, 13, 15, 17, 20, 22,
	94	+ 25, 28, 32, 36, 40, 45, 50, 56, 63, 71,
	95	+ 80, 90, 101, 113, 127, 144, 162, 182, 203, 226,
	96	+ 255, 255
	97	+ , 255, 255, 255, 255, 255, 255, 255, 255, 255, 255, 255, 255
98	98	};
99	99
100		−static const int8_t g_kiBetaTable[52+24] = { //this table refers to Table 8-16 in H.264/AVC standard
101		− 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0,
102		− 0, 0, 0, 0, 0, 0, 0, 0, 0, 0,
103		− 0, 0, 0, 0, 0, 0, 2, 2, 2, 3,
104		− 3, 3, 3, 4, 4, 4, 6, 6, 7, 7,
105		− 8, 8, 9, 9, 10, 10, 11, 11, 12, 12,
106		− 13, 13, 14, 14, 15, 15, 16, 16, 17, 17,
107		− 18, 18
108		− ,18, 18,18, 18,18, 18,18, 18,18, 18,18, 18
	100	+static const int8_t g_kiBetaTable[52 + 24] = { //this table refers to Table 8-16 in H.264/AVC standard
	101	+ 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0,
	102	+ 0, 0, 0, 0, 0, 0, 0, 0, 0, 0,
	103	+ 0, 0, 0, 0, 0, 0, 2, 2, 2, 3,
	104	+ 3, 3, 3, 4, 4, 4, 6, 6, 7, 7,
	105	+ 8, 8, 9, 9, 10, 10, 11, 11, 12, 12,
	106	+ 13, 13, 14, 14, 15, 15, 16, 16, 17, 17,
	107	+ 18, 18
	108	+ , 18, 18, 18, 18, 18, 18, 18, 18, 18, 18, 18, 18
109	109	};
110	110
111		−static const int8_t g_kiTc0Table[52+24][4] = { //this table refers Table 8-17 in H.264/AVC standard
112		− { -1, 0, 0, 0 }, { -1, 0, 0, 0 }, { -1, 0, 0, 0 }, { -1, 0, 0, 0 }, { -1, 0, 0, 0 }, { -1, 0, 0, 0 },
113		− { -1, 0, 0, 0 }, { -1, 0, 0, 0 }, { -1, 0, 0, 0 }, { -1, 0, 0, 0 }, { -1, 0, 0, 0 }, { -1, 0, 0, 0 },
114		− { -1, 0, 0, 0 }, { -1, 0, 0, 0 }, { -1, 0, 0, 0 }, { -1, 0, 0, 0 }, { -1, 0, 0, 0 }, { -1, 0, 0, 0 },
115		− { -1, 0, 0, 0 }, { -1, 0, 0, 0 }, { -1, 0, 0, 0 }, { -1, 0, 0, 0 }, { -1, 0, 0, 0 }, { -1, 0, 0, 0 },
116		− { -1, 0, 0, 0 }, { -1, 0, 0, 0 }, { -1, 0, 0, 0 }, { -1, 0, 0, 0 }, { -1, 0, 0, 0 }, { -1, 0, 0, 1 },
117		− { -1, 0, 0, 1 }, { -1, 0, 0, 1 }, { -1, 0, 0, 1 }, { -1, 0, 1, 1 }, { -1, 0, 1, 1 }, { -1, 1, 1, 1 },
118		− { -1, 1, 1, 1 }, { -1, 1, 1, 1 }, { -1, 1, 1, 1 }, { -1, 1, 1, 2 }, { -1, 1, 1, 2 }, { -1, 1, 1, 2 },
119		− { -1, 1, 1, 2 }, { -1, 1, 2, 3 }, { -1, 1, 2, 3 }, { -1, 2, 2, 3 }, { -1, 2, 2, 4 }, { -1, 2, 3, 4 },
120		− { -1, 2, 3, 4 }, { -1, 3, 3, 5 }, { -1, 3, 4, 6 }, { -1, 3, 4, 6 }, { -1, 4, 5, 7 }, { -1, 4, 5, 8 },
121		− { -1, 4, 6, 9 }, { -1, 5, 7,10 }, { -1, 6, 8,11 }, { -1, 6, 8,13 }, { -1, 7,10,14 }, { -1, 8,11,16 },
122		− { -1, 9,12,18 }, { -1, 10,13,20 }, {-1,11,15,23 }, { -1,13,17,25 }
123		− ,{ -1,13,17,25 },{ -1,13,17,25 },{ -1,13,17,25 },{ -1,13,17,25 },{ -1,13,17,25 },{ -1,13,17,25 }
124		− ,{ -1,13,17,25 },{ -1,13,17,25 },{ -1,13,17,25 },{ -1,13,17,25 },{ -1,13,17,25 },{ -1,13,17,25 }
	111	+static const int8_t g_kiTc0Table[52 + 24][4] = { //this table refers Table 8-17 in H.264/AVC standard
	112	+ { -1, 0, 0, 0 }, { -1, 0, 0, 0 }, { -1, 0, 0, 0 }, { -1, 0, 0, 0 }, { -1, 0, 0, 0 }, { -1, 0, 0, 0 },
	113	+ { -1, 0, 0, 0 }, { -1, 0, 0, 0 }, { -1, 0, 0, 0 }, { -1, 0, 0, 0 }, { -1, 0, 0, 0 }, { -1, 0, 0, 0 },
	114	+ { -1, 0, 0, 0 }, { -1, 0, 0, 0 }, { -1, 0, 0, 0 }, { -1, 0, 0, 0 }, { -1, 0, 0, 0 }, { -1, 0, 0, 0 },
	115	+ { -1, 0, 0, 0 }, { -1, 0, 0, 0 }, { -1, 0, 0, 0 }, { -1, 0, 0, 0 }, { -1, 0, 0, 0 }, { -1, 0, 0, 0 },
	116	+ { -1, 0, 0, 0 }, { -1, 0, 0, 0 }, { -1, 0, 0, 0 }, { -1, 0, 0, 0 }, { -1, 0, 0, 0 }, { -1, 0, 0, 1 },
	117	+ { -1, 0, 0, 1 }, { -1, 0, 0, 1 }, { -1, 0, 0, 1 }, { -1, 0, 1, 1 }, { -1, 0, 1, 1 }, { -1, 1, 1, 1 },
	118	+ { -1, 1, 1, 1 }, { -1, 1, 1, 1 }, { -1, 1, 1, 1 }, { -1, 1, 1, 2 }, { -1, 1, 1, 2 }, { -1, 1, 1, 2 },
	119	+ { -1, 1, 1, 2 }, { -1, 1, 2, 3 }, { -1, 1, 2, 3 }, { -1, 2, 2, 3 }, { -1, 2, 2, 4 }, { -1, 2, 3, 4 },
	120	+ { -1, 2, 3, 4 }, { -1, 3, 3, 5 }, { -1, 3, 4, 6 }, { -1, 3, 4, 6 }, { -1, 4, 5, 7 }, { -1, 4, 5, 8 },
	121	+ { -1, 4, 6, 9 }, { -1, 5, 7, 10 }, { -1, 6, 8, 11 }, { -1, 6, 8, 13 }, { -1, 7, 10, 14 }, { -1, 8, 11, 16 },
	122	+ { -1, 9, 12, 18 }, { -1, 10, 13, 20 }, { -1, 11, 15, 23 }, { -1, 13, 17, 25 }
	123	+ , { -1, 13, 17, 25 }, { -1, 13, 17, 25 }, { -1, 13, 17, 25 }, { -1, 13, 17, 25 }, { -1, 13, 17, 25 }, { -1, 13, 17, 25 }
	124	+ , { -1, 13, 17, 25 }, { -1, 13, 17, 25 }, { -1, 13, 17, 25 }, { -1, 13, 17, 25 }, { -1, 13, 17, 25 }, { -1, 13, 17, 25 }
125	125	};
126	126
127		−static const uint8_t g_kuiTableBIdx[2][8] =
128		−{
129		− {0, 4, 8, 12,
130		− 3, 7, 11, 15},
	127	+static const uint8_t g_kuiTableBIdx[2][8] = {
	128	+ {
	129	+ 0, 4, 8, 12,
	130	+ 3, 7, 11, 15
	131	+ },
131	132
132		− {0, 1, 2, 3 ,
133		− 12, 13, 14, 15},
	133	+ {
	134	+ 0, 1, 2, 3 ,
	135	+ 12, 13, 14, 15
	136	+ },
134	137	};
135	138
136	139	#define TC0_TBL_LOOKUP(tc, iIndexA, pBS, bChroma) \
@@ -141,761 +144,681 @@ static const uint8_t g_kuiTableBIdx[2][8] = · 4 unmodified lines
141	144	tc[3] = g_kiTc0Table(iIndexA)[pBS[3]] + bChroma;\
142	145	}
143	146
144		−void_t inline DeblockingBSInsideMBAvsbase( int8_t* pNnzTab, uint8_t nBS[2][4][4], int32_t iLShiftFactor )
145		−{
146		− uint32_t uiNnz32b0, uiNnz32b1, uiNnz32b2, uiNnz32b3;
147		− FORCE_STACK_ALIGN_1D( uint8_t, uiBsx3, 4, 4 );
148		−
149		− uiNnz32b0 = (uint32_t )(pNnzTab+0);
150		− uiNnz32b1 = (uint32_t )(pNnzTab+4);
151		− uiNnz32b2 = (uint32_t )(pNnzTab+8);
152		− uiNnz32b3 = (uint32_t )(pNnzTab+12);
153		−
154		− (uint32_t )uiBsx3 = (uiNnz32b0\|(uiNnz32b0>>8))<<iLShiftFactor;
155		− nBS[0][1][0] = uiBsx3[0];
156		− nBS[0][2][0] = uiBsx3[1];
157		− nBS[0][3][0] = uiBsx3[2];
158		−
159		− (uint32_t )uiBsx3 = (uiNnz32b1\|(uiNnz32b1>>8))<<iLShiftFactor;
160		− nBS[0][1][1] = uiBsx3[0];
161		− nBS[0][2][1] = uiBsx3[1];
162		− nBS[0][3][1] = uiBsx3[2];
163		− (uint32_t )nBS[1][1] = (uiNnz32b0\|uiNnz32b1)<<iLShiftFactor;
164		−
165		− (uint32_t )uiBsx3 = (uiNnz32b2\|(uiNnz32b2>>8))<<iLShiftFactor;
166		− nBS[0][1][2] = uiBsx3[0];
167		− nBS[0][2][2] = uiBsx3[1];
168		− nBS[0][3][2] = uiBsx3[2];
169		− (uint32_t )nBS[1][2] = (uiNnz32b1\|uiNnz32b2)<<iLShiftFactor;
170		−
171		− (uint32_t )uiBsx3 = (uiNnz32b3\|(uiNnz32b3>>8))<<iLShiftFactor;
172		− nBS[0][1][3] = uiBsx3[0];
173		− nBS[0][2][3] = uiBsx3[1];
174		− nBS[0][3][3] = uiBsx3[2];
175		− (uint32_t )nBS[1][3] = (uiNnz32b2\|uiNnz32b3)<<iLShiftFactor;
176		−
177		−}
178		−
179		−void_t static inline DeblockingBSInsideMBNormal( PDqLayer pCurDqLayer, uint8_t nBS[2][4][4], int8_t* pNnzTab, int32_t iMbXy )
180		−{
181		− uint32_t uiNnz32b0, uiNnz32b1, uiNnz32b2, uiNnz32b3;
182		− int8_t* iRefIndex = pCurDqLayer->pRefIndex[LIST_0][iMbXy];
183		− FORCE_STACK_ALIGN_1D( uint8_t, uiBsx4, 4, 4 );
184		−
185		− uiNnz32b0 = (uint32_t )(pNnzTab+0);
186		− uiNnz32b1 = (uint32_t )(pNnzTab+4);
187		− uiNnz32b2 = (uint32_t )(pNnzTab+8);
188		− uiNnz32b3 = (uint32_t )(pNnzTab+12);
189		−
190		− (uint32_t )uiBsx4 = (uiNnz32b0\|(uiNnz32b0>>8));
191		− nBS[0][1][0] = BS_EDGE(uiBsx4[0], iRefIndex, pCurDqLayer->pMv[LIST_0][iMbXy], 1, 0);
192		− nBS[0][2][0] = BS_EDGE(uiBsx4[1], iRefIndex, pCurDqLayer->pMv[LIST_0][iMbXy], 2, 1);
193		− nBS[0][3][0] = BS_EDGE(uiBsx4[2], iRefIndex, pCurDqLayer->pMv[LIST_0][iMbXy], 3, 2);
194		−
195		− (uint32_t )uiBsx4 = (uiNnz32b1\|(uiNnz32b1>>8));
196		− nBS[0][1][1] = BS_EDGE(uiBsx4[0], iRefIndex, pCurDqLayer->pMv[LIST_0][iMbXy], 5, 4);
197		− nBS[0][2][1] = BS_EDGE(uiBsx4[1], iRefIndex, pCurDqLayer->pMv[LIST_0][iMbXy], 6, 5);
198		− nBS[0][3][1] = BS_EDGE(uiBsx4[2], iRefIndex, pCurDqLayer->pMv[LIST_0][iMbXy], 7, 6);
199		−
200		− (uint32_t )uiBsx4 = (uiNnz32b2\|(uiNnz32b2>>8));
201		− nBS[0][1][2] = BS_EDGE(uiBsx4[0], iRefIndex, pCurDqLayer->pMv[LIST_0][iMbXy], 9, 8);
202		− nBS[0][2][2] = BS_EDGE(uiBsx4[1], iRefIndex, pCurDqLayer->pMv[LIST_0][iMbXy], 10,9);
203		− nBS[0][3][2] = BS_EDGE(uiBsx4[2], iRefIndex, pCurDqLayer->pMv[LIST_0][iMbXy], 11,10);
204		−
205		− (uint32_t )uiBsx4 = (uiNnz32b3\|(uiNnz32b3>>8));
206		− nBS[0][1][3] = BS_EDGE(uiBsx4[0], iRefIndex, pCurDqLayer->pMv[LIST_0][iMbXy], 13,12);
207		− nBS[0][2][3] = BS_EDGE(uiBsx4[1], iRefIndex, pCurDqLayer->pMv[LIST_0][iMbXy], 14,13);
208		− nBS[0][3][3] = BS_EDGE(uiBsx4[2], iRefIndex, pCurDqLayer->pMv[LIST_0][iMbXy], 15,14);
209		−
210		− // horizontal
211		− (uint32_t )uiBsx4 = (uiNnz32b0\|uiNnz32b1);
212		− nBS[1][1][0] = BS_EDGE(uiBsx4[0], iRefIndex, pCurDqLayer->pMv[LIST_0][iMbXy], 4, 0);
213		− nBS[1][1][1] = BS_EDGE(uiBsx4[1], iRefIndex, pCurDqLayer->pMv[LIST_0][iMbXy], 5, 1);
214		− nBS[1][1][2] = BS_EDGE(uiBsx4[2], iRefIndex, pCurDqLayer->pMv[LIST_0][iMbXy], 6, 2);
215		− nBS[1][1][3] = BS_EDGE(uiBsx4[3], iRefIndex, pCurDqLayer->pMv[LIST_0][iMbXy], 7, 3);
216		−
217		− (uint32_t )uiBsx4 = (uiNnz32b1\|uiNnz32b2);
218		− nBS[1][2][0] = BS_EDGE(uiBsx4[0], iRefIndex, pCurDqLayer->pMv[LIST_0][iMbXy], 8, 4);
219		− nBS[1][2][1] = BS_EDGE(uiBsx4[1], iRefIndex, pCurDqLayer->pMv[LIST_0][iMbXy], 9, 5);
220		− nBS[1][2][2] = BS_EDGE(uiBsx4[2], iRefIndex, pCurDqLayer->pMv[LIST_0][iMbXy], 10, 6);
221		− nBS[1][2][3] = BS_EDGE(uiBsx4[3], iRefIndex, pCurDqLayer->pMv[LIST_0][iMbXy], 11, 7);
222		−
223		− (uint32_t )uiBsx4 = (uiNnz32b2\|uiNnz32b3);
224		− nBS[1][3][0] = BS_EDGE(uiBsx4[0], iRefIndex, pCurDqLayer->pMv[LIST_0][iMbXy], 12, 8);
225		− nBS[1][3][1] = BS_EDGE(uiBsx4[1], iRefIndex, pCurDqLayer->pMv[LIST_0][iMbXy], 13, 9);
226		− nBS[1][3][2] = BS_EDGE(uiBsx4[2], iRefIndex, pCurDqLayer->pMv[LIST_0][iMbXy], 14, 10);
227		− nBS[1][3][3] = BS_EDGE(uiBsx4[3], iRefIndex, pCurDqLayer->pMv[LIST_0][iMbXy], 15, 11);
228		−}
229		−
230		−uint32_t DeblockingBsMarginalMBAvcbase( PDqLayer pCurDqLayer, int32_t iEdge, int32_t iNeighMb, int32_t iMbXy)
231		−{
232		− int32_t i;
233		− uint32_t uiBSx4;
234		− //uint8_t* bS = static_cast<uint8_t*>(&uiBSx4);
235		− uint8_t* pBS = (uint8_t*)(&uiBSx4);
236		− uint32_t uiBIdx = (uint32_t )(&g_kuiTableBIdx[iEdge][0]);
237		− uint32_t uiBnIdx = (uint32_t )(&g_kuiTableBIdx[iEdge][4]);
238		−
239		− for( i = 0; i < 4; i++ )
240		− {
241		− if( pCurDqLayer->pNzc[iMbXy][uiBIdx&0xff] \| pCurDqLayer->pNzc[iNeighMb][uiBnIdx&0xff] )
242		− {
243		− pBS[i] = 2;
244		− }
245		− else
246		− {
247		− pBS[i] = MB_BS_MV(pCurDqLayer->pRefIndex[LIST_0], pCurDqLayer->pMv[LIST_0], iMbXy, iNeighMb, (uiBIdx&0xff), (uiBnIdx&0xff));
248		− }
249		− uiBIdx = uiBIdx >> 8;
250		− uiBnIdx = uiBnIdx >> 8;
251		− }
252		− return uiBSx4;
253		−}
254		−int32_t DeblockingAvailableNoInterlayer( PDqLayer pCurDqLayer, int32_t iFilterIdc )
255		−{
256		− int32_t iMbY = pCurDqLayer->iMbY;
257		− int32_t iMbX = pCurDqLayer->iMbX;
258		− int32_t iMbXy = pCurDqLayer->iMbXyIndex;
259		− BOOL_T bLeftFlag = FALSE;
260		− BOOL_T bTopFlag = FALSE;
261		−
262		− if ( 2 == iFilterIdc )
263		− {
264		− bLeftFlag = ( iMbX > 0 ) && ( pCurDqLayer->pSliceIdc[iMbXy] == pCurDqLayer->pSliceIdc[iMbXy-1] );
265		− bTopFlag = ( iMbY > 0 ) && ( pCurDqLayer->pSliceIdc[iMbXy] == pCurDqLayer->pSliceIdc[iMbXy-pCurDqLayer->iMbWidth] );
266		− }
267		− else //if ( 0 == iFilterIdc )
268		− {
269		− bLeftFlag = ( iMbX > 0 );
270		− bTopFlag = ( iMbY > 0 );
271		− }
272		− return (bLeftFlag<<LEFT_FLAG_BIT)\|(bTopFlag<<TOP_FLAG_BIT);
273		−}
274		−
275		−void_t FilteringEdgeLumaH(SDeblockingFilter* pFilter, uint8_t* pPix, int32_t iStride, uint8_t* pBS )
276		−{
277		− int32_t iIndexA;
278		− int32_t iAlpha;
279		− int32_t iBeta;
280		− FORCE_STACK_ALIGN_1D( int8_t, tc, 4, 16 );
281		−
282		− GET_ALPHA_BETA_FROM_QP(pFilter->iLumaQP, pFilter->iSliceAlphaC0Offset, pFilter->iSliceBetaOffset, iIndexA, iAlpha, iBeta);
283		−
284		− if( iAlpha \| iBeta )
285		− {
286		− TC0_TBL_LOOKUP(tc, iIndexA, pBS, 0);
287		− pFilter->pLoopf->pfLumaDeblockingLT4Ver(pPix, iStride, iAlpha, iBeta, tc);
288		− }
289		− return;
290		−}
291		−
292		−
293		−void_t FilteringEdgeLumaV(SDeblockingFilter* pFilter, uint8_t* pPix, int32_t iStride, uint8_t* pBS )
294		−{
295		− int32_t iIndexA;
296		− int32_t iAlpha;
297		− int32_t iBeta;
298		− FORCE_STACK_ALIGN_1D( int8_t, tc, 4, 16 );
299		−
300		− GET_ALPHA_BETA_FROM_QP(pFilter->iLumaQP, pFilter->iSliceAlphaC0Offset, pFilter->iSliceBetaOffset, iIndexA, iAlpha, iBeta);
301		−
302		− if( iAlpha \| iBeta )
303		− {
304		− TC0_TBL_LOOKUP(tc, iIndexA, pBS, 0);
305		− pFilter->pLoopf->pfLumaDeblockingLT4Hor(pPix, iStride, iAlpha, iBeta, tc);
306		− }
307		− return;
308		−}
309		−
310		−
311		−void_t FilteringEdgeLumaIntraH( SDeblockingFilter* pFilter, uint8_t* pPix, int32_t iStride, uint8_t* pBS )
312		−{
313		− int32_t iIndexA;
314		− int32_t iAlpha;
315		− int32_t iBeta;
316		−
317		− GET_ALPHA_BETA_FROM_QP(pFilter->iLumaQP, pFilter->iSliceAlphaC0Offset, pFilter->iSliceBetaOffset, iIndexA, iAlpha, iBeta);
318		−
319		− if( iAlpha \| iBeta )
320		− {
321		− pFilter->pLoopf->pfLumaDeblockingEQ4Ver(pPix, iStride, iAlpha, iBeta);
322		− }
323		− return;
324		−}
325		−
326		−void_t FilteringEdgeLumaIntraV( SDeblockingFilter* pFilter, uint8_t* pPix, int32_t iStride, uint8_t* pBS )
327		−{
328		− int32_t iIndexA;
329		− int32_t iAlpha;
330		− int32_t iBeta;
331		−
332		− GET_ALPHA_BETA_FROM_QP(pFilter->iLumaQP, pFilter->iSliceAlphaC0Offset, pFilter->iSliceBetaOffset, iIndexA, iAlpha, iBeta);
333		−
334		− if( iAlpha \| iBeta )
335		− {
336		− pFilter->pLoopf->pfLumaDeblockingEQ4Hor(pPix, iStride, iAlpha, iBeta);
337		− }
338		− return;
339		−}
340		−void_t FilteringEdgeChromaH( SDeblockingFilter* pFilter, uint8_t* pPixCb, uint8_t* pPixCr, int32_t iStride, uint8_t* pBS )
341		−{
342		− int32_t iIndexA;
343		− int32_t iAlpha;
344		− int32_t iBeta;
345		− FORCE_STACK_ALIGN_1D( int8_t, tc, 4, 16 );
346		−
347		− GET_ALPHA_BETA_FROM_QP(pFilter->iChromaQP, pFilter->iSliceAlphaC0Offset, pFilter->iSliceBetaOffset, iIndexA, iAlpha, iBeta);
348		−
349		− if( iAlpha \| iBeta )
350		− {
351		− TC0_TBL_LOOKUP(tc, iIndexA, pBS, 1);
352		− pFilter->pLoopf->pfChromaDeblockingLT4Ver(pPixCb, pPixCr, iStride,iAlpha, iBeta, tc);
353		− }
354		− return;
355		−}
356		−void_t FilteringEdgeChromaV( SDeblockingFilter* pFilter, uint8_t* pPixCb, uint8_t* pPixCr, int32_t iStride, uint8_t* pBS )
357		−{
358		− int32_t iIndexA;
359		− int32_t iAlpha;
360		− int32_t iBeta;
361		− FORCE_STACK_ALIGN_1D( int8_t, tc, 4, 16 );
362		−
363		− GET_ALPHA_BETA_FROM_QP(pFilter->iChromaQP, pFilter->iSliceAlphaC0Offset, pFilter->iSliceBetaOffset, iIndexA, iAlpha, iBeta);
364		−
365		− if( iAlpha \| iBeta )
366		− {
367		− TC0_TBL_LOOKUP(tc, iIndexA, pBS, 1);
368		− pFilter->pLoopf->pfChromaDeblockingLT4Hor(pPixCb, pPixCr, iStride, iAlpha, iBeta, tc);
369		− }
370		− return;
371		−}
372		−
373		−void_t FilteringEdgeChromaIntraH( SDeblockingFilter* pFilter, uint8_t* pPixCb, uint8_t* pPixCr, int32_t iStride, uint8_t* pBS )
374		−{
375		− int32_t iIndexA;
376		− int32_t iAlpha;
377		− int32_t iBeta;
378		−
379		− GET_ALPHA_BETA_FROM_QP(pFilter->iChromaQP, pFilter->iSliceAlphaC0Offset, pFilter->iSliceBetaOffset, iIndexA, iAlpha, iBeta);
380		−
381		− if( iAlpha \| iBeta )
382		− {
383		− pFilter->pLoopf->pfChromaDeblockingEQ4Ver(pPixCb, pPixCr, iStride, iAlpha, iBeta);
384		− }
385		− return;
386		−}
387		−
388		−void_t FilteringEdgeChromaIntraV( SDeblockingFilter* pFilter, uint8_t* pPixCb, uint8_t* pPixCr, int32_t iStride, uint8_t* pBS )
389		−{
390		− int32_t iIndexA;
391		− int32_t iAlpha;
392		− int32_t iBeta;
393		−
394		− GET_ALPHA_BETA_FROM_QP(pFilter->iChromaQP, pFilter->iSliceAlphaC0Offset, pFilter->iSliceBetaOffset, iIndexA, iAlpha, iBeta);
395		−
396		− if( iAlpha \| iBeta )
397		− {
398		− pFilter->pLoopf->pfChromaDeblockinEQ4Hor(pPixCb, pPixCr, iStride, iAlpha, iBeta);
399		− }
400		− return;
401		−}
402		−
403		−
404		−void_t DeblockingInterMb( PDqLayer pCurDqLayer, PDeblockingFilter pFilter, uint8_t nBS[2][4][4], int32_t iBoundryFlag )
405		−{
406		− int32_t iMbXyIndex = pCurDqLayer->iMbXyIndex;
407		− int32_t iMbX = pCurDqLayer->iMbX;
408		− int32_t iMbY = pCurDqLayer->iMbY;
409		−
410		− int32_t iCurLumaQp = pCurDqLayer->pLumaQp[iMbXyIndex];
411		− int32_t iCurChromaQp = pCurDqLayer->pChromaQp[iMbXyIndex];
412		− int32_t iLineSize = pFilter->iCsStride[0];
413		− int32_t iLineSizeUV = pFilter->iCsStride[1];
414		−
415		− uint8_t * pDestY, * pDestCb, * pDestCr;
416		− pDestY = pFilter->pCsData[0] + ((iMbY * iLineSize + iMbX) << 4);
417		− pDestCb = pFilter->pCsData[1] + ((iMbY * iLineSizeUV + iMbX) << 3);
418		− pDestCr = pFilter->pCsData[2] + ((iMbY * iLineSizeUV + iMbX) << 3);
419		−
420		− if( iBoundryFlag & LEFT_FLAG_MASK)
421		− {
422		− int32_t iLeftXyIndex = iMbXyIndex - 1;
423		− pFilter->iLumaQP = (iCurLumaQp + pCurDqLayer->pLumaQp[iLeftXyIndex] + 1) >> 1;
424		− pFilter->iChromaQP = (iCurChromaQp + pCurDqLayer->pChromaQp[iLeftXyIndex]+ 1) >> 1;
425		−
426		− if( nBS[0][0][0] == 0x04 )
427		− {
428		− FilteringEdgeLumaIntraV( pFilter, pDestY, iLineSize, NULL );
429		− FilteringEdgeChromaIntraV( pFilter, pDestCb, pDestCr, iLineSizeUV, NULL );
430		− }
431		− else
432		− {
433		− if((uint32_t )nBS[0][0] != 0)
434		− {
435		− FilteringEdgeLumaV( pFilter, pDestY, iLineSize, nBS[0][0] );
436		− FilteringEdgeChromaV( pFilter, pDestCb, pDestCr, iLineSizeUV, nBS[0][0] );
437		− }
438		− }
439		− }
440		−
441		− pFilter->iLumaQP = iCurLumaQp;
442		− pFilter->iChromaQP = iCurChromaQp;
443		−
444		− if((uint32_t )nBS[0][1] != 0)
445		− {
446		− FilteringEdgeLumaV( pFilter, &pDestY[1<<2], iLineSize, nBS[0][1]);
447		− }
448		−
449		− if((uint32_t )nBS[0][2] != 0)
450		− {
451		− FilteringEdgeLumaV( pFilter, &pDestY[2<<2], iLineSize, nBS[0][2]);
452		− FilteringEdgeChromaV( pFilter, &pDestCb[2<<1], &pDestCr[2<<1], iLineSizeUV, nBS[0][2] );
453		− }
454		−
455		− if((uint32_t )nBS[0][3] != 0)
456		− {
457		− FilteringEdgeLumaV( pFilter, &pDestY[3<<2], iLineSize, nBS[0][3] );
458		− }
459		−
460		− if( iBoundryFlag & TOP_FLAG_MASK)
461		− {
462		− int32_t iTopXyIndex = iMbXyIndex - pCurDqLayer->iMbWidth;
463		− pFilter->iLumaQP = (iCurLumaQp + pCurDqLayer->pLumaQp[iTopXyIndex] + 1) >> 1;
464		− pFilter->iChromaQP = (iCurChromaQp + pCurDqLayer->pChromaQp[iTopXyIndex] + 1) >> 1;
465		−
466		− if( nBS[1][0][0] == 0x04)
467		− {
468		− FilteringEdgeLumaIntraH( pFilter, pDestY, iLineSize, NULL );
469		− FilteringEdgeChromaIntraH( pFilter, pDestCb, pDestCr, iLineSizeUV, NULL );
470		− }
471		− else
472		− {
473		− if((uint32_t )nBS[1][0] != 0)
474		− {
475		− FilteringEdgeLumaH( pFilter, pDestY, iLineSize, nBS[1][0] );
476		− FilteringEdgeChromaH( pFilter, pDestCb, pDestCr, iLineSizeUV, nBS[1][0] );
477		− }
478		− }
479		− }
480		−
481		− pFilter->iLumaQP = iCurLumaQp;
482		− pFilter->iChromaQP = iCurChromaQp;
483		−
484		− if((uint32_t )nBS[1][1] != 0)
485		− {
486		− FilteringEdgeLumaH( pFilter, &pDestY[(1<<2)*iLineSize], iLineSize, nBS[1][1] );
487		− }
488		−
489		− if((uint32_t )nBS[1][2] != 0)
490		− {
491		− FilteringEdgeLumaH( pFilter, &pDestY[(2<<2)*iLineSize], iLineSize, nBS[1][2] );
492		− FilteringEdgeChromaH( pFilter, &pDestCb[(2<<1)iLineSizeUV], &pDestCr[(2<<1)iLineSizeUV], iLineSizeUV, nBS[1][2] );
493		− }
494		−
495		− if((uint32_t )nBS[1][3] != 0)
496		− {
497		− FilteringEdgeLumaH( pFilter, &pDestY[(3<<2)*iLineSize], iLineSize, nBS[1][3] );
498		− }
499		−}
500		−
501		−void_t /__FASTCALL/ FilteringEdgeLumaHV( PDqLayer pCurDqLayer, PDeblockingFilter pFilter, int32_t iBoundryFlag )
502		−{
503		− int32_t iMbXyIndex = pCurDqLayer->iMbXyIndex;
504		− int32_t iMbX = pCurDqLayer->iMbX;
505		− int32_t iMbY = pCurDqLayer->iMbY;
506		− int32_t iMbWidth = pCurDqLayer->iMbWidth;
507		− int32_t iLineSize = pFilter->iCsStride[0];
508		−
509		− uint8_t *pDestY;
510		− int32_t iCurQp;
511		− int32_t iIndexA, iAlpha, iBeta;
512		−
513		− FORCE_STACK_ALIGN_1D(int8_t, iTc, 4, 16 );
514		− FORCE_STACK_ALIGN_1D(uint8_t, uiBSx4, 4, 4 );
515		−
516		− pDestY = pFilter->pCsData[0] + ((iMbY * iLineSize + iMbX) << 4);
517		− iCurQp = pCurDqLayer->pLumaQp[iMbXyIndex];
518		−
519		− (uint32_t)uiBSx4 = 0x03030303;
520		−
521		− // luma v
522		− if( iBoundryFlag & LEFT_FLAG_MASK)
523		− {
524		− pFilter->iLumaQP = ( iCurQp + pCurDqLayer->pLumaQp[iMbXyIndex-1] + 1 ) >> 1;
525		− FilteringEdgeLumaIntraV( pFilter, pDestY, iLineSize, NULL );
526		− }
527		−
528		− pFilter->iLumaQP = iCurQp;
529		− GET_ALPHA_BETA_FROM_QP(pFilter->iLumaQP, pFilter->iSliceAlphaC0Offset, pFilter->iSliceBetaOffset, iIndexA, iAlpha, iBeta);
530		− if( iAlpha \| iBeta )
531		− {
532		− TC0_TBL_LOOKUP(iTc, iIndexA, uiBSx4, 0);
533		− pFilter->pLoopf->pfLumaDeblockingLT4Hor( &pDestY[1 << 2],iLineSize,iAlpha,iBeta,iTc );
534		− pFilter->pLoopf->pfLumaDeblockingLT4Hor( &pDestY[2 << 2],iLineSize,iAlpha,iBeta,iTc );
535		− pFilter->pLoopf->pfLumaDeblockingLT4Hor( &pDestY[3 << 2],iLineSize,iAlpha,iBeta,iTc );
536		− }
537		−
538		− // luma h
539		− if( iBoundryFlag & TOP_FLAG_MASK)
540		− {
541		− pFilter->iLumaQP = ( iCurQp + pCurDqLayer->pLumaQp[iMbXyIndex-iMbWidth] + 1 ) >> 1;
542		− FilteringEdgeLumaIntraH( pFilter, pDestY, iLineSize, NULL );
543		− }
544		−
545		− pFilter->iLumaQP = iCurQp;
546		− if( iAlpha \| iBeta )
547		− {
548		− pFilter->pLoopf->pfLumaDeblockingLT4Ver( &pDestY[(1<<2)*iLineSize],iLineSize,iAlpha,iBeta,iTc );
549		− pFilter->pLoopf->pfLumaDeblockingLT4Ver( &pDestY[(2<<2)*iLineSize],iLineSize,iAlpha,iBeta,iTc );
550		− pFilter->pLoopf->pfLumaDeblockingLT4Ver( &pDestY[(3<<2)*iLineSize],iLineSize,iAlpha,iBeta,iTc );
551		− }
552		−}
553		−void_t /__FASTCALL/ FilteringEdgeChromaHV( PDqLayer pCurDqLayer, PDeblockingFilter pFilter, int32_t iBoundryFlag )
554		−{
555		− int32_t iMbXyIndex = pCurDqLayer->iMbXyIndex;
556		− int32_t iMbX = pCurDqLayer->iMbX;
557		− int32_t iMbY = pCurDqLayer->iMbY;
558		− int32_t iMbWidth = pCurDqLayer->iMbWidth;
559		− int32_t iLineSize = pFilter->iCsStride[1];
560		−
561		− uint8_t pDestCb, pDestCr;
562		− int32_t iCurQp;
563		− int32_t iIndexA, iAlpha, iBeta;
564		−
565		− FORCE_STACK_ALIGN_1D( int8_t, iTc, 4, 16 );
566		− FORCE_STACK_ALIGN_1D( uint8_t, uiBSx4, 4, 4 );
567		−
568		− pDestCb = pFilter->pCsData[1] + ((iMbY * iLineSize + iMbX) << 3);
569		− pDestCr = pFilter->pCsData[2] + ((iMbY * iLineSize + iMbX) << 3);
570		− iCurQp = pCurDqLayer->pChromaQp[iMbXyIndex];
571		− (uint32_t)uiBSx4 = 0x03030303;
572		−
573		− // chroma v
574		− if( iBoundryFlag & LEFT_FLAG_MASK)
575		− {
576		− pFilter->iChromaQP = ( iCurQp + pCurDqLayer->pChromaQp[iMbXyIndex-1] + 1 ) >> 1;
577		− FilteringEdgeChromaIntraV( pFilter, pDestCb, pDestCr, iLineSize, NULL);
578		− }
579		−
580		− pFilter->iChromaQP = iCurQp;
581		− GET_ALPHA_BETA_FROM_QP(pFilter->iChromaQP, pFilter->iSliceAlphaC0Offset, pFilter->iSliceBetaOffset, iIndexA, iAlpha, iBeta);
582		− if( iAlpha \| iBeta )
583		− {
584		− TC0_TBL_LOOKUP(iTc, iIndexA, uiBSx4, 1);
585		− pFilter->pLoopf->pfChromaDeblockingLT4Hor( &pDestCb[2 << 1],&pDestCr[2 << 1],iLineSize,iAlpha,iBeta,iTc );
586		− }
587		−
588		− // chroma h
589		− if( iBoundryFlag & TOP_FLAG_MASK)
590		− {
591		− pFilter->iChromaQP = ( iCurQp + pCurDqLayer->pChromaQp[iMbXyIndex-iMbWidth] + 1 ) >> 1;
592		− FilteringEdgeChromaIntraH( pFilter, pDestCb, pDestCr, iLineSize, NULL);
593		− }
594		−
595		− pFilter->iChromaQP = iCurQp;
596		− if( iAlpha \| iBeta )
597		− {
598		− pFilter->pLoopf->pfChromaDeblockingLT4Ver( &pDestCb[(2<<1)iLineSize],&pDestCr[(2<<1)iLineSize],iLineSize,iAlpha,iBeta,iTc );
599		− }
	147	+void_t inline DeblockingBSInsideMBAvsbase (int8_t* pNnzTab, uint8_t nBS[2][4][4], int32_t iLShiftFactor) {
	148	+ uint32_t uiNnz32b0, uiNnz32b1, uiNnz32b2, uiNnz32b3;
	149	+ FORCE_STACK_ALIGN_1D (uint8_t, uiBsx3, 4, 4);
	150	+
	151	+ uiNnz32b0 = * (uint32_t*) (pNnzTab + 0);
	152	+ uiNnz32b1 = * (uint32_t*) (pNnzTab + 4);
	153	+ uiNnz32b2 = * (uint32_t*) (pNnzTab + 8);
	154	+ uiNnz32b3 = * (uint32_t*) (pNnzTab + 12);
	155	+
	156	+ * (uint32_t*)uiBsx3 = (uiNnz32b0 \| (uiNnz32b0 >> 8)) << iLShiftFactor;
	157	+ nBS[0][1][0] = uiBsx3[0];
	158	+ nBS[0][2][0] = uiBsx3[1];
	159	+ nBS[0][3][0] = uiBsx3[2];
	160	+
	161	+ * (uint32_t*)uiBsx3 = (uiNnz32b1 \| (uiNnz32b1 >> 8)) << iLShiftFactor;
	162	+ nBS[0][1][1] = uiBsx3[0];
	163	+ nBS[0][2][1] = uiBsx3[1];
	164	+ nBS[0][3][1] = uiBsx3[2];
	165	+ * (uint32_t*)nBS[1][1] = (uiNnz32b0 \| uiNnz32b1) << iLShiftFactor;
	166	+
	167	+ * (uint32_t*)uiBsx3 = (uiNnz32b2 \| (uiNnz32b2 >> 8)) << iLShiftFactor;
	168	+ nBS[0][1][2] = uiBsx3[0];
	169	+ nBS[0][2][2] = uiBsx3[1];
	170	+ nBS[0][3][2] = uiBsx3[2];
	171	+ * (uint32_t*)nBS[1][2] = (uiNnz32b1 \| uiNnz32b2) << iLShiftFactor;
	172	+
	173	+ * (uint32_t*)uiBsx3 = (uiNnz32b3 \| (uiNnz32b3 >> 8)) << iLShiftFactor;
	174	+ nBS[0][1][3] = uiBsx3[0];
	175	+ nBS[0][2][3] = uiBsx3[1];
	176	+ nBS[0][3][3] = uiBsx3[2];
	177	+ * (uint32_t*)nBS[1][3] = (uiNnz32b2 \| uiNnz32b3) << iLShiftFactor;
	178	+
	179	+}
	180	+
	181	+void_t static inline DeblockingBSInsideMBNormal (PDqLayer pCurDqLayer, uint8_t nBS[2][4][4], int8_t* pNnzTab,
	182	+ int32_t iMbXy) {
	183	+ uint32_t uiNnz32b0, uiNnz32b1, uiNnz32b2, uiNnz32b3;
	184	+ int8_t* iRefIndex = pCurDqLayer->pRefIndex[LIST_0][iMbXy];
	185	+ FORCE_STACK_ALIGN_1D (uint8_t, uiBsx4, 4, 4);
	186	+
	187	+ uiNnz32b0 = * (uint32_t*) (pNnzTab + 0);
	188	+ uiNnz32b1 = * (uint32_t*) (pNnzTab + 4);
	189	+ uiNnz32b2 = * (uint32_t*) (pNnzTab + 8);
	190	+ uiNnz32b3 = * (uint32_t*) (pNnzTab + 12);
	191	+
	192	+ * (uint32_t*)uiBsx4 = (uiNnz32b0 \| (uiNnz32b0 >> 8));
	193	+ nBS[0][1][0] = BS_EDGE (uiBsx4[0], iRefIndex, pCurDqLayer->pMv[LIST_0][iMbXy], 1, 0);
	194	+ nBS[0][2][0] = BS_EDGE (uiBsx4[1], iRefIndex, pCurDqLayer->pMv[LIST_0][iMbXy], 2, 1);
	195	+ nBS[0][3][0] = BS_EDGE (uiBsx4[2], iRefIndex, pCurDqLayer->pMv[LIST_0][iMbXy], 3, 2);
	196	+
	197	+ * (uint32_t*)uiBsx4 = (uiNnz32b1 \| (uiNnz32b1 >> 8));
	198	+ nBS[0][1][1] = BS_EDGE (uiBsx4[0], iRefIndex, pCurDqLayer->pMv[LIST_0][iMbXy], 5, 4);
	199	+ nBS[0][2][1] = BS_EDGE (uiBsx4[1], iRefIndex, pCurDqLayer->pMv[LIST_0][iMbXy], 6, 5);
	200	+ nBS[0][3][1] = BS_EDGE (uiBsx4[2], iRefIndex, pCurDqLayer->pMv[LIST_0][iMbXy], 7, 6);
	201	+
	202	+ * (uint32_t*)uiBsx4 = (uiNnz32b2 \| (uiNnz32b2 >> 8));
	203	+ nBS[0][1][2] = BS_EDGE (uiBsx4[0], iRefIndex, pCurDqLayer->pMv[LIST_0][iMbXy], 9, 8);
	204	+ nBS[0][2][2] = BS_EDGE (uiBsx4[1], iRefIndex, pCurDqLayer->pMv[LIST_0][iMbXy], 10, 9);
	205	+ nBS[0][3][2] = BS_EDGE (uiBsx4[2], iRefIndex, pCurDqLayer->pMv[LIST_0][iMbXy], 11, 10);
	206	+
	207	+ * (uint32_t*)uiBsx4 = (uiNnz32b3 \| (uiNnz32b3 >> 8));
	208	+ nBS[0][1][3] = BS_EDGE (uiBsx4[0], iRefIndex, pCurDqLayer->pMv[LIST_0][iMbXy], 13, 12);
	209	+ nBS[0][2][3] = BS_EDGE (uiBsx4[1], iRefIndex, pCurDqLayer->pMv[LIST_0][iMbXy], 14, 13);
	210	+ nBS[0][3][3] = BS_EDGE (uiBsx4[2], iRefIndex, pCurDqLayer->pMv[LIST_0][iMbXy], 15, 14);
	211	+
	212	+ // horizontal
	213	+ * (uint32_t*)uiBsx4 = (uiNnz32b0 \| uiNnz32b1);
	214	+ nBS[1][1][0] = BS_EDGE (uiBsx4[0], iRefIndex, pCurDqLayer->pMv[LIST_0][iMbXy], 4, 0);
	215	+ nBS[1][1][1] = BS_EDGE (uiBsx4[1], iRefIndex, pCurDqLayer->pMv[LIST_0][iMbXy], 5, 1);
	216	+ nBS[1][1][2] = BS_EDGE (uiBsx4[2], iRefIndex, pCurDqLayer->pMv[LIST_0][iMbXy], 6, 2);
	217	+ nBS[1][1][3] = BS_EDGE (uiBsx4[3], iRefIndex, pCurDqLayer->pMv[LIST_0][iMbXy], 7, 3);
	218	+
	219	+ * (uint32_t*)uiBsx4 = (uiNnz32b1 \| uiNnz32b2);
	220	+ nBS[1][2][0] = BS_EDGE (uiBsx4[0], iRefIndex, pCurDqLayer->pMv[LIST_0][iMbXy], 8, 4);
	221	+ nBS[1][2][1] = BS_EDGE (uiBsx4[1], iRefIndex, pCurDqLayer->pMv[LIST_0][iMbXy], 9, 5);
	222	+ nBS[1][2][2] = BS_EDGE (uiBsx4[2], iRefIndex, pCurDqLayer->pMv[LIST_0][iMbXy], 10, 6);
	223	+ nBS[1][2][3] = BS_EDGE (uiBsx4[3], iRefIndex, pCurDqLayer->pMv[LIST_0][iMbXy], 11, 7);
	224	+
	225	+ * (uint32_t*)uiBsx4 = (uiNnz32b2 \| uiNnz32b3);
	226	+ nBS[1][3][0] = BS_EDGE (uiBsx4[0], iRefIndex, pCurDqLayer->pMv[LIST_0][iMbXy], 12, 8);
	227	+ nBS[1][3][1] = BS_EDGE (uiBsx4[1], iRefIndex, pCurDqLayer->pMv[LIST_0][iMbXy], 13, 9);
	228	+ nBS[1][3][2] = BS_EDGE (uiBsx4[2], iRefIndex, pCurDqLayer->pMv[LIST_0][iMbXy], 14, 10);
	229	+ nBS[1][3][3] = BS_EDGE (uiBsx4[3], iRefIndex, pCurDqLayer->pMv[LIST_0][iMbXy], 15, 11);
	230	+}
	231	+
	232	+uint32_t DeblockingBsMarginalMBAvcbase (PDqLayer pCurDqLayer, int32_t iEdge, int32_t iNeighMb, int32_t iMbXy) {
	233	+ int32_t i;
	234	+ uint32_t uiBSx4;
	235	+ //uint8_t* bS = static_cast<uint8_t*>(&uiBSx4);
	236	+ uint8_t* pBS = (uint8_t*) (&uiBSx4);
	237	+ uint32_t uiBIdx = * (uint32_t*) (&g_kuiTableBIdx[iEdge][0]);
	238	+ uint32_t uiBnIdx = * (uint32_t*) (&g_kuiTableBIdx[iEdge][4]);
	239	+
	240	+ for (i = 0; i < 4; i++) {
	241	+ if (pCurDqLayer->pNzc[iMbXy][uiBIdx & 0xff] \| pCurDqLayer->pNzc[iNeighMb][uiBnIdx & 0xff]) {
	242	+ pBS[i] = 2;
	243	+ } else {
	244	+ pBS[i] = MB_BS_MV (pCurDqLayer->pRefIndex[LIST_0], pCurDqLayer->pMv[LIST_0], iMbXy, iNeighMb, (uiBIdx & 0xff),
	245	+ (uiBnIdx & 0xff));
	246	+ }
	247	+ uiBIdx = uiBIdx >> 8;
	248	+ uiBnIdx = uiBnIdx >> 8;
	249	+ }
	250	+ return uiBSx4;
	251	+}
	252	+int32_t DeblockingAvailableNoInterlayer (PDqLayer pCurDqLayer, int32_t iFilterIdc) {
	253	+ int32_t iMbY = pCurDqLayer->iMbY;
	254	+ int32_t iMbX = pCurDqLayer->iMbX;
	255	+ int32_t iMbXy = pCurDqLayer->iMbXyIndex;
	256	+ BOOL_T bLeftFlag = FALSE;
	257	+ BOOL_T bTopFlag = FALSE;
	258	+
	259	+ if (2 == iFilterIdc) {
	260	+ bLeftFlag = (iMbX > 0) && (pCurDqLayer->pSliceIdc[iMbXy] == pCurDqLayer->pSliceIdc[iMbXy - 1]);
	261	+ bTopFlag = (iMbY > 0) && (pCurDqLayer->pSliceIdc[iMbXy] == pCurDqLayer->pSliceIdc[iMbXy - pCurDqLayer->iMbWidth]);
	262	+ } else { //if ( 0 == iFilterIdc )
	263	+ bLeftFlag = (iMbX > 0);
	264	+ bTopFlag = (iMbY > 0);
	265	+ }
	266	+ return (bLeftFlag << LEFT_FLAG_BIT) \| (bTopFlag << TOP_FLAG_BIT);
	267	+}
	268	+
	269	+void_t FilteringEdgeLumaH (SDeblockingFilter* pFilter, uint8_t* pPix, int32_t iStride, uint8_t* pBS) {
	270	+ int32_t iIndexA;
	271	+ int32_t iAlpha;
	272	+ int32_t iBeta;
	273	+ FORCE_STACK_ALIGN_1D (int8_t, tc, 4, 16);
	274	+
	275	+ GET_ALPHA_BETA_FROM_QP (pFilter->iLumaQP, pFilter->iSliceAlphaC0Offset, pFilter->iSliceBetaOffset, iIndexA, iAlpha,
	276	+ iBeta);
	277	+
	278	+ if (iAlpha \| iBeta) {
	279	+ TC0_TBL_LOOKUP (tc, iIndexA, pBS, 0);
	280	+ pFilter->pLoopf->pfLumaDeblockingLT4Ver (pPix, iStride, iAlpha, iBeta, tc);
	281	+ }
	282	+ return;
	283	+}
	284	+
	285	+
	286	+void_t FilteringEdgeLumaV (SDeblockingFilter* pFilter, uint8_t* pPix, int32_t iStride, uint8_t* pBS) {
	287	+ int32_t iIndexA;
	288	+ int32_t iAlpha;
	289	+ int32_t iBeta;
	290	+ FORCE_STACK_ALIGN_1D (int8_t, tc, 4, 16);
	291	+
	292	+ GET_ALPHA_BETA_FROM_QP (pFilter->iLumaQP, pFilter->iSliceAlphaC0Offset, pFilter->iSliceBetaOffset, iIndexA, iAlpha,
	293	+ iBeta);
	294	+
	295	+ if (iAlpha \| iBeta) {
	296	+ TC0_TBL_LOOKUP (tc, iIndexA, pBS, 0);
	297	+ pFilter->pLoopf->pfLumaDeblockingLT4Hor (pPix, iStride, iAlpha, iBeta, tc);
	298	+ }
	299	+ return;
	300	+}
	301	+
	302	+
	303	+void_t FilteringEdgeLumaIntraH (SDeblockingFilter* pFilter, uint8_t* pPix, int32_t iStride, uint8_t* pBS) {
	304	+ int32_t iIndexA;
	305	+ int32_t iAlpha;
	306	+ int32_t iBeta;
	307	+
	308	+ GET_ALPHA_BETA_FROM_QP (pFilter->iLumaQP, pFilter->iSliceAlphaC0Offset, pFilter->iSliceBetaOffset, iIndexA, iAlpha,
	309	+ iBeta);
	310	+
	311	+ if (iAlpha \| iBeta) {
	312	+ pFilter->pLoopf->pfLumaDeblockingEQ4Ver (pPix, iStride, iAlpha, iBeta);
	313	+ }
	314	+ return;
	315	+}
	316	+
	317	+void_t FilteringEdgeLumaIntraV (SDeblockingFilter* pFilter, uint8_t* pPix, int32_t iStride, uint8_t* pBS) {
	318	+ int32_t iIndexA;
	319	+ int32_t iAlpha;
	320	+ int32_t iBeta;
	321	+
	322	+ GET_ALPHA_BETA_FROM_QP (pFilter->iLumaQP, pFilter->iSliceAlphaC0Offset, pFilter->iSliceBetaOffset, iIndexA, iAlpha,
	323	+ iBeta);
	324	+
	325	+ if (iAlpha \| iBeta) {
	326	+ pFilter->pLoopf->pfLumaDeblockingEQ4Hor (pPix, iStride, iAlpha, iBeta);
	327	+ }
	328	+ return;
	329	+}
	330	+void_t FilteringEdgeChromaH (SDeblockingFilter* pFilter, uint8_t* pPixCb, uint8_t* pPixCr, int32_t iStride,
	331	+ uint8_t* pBS) {
	332	+ int32_t iIndexA;
	333	+ int32_t iAlpha;
	334	+ int32_t iBeta;
	335	+ FORCE_STACK_ALIGN_1D (int8_t, tc, 4, 16);
	336	+
	337	+ GET_ALPHA_BETA_FROM_QP (pFilter->iChromaQP, pFilter->iSliceAlphaC0Offset, pFilter->iSliceBetaOffset, iIndexA, iAlpha,
	338	+ iBeta);
	339	+
	340	+ if (iAlpha \| iBeta) {
	341	+ TC0_TBL_LOOKUP (tc, iIndexA, pBS, 1);
	342	+ pFilter->pLoopf->pfChromaDeblockingLT4Ver (pPixCb, pPixCr, iStride, iAlpha, iBeta, tc);
	343	+ }
	344	+ return;
	345	+}
	346	+void_t FilteringEdgeChromaV (SDeblockingFilter* pFilter, uint8_t* pPixCb, uint8_t* pPixCr, int32_t iStride,
	347	+ uint8_t* pBS) {
	348	+ int32_t iIndexA;
	349	+ int32_t iAlpha;
	350	+ int32_t iBeta;
	351	+ FORCE_STACK_ALIGN_1D (int8_t, tc, 4, 16);
	352	+
	353	+ GET_ALPHA_BETA_FROM_QP (pFilter->iChromaQP, pFilter->iSliceAlphaC0Offset, pFilter->iSliceBetaOffset, iIndexA, iAlpha,
	354	+ iBeta);
	355	+
	356	+ if (iAlpha \| iBeta) {
	357	+ TC0_TBL_LOOKUP (tc, iIndexA, pBS, 1);
	358	+ pFilter->pLoopf->pfChromaDeblockingLT4Hor (pPixCb, pPixCr, iStride, iAlpha, iBeta, tc);
	359	+ }
	360	+ return;
	361	+}
	362	+
	363	+void_t FilteringEdgeChromaIntraH (SDeblockingFilter* pFilter, uint8_t* pPixCb, uint8_t* pPixCr, int32_t iStride,
	364	+ uint8_t* pBS) {
	365	+ int32_t iIndexA;
	366	+ int32_t iAlpha;
	367	+ int32_t iBeta;
	368	+
	369	+ GET_ALPHA_BETA_FROM_QP (pFilter->iChromaQP, pFilter->iSliceAlphaC0Offset, pFilter->iSliceBetaOffset, iIndexA, iAlpha,
	370	+ iBeta);
	371	+
	372	+ if (iAlpha \| iBeta) {
	373	+ pFilter->pLoopf->pfChromaDeblockingEQ4Ver (pPixCb, pPixCr, iStride, iAlpha, iBeta);
	374	+ }
	375	+ return;
	376	+}
	377	+
	378	+void_t FilteringEdgeChromaIntraV (SDeblockingFilter* pFilter, uint8_t* pPixCb, uint8_t* pPixCr, int32_t iStride,
	379	+ uint8_t* pBS) {
	380	+ int32_t iIndexA;
	381	+ int32_t iAlpha;
	382	+ int32_t iBeta;
	383	+
	384	+ GET_ALPHA_BETA_FROM_QP (pFilter->iChromaQP, pFilter->iSliceAlphaC0Offset, pFilter->iSliceBetaOffset, iIndexA, iAlpha,
	385	+ iBeta);
	386	+
	387	+ if (iAlpha \| iBeta) {
	388	+ pFilter->pLoopf->pfChromaDeblockinEQ4Hor (pPixCb, pPixCr, iStride, iAlpha, iBeta);
	389	+ }
	390	+ return;
	391	+}
	392	+
	393	+
	394	+void_t DeblockingInterMb (PDqLayer pCurDqLayer, PDeblockingFilter pFilter, uint8_t nBS[2][4][4],
	395	+ int32_t iBoundryFlag) {
	396	+ int32_t iMbXyIndex = pCurDqLayer->iMbXyIndex;
	397	+ int32_t iMbX = pCurDqLayer->iMbX;
	398	+ int32_t iMbY = pCurDqLayer->iMbY;
	399	+
	400	+ int32_t iCurLumaQp = pCurDqLayer->pLumaQp[iMbXyIndex];
	401	+ int32_t iCurChromaQp = pCurDqLayer->pChromaQp[iMbXyIndex];
	402	+ int32_t iLineSize = pFilter->iCsStride[0];
	403	+ int32_t iLineSizeUV = pFilter->iCsStride[1];
	404	+
	405	+ uint8_t* pDestY, * pDestCb, * pDestCr;
	406	+ pDestY = pFilter->pCsData[0] + ((iMbY * iLineSize + iMbX) << 4);
	407	+ pDestCb = pFilter->pCsData[1] + ((iMbY * iLineSizeUV + iMbX) << 3);
	408	+ pDestCr = pFilter->pCsData[2] + ((iMbY * iLineSizeUV + iMbX) << 3);
	409	+
	410	+ if (iBoundryFlag & LEFT_FLAG_MASK) {
	411	+ int32_t iLeftXyIndex = iMbXyIndex - 1;
	412	+ pFilter->iLumaQP = (iCurLumaQp + pCurDqLayer->pLumaQp[iLeftXyIndex] + 1) >> 1;
	413	+ pFilter->iChromaQP = (iCurChromaQp + pCurDqLayer->pChromaQp[iLeftXyIndex] + 1) >> 1;
	414	+
	415	+ if (nBS[0][0][0] == 0x04) {
	416	+ FilteringEdgeLumaIntraV (pFilter, pDestY, iLineSize, NULL);
	417	+ FilteringEdgeChromaIntraV (pFilter, pDestCb, pDestCr, iLineSizeUV, NULL);
	418	+ } else {
	419	+ if (* (uint32_t*)nBS[0][0] != 0) {
	420	+ FilteringEdgeLumaV (pFilter, pDestY, iLineSize, nBS[0][0]);
	421	+ FilteringEdgeChromaV (pFilter, pDestCb, pDestCr, iLineSizeUV, nBS[0][0]);
	422	+ }
	423	+ }
	424	+ }
	425	+
	426	+ pFilter->iLumaQP = iCurLumaQp;
	427	+ pFilter->iChromaQP = iCurChromaQp;
	428	+
	429	+ if (* (uint32_t*)nBS[0][1] != 0) {
	430	+ FilteringEdgeLumaV (pFilter, &pDestY[1 << 2], iLineSize, nBS[0][1]);
	431	+ }
	432	+
	433	+ if (* (uint32_t*)nBS[0][2] != 0) {
	434	+ FilteringEdgeLumaV (pFilter, &pDestY[2 << 2], iLineSize, nBS[0][2]);
	435	+ FilteringEdgeChromaV (pFilter, &pDestCb[2 << 1], &pDestCr[2 << 1], iLineSizeUV, nBS[0][2]);
	436	+ }
	437	+
	438	+ if (* (uint32_t*)nBS[0][3] != 0) {
	439	+ FilteringEdgeLumaV (pFilter, &pDestY[3 << 2], iLineSize, nBS[0][3]);
	440	+ }
	441	+
	442	+ if (iBoundryFlag & TOP_FLAG_MASK) {
	443	+ int32_t iTopXyIndex = iMbXyIndex - pCurDqLayer->iMbWidth;
	444	+ pFilter->iLumaQP = (iCurLumaQp + pCurDqLayer->pLumaQp[iTopXyIndex] + 1) >> 1;
	445	+ pFilter->iChromaQP = (iCurChromaQp + pCurDqLayer->pChromaQp[iTopXyIndex] + 1) >> 1;
	446	+
	447	+ if (nBS[1][0][0] == 0x04) {
	448	+ FilteringEdgeLumaIntraH (pFilter, pDestY, iLineSize, NULL);
	449	+ FilteringEdgeChromaIntraH (pFilter, pDestCb, pDestCr, iLineSizeUV, NULL);
	450	+ } else {
	451	+ if (* (uint32_t*)nBS[1][0] != 0) {
	452	+ FilteringEdgeLumaH (pFilter, pDestY, iLineSize, nBS[1][0]);
	453	+ FilteringEdgeChromaH (pFilter, pDestCb, pDestCr, iLineSizeUV, nBS[1][0]);
	454	+ }
	455	+ }
	456	+ }
	457	+
	458	+ pFilter->iLumaQP = iCurLumaQp;
	459	+ pFilter->iChromaQP = iCurChromaQp;
	460	+
	461	+ if (* (uint32_t*)nBS[1][1] != 0) {
	462	+ FilteringEdgeLumaH (pFilter, &pDestY[ (1 << 2)*iLineSize], iLineSize, nBS[1][1]);
	463	+ }
	464	+
	465	+ if (* (uint32_t*)nBS[1][2] != 0) {
	466	+ FilteringEdgeLumaH (pFilter, &pDestY[ (2 << 2)*iLineSize], iLineSize, nBS[1][2]);
	467	+ FilteringEdgeChromaH (pFilter, &pDestCb[ (2 << 1)iLineSizeUV], &pDestCr[ (2 << 1)iLineSizeUV], iLineSizeUV,
	468	+ nBS[1][2]);
	469	+ }
	470	+
	471	+ if (* (uint32_t*)nBS[1][3] != 0) {
	472	+ FilteringEdgeLumaH (pFilter, &pDestY[ (3 << 2)*iLineSize], iLineSize, nBS[1][3]);
	473	+ }
	474	+}
	475	+
	476	+void_t /__FASTCALL/ FilteringEdgeLumaHV (PDqLayer pCurDqLayer, PDeblockingFilter pFilter, int32_t iBoundryFlag) {
	477	+ int32_t iMbXyIndex = pCurDqLayer->iMbXyIndex;
	478	+ int32_t iMbX = pCurDqLayer->iMbX;
	479	+ int32_t iMbY = pCurDqLayer->iMbY;
	480	+ int32_t iMbWidth = pCurDqLayer->iMbWidth;
	481	+ int32_t iLineSize = pFilter->iCsStride[0];
	482	+
	483	+ uint8_t* pDestY;
	484	+ int32_t iCurQp;
	485	+ int32_t iIndexA, iAlpha, iBeta;
	486	+
	487	+ FORCE_STACK_ALIGN_1D (int8_t, iTc, 4, 16);
	488	+ FORCE_STACK_ALIGN_1D (uint8_t, uiBSx4, 4, 4);
	489	+
	490	+ pDestY = pFilter->pCsData[0] + ((iMbY * iLineSize + iMbX) << 4);
	491	+ iCurQp = pCurDqLayer->pLumaQp[iMbXyIndex];
	492	+
	493	+ * (uint32_t*)uiBSx4 = 0x03030303;
	494	+
	495	+ // luma v
	496	+ if (iBoundryFlag & LEFT_FLAG_MASK) {
	497	+ pFilter->iLumaQP = (iCurQp + pCurDqLayer->pLumaQp[iMbXyIndex - 1] + 1) >> 1;
	498	+ FilteringEdgeLumaIntraV (pFilter, pDestY, iLineSize, NULL);
	499	+ }
	500	+
	501	+ pFilter->iLumaQP = iCurQp;
	502	+ GET_ALPHA_BETA_FROM_QP (pFilter->iLumaQP, pFilter->iSliceAlphaC0Offset, pFilter->iSliceBetaOffset, iIndexA, iAlpha,
	503	+ iBeta);
	504	+ if (iAlpha \| iBeta) {
	505	+ TC0_TBL_LOOKUP (iTc, iIndexA, uiBSx4, 0);
	506	+ pFilter->pLoopf->pfLumaDeblockingLT4Hor (&pDestY[1 << 2], iLineSize, iAlpha, iBeta, iTc);
	507	+ pFilter->pLoopf->pfLumaDeblockingLT4Hor (&pDestY[2 << 2], iLineSize, iAlpha, iBeta, iTc);
	508	+ pFilter->pLoopf->pfLumaDeblockingLT4Hor (&pDestY[3 << 2], iLineSize, iAlpha, iBeta, iTc);
	509	+ }
	510	+
	511	+ // luma h
	512	+ if (iBoundryFlag & TOP_FLAG_MASK) {
	513	+ pFilter->iLumaQP = (iCurQp + pCurDqLayer->pLumaQp[iMbXyIndex - iMbWidth] + 1) >> 1;
	514	+ FilteringEdgeLumaIntraH (pFilter, pDestY, iLineSize, NULL);
	515	+ }
	516	+
	517	+ pFilter->iLumaQP = iCurQp;
	518	+ if (iAlpha \| iBeta) {
	519	+ pFilter->pLoopf->pfLumaDeblockingLT4Ver (&pDestY[ (1 << 2)*iLineSize], iLineSize, iAlpha, iBeta, iTc);
	520	+ pFilter->pLoopf->pfLumaDeblockingLT4Ver (&pDestY[ (2 << 2)*iLineSize], iLineSize, iAlpha, iBeta, iTc);
	521	+ pFilter->pLoopf->pfLumaDeblockingLT4Ver (&pDestY[ (3 << 2)*iLineSize], iLineSize, iAlpha, iBeta, iTc);
	522	+ }
	523	+}
	524	+void_t /__FASTCALL/ FilteringEdgeChromaHV (PDqLayer pCurDqLayer, PDeblockingFilter pFilter, int32_t iBoundryFlag) {
	525	+ int32_t iMbXyIndex = pCurDqLayer->iMbXyIndex;
	526	+ int32_t iMbX = pCurDqLayer->iMbX;
	527	+ int32_t iMbY = pCurDqLayer->iMbY;
	528	+ int32_t iMbWidth = pCurDqLayer->iMbWidth;
	529	+ int32_t iLineSize = pFilter->iCsStride[1];
	530	+
	531	+ uint8_t* pDestCb, *pDestCr;
	532	+ int32_t iCurQp;
	533	+ int32_t iIndexA, iAlpha, iBeta;
	534	+
	535	+ FORCE_STACK_ALIGN_1D (int8_t, iTc, 4, 16);
	536	+ FORCE_STACK_ALIGN_1D (uint8_t, uiBSx4, 4, 4);
	537	+
	538	+ pDestCb = pFilter->pCsData[1] + ((iMbY * iLineSize + iMbX) << 3);
	539	+ pDestCr = pFilter->pCsData[2] + ((iMbY * iLineSize + iMbX) << 3);
	540	+ iCurQp = pCurDqLayer->pChromaQp[iMbXyIndex];
	541	+ * (uint32_t*)uiBSx4 = 0x03030303;
	542	+
	543	+ // chroma v
	544	+ if (iBoundryFlag & LEFT_FLAG_MASK) {
	545	+ pFilter->iChromaQP = (iCurQp + pCurDqLayer->pChromaQp[iMbXyIndex - 1] + 1) >> 1;
	546	+ FilteringEdgeChromaIntraV (pFilter, pDestCb, pDestCr, iLineSize, NULL);
	547	+ }
	548	+
	549	+ pFilter->iChromaQP = iCurQp;
	550	+ GET_ALPHA_BETA_FROM_QP (pFilter->iChromaQP, pFilter->iSliceAlphaC0Offset, pFilter->iSliceBetaOffset, iIndexA, iAlpha,
	551	+ iBeta);
	552	+ if (iAlpha \| iBeta) {
	553	+ TC0_TBL_LOOKUP (iTc, iIndexA, uiBSx4, 1);
	554	+ pFilter->pLoopf->pfChromaDeblockingLT4Hor (&pDestCb[2 << 1], &pDestCr[2 << 1], iLineSize, iAlpha, iBeta, iTc);
	555	+ }
	556	+
	557	+ // chroma h
	558	+ if (iBoundryFlag & TOP_FLAG_MASK) {
	559	+ pFilter->iChromaQP = (iCurQp + pCurDqLayer->pChromaQp[iMbXyIndex - iMbWidth] + 1) >> 1;
	560	+ FilteringEdgeChromaIntraH (pFilter, pDestCb, pDestCr, iLineSize, NULL);
	561	+ }
	562	+
	563	+ pFilter->iChromaQP = iCurQp;
	564	+ if (iAlpha \| iBeta) {
	565	+ pFilter->pLoopf->pfChromaDeblockingLT4Ver (&pDestCb[ (2 << 1)iLineSize], &pDestCr[ (2 << 1)iLineSize], iLineSize,
	566	+ iAlpha, iBeta, iTc);
	567	+ }
600	568	}
601	569
602	570	// merge h&v lookup table operation to save performance
603		−void_t DeblockingIntraMb( PDqLayer pCurDqLayer, PDeblockingFilter pFilter, int32_t iBoundryFlag )
604		−{
605		− FilteringEdgeLumaHV(pCurDqLayer, pFilter, iBoundryFlag);
606		− FilteringEdgeChromaHV(pCurDqLayer, pFilter, iBoundryFlag);
607		−}
608		−
609		−void_t WelsDeblockingMb( PDqLayer pCurDqLayer, PDeblockingFilter pFilter, int32_t iBoundryFlag )
610		−{
611		− uint8_t nBS[2][4][4] = { 0 };
612		−
613		− int32_t iMbXyIndex = pCurDqLayer->iMbXyIndex;
614		− int32_t iCurMbType = pCurDqLayer->pMbType[iMbXyIndex];
615		− int32_t iMbNb;
616		−
617		− switch( iCurMbType )
618		− {
619		− case MB_TYPE_INTRA4x4:
620		− case MB_TYPE_INTRA16x16:
621		− case MB_TYPE_INTRA_PCM:
622		− DeblockingIntraMb( pCurDqLayer, pFilter, iBoundryFlag );
623		− break;
624		− default:
625		−
626		− if(iBoundryFlag & LEFT_FLAG_MASK)
627		− {
628		− iMbNb = iMbXyIndex - 1;
629		− (uint32_t)nBS[0][0] = IS_INTRA(pCurDqLayer->pMbType[iMbNb])?0x04040404:DeblockingBsMarginalMBAvcbase( pCurDqLayer, 0, iMbNb, iMbXyIndex);
	571	+void_t DeblockingIntraMb (PDqLayer pCurDqLayer, PDeblockingFilter pFilter, int32_t iBoundryFlag) {
	572	+ FilteringEdgeLumaHV (pCurDqLayer, pFilter, iBoundryFlag);
	573	+ FilteringEdgeChromaHV (pCurDqLayer, pFilter, iBoundryFlag);
	574	+}
	575	+
	576	+void_t WelsDeblockingMb (PDqLayer pCurDqLayer, PDeblockingFilter pFilter, int32_t iBoundryFlag) {
	577	+ uint8_t nBS[2][4][4] = { 0 };
	578	+
	579	+ int32_t iMbXyIndex = pCurDqLayer->iMbXyIndex;
	580	+ int32_t iCurMbType = pCurDqLayer->pMbType[iMbXyIndex];
	581	+ int32_t iMbNb;
	582	+
	583	+ switch (iCurMbType) {
	584	+ case MB_TYPE_INTRA4x4:
	585	+ case MB_TYPE_INTRA16x16:
	586	+ case MB_TYPE_INTRA_PCM:
	587	+ DeblockingIntraMb (pCurDqLayer, pFilter, iBoundryFlag);
	588	+ break;
	589	+ default:
	590	+
	591	+ if (iBoundryFlag & LEFT_FLAG_MASK) {
	592	+ iMbNb = iMbXyIndex - 1;
	593	+ * (uint32_t*)nBS[0][0] = IS_INTRA (pCurDqLayer->pMbType[iMbNb]) ? 0x04040404 : DeblockingBsMarginalMBAvcbase (
	594	+ pCurDqLayer, 0, iMbNb, iMbXyIndex);
	595	+ } else {
	596	+ * (uint32_t*)nBS[0][0] = 0;
	597	+ }
	598	+ if (iBoundryFlag & TOP_FLAG_MASK) {
	599	+ iMbNb = iMbXyIndex - pCurDqLayer->iMbWidth;
	600	+ * (uint32_t*)nBS[1][0] = IS_INTRA (pCurDqLayer->pMbType[iMbNb]) ? 0x04040404 : DeblockingBsMarginalMBAvcbase (
	601	+ pCurDqLayer, 1, iMbNb, iMbXyIndex);
	602	+ } else {
	603	+ * (uint32_t*)nBS[1][0] = 0;
	604	+ }
	605	+ //SKIP MB_16x16 or others
	606	+ if (iCurMbType != MB_TYPE_SKIP) {
	607	+ if (iCurMbType == MB_TYPE_16x16) {
	608	+ DeblockingBSInsideMBAvsbase (pCurDqLayer->pNzc[iMbXyIndex], nBS, 1);
	609	+ } else {
	610	+ DeblockingBSInsideMBNormal (pCurDqLayer, nBS, pCurDqLayer->pNzc[iMbXyIndex], iMbXyIndex);
	611	+ }
	612	+ } else {
	613	+ * (uint32_t)nBS[0][1] = (uint32_t)nBS[0][2] = (uint32_t*)nBS[0][3] =
	614	+ * (uint32_t)nBS[1][1] = (uint32_t)nBS[1][2] = (uint32_t*)nBS[1][3] = 0;
	615	+ }
	616	+ DeblockingInterMb (pCurDqLayer, pFilter, nBS, iBoundryFlag);
	617	+ break;
	618	+ }
	619	+}
	620	+
	621	+// C code only
	622	+void_t DeblockLumaLt4_c (uint8_t* pPix, int32_t iStrideX, int32_t iStrideY, int32_t iAlpha, int32_t iBeta,
	623	+ int8_t* pTc) {
	624	+ for (int32_t i = 0; i < 16; i++) {
	625	+ int32_t iTc0 = pTc[i >> 2];
	626	+ if (iTc0 >= 0) {
	627	+ int32_t p0 = pPix[-iStrideX];
	628	+ int32_t p1 = pPix[-2 * iStrideX];
	629	+ int32_t p2 = pPix[-3 * iStrideX];
	630	+ int32_t q0 = pPix[0];
	631	+ int32_t q1 = pPix[iStrideX];
	632	+ int32_t q2 = pPix[2 * iStrideX];
	633	+ bool_t bDetaP0Q0 = WELS_ABS (p0 - q0) < iAlpha;
	634	+ bool_t bDetaP1P0 = WELS_ABS (p1 - p0) < iBeta;
	635	+ bool_t bDetaQ1Q0 = WELS_ABS (q1 - q0) < iBeta;
	636	+ int32_t iTc = iTc0;
	637	+ if (bDetaP0Q0 && bDetaP1P0 && bDetaQ1Q0) {
	638	+ bool_t bDetaP2P0 = WELS_ABS (p2 - p0) < iBeta;
	639	+ bool_t bDetaQ2Q0 = WELS_ABS (q2 - q0) < iBeta;
	640	+ if (bDetaP2P0) {
	641	+ pPix[-2 * iStrideX] = p1 + WELS_CLIP3 ((p2 + ((p0 + q0 + 1) >> 1) - (p1 << 1)) >> 1, -iTc0, iTc0);
	642	+ iTc++;
	643	+ }
	644	+ if (bDetaQ2Q0) {
	645	+ pPix[iStrideX] = q1 + WELS_CLIP3 ((q2 + ((p0 + q0 + 1) >> 1) - (q1 << 1)) >> 1, -iTc0, iTc0);
	646	+ iTc++;
630	647	}
631		− else
632		− {
633		− (uint32_t)nBS[0][0] = 0;
634		− }
635		− if(iBoundryFlag & TOP_FLAG_MASK)
636		− {
637		− iMbNb = iMbXyIndex - pCurDqLayer->iMbWidth;
638		− (uint32_t)nBS[1][0] = IS_INTRA(pCurDqLayer->pMbType[iMbNb])?0x04040404:DeblockingBsMarginalMBAvcbase( pCurDqLayer, 1, iMbNb, iMbXyIndex);
	648	+ int32_t iDeta = WELS_CLIP3 ((((q0 - p0) << 2) + (p1 - q1) + 4) >> 3, -iTc, iTc);
	649	+ pPix[-iStrideX] = WELS_CLIP1 (p0 + iDeta); /* p0' */
	650	+ pPix[0] = WELS_CLIP1 (q0 - iDeta); /* q0' */
	651	+ }
	652	+ }
	653	+ pPix += iStrideY;
	654	+ }
	655	+}
	656	+void_t DeblockLumaEq4_c (uint8_t* pPix, int32_t iStrideX, int32_t iStrideY, int32_t iAlpha, int32_t iBeta) {
	657	+ int32_t p0, p1, p2, q0, q1, q2;
	658	+ int32_t iDetaP0Q0;
	659	+ bool_t bDetaP1P0, bDetaQ1Q0;
	660	+ for (int32_t i = 0; i < 16; i++) {
	661	+ p0 = pPix[-iStrideX];
	662	+ p1 = pPix[-2 * iStrideX];
	663	+ p2 = pPix[-3 * iStrideX];
	664	+ q0 = pPix[0];
	665	+ q1 = pPix[iStrideX];
	666	+ q2 = pPix[2 * iStrideX];
	667	+ iDetaP0Q0 = WELS_ABS (p0 - q0);
	668	+ bDetaP1P0 = WELS_ABS (p1 - p0) < iBeta;
	669	+ bDetaQ1Q0 = WELS_ABS (q1 - q0) < iBeta;
	670	+ if ((iDetaP0Q0 < iAlpha) && bDetaP1P0 && bDetaQ1Q0) {
	671	+ if (iDetaP0Q0 < ((iAlpha >> 2) + 2)) {
	672	+ bool_t bDetaP2P0 = WELS_ABS (p2 - p0) < iBeta;
	673	+ bool_t bDetaQ2Q0 = WELS_ABS (q2 - q0) < iBeta;
	674	+ if (bDetaP2P0) {
	675	+ const int32_t p3 = pPix[-4 * iStrideX];
	676	+ pPix[-iStrideX] = (p2 + (p1 << 1) + (p0 << 1) + (q0 << 1) + q1 + 4) >> 3; //p0
	677	+ pPix[-2 * iStrideX] = (p2 + p1 + p0 + q0 + 2) >> 2; //p1
	678	+ pPix[-3 * iStrideX] = ((p3 << 1) + p2 + (p2 << 1) + p1 + p0 + q0 + 4) >> 3;//p2
	679	+ } else {
	680	+ pPix[-1 * iStrideX] = ((p1 << 1) + p0 + q1 + 2) >> 2; //p0
639	681	}
640		− else
641		− {
642		− (uint32_t)nBS[1][0] = 0;
643		− }
644		− //SKIP MB_16x16 or others
645		− if( iCurMbType != MB_TYPE_SKIP )
646		− {
647		− if( iCurMbType == MB_TYPE_16x16 )
648		− {
649		− DeblockingBSInsideMBAvsbase( pCurDqLayer->pNzc[iMbXyIndex], nBS, 1 );
650		− }
651		− else
652		− {
653		− DeblockingBSInsideMBNormal(pCurDqLayer, nBS, pCurDqLayer->pNzc[iMbXyIndex], iMbXyIndex);
654		− }
655		− }
656		− else
657		− {
658		− (uint32_t)nBS[0][1] = (uint32_t)nBS[0][2] = (uint32_t)nBS[0][3] =
659		− (uint32_t)nBS[1][1] = (uint32_t)nBS[1][2] = (uint32_t)nBS[1][3] = 0;
660		− }
661		− DeblockingInterMb( pCurDqLayer, pFilter, nBS, iBoundryFlag );
662		− break;
663		− }
	682	+ if (bDetaQ2Q0) {
	683	+ const int32_t q3 = pPix[3 * iStrideX];
	684	+ pPix[0] = (p1 + (p0 << 1) + (q0 << 1) + (q1 << 1) + q2 + 4) >> 3; //q0
	685	+ pPix[iStrideX] = (p0 + q0 + q1 + q2 + 2) >> 2; //q1
	686	+ pPix[2 * iStrideX] = ((q3 << 1) + q2 + (q2 << 1) + q1 + q0 + p0 + 4) >> 3;//q2
	687	+ } else {
	688	+ pPix[0] = ((q1 << 1) + q0 + p1 + 2) >> 2; //q0
	689	+ }
	690	+ } else {
	691	+ pPix[-iStrideX] = ((p1 << 1) + p0 + q1 + 2) >> 2; //p0
	692	+ pPix[ 0] = ((q1 << 1) + q0 + p1 + 2) >> 2; //q0
	693	+ }
	694	+ }
	695	+ pPix += iStrideY;
	696	+ }
	697	+}
	698	+void_t DeblockLumaLt4V_c (uint8_t* pPix, int32_t iStride, int32_t iAlpha, int32_t iBeta, int8_t* tc) {
	699	+ DeblockLumaLt4_c (pPix, iStride, 1, iAlpha, iBeta, tc);
	700	+}
	701	+void_t DeblockLumaLt4H_c (uint8_t* pPix, int32_t iStride, int32_t iAlpha, int32_t iBeta, int8_t* tc) {
	702	+ DeblockLumaLt4_c (pPix, 1, iStride, iAlpha, iBeta, tc);
	703	+}
	704	+void_t DeblockLumaEq4V_c (uint8_t* pPix, int32_t iStride, int32_t iAlpha, int32_t iBeta) {
	705	+ DeblockLumaEq4_c (pPix, iStride, 1, iAlpha, iBeta);
	706	+}
	707	+void_t DeblockLumaEq4H_c (uint8_t* pPix, int32_t iStride, int32_t iAlpha, int32_t iBeta) {
	708	+ DeblockLumaEq4_c (pPix, 1, iStride, iAlpha, iBeta);
	709	+}
	710	+void_t DeblockChromaLt4_c (uint8_t* pPixCb, uint8_t* pPixCr, int32_t iStrideX, int32_t iStrideY, int32_t iAlpha,
	711	+ int32_t iBeta, int8_t* pTc) {
	712	+ int32_t p0, p1, q0, q1, iDeta;
	713	+ bool_t bDetaP0Q0, bDetaP1P0, bDetaQ1Q0;
	714	+
	715	+ for (int32_t i = 0; i < 8; i++) {
	716	+ int32_t iTc0 = pTc[i >> 1];
	717	+ if (iTc0 > 0) {
	718	+ p0 = pPixCb[-iStrideX];
	719	+ p1 = pPixCb[-2 * iStrideX];
	720	+ q0 = pPixCb[0];
	721	+ q1 = pPixCb[iStrideX];
	722	+
	723	+ bDetaP0Q0 = WELS_ABS (p0 - q0) < iAlpha;
	724	+ bDetaP1P0 = WELS_ABS (p1 - p0) < iBeta;
	725	+ bDetaQ1Q0 = WELS_ABS (q1 - q0) < iBeta;
	726	+ if (bDetaP0Q0 && bDetaP1P0 && bDetaQ1Q0) {
	727	+ iDeta = WELS_CLIP3 ((((q0 - p0) << 2) + (p1 - q1) + 4) >> 3, -iTc0, iTc0);
	728	+ pPixCb[-iStrideX] = WELS_CLIP1 (p0 + iDeta); /* p0' */
	729	+ pPixCb[0] = WELS_CLIP1 (q0 - iDeta); /* q0' */
	730	+ }
	731	+
	732	+
	733	+ p0 = pPixCr[-iStrideX];
	734	+ p1 = pPixCr[-2 * iStrideX];
	735	+ q0 = pPixCr[0];
	736	+ q1 = pPixCr[iStrideX];
	737	+
	738	+ bDetaP0Q0 = WELS_ABS (p0 - q0) < iAlpha;
	739	+ bDetaP1P0 = WELS_ABS (p1 - p0) < iBeta;
	740	+ bDetaQ1Q0 = WELS_ABS (q1 - q0) < iBeta;
	741	+
	742	+ if (bDetaP0Q0 && bDetaP1P0 && bDetaQ1Q0) {
	743	+ iDeta = WELS_CLIP3 ((((q0 - p0) << 2) + (p1 - q1) + 4) >> 3, -iTc0, iTc0);
	744	+ pPixCr[-iStrideX] = WELS_CLIP1 (p0 + iDeta); /* p0' */
	745	+ pPixCr[0] = WELS_CLIP1 (q0 - iDeta); /* q0' */
	746	+ }
	747	+ }
	748	+ pPixCb += iStrideY;
	749	+ pPixCr += iStrideY;
	750	+ }
	751	+}
	752	+void_t DeblockChromaEq4_c (uint8_t* pPixCb, uint8_t* pPixCr, int32_t iStrideX, int32_t iStrideY, int32_t iAlpha,
	753	+ int32_t iBeta) {
	754	+ int32_t i = 0, d = 0;
	755	+ int32_t p0, p1, q0, q1;
	756	+ bool_t bDetaP0Q0, bDetaP1P0, bDetaQ1Q0;
	757	+ for (int32_t i = 0; i < 8; i++) {
	758	+ //cb
	759	+ p0 = pPixCb[-iStrideX];
	760	+ p1 = pPixCb[-2 * iStrideX];
	761	+ q0 = pPixCb[0];
	762	+ q1 = pPixCb[iStrideX];
	763	+ bDetaP0Q0 = WELS_ABS (p0 - q0) < iAlpha;
	764	+ bDetaP1P0 = WELS_ABS (p1 - p0) < iBeta;
	765	+ bDetaQ1Q0 = WELS_ABS (q1 - q0) < iBeta;
	766	+ if (bDetaP0Q0 && bDetaP1P0 && bDetaQ1Q0) {
	767	+ pPixCb[-iStrideX] = ((p1 << 1) + p0 + q1 + 2) >> 2; /* p0' */
	768	+ pPixCb[0] = ((q1 << 1) + q0 + p1 + 2) >> 2; /* q0' */
	769	+ }
	770	+
	771	+ //cr
	772	+ p0 = pPixCr[-iStrideX];
	773	+ p1 = pPixCr[-2 * iStrideX];
	774	+ q0 = pPixCr[0];
	775	+ q1 = pPixCr[iStrideX];
	776	+ bDetaP0Q0 = WELS_ABS (p0 - q0) < iAlpha;
	777	+ bDetaP1P0 = WELS_ABS (p1 - p0) < iBeta;
	778	+ bDetaQ1Q0 = WELS_ABS (q1 - q0) < iBeta;
	779	+ if (bDetaP0Q0 && bDetaP1P0 && bDetaQ1Q0) {
	780	+ pPixCr[-iStrideX] = ((p1 << 1) + p0 + q1 + 2) >> 2; /* p0' */
	781	+ pPixCr[0] = ((q1 << 1) + q0 + p1 + 2) >> 2; /* q0' */
	782	+ }
	783	+ pPixCr += iStrideY;
	784	+ pPixCb += iStrideY;
	785	+ }
	786	+}
	787	+void_t DeblockChromaLt4V_c (uint8_t* pPixCb, uint8_t* pPixCr, int32_t iStride, int32_t iAlpha, int32_t iBeta,
	788	+ int8_t* tc) {
	789	+ DeblockChromaLt4_c (pPixCb, pPixCr, iStride, 1, iAlpha, iBeta, tc);
	790	+}
	791	+void_t DeblockChromaLt4H_c (uint8_t* pPixCb, uint8_t* pPixCr, int32_t iStride, int32_t iAlpha, int32_t iBeta,
	792	+ int8_t* tc) {
	793	+ DeblockChromaLt4_c (pPixCb, pPixCr, 1, iStride, iAlpha, iBeta, tc);
	794	+}
	795	+void_t DeblockChromaEq4V_c (uint8_t* pPixCb, uint8_t* pPixCr, int32_t iStride, int32_t iAlpha, int32_t iBeta) {
	796	+ DeblockChromaEq4_c (pPixCb, pPixCr, iStride, 1, iAlpha, iBeta);
	797	+}
	798	+void_t DeblockChromaEq4H_c (uint8_t* pPixCb, uint8_t* pPixCr, int32_t iStride, int32_t iAlpha, int32_t iBeta) {
	799	+ DeblockChromaEq4_c (pPixCb, pPixCr, 1, iStride, iAlpha, iBeta);
	800	+}
	801	+
	802	+#ifdef X86_ASM
	803	+extern "C" {
	804	+ void DeblockLumaLt4H_sse2 (uint8_t* pPixY, int32_t iStride, int32_t iAlpha, int32_t iBeta, int8_t* pTc) {
	805	+ FORCE_STACK_ALIGN_1D (uint8_t, uiBuf, 16 * 8, 16);
	806	+
	807	+ DeblockLumaTransposeH2V_sse2 (pPixY - 4, iStride, &uiBuf[0]);
	808	+ DeblockLumaLt4V_sse2 (&uiBuf[4 * 16], 16, iAlpha, iBeta, pTc);
	809	+ DeblockLumaTransposeV2H_sse2 (pPixY - 4, iStride, &uiBuf[0]);
	810	+ }
	811	+
	812	+ void DeblockLumaEq4H_sse2 (uint8_t* pPixY, int32_t iStride, int32_t iAlpha, int32_t iBeta) {
	813	+ FORCE_STACK_ALIGN_1D (uint8_t, uiBuf, 16 * 8, 16);
	814	+
	815	+ DeblockLumaTransposeH2V_sse2 (pPixY - 4, iStride, &uiBuf[0]);
	816	+ DeblockLumaEq4V_sse2 (&uiBuf[4 * 16], 16, iAlpha, iBeta);
	817	+ DeblockLumaTransposeV2H_sse2 (pPixY - 4, iStride, &uiBuf[0]);
	818	+ }
	819	+
664	820	}
665	821
666		−// C code only
667		−void_t DeblockLumaLt4_c( uint8_t pPix, int32_t iStrideX,int32_t iStrideY, int32_t iAlpha, int32_t iBeta, int8_t pTc )
668		−{
669		− for( int32_t i = 0;i<16;i++)
670		− {
671		− int32_t iTc0 = pTc[i>>2];
672		− if(iTc0>=0)
673		− {
674		− int32_t p0 = pPix[-iStrideX];
675		− int32_t p1 = pPix[-2*iStrideX];
676		− int32_t p2 = pPix[-3*iStrideX];
677		− int32_t q0 = pPix[0];
678		− int32_t q1 = pPix[iStrideX];
679		− int32_t q2 = pPix[2*iStrideX];
680		− bool_t bDetaP0Q0= WELS_ABS( p0 - q0 )<iAlpha;
681		− bool_t bDetaP1P0 = WELS_ABS( p1 - p0 ) < iBeta;
682		− bool_t bDetaQ1Q0 = WELS_ABS( q1 - q0 ) < iBeta;
683		− int32_t iTc = iTc0;
684		− if ( bDetaP0Q0&& bDetaP1P0 && bDetaQ1Q0 )
685		− {
686		− bool_t bDetaP2P0 = WELS_ABS( p2 - p0 ) < iBeta;
687		− bool_t bDetaQ2Q0 = WELS_ABS( q2 - q0 ) < iBeta;
688		− if ( bDetaP2P0)
689		− {
690		− pPix[-2*iStrideX] = p1 + WELS_CLIP3( ( p2 + ( ( p0 + q0 + 1 ) >> 1 ) - ( p1 << 1 ) ) >> 1, -iTc0, iTc0 );
691		− iTc++;
692		− }
693		− if (bDetaQ2Q0)
694		− {
695		− pPix[iStrideX] = q1 + WELS_CLIP3( ( q2 + ( ( p0 + q0 + 1 ) >> 1 ) - ( q1 << 1 ) ) >> 1, -iTc0, iTc0 );
696		− iTc++;
697		− }
698		− int32_t iDeta = WELS_CLIP3( (((q0 - p0 ) << 2) + (p1 - q1) + 4) >> 3, -iTc, iTc );
699		− pPix[-iStrideX] = WELS_CLIP1( p0 + iDeta ); /* p0' */
700		− pPix[0] = WELS_CLIP1( q0 - iDeta ); /* q0' */
701		− }
702		− }
703		− pPix +=iStrideY;
704		− }
705		−}
706		−void_t DeblockLumaEq4_c( uint8_t *pPix, int32_t iStrideX, int32_t iStrideY, int32_t iAlpha, int32_t iBeta )
707		−{
708		− int32_t p0,p1,p2,q0,q1,q2;
709		− int32_t iDetaP0Q0;
710		− bool_t bDetaP1P0,bDetaQ1Q0;
711		− for (int32_t i = 0;i<16;i++)
712		− {
713		− p0 = pPix[-iStrideX];
714		− p1 = pPix[-2*iStrideX];
715		− p2 = pPix[-3*iStrideX];
716		− q0 = pPix[0];
717		− q1 = pPix[iStrideX];
718		− q2 = pPix[2*iStrideX];
719		− iDetaP0Q0 = WELS_ABS( p0 - q0 );
720		− bDetaP1P0 = WELS_ABS( p1 - p0 ) < iBeta;
721		− bDetaQ1Q0 = WELS_ABS( q1 - q0 ) < iBeta;
722		− if ((iDetaP0Q0<iAlpha) && bDetaP1P0 &&bDetaQ1Q0)
723		− {
724		− if (iDetaP0Q0< (( iAlpha >> 2 ) + 2 ) )
725		− {
726		− bool_t bDetaP2P0 = WELS_ABS( p2 - p0 ) < iBeta;
727		− bool_t bDetaQ2Q0 = WELS_ABS( q2 - q0 ) < iBeta;
728		− if(bDetaP2P0)
729		− {
730		− const int32_t p3 = pPix[-4*iStrideX];
731		− pPix[-iStrideX] = ( p2 + (p1 << 1) + (p0 << 1) + (q0 << 1) + q1 + 4 ) >> 3; //p0
732		− pPix[-2*iStrideX] = ( p2 + p1 + p0 + q0 + 2 ) >> 2; //p1
733		− pPix[-3*iStrideX] = ( (p3 << 1) + p2 + (p2 << 1) + p1 + p0 + q0 + 4 ) >> 3;//p2
734		− }
735		− else
736		− {
737		− pPix[-1*iStrideX] = ( (p1 << 1) + p0 + q1 + 2 ) >> 2; //p0
738		− }
739		− if (bDetaQ2Q0)
740		− {
741		− const int32_t q3 = pPix[3*iStrideX];
742		− pPix[0] = ( p1 + (p0 << 1) + (q0 << 1) + (q1 << 1) + q2 + 4 ) >> 3; //q0
743		− pPix[iStrideX] = ( p0 + q0 + q1 + q2 + 2 ) >> 2; //q1
744		− pPix[2*iStrideX] = ( (q3 << 1) + q2 + (q2 << 1) + q1 + q0 + p0 + 4 ) >> 3;//q2
745		− }
746		− else
747		− {
748		− pPix[0] = ( (q1 << 1) + q0 + p1 + 2 ) >> 2; //q0
749		− }
750		− }
751		− else
752		− {
753		− pPix[-iStrideX] = ( (p1 << 1) + p0 + q1 + 2 ) >> 2; //p0
754		− pPix[ 0] = ( (q1 << 1) + q0 + p1 + 2 ) >> 2; //q0
755		− }
756		− }
757		− pPix += iStrideY;
758		− }
759		−}
760		−void_t DeblockLumaLt4V_c( uint8_t pPix, int32_t iStride, int32_t iAlpha, int32_t iBeta, int8_t tc )
761		−{
762		− DeblockLumaLt4_c( pPix, iStride, 1, iAlpha, iBeta, tc );
763		−}
764		−void_t DeblockLumaLt4H_c( uint8_t pPix, int32_t iStride, int32_t iAlpha, int32_t iBeta, int8_t tc )
765		−{
766		− DeblockLumaLt4_c( pPix, 1, iStride, iAlpha, iBeta, tc );
767		−}
768		−void_t DeblockLumaEq4V_c( uint8_t *pPix, int32_t iStride, int32_t iAlpha, int32_t iBeta )
769		−{
770		− DeblockLumaEq4_c( pPix, iStride, 1, iAlpha, iBeta);
771		−}
772		−void_t DeblockLumaEq4H_c( uint8_t *pPix, int32_t iStride, int32_t iAlpha, int32_t iBeta )
773		−{
774		− DeblockLumaEq4_c( pPix, 1, iStride, iAlpha, iBeta );
775		−}
776		−void_t DeblockChromaLt4_c( uint8_t pPixCb, uint8_t pPixCr, int32_t iStrideX, int32_t iStrideY, int32_t iAlpha, int32_t iBeta, int8_t *pTc )
777		−{
778		− int32_t p0, p1, q0, q1,iDeta;
779		− bool_t bDetaP0Q0,bDetaP1P0,bDetaQ1Q0;
780		−
781		− for(int32_t i = 0;i<8;i++)
782		− {
783		− int32_t iTc0 = pTc[i>>1];
784		− if(iTc0 >0)
785		− {
786		− p0 = pPixCb[-iStrideX];
787		− p1 = pPixCb[-2*iStrideX];
788		− q0 = pPixCb[0];
789		− q1 = pPixCb[iStrideX];
790		−
791		− bDetaP0Q0 = WELS_ABS( p0 - q0 ) < iAlpha;
792		− bDetaP1P0 = WELS_ABS( p1 - p0 ) < iBeta;
793		− bDetaQ1Q0 = WELS_ABS( q1 - q0 ) < iBeta;
794		− if( bDetaP0Q0&&bDetaP1P0 && bDetaQ1Q0)
795		− {
796		− iDeta = WELS_CLIP3( (((q0 - p0 ) << 2) + (p1 - q1) + 4) >> 3, -iTc0, iTc0 );
797		− pPixCb[-iStrideX] = WELS_CLIP1( p0 + iDeta ); /* p0' */
798		− pPixCb[0] = WELS_CLIP1( q0 - iDeta ); /* q0' */
799		− }
800		−
801		−
802		− p0 = pPixCr[-iStrideX];
803		− p1 = pPixCr[-2*iStrideX];
804		− q0 = pPixCr[0];
805		− q1 = pPixCr[iStrideX];
806		−
807		− bDetaP0Q0 = WELS_ABS( p0 - q0 ) < iAlpha;
808		− bDetaP1P0 = WELS_ABS( p1 - p0 ) < iBeta;
809		− bDetaQ1Q0 = WELS_ABS( q1 - q0 ) < iBeta;
810		−
811		− if( bDetaP0Q0&&bDetaP1P0 && bDetaQ1Q0)
812		− {
813		− iDeta = WELS_CLIP3( (((q0 - p0 ) << 2) + (p1 - q1) + 4) >> 3, -iTc0, iTc0 );
814		− pPixCr[-iStrideX] = WELS_CLIP1( p0 + iDeta ); /* p0' */
815		− pPixCr[0] = WELS_CLIP1( q0 - iDeta ); /* q0' */
816		− }
817		− }
818		− pPixCb += iStrideY;
819		− pPixCr += iStrideY;
820		− }
821		−}
822		−void_t DeblockChromaEq4_c( uint8_t pPixCb, uint8_t pPixCr, int32_t iStrideX, int32_t iStrideY, int32_t iAlpha, int32_t iBeta )
823		−{
824		− int32_t i = 0, d = 0;
825		− int32_t p0, p1, q0, q1;
826		− bool_t bDetaP0Q0,bDetaP1P0,bDetaQ1Q0;
827		− for(int32_t i =0;i<8;i++)
828		− {
829		− //cb
830		− p0 = pPixCb[-iStrideX];
831		− p1 = pPixCb[-2*iStrideX];
832		− q0 = pPixCb[0];
833		− q1 = pPixCb[iStrideX];
834		− bDetaP0Q0 = WELS_ABS( p0 - q0 ) < iAlpha;
835		− bDetaP1P0 = WELS_ABS( p1 - p0 ) < iBeta;
836		− bDetaQ1Q0 = WELS_ABS( q1 - q0 ) < iBeta;
837		− if(bDetaP0Q0&&bDetaP1P0&&bDetaQ1Q0)
838		− {
839		− pPixCb[-iStrideX] = ( (p1 << 1) + p0 + q1 + 2 ) >> 2; /* p0' */
840		− pPixCb[0] = ( (q1 << 1) + q0 + p1 + 2 ) >> 2; /* q0' */
841		− }
842		−
843		− //cr
844		− p0 = pPixCr[-iStrideX];
845		− p1 = pPixCr[-2*iStrideX];
846		− q0 = pPixCr[0];
847		− q1 = pPixCr[iStrideX];
848		− bDetaP0Q0 = WELS_ABS( p0 - q0 ) < iAlpha;
849		− bDetaP1P0 = WELS_ABS( p1 - p0 ) < iBeta;
850		− bDetaQ1Q0 = WELS_ABS( q1 - q0 ) < iBeta;
851		− if(bDetaP0Q0&&bDetaP1P0&&bDetaQ1Q0)
852		− {
853		− pPixCr[-iStrideX] = ( (p1 << 1) + p0 + q1 + 2 ) >> 2; /* p0' */
854		− pPixCr[0] = ( (q1 << 1) + q0 + p1 + 2 ) >> 2; /* q0' */
855		− }
856		− pPixCr += iStrideY;
857		− pPixCb += iStrideY;
858		− }
859		−}
860		−void_t DeblockChromaLt4V_c( uint8_t pPixCb, uint8_t pPixCr, int32_t iStride, int32_t iAlpha, int32_t iBeta, int8_t *tc )
861		−{
862		− DeblockChromaLt4_c( pPixCb, pPixCr, iStride, 1, iAlpha, iBeta, tc );
863		−}
864		−void_t DeblockChromaLt4H_c( uint8_t pPixCb, uint8_t pPixCr, int32_t iStride, int32_t iAlpha, int32_t iBeta, int8_t *tc )
865		−{
866		− DeblockChromaLt4_c( pPixCb, pPixCr, 1, iStride, iAlpha, iBeta, tc );
867		−}
868		−void_t DeblockChromaEq4V_c( uint8_t pPixCb, uint8_t pPixCr, int32_t iStride, int32_t iAlpha, int32_t iBeta )
869		−{
870		− DeblockChromaEq4_c( pPixCb, pPixCr, iStride, 1, iAlpha, iBeta );
871		−}
872		−void_t DeblockChromaEq4H_c( uint8_t pPixCb, uint8_t pPixCr, int32_t iStride, int32_t iAlpha, int32_t iBeta )
873		−{
874		− DeblockChromaEq4_c( pPixCb, pPixCr, 1, iStride, iAlpha, iBeta );
875		−}
876		−
877		−#ifdef X86_ASM
878		−extern "C" {
879		−void DeblockLumaLt4H_sse2(uint8_t pPixY, int32_t iStride, int32_t iAlpha, int32_t iBeta, int8_t pTc)
880		−{
881		− FORCE_STACK_ALIGN_1D(uint8_t, uiBuf, 16*8, 16);
882		−
883		− DeblockLumaTransposeH2V_sse2(pPixY - 4, iStride, &uiBuf[0]);
884		− DeblockLumaLt4V_sse2(&uiBuf[4*16], 16, iAlpha, iBeta, pTc);
885		− DeblockLumaTransposeV2H_sse2(pPixY - 4, iStride, &uiBuf[0]);
886		−}
887		−
888		−void DeblockLumaEq4H_sse2(uint8_t *pPixY, int32_t iStride, int32_t iAlpha, int32_t iBeta)
889		−{
890		− FORCE_STACK_ALIGN_1D(uint8_t, uiBuf, 16*8, 16);
891		−
892		− DeblockLumaTransposeH2V_sse2(pPixY - 4, iStride, &uiBuf[0]);
893		− DeblockLumaEq4V_sse2(&uiBuf[4*16], 16, iAlpha, iBeta);
894		− DeblockLumaTransposeV2H_sse2(pPixY - 4, iStride, &uiBuf[0]);
895		−}
896		−
897		−}
898		−
899	822	#endif
900	823	/*!
901	824	* \brief AVC slice deblocking filtering target layer
@@ -904,108 +827,99 @@ void DeblockLumaEq4H_sse2(uint8_t *pPixY, int32_t iStride, int32_t iAlpha, int32 · 2 unmodified lines
904	827	*
905	828	* \return NONE
906	829	*/
907		−void_t WelsDeblockingFilterSlice( PWelsDecoderContext pCtx, PDeblockingFilterMbFunc pDeblockMb )
908		−{
909		− PDqLayer pCurDqLayer = pCtx->pCurDqLayer;
910		− PSliceHeaderExt pSliceHeaderExt = &pCurDqLayer->sLayerInfo.sSliceInLayer.sSliceHeaderExt;
911		− int32_t iMbWidth = pCurDqLayer->iMbWidth;
912		− int32_t iTotalMbCount = pSliceHeaderExt->sSliceHeader.pSps->uiTotalMbCount;
913		−
914		− SDeblockingFilter pFilter = {0};
915		−
916		− PFmo pFmo = pCtx->pFmo;
917		− int32_t iNextMbXyIndex = 0;
918		− int32_t iTotalNumMb = pCurDqLayer->sLayerInfo.sSliceInLayer.iTotalMbInCurSlice;
919		− int32_t iCountNumMb = 0;
920		− int32_t iBoundryFlag;
921		− int32_t iFilterIdc = pCurDqLayer->sLayerInfo.sSliceInLayer.sSliceHeaderExt.sSliceHeader.uiDisableDeblockingFilterIdc;
922		−
923		− /* Step1: parameters set */
924		− pFilter.pCsData[0] = pCtx->pDec->pData[0];
925		− pFilter.pCsData[1] = pCtx->pDec->pData[1];
926		− pFilter.pCsData[2] = pCtx->pDec->pData[2];
927		−
928		− pFilter.iCsStride[0] = pCtx->pDec->iLinesize[0];
929		− pFilter.iCsStride[1] = pCtx->pDec->iLinesize[1];
930		−
931		− pFilter.eSliceType = (ESliceType) pCurDqLayer->sLayerInfo.sSliceInLayer.eSliceType;
932		−
933		− pFilter.iSliceAlphaC0Offset = pSliceHeaderExt->sSliceHeader.iSliceAlphaC0Offset;
934		− pFilter.iSliceBetaOffset = pSliceHeaderExt->sSliceHeader.iSliceBetaOffset;
935		−
936		− pFilter.pLoopf = &pCtx->sDeblockingFunc;
937		−
938		− /* Step2: macroblock deblocking */
939		− if( 0 == iFilterIdc \|\| 2 == iFilterIdc )
940		− {
941		− iNextMbXyIndex = pSliceHeaderExt->sSliceHeader.iFirstMbInSlice;
942		− pCurDqLayer->iMbX = iNextMbXyIndex % iMbWidth;
943		− pCurDqLayer->iMbY = iNextMbXyIndex / iMbWidth;
944		− pCurDqLayer->iMbXyIndex = iNextMbXyIndex;
945		−
946		− do
947		− {
948		− iBoundryFlag = DeblockingAvailableNoInterlayer(pCurDqLayer, iFilterIdc);
949		−
950		− pDeblockMb( pCurDqLayer, &pFilter, iBoundryFlag );
951		−
952		− ++iCountNumMb;
953		− if ( iCountNumMb >= iTotalNumMb )
954		− {
955		− break;
956		− }
957		−
958		− if ( pSliceHeaderExt->sSliceHeader.pPps->uiNumSliceGroups > 1 )
959		− {
960		− iNextMbXyIndex = FmoNextMb( pFmo, iNextMbXyIndex );
961		− }
962		− else
963		− {
964		− ++iNextMbXyIndex;
965		− }
966		− if ( -1 == iNextMbXyIndex \|\| iNextMbXyIndex >= iTotalMbCount ) // slice group boundary or end of a frame
967		− {
968		− break;
969		− }
970		−
971		− pCurDqLayer->iMbX = iNextMbXyIndex % iMbWidth;
972		− pCurDqLayer->iMbY = iNextMbXyIndex / iMbWidth;
973		− pCurDqLayer->iMbXyIndex = iNextMbXyIndex;
974		− } while ( 1 );
975		− }
	830	+void_t WelsDeblockingFilterSlice (PWelsDecoderContext pCtx, PDeblockingFilterMbFunc pDeblockMb) {
	831	+ PDqLayer pCurDqLayer = pCtx->pCurDqLayer;
	832	+ PSliceHeaderExt pSliceHeaderExt = &pCurDqLayer->sLayerInfo.sSliceInLayer.sSliceHeaderExt;
	833	+ int32_t iMbWidth = pCurDqLayer->iMbWidth;
	834	+ int32_t iTotalMbCount = pSliceHeaderExt->sSliceHeader.pSps->uiTotalMbCount;
	835	+
	836	+ SDeblockingFilter pFilter = {0};
	837	+
	838	+ PFmo pFmo = pCtx->pFmo;
	839	+ int32_t iNextMbXyIndex = 0;
	840	+ int32_t iTotalNumMb = pCurDqLayer->sLayerInfo.sSliceInLayer.iTotalMbInCurSlice;
	841	+ int32_t iCountNumMb = 0;
	842	+ int32_t iBoundryFlag;
	843	+ int32_t iFilterIdc = pCurDqLayer->sLayerInfo.sSliceInLayer.sSliceHeaderExt.sSliceHeader.uiDisableDeblockingFilterIdc;
	844	+
	845	+ /* Step1: parameters set */
	846	+ pFilter.pCsData[0] = pCtx->pDec->pData[0];
	847	+ pFilter.pCsData[1] = pCtx->pDec->pData[1];
	848	+ pFilter.pCsData[2] = pCtx->pDec->pData[2];
	849	+
	850	+ pFilter.iCsStride[0] = pCtx->pDec->iLinesize[0];
	851	+ pFilter.iCsStride[1] = pCtx->pDec->iLinesize[1];
	852	+
	853	+ pFilter.eSliceType = (ESliceType) pCurDqLayer->sLayerInfo.sSliceInLayer.eSliceType;
	854	+
	855	+ pFilter.iSliceAlphaC0Offset = pSliceHeaderExt->sSliceHeader.iSliceAlphaC0Offset;
	856	+ pFilter.iSliceBetaOffset = pSliceHeaderExt->sSliceHeader.iSliceBetaOffset;
	857	+
	858	+ pFilter.pLoopf = &pCtx->sDeblockingFunc;
	859	+
	860	+ /* Step2: macroblock deblocking */
	861	+ if (0 == iFilterIdc \|\| 2 == iFilterIdc) {
	862	+ iNextMbXyIndex = pSliceHeaderExt->sSliceHeader.iFirstMbInSlice;
	863	+ pCurDqLayer->iMbX = iNextMbXyIndex % iMbWidth;
	864	+ pCurDqLayer->iMbY = iNextMbXyIndex / iMbWidth;
	865	+ pCurDqLayer->iMbXyIndex = iNextMbXyIndex;
	866	+
	867	+ do {
	868	+ iBoundryFlag = DeblockingAvailableNoInterlayer (pCurDqLayer, iFilterIdc);
	869	+
	870	+ pDeblockMb (pCurDqLayer, &pFilter, iBoundryFlag);
	871	+
	872	+ ++iCountNumMb;
	873	+ if (iCountNumMb >= iTotalNumMb) {
	874	+ break;
	875	+ }
	876	+
	877	+ if (pSliceHeaderExt->sSliceHeader.pPps->uiNumSliceGroups > 1) {
	878	+ iNextMbXyIndex = FmoNextMb (pFmo, iNextMbXyIndex);
	879	+ } else {
	880	+ ++iNextMbXyIndex;
	881	+ }
	882	+ if (-1 == iNextMbXyIndex \|\| iNextMbXyIndex >= iTotalMbCount) { // slice group boundary or end of a frame
	883	+ break;
	884	+ }
	885	+
	886	+ pCurDqLayer->iMbX = iNextMbXyIndex % iMbWidth;
	887	+ pCurDqLayer->iMbY = iNextMbXyIndex / iMbWidth;
	888	+ pCurDqLayer->iMbXyIndex = iNextMbXyIndex;
	889	+ } while (1);
	890	+ }
976	891	}
977	892	/*!
978		− * \brief deblocking module initialize
	893	+ * \brief deblocking module initialize
979	894	*
980	895	* \param pf
981	896	* cpu
982	897	*
983	898	* \return NONE
984	899	*/
985		−
986		−void_t DeblockingInit( SDeblockingFunc *pFunc, int32_t iCpu )
987		−{
988		− pFunc->pfLumaDeblockingLT4Ver = DeblockLumaLt4V_c;
989		− pFunc->pfLumaDeblockingEQ4Ver = DeblockLumaEq4V_c;
990		− pFunc->pfLumaDeblockingLT4Hor = DeblockLumaLt4H_c;
991		− pFunc->pfLumaDeblockingEQ4Hor = DeblockLumaEq4H_c;
992		−
993		− pFunc->pfChromaDeblockingLT4Ver = DeblockChromaLt4V_c;
994		− pFunc->pfChromaDeblockingEQ4Ver = DeblockChromaEq4V_c;
995		− pFunc->pfChromaDeblockingLT4Hor = DeblockChromaLt4H_c;
996		− pFunc->pfChromaDeblockinEQ4Hor = DeblockChromaEq4H_c;
997		−
998		−#ifdef X86_ASM
999		− if( iCpu & WELS_CPU_SSE2 ){
1000		− pFunc->pfLumaDeblockingLT4Ver = DeblockLumaLt4V_sse2;
1001		− pFunc->pfLumaDeblockingEQ4Ver = DeblockLumaEq4V_sse2;
1002		− pFunc->pfLumaDeblockingLT4Hor = DeblockLumaLt4H_sse2;
1003		− pFunc->pfLumaDeblockingEQ4Hor = DeblockLumaEq4H_sse2;
1004		− pFunc->pfChromaDeblockingLT4Ver = DeblockChromaLt4V_sse2;
1005		− pFunc->pfChromaDeblockingEQ4Ver = DeblockChromaEq4V_sse2;
1006		− pFunc->pfChromaDeblockingLT4Hor = DeblockChromaLt4H_sse2;
1007		− pFunc->pfChromaDeblockinEQ4Hor = DeblockChromaEq4H_sse2;
1008		− }
	900	+
	901	+void_t DeblockingInit (SDeblockingFunc* pFunc, int32_t iCpu) {
	902	+ pFunc->pfLumaDeblockingLT4Ver = DeblockLumaLt4V_c;
	903	+ pFunc->pfLumaDeblockingEQ4Ver = DeblockLumaEq4V_c;
	904	+ pFunc->pfLumaDeblockingLT4Hor = DeblockLumaLt4H_c;
	905	+ pFunc->pfLumaDeblockingEQ4Hor = DeblockLumaEq4H_c;
	906	+
	907	+ pFunc->pfChromaDeblockingLT4Ver = DeblockChromaLt4V_c;
	908	+ pFunc->pfChromaDeblockingEQ4Ver = DeblockChromaEq4V_c;
	909	+ pFunc->pfChromaDeblockingLT4Hor = DeblockChromaLt4H_c;
	910	+ pFunc->pfChromaDeblockinEQ4Hor = DeblockChromaEq4H_c;
	911	+
	912	+#ifdef X86_ASM
	913	+ if (iCpu & WELS_CPU_SSE2) {
	914	+ pFunc->pfLumaDeblockingLT4Ver = DeblockLumaLt4V_sse2;
	915	+ pFunc->pfLumaDeblockingEQ4Ver = DeblockLumaEq4V_sse2;
	916	+ pFunc->pfLumaDeblockingLT4Hor = DeblockLumaLt4H_sse2;
	917	+ pFunc->pfLumaDeblockingEQ4Hor = DeblockLumaEq4H_sse2;
	918	+ pFunc->pfChromaDeblockingLT4Ver = DeblockChromaLt4V_sse2;
	919	+ pFunc->pfChromaDeblockingEQ4Ver = DeblockChromaEq4V_sse2;
	920	+ pFunc->pfChromaDeblockingLT4Hor = DeblockChromaLt4H_sse2;
	921	+ pFunc->pfChromaDeblockinEQ4Hor = DeblockChromaEq4H_sse2;
	922	+ }
1009	923	#endif
1010	924
1011	925	}

Pretty printed the C++ code with astyle

Keyboard shortcuts

General

Issues and pull requests

Issues, pull requests and Inbox

Repository code