1 | /* The copyright in this software is being made available under the BSD |
---|
2 | * License, included below. This software may be subject to other third party |
---|
3 | * and contributor rights, including patent rights, and no such rights are |
---|
4 | * granted under this license. |
---|
5 | * |
---|
6 | * Copyright (c) 2010-2015, ITU/ISO/IEC |
---|
7 | * All rights reserved. |
---|
8 | * |
---|
9 | * Redistribution and use in source and binary forms, with or without |
---|
10 | * modification, are permitted provided that the following conditions are met: |
---|
11 | * |
---|
12 | * * Redistributions of source code must retain the above copyright notice, |
---|
13 | * this list of conditions and the following disclaimer. |
---|
14 | * * Redistributions in binary form must reproduce the above copyright notice, |
---|
15 | * this list of conditions and the following disclaimer in the documentation |
---|
16 | * and/or other materials provided with the distribution. |
---|
17 | * * Neither the name of the ISO/IEC nor the names of its contributors may |
---|
18 | * be used to endorse or promote products derived from this software without |
---|
19 | * specific prior written permission. |
---|
20 | * |
---|
21 | * THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS "AS IS" |
---|
22 | * AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE |
---|
23 | * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE |
---|
24 | * ARE DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT HOLDER OR CONTRIBUTORS |
---|
25 | * BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR |
---|
26 | * CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF |
---|
27 | * SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS |
---|
28 | * INTERRUPTION) HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN |
---|
29 | * CONTRACT, STRICT LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) |
---|
30 | * ARISING IN ANY WAY OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF |
---|
31 | * THE POSSIBILITY OF SUCH DAMAGE. |
---|
32 | */ |
---|
33 | |
---|
34 | |
---|
35 | |
---|
36 | /** \file TRenInterpFilter.h |
---|
37 | \brief interpolation filter class (header), (From HM 3.0 TComPredFilter)) |
---|
38 | */ |
---|
39 | |
---|
40 | #ifndef __TRENINTERPFILTER__ |
---|
41 | #define __TRENINTERPFILTER__ |
---|
42 | |
---|
43 | // Include files |
---|
44 | #include "TLibCommon/CommonDef.h" |
---|
45 | #include "assert.h" |
---|
46 | #if NH_3D |
---|
47 | |
---|
48 | // ==================================================================================================================== |
---|
49 | // Constants |
---|
50 | // ==================================================================================================================== |
---|
51 | |
---|
52 | // Local type definitions |
---|
53 | #define HAL_IDX 1 |
---|
54 | #define QU0_IDX 0 |
---|
55 | #define QU1_IDX 2 |
---|
56 | |
---|
57 | // ==================================================================================================================== |
---|
58 | // Class definition |
---|
59 | // ==================================================================================================================== |
---|
60 | |
---|
61 | /// interpolation filter class |
---|
62 | template<UInt bitDepth> |
---|
63 | class TRenInterpFilter |
---|
64 | { |
---|
65 | public: |
---|
66 | TRenInterpFilter(); |
---|
67 | |
---|
68 | // DIF filter interface (for half & quarter) |
---|
69 | __inline Void xCTI_FilterHalfHor(Pel* piSrc, Int iSrcStride, Int iSrcStep, Int iWidth, Int iHeight, Int iDstStride, Int iDstStep, Pel*& rpiDst); |
---|
70 | __inline Void xCTI_FilterHalfHor(Int* piSrc, Int iSrcStride, Int iSrcStep, Int iWidth, Int iHeight, Int iDstStride, Int iDstStep, Pel*& rpiDst); |
---|
71 | |
---|
72 | __inline Void xCTI_FilterQuarter0Hor(Pel* piSrc, Int iSrcStride, Int iSrcStep, Int iWidth, Int iHeight, Int iDstStride, Int iDstStep, Pel*& rpiDst); |
---|
73 | __inline Void xCTI_FilterQuarter0Hor(Int* piSrc, Int iSrcStride, Int iSrcStep, Int iWidth, Int iHeight, Int iDstStride, Int iDstStep, Pel*& rpiDst); |
---|
74 | |
---|
75 | __inline Void xCTI_FilterQuarter1Hor(Pel* piSrc, Int iSrcStride, Int iSrcStep, Int iWidth, Int iHeight, Int iDstStride, Int iDstStep, Pel*& rpiDst); |
---|
76 | __inline Void xCTI_FilterQuarter1Hor(Int* piSrc, Int iSrcStride, Int iSrcStep, Int iWidth, Int iHeight, Int iDstStride, Int iDstStep, Pel*& rpiDst); |
---|
77 | |
---|
78 | __inline Void xCTI_FilterHalfVer (Pel* piSrc, Int iSrcStride, Int iSrcStep, Int iWidth, Int iHeight, Int iDstStride, Int iDstStep, Int*& rpiDst, Int iDstStridePel, Pel*& rpiDstPel ); |
---|
79 | __inline Void xCTI_FilterHalfVer (Pel* piSrc, Int iSrcStride, Int iSrcStep, Int iWidth, Int iHeight, Int iDstStride, Int iDstStep, Int*& rpiDst ); |
---|
80 | __inline Void xCTI_FilterHalfVer (Pel* piSrc, Int iSrcStride, Int iSrcStep, Int iWidth, Int iHeight, Int iDstStride, Int iDstStep, Pel*& rpiDst ); |
---|
81 | |
---|
82 | __inline Void xCTI_FilterQuarter0Ver (Pel* piSrc, Int iSrcStride, Int iSrcStep, Int iWidth, Int iHeight, Int iDstStride, Int iDstStep, Int*& rpiDst ); |
---|
83 | __inline Void xCTI_FilterQuarter0Ver (Pel* piSrc, Int iSrcStride, Int iSrcStep, Int iWidth, Int iHeight, Int iDstStride, Int iDstStep, Pel*& rpiDst ); |
---|
84 | |
---|
85 | __inline Void xCTI_FilterQuarter1Ver (Pel* piSrc, Int iSrcStride, Int iSrcStep, Int iWidth, Int iHeight, Int iDstStride, Int iDstStep, Int*& rpiDst ); |
---|
86 | __inline Void xCTI_FilterQuarter1Ver (Pel* piSrc, Int iSrcStride, Int iSrcStep, Int iWidth, Int iHeight, Int iDstStride, Int iDstStep, Pel*& rpiDst ); |
---|
87 | |
---|
88 | __inline Void xCTI_Filter2DVerC (Pel* piSrc, Int iSrcStride, Int iWidth, Int iHeight, Int iDstStride, Int*& rpiDst, Int iMv); |
---|
89 | __inline Void xCTI_Filter2DHorC (Int* piSrc, Int iSrcStride, Int iWidth, Int iHeight, Int iDstStride, Pel*& rpiDst, Int iMV); |
---|
90 | __inline Void xCTI_Filter1DHorC (Pel* piSrc, Int iSrcStride, Int iWidth, Int iHeight, Int iDstStride, Pel*& rpiDst, Int iMV); |
---|
91 | __inline Void xCTI_Filter1DVerC (Pel* piSrc, Int iSrcStride, Int iWidth, Int iHeight, Int iDstStride, Pel*& rpiDst, Int iMV); |
---|
92 | |
---|
93 | __inline Int xCTI_Filter_VPS04_C_HAL( Pel* pSrc, Int iStride ); |
---|
94 | __inline Int xCTI_Filter_VIS04_C_HAL( Int* pSrc, Int iStride ); |
---|
95 | __inline Int xCTI_Filter_VP04_C_OCT0( Pel* pSrc, Int iStride ); |
---|
96 | __inline Int xCTI_Filter_VI04_C_OCT0( Int* pSrc, Int iStride ); |
---|
97 | __inline Int xCTI_Filter_VP04_C_QUA0( Pel* pSrc, Int iStride ); |
---|
98 | __inline Int xCTI_Filter_VI04_C_QUA0( Int* pSrc, Int iStride ); |
---|
99 | __inline Int xCTI_Filter_VP04_C_OCT1( Pel* pSrc, Int iStride ); |
---|
100 | __inline Int xCTI_Filter_VI04_C_OCT1( Int* pSrc, Int iStride ); |
---|
101 | __inline Int xCTI_Filter_VP04_C_OCT2( Pel* pSrc, Int iStride ); |
---|
102 | __inline Int xCTI_Filter_VI04_C_OCT2( Int* pSrc, Int iStride ); |
---|
103 | __inline Int xCTI_Filter_VP04_C_QUA1( Pel* pSrc, Int iStride ); |
---|
104 | __inline Int xCTI_Filter_VI04_C_QUA1( Int* pSrc, Int iStride ); |
---|
105 | __inline Int xCTI_Filter_VP04_C_OCT3( Pel* pSrc, Int iStride ); |
---|
106 | __inline Int xCTI_Filter_VI04_C_OCT3( Int* pSrc, Int iStride ); |
---|
107 | private: |
---|
108 | __inline Int xClipY( Pel x ) { return std::min<Pel>(Pel((1 << bitDepth)-1), std::max<Pel>( Pel(0), x)); } |
---|
109 | __inline Int xClipC( Pel x ) { return std::min<Pel>(Pel((1 << bitDepth)-1), std::max<Pel>( Pel(0), x)); } |
---|
110 | |
---|
111 | }; |
---|
112 | |
---|
113 | |
---|
114 | // ------------------------------------------------------------------------------------------------ |
---|
115 | // DCTIF filters |
---|
116 | // ------------------------------------------------------------------------------------------------ |
---|
117 | template<UInt bitDepth> |
---|
118 | __inline Void TRenInterpFilter<bitDepth>::xCTI_FilterHalfHor(Pel* piSrc, Int iSrcStride, Int iSrcStep, Int iWidth, Int iHeight, Int iDstStride, Int iDstStep, Pel*& rpiDst) |
---|
119 | { |
---|
120 | Pel* piDst = rpiDst; |
---|
121 | Int iSum; |
---|
122 | Pel* piSrcTmp; |
---|
123 | Int iSrcStep2 = iSrcStep*2; |
---|
124 | Int iSrcStep3 = iSrcStep*3; |
---|
125 | Int iSrcStep4 = iSrcStep*4; |
---|
126 | Int iSrcStep5 = iSrcStep*5; |
---|
127 | Int iSrcStep6 = iSrcStep*6; |
---|
128 | Int iSrcStep7 = iSrcStep*7; |
---|
129 | |
---|
130 | Int iTmp0, iTmp1, iTmp2, iTmp3, iTmpA; |
---|
131 | |
---|
132 | for ( Int y = iHeight; y != 0; y-- ) |
---|
133 | { |
---|
134 | piSrcTmp = &piSrc[ -3*iSrcStep ]; |
---|
135 | for ( Int x = 0; x < iWidth; x++ ) |
---|
136 | { |
---|
137 | // { -1,4,-11,40,40,-11,4,-1 } |
---|
138 | iTmp0 = piSrcTmp[ 0]+piSrcTmp[iSrcStep7]; |
---|
139 | iTmp1 = piSrcTmp[iSrcStep]+piSrcTmp[iSrcStep6]; |
---|
140 | iTmp2 = piSrcTmp[iSrcStep2]+piSrcTmp[iSrcStep5]; |
---|
141 | iTmp3 = piSrcTmp[iSrcStep3]+piSrcTmp[iSrcStep4]; |
---|
142 | |
---|
143 | iTmpA = (iTmp3 << 2) - iTmp2; |
---|
144 | |
---|
145 | iSum = ( iTmp1 << 2 ) |
---|
146 | + ( iTmpA << 3 ) |
---|
147 | + ( iTmpA << 1 ) |
---|
148 | - iTmp0 - iTmp2; |
---|
149 | |
---|
150 | piDst [x * iDstStep] = xClipY( (iSum + 32) >> 6 ); |
---|
151 | piSrcTmp += iSrcStep; |
---|
152 | } |
---|
153 | piSrc += iSrcStride; |
---|
154 | piDst += iDstStride; |
---|
155 | } |
---|
156 | return; |
---|
157 | } |
---|
158 | |
---|
159 | template<UInt bitDepth> |
---|
160 | __inline Void TRenInterpFilter<bitDepth>::xCTI_FilterHalfHor(Int* piSrc, Int iSrcStride, Int iSrcStep, Int iWidth, Int iHeight, Int iDstStride, Int iDstStep, Pel*& rpiDst) |
---|
161 | { |
---|
162 | Pel* piDst = rpiDst; |
---|
163 | Int iSum; |
---|
164 | Int* piSrcTmp; |
---|
165 | Int iSrcStep2 = iSrcStep*2; |
---|
166 | Int iSrcStep3 = iSrcStep*3; |
---|
167 | Int iSrcStep4 = iSrcStep*4; |
---|
168 | Int iSrcStep5 = iSrcStep*5; |
---|
169 | Int iSrcStep6 = iSrcStep*6; |
---|
170 | Int iSrcStep7 = iSrcStep*7; |
---|
171 | |
---|
172 | Int iTmp0, iTmp1, iTmp2, iTmp3, iTmpA; |
---|
173 | |
---|
174 | for ( Int y = iHeight; y != 0; y-- ) |
---|
175 | { |
---|
176 | piSrcTmp = &piSrc[ -3*iSrcStep ]; |
---|
177 | for ( Int x = 0; x < iWidth; x++ ) |
---|
178 | { |
---|
179 | // { -1,4,-11,40,40,-11,4,-1 } |
---|
180 | iTmp0 = piSrcTmp[ 0]+piSrcTmp[iSrcStep7]; |
---|
181 | iTmp1 = piSrcTmp[iSrcStep ]+piSrcTmp[iSrcStep6]; |
---|
182 | iTmp2 = piSrcTmp[iSrcStep2]+piSrcTmp[iSrcStep5]; |
---|
183 | iTmp3 = piSrcTmp[iSrcStep3]+piSrcTmp[iSrcStep4]; |
---|
184 | |
---|
185 | iTmpA = (iTmp3 << 2) - iTmp2; |
---|
186 | |
---|
187 | iSum = ( iTmp1 << 2 ) |
---|
188 | + ( iTmpA << 3 ) |
---|
189 | + ( iTmpA << 1 ) |
---|
190 | - iTmp0 - iTmp2; |
---|
191 | |
---|
192 | piDst [x * iDstStep] = xClipY( (iSum + 2048) >> 12 ); |
---|
193 | piSrcTmp += iSrcStep; |
---|
194 | } |
---|
195 | piSrc += iSrcStride; |
---|
196 | piDst += iDstStride; |
---|
197 | } |
---|
198 | return; |
---|
199 | } |
---|
200 | |
---|
201 | template<UInt bitDepth> |
---|
202 | __inline Void TRenInterpFilter<bitDepth>::xCTI_FilterQuarter0Hor(Pel* piSrc, Int iSrcStride, Int iSrcStep, Int iWidth, Int iHeight, Int iDstStride, Int iDstStep, Pel*& rpiDst) |
---|
203 | { |
---|
204 | Pel* piDst = rpiDst; |
---|
205 | Int iSum; |
---|
206 | Pel* piSrcTmp; |
---|
207 | Int iSrcStep2 = iSrcStep*2; |
---|
208 | Int iSrcStep3 = iSrcStep*3; |
---|
209 | Int iSrcStep4 = iSrcStep*4; |
---|
210 | Int iSrcStep5 = iSrcStep*5; |
---|
211 | Int iSrcStep6 = iSrcStep*6; |
---|
212 | Int iSrcStep7 = iSrcStep*7; |
---|
213 | |
---|
214 | Int iTmp1, iTmp2; |
---|
215 | |
---|
216 | for ( Int y = iHeight; y != 0; y-- ) |
---|
217 | { |
---|
218 | piSrcTmp = &piSrc[ -3*iSrcStep ]; |
---|
219 | for ( Int x = 0; x < iWidth; x++ ) |
---|
220 | { |
---|
221 | // {-1, 4, -10, 57, 19, -7, 3, -1 }, |
---|
222 | |
---|
223 | iTmp1 = piSrcTmp[iSrcStep3] + piSrcTmp[iSrcStep5]; |
---|
224 | iTmp2 = piSrcTmp[iSrcStep6] + piSrcTmp[iSrcStep4]; |
---|
225 | |
---|
226 | iSum = iTmp1 + iTmp2 - piSrcTmp[0] - piSrcTmp[iSrcStep7] |
---|
227 | - ( ( piSrcTmp[iSrcStep2] - iTmp2 ) << 1 ) |
---|
228 | + ( piSrcTmp[iSrcStep] << 2 ) |
---|
229 | - ( ( piSrcTmp[iSrcStep2] + iTmp1 ) << 3 ) |
---|
230 | + ( piSrcTmp[iSrcStep4] << 4 ); |
---|
231 | |
---|
232 | piDst [x * iDstStep] = xClipY(( (iSum + 32) >> 6 )+ piSrcTmp[iSrcStep3]); |
---|
233 | piSrcTmp += iSrcStep; |
---|
234 | } |
---|
235 | piSrc += iSrcStride; |
---|
236 | piDst += iDstStride; |
---|
237 | } |
---|
238 | return; |
---|
239 | } |
---|
240 | |
---|
241 | template<UInt bitDepth> |
---|
242 | __inline Void TRenInterpFilter<bitDepth>::xCTI_FilterQuarter0Hor(Int* piSrc, Int iSrcStride, Int iSrcStep, Int iWidth, Int iHeight, Int iDstStride, Int iDstStep, Pel*& rpiDst) |
---|
243 | { |
---|
244 | Pel* piDst = rpiDst; |
---|
245 | Int iSum; |
---|
246 | Int* piSrcTmp; |
---|
247 | Int iSrcStep2 = iSrcStep*2; |
---|
248 | Int iSrcStep3 = iSrcStep*3; |
---|
249 | Int iSrcStep4 = iSrcStep*4; |
---|
250 | Int iSrcStep5 = iSrcStep*5; |
---|
251 | Int iSrcStep6 = iSrcStep*6; |
---|
252 | Int iSrcStep7 = iSrcStep*7; |
---|
253 | |
---|
254 | Int iTmp1, iTmp2; |
---|
255 | |
---|
256 | for ( Int y = iHeight; y != 0; y-- ) |
---|
257 | { |
---|
258 | piSrcTmp = &piSrc[ -3*iSrcStep ]; |
---|
259 | for ( Int x = 0; x < iWidth; x++ ) |
---|
260 | { |
---|
261 | // {-1, 4, -10, 57, 19, -7, 3, -1 }, |
---|
262 | |
---|
263 | iTmp1 = piSrcTmp[iSrcStep3] + piSrcTmp[iSrcStep5]; |
---|
264 | iTmp2 = piSrcTmp[iSrcStep6] + piSrcTmp[iSrcStep4]; |
---|
265 | |
---|
266 | iSum = iTmp1 + iTmp2 - piSrcTmp[0] - piSrcTmp[iSrcStep7] |
---|
267 | - ( ( piSrcTmp[iSrcStep2] - iTmp2 ) << 1 ) |
---|
268 | + ( piSrcTmp[iSrcStep] << 2 ) |
---|
269 | - ( ( piSrcTmp[iSrcStep2] + iTmp1 ) << 3 ) |
---|
270 | + ( piSrcTmp[iSrcStep4] << 4 ) |
---|
271 | + ( piSrcTmp[iSrcStep3] << 6 ); |
---|
272 | |
---|
273 | piDst [x * iDstStep] = xClipY( (iSum + 2048) >> 12 ); |
---|
274 | piSrcTmp += iSrcStep; |
---|
275 | } |
---|
276 | piSrc += iSrcStride; |
---|
277 | piDst += iDstStride; |
---|
278 | } |
---|
279 | return; |
---|
280 | } |
---|
281 | |
---|
282 | template<UInt bitDepth> |
---|
283 | __inline Void TRenInterpFilter<bitDepth>::xCTI_FilterQuarter1Hor(Pel* piSrc, Int iSrcStride, Int iSrcStep, Int iWidth, Int iHeight, Int iDstStride, Int iDstStep, Pel*& rpiDst) |
---|
284 | { |
---|
285 | Pel* piDst = rpiDst; |
---|
286 | Int iSum; |
---|
287 | Pel* piSrcTmp; |
---|
288 | Int iSrcStep2 = iSrcStep*2; |
---|
289 | Int iSrcStep3 = iSrcStep*3; |
---|
290 | Int iSrcStep4 = iSrcStep*4; |
---|
291 | Int iSrcStep5 = iSrcStep*5; |
---|
292 | Int iSrcStep6 = iSrcStep*6; |
---|
293 | Int iSrcStep7 = iSrcStep*7; |
---|
294 | |
---|
295 | Int iTmp1, iTmp2; |
---|
296 | for ( Int y = iHeight; y != 0; y-- ) |
---|
297 | { |
---|
298 | piSrcTmp = &piSrc[ -3*iSrcStep ]; |
---|
299 | for ( Int x = 0; x < iWidth; x++ ) |
---|
300 | { |
---|
301 | // {-1, 3, -7, 19, 57, -10, 4, -1 }, |
---|
302 | |
---|
303 | iTmp1 = piSrcTmp[iSrcStep4] + piSrcTmp[iSrcStep2]; |
---|
304 | iTmp2 = piSrcTmp[iSrcStep ] + piSrcTmp[iSrcStep3]; |
---|
305 | |
---|
306 | iSum = iTmp1 + iTmp2 - piSrcTmp[0] - piSrcTmp[iSrcStep7] |
---|
307 | - ( ( piSrcTmp[iSrcStep5] - iTmp2 ) << 1 ) |
---|
308 | + ( piSrcTmp[iSrcStep6] << 2 ) |
---|
309 | - ( ( piSrcTmp[iSrcStep5] + iTmp1 ) << 3 ) |
---|
310 | + ( piSrcTmp[iSrcStep3] << 4 ); |
---|
311 | |
---|
312 | piDst [x * iDstStep] = xClipY( ((iSum + 32) >> 6) + piSrcTmp[iSrcStep4] ); |
---|
313 | piSrcTmp += iSrcStep; |
---|
314 | } |
---|
315 | piSrc += iSrcStride; |
---|
316 | piDst += iDstStride; |
---|
317 | } |
---|
318 | return; |
---|
319 | } |
---|
320 | |
---|
321 | template<UInt bitDepth> |
---|
322 | __inline Void TRenInterpFilter<bitDepth>::xCTI_FilterQuarter1Hor(Int* piSrc, Int iSrcStride, Int iSrcStep, Int iWidth, Int iHeight, Int iDstStride, Int iDstStep, Pel*& rpiDst) |
---|
323 | { |
---|
324 | Pel* piDst = rpiDst; |
---|
325 | Int iSum; |
---|
326 | Int* piSrcTmp; |
---|
327 | Int iSrcStep2 = iSrcStep*2; |
---|
328 | Int iSrcStep3 = iSrcStep*3; |
---|
329 | Int iSrcStep4 = iSrcStep*4; |
---|
330 | Int iSrcStep5 = iSrcStep*5; |
---|
331 | Int iSrcStep6 = iSrcStep*6; |
---|
332 | Int iSrcStep7 = iSrcStep*7; |
---|
333 | |
---|
334 | Int iTmp1, iTmp2; |
---|
335 | for ( Int y = iHeight; y != 0; y-- ) |
---|
336 | { |
---|
337 | piSrcTmp = &piSrc[ -3*iSrcStep ]; |
---|
338 | for ( Int x = 0; x < iWidth; x++ ) |
---|
339 | { |
---|
340 | // {-1, 3, -7, 19, 57, -10, 4, -1 }, |
---|
341 | |
---|
342 | iTmp1 = piSrcTmp[iSrcStep4] + piSrcTmp[iSrcStep2]; |
---|
343 | iTmp2 = piSrcTmp[iSrcStep ] + piSrcTmp[iSrcStep3]; |
---|
344 | |
---|
345 | iSum = iTmp1 + iTmp2 - piSrcTmp[0] - piSrcTmp[iSrcStep7] |
---|
346 | - ( ( piSrcTmp[iSrcStep5] - iTmp2 ) << 1 ) |
---|
347 | + ( piSrcTmp[iSrcStep6] << 2 ) |
---|
348 | - ( ( piSrcTmp[iSrcStep5] + iTmp1 ) << 3 ) |
---|
349 | + ( piSrcTmp[iSrcStep3] << 4 ) |
---|
350 | + ( piSrcTmp[iSrcStep4] << 6 ); |
---|
351 | |
---|
352 | piDst [x * iDstStep] = xClipY( (iSum + 2048) >> 12 ); |
---|
353 | piSrcTmp += iSrcStep; |
---|
354 | } |
---|
355 | piSrc += iSrcStride; |
---|
356 | piDst += iDstStride; |
---|
357 | } |
---|
358 | return; |
---|
359 | } |
---|
360 | |
---|
361 | template<UInt bitDepth> |
---|
362 | __inline Void TRenInterpFilter<bitDepth>::xCTI_FilterHalfVer (Pel* piSrc, Int iSrcStride, Int iSrcStep, Int iWidth, Int iHeight, Int iDstStride, Int iDstStep, Int*& rpiDst, Int iDstStridePel, Pel*& rpiDstPel ) |
---|
363 | { |
---|
364 | Int* piDst = rpiDst; |
---|
365 | Pel* piDstPel = rpiDstPel; |
---|
366 | Int iSum; |
---|
367 | Pel* piSrcTmp; |
---|
368 | Int iSrcStride2 = iSrcStride*2; |
---|
369 | Int iSrcStride3 = iSrcStride*3; |
---|
370 | Int iSrcStride4 = iSrcStride*4; |
---|
371 | Int iSrcStride5 = iSrcStride*5; |
---|
372 | Int iSrcStride6 = iSrcStride*6; |
---|
373 | Int iSrcStride7 = iSrcStride*7; |
---|
374 | |
---|
375 | Int iTmp0, iTmp1, iTmp2, iTmp3, iTmpA; |
---|
376 | for ( Int y = iHeight; y != 0; y-- ) |
---|
377 | { |
---|
378 | piSrcTmp = &piSrc[ -3*iSrcStride ]; |
---|
379 | for ( Int x = 0; x < iWidth; x++ ) |
---|
380 | { |
---|
381 | // { -1,4,-11,40,40,-11,4,-1 } |
---|
382 | iTmp0 = piSrcTmp[ 0]+piSrcTmp[iSrcStride7]; |
---|
383 | iTmp1 = piSrcTmp[iSrcStride ]+piSrcTmp[iSrcStride6]; |
---|
384 | iTmp2 = piSrcTmp[iSrcStride2]+piSrcTmp[iSrcStride5]; |
---|
385 | iTmp3 = piSrcTmp[iSrcStride3]+piSrcTmp[iSrcStride4]; |
---|
386 | |
---|
387 | iTmpA = (iTmp3 << 2) - iTmp2; |
---|
388 | |
---|
389 | iSum = ( iTmp1 << 2 ) |
---|
390 | + ( iTmpA << 3 ) |
---|
391 | + ( iTmpA << 1 ) |
---|
392 | - iTmp0 - iTmp2; |
---|
393 | |
---|
394 | piDst[x * iDstStep] = iSum; |
---|
395 | piDstPel[x * iDstStep] = xClipY( (iSum + 32) >> 6 ); |
---|
396 | piSrcTmp += iSrcStep; |
---|
397 | } |
---|
398 | piSrc += iSrcStride; |
---|
399 | piDst += iDstStride; |
---|
400 | piDstPel += iDstStridePel; |
---|
401 | } |
---|
402 | return; |
---|
403 | } |
---|
404 | |
---|
405 | template<UInt bitDepth> |
---|
406 | __inline Void TRenInterpFilter<bitDepth>::xCTI_FilterHalfVer (Pel* piSrc, Int iSrcStride, Int iSrcStep, Int iWidth, Int iHeight, Int iDstStride, Int iDstStep, Int*& rpiDst) |
---|
407 | { |
---|
408 | Int* piDst = rpiDst; |
---|
409 | Int iSum; |
---|
410 | Pel* piSrcTmp; |
---|
411 | Int iSrcStride2 = iSrcStride*2; |
---|
412 | Int iSrcStride3 = iSrcStride*3; |
---|
413 | Int iSrcStride4 = iSrcStride*4; |
---|
414 | Int iSrcStride5 = iSrcStride*5; |
---|
415 | Int iSrcStride6 = iSrcStride*6; |
---|
416 | Int iSrcStride7 = iSrcStride*7; |
---|
417 | |
---|
418 | Int iTmp0, iTmp1, iTmp2, iTmp3, iTmpA; |
---|
419 | for ( Int y = iHeight; y != 0; y-- ) |
---|
420 | { |
---|
421 | piSrcTmp = &piSrc[ -3*iSrcStride ]; |
---|
422 | for ( Int x = 0; x < iWidth; x++ ) |
---|
423 | { |
---|
424 | // { -1,4,-11,40,40,-11,4,-1 } |
---|
425 | iTmp0 = piSrcTmp[ 0]+piSrcTmp[iSrcStride7]; |
---|
426 | iTmp1 = piSrcTmp[iSrcStride ]+piSrcTmp[iSrcStride6]; |
---|
427 | iTmp2 = piSrcTmp[iSrcStride2]+piSrcTmp[iSrcStride5]; |
---|
428 | iTmp3 = piSrcTmp[iSrcStride3]+piSrcTmp[iSrcStride4]; |
---|
429 | |
---|
430 | iTmpA = (iTmp3 << 2) - iTmp2; |
---|
431 | |
---|
432 | iSum = ( iTmp1 << 2 ) |
---|
433 | + ( iTmpA << 3 ) |
---|
434 | + ( iTmpA << 1 ) |
---|
435 | - iTmp0 - iTmp2; |
---|
436 | |
---|
437 | piDst[x * iDstStep] = iSum; |
---|
438 | piSrcTmp += iSrcStep; |
---|
439 | } |
---|
440 | piSrc += iSrcStride; |
---|
441 | piDst += iDstStride; |
---|
442 | } |
---|
443 | return; |
---|
444 | } |
---|
445 | |
---|
446 | template<UInt bitDepth> |
---|
447 | __inline Void TRenInterpFilter<bitDepth>::xCTI_FilterHalfVer (Pel* piSrc, Int iSrcStride, Int iSrcStep, Int iWidth, Int iHeight, Int iDstStride, Int iDstStep, Pel*& rpiDst) |
---|
448 | { |
---|
449 | Pel* piDst = rpiDst; |
---|
450 | Int iSum; |
---|
451 | Pel* piSrcTmp; |
---|
452 | |
---|
453 | Int iSrcStride2 = iSrcStride*2; |
---|
454 | Int iSrcStride3 = iSrcStride*3; |
---|
455 | Int iSrcStride4 = iSrcStride*4; |
---|
456 | Int iSrcStride5 = iSrcStride*5; |
---|
457 | Int iSrcStride6 = iSrcStride*6; |
---|
458 | Int iSrcStride7 = iSrcStride*7; |
---|
459 | |
---|
460 | Int iTmp0, iTmp1, iTmp2, iTmp3, iTmpA; |
---|
461 | for ( Int y = iHeight; y != 0; y-- ) |
---|
462 | { |
---|
463 | piSrcTmp = &piSrc[ -3*iSrcStride ]; |
---|
464 | for ( Int x = 0; x < iWidth; x++ ) |
---|
465 | { |
---|
466 | // { -1,4,-11,40,40,-11,4,-1 } |
---|
467 | iTmp0 = piSrcTmp[ 0]+piSrcTmp[iSrcStride7]; |
---|
468 | iTmp1 = piSrcTmp[iSrcStride ]+piSrcTmp[iSrcStride6]; |
---|
469 | iTmp2 = piSrcTmp[iSrcStride2]+piSrcTmp[iSrcStride5]; |
---|
470 | iTmp3 = piSrcTmp[iSrcStride3]+piSrcTmp[iSrcStride4]; |
---|
471 | |
---|
472 | iTmpA = (iTmp3 << 2) - iTmp2; |
---|
473 | |
---|
474 | iSum = ( iTmp1 << 2 ) |
---|
475 | + ( iTmpA << 3 ) |
---|
476 | + ( iTmpA << 1 ) |
---|
477 | - iTmp0 - iTmp2; |
---|
478 | |
---|
479 | piDst[x * iDstStep] = xClipY( (iSum + 32) >> 6 ); |
---|
480 | piSrcTmp += iSrcStep; |
---|
481 | } |
---|
482 | piSrc += iSrcStride; |
---|
483 | piDst += iDstStride; |
---|
484 | } |
---|
485 | return; |
---|
486 | } |
---|
487 | |
---|
488 | template<UInt bitDepth> |
---|
489 | __inline Void TRenInterpFilter<bitDepth>::xCTI_FilterQuarter0Ver (Pel* piSrc, Int iSrcStride, Int iSrcStep, Int iWidth, Int iHeight, Int iDstStride, Int iDstStep, Int*& rpiDst) |
---|
490 | { |
---|
491 | Int* piDst = rpiDst; |
---|
492 | Int iSum; |
---|
493 | Pel* piSrcTmp; |
---|
494 | Int iSrcStride2 = iSrcStride*2; |
---|
495 | Int iSrcStride3 = iSrcStride*3; |
---|
496 | Int iSrcStride4 = iSrcStride*4; |
---|
497 | Int iSrcStride5 = iSrcStride*5; |
---|
498 | Int iSrcStride6 = iSrcStride*6; |
---|
499 | Int iSrcStride7 = iSrcStride*7; |
---|
500 | |
---|
501 | Int iTmp1, iTmp2; |
---|
502 | for ( Int y = iHeight; y != 0; y-- ) |
---|
503 | { |
---|
504 | piSrcTmp = &piSrc[ -3*iSrcStride ]; |
---|
505 | for ( Int x = 0; x < iWidth; x++ ) |
---|
506 | { |
---|
507 | // {-1, 4, -10, 57, 19, -7, 3, -1 }, |
---|
508 | |
---|
509 | iTmp1 = piSrcTmp[iSrcStride3] + piSrcTmp[iSrcStride5]; |
---|
510 | iTmp2 = piSrcTmp[iSrcStride6] + piSrcTmp[iSrcStride4]; |
---|
511 | |
---|
512 | iSum = iTmp1 + iTmp2 - piSrcTmp[0] - piSrcTmp[iSrcStride7] |
---|
513 | - ( ( piSrcTmp[iSrcStride2] - iTmp2 ) << 1 ) |
---|
514 | + ( piSrcTmp[iSrcStride] << 2 ) |
---|
515 | - ( ( piSrcTmp[iSrcStride2] + iTmp1 ) << 3 ) |
---|
516 | + ( piSrcTmp[iSrcStride4] << 4 ) |
---|
517 | + ( piSrcTmp[iSrcStride3] << 6 ); |
---|
518 | |
---|
519 | piDst[x * iDstStep] = iSum; |
---|
520 | piSrcTmp += iSrcStep; |
---|
521 | } |
---|
522 | piSrc += iSrcStride; |
---|
523 | piDst += iDstStride; |
---|
524 | } |
---|
525 | return; |
---|
526 | } |
---|
527 | |
---|
528 | template<UInt bitDepth> |
---|
529 | __inline Void TRenInterpFilter<bitDepth>::xCTI_FilterQuarter0Ver (Pel* piSrc, Int iSrcStride, Int iSrcStep, Int iWidth, Int iHeight, Int iDstStride, Int iDstStep, Pel*& rpiDst) |
---|
530 | { |
---|
531 | Pel* piDst = rpiDst; |
---|
532 | Int iSum; |
---|
533 | Pel* piSrcTmp; |
---|
534 | |
---|
535 | Int iSrcStride2 = iSrcStride*2; |
---|
536 | Int iSrcStride3 = iSrcStride*3; |
---|
537 | Int iSrcStride4 = iSrcStride*4; |
---|
538 | Int iSrcStride5 = iSrcStride*5; |
---|
539 | Int iSrcStride6 = iSrcStride*6; |
---|
540 | Int iSrcStride7 = iSrcStride*7; |
---|
541 | |
---|
542 | Int iTmp1, iTmp2; |
---|
543 | for ( Int y = iHeight; y != 0; y-- ) |
---|
544 | { |
---|
545 | piSrcTmp = &piSrc[ -3*iSrcStride ]; |
---|
546 | for ( Int x = 0; x < iWidth; x++ ) |
---|
547 | { |
---|
548 | // {-1, 4, -10, 57, 19, -7, 3, -1 }, |
---|
549 | |
---|
550 | iTmp1 = piSrcTmp[iSrcStride3] + piSrcTmp[iSrcStride5]; |
---|
551 | iTmp2 = piSrcTmp[iSrcStride6] + piSrcTmp[iSrcStride4]; |
---|
552 | |
---|
553 | iSum = iTmp1 + iTmp2 - piSrcTmp[0] - piSrcTmp[iSrcStride7] |
---|
554 | - ( ( piSrcTmp[iSrcStride2] - iTmp2 ) << 1 ) |
---|
555 | + ( piSrcTmp[iSrcStride] << 2 ) |
---|
556 | - ( ( piSrcTmp[iSrcStride2] + iTmp1 ) << 3 ) |
---|
557 | + ( piSrcTmp[iSrcStride4] << 4 ); |
---|
558 | |
---|
559 | piDst[x * iDstStep] = xClipY( ((iSum + 32) >> 6) + piSrcTmp[iSrcStride3] ); |
---|
560 | piSrcTmp += iSrcStep; |
---|
561 | } |
---|
562 | piSrc += iSrcStride; |
---|
563 | piDst += iDstStride; |
---|
564 | } |
---|
565 | return; |
---|
566 | } |
---|
567 | |
---|
568 | template<UInt bitDepth> |
---|
569 | __inline Void TRenInterpFilter<bitDepth>::xCTI_FilterQuarter1Ver (Pel* piSrc, Int iSrcStride, Int iSrcStep, Int iWidth, Int iHeight, Int iDstStride, Int iDstStep, Int*& rpiDst) |
---|
570 | { |
---|
571 | Int* piDst = rpiDst; |
---|
572 | Int iSum; |
---|
573 | Pel* piSrcTmp; |
---|
574 | Int iSrcStride2 = iSrcStride*2; |
---|
575 | Int iSrcStride3 = iSrcStride*3; |
---|
576 | Int iSrcStride4 = iSrcStride*4; |
---|
577 | Int iSrcStride5 = iSrcStride*5; |
---|
578 | Int iSrcStride6 = iSrcStride*6; |
---|
579 | Int iSrcStride7 = iSrcStride*7; |
---|
580 | |
---|
581 | Int iTmp1, iTmp2; |
---|
582 | |
---|
583 | for ( Int y = iHeight; y != 0; y-- ) |
---|
584 | { |
---|
585 | piSrcTmp = &piSrc[ -3*iSrcStride ]; |
---|
586 | for ( Int x = 0; x < iWidth; x++ ) |
---|
587 | { |
---|
588 | /// {-1, 3, -7, 19, 57, -10, 4, -1 }, |
---|
589 | iTmp1 = piSrcTmp[iSrcStride4] + piSrcTmp[iSrcStride2]; |
---|
590 | iTmp2 = piSrcTmp[iSrcStride ] + piSrcTmp[iSrcStride3]; |
---|
591 | |
---|
592 | iSum = iTmp1 + iTmp2 - piSrcTmp[0] - piSrcTmp[iSrcStride7] |
---|
593 | - ( ( piSrcTmp[iSrcStride5] - iTmp2 ) << 1 ) |
---|
594 | + ( piSrcTmp[iSrcStride6] << 2 ) |
---|
595 | - ( ( piSrcTmp[iSrcStride5] + iTmp1 ) << 3 ) |
---|
596 | + ( piSrcTmp[iSrcStride3] << 4 ) |
---|
597 | + ( piSrcTmp[iSrcStride4] << 6 ); |
---|
598 | |
---|
599 | piDst[x * iDstStep] = iSum; |
---|
600 | piSrcTmp += iSrcStep; |
---|
601 | } |
---|
602 | piSrc += iSrcStride; |
---|
603 | piDst += iDstStride; |
---|
604 | } |
---|
605 | return; |
---|
606 | } |
---|
607 | |
---|
608 | template<UInt bitDepth> |
---|
609 | __inline Void TRenInterpFilter<bitDepth>::xCTI_FilterQuarter1Ver (Pel* piSrc, Int iSrcStride, Int iSrcStep, Int iWidth, Int iHeight, Int iDstStride, Int iDstStep, Pel*& rpiDst) |
---|
610 | { |
---|
611 | Pel* piDst = rpiDst; |
---|
612 | Int iSum; |
---|
613 | Pel* piSrcTmp; |
---|
614 | Int iSrcStride2 = iSrcStride*2; |
---|
615 | Int iSrcStride3 = iSrcStride*3; |
---|
616 | Int iSrcStride4 = iSrcStride*4; |
---|
617 | Int iSrcStride5 = iSrcStride*5; |
---|
618 | Int iSrcStride6 = iSrcStride*6; |
---|
619 | Int iSrcStride7 = iSrcStride*7; |
---|
620 | |
---|
621 | Int iTmp1, iTmp2; |
---|
622 | |
---|
623 | for ( Int y = iHeight; y != 0; y-- ) |
---|
624 | { |
---|
625 | piSrcTmp = &piSrc[ -3*iSrcStride ]; |
---|
626 | for ( Int x = 0; x < iWidth; x++ ) |
---|
627 | { |
---|
628 | /// {-1, 3, -7, 19, 57, -10, 4, -1 }, |
---|
629 | iTmp1 = piSrcTmp[iSrcStride4] + piSrcTmp[iSrcStride2]; |
---|
630 | iTmp2 = piSrcTmp[iSrcStride ] + piSrcTmp[iSrcStride3]; |
---|
631 | |
---|
632 | iSum = iTmp1 + iTmp2 - piSrcTmp[0] - piSrcTmp[iSrcStride7] |
---|
633 | - ( ( piSrcTmp[iSrcStride5] - iTmp2 ) << 1 ) |
---|
634 | + ( piSrcTmp[iSrcStride6] << 2 ) |
---|
635 | - ( ( piSrcTmp[iSrcStride5] + iTmp1 ) << 3 ) |
---|
636 | + ( piSrcTmp[iSrcStride3] << 4 ); |
---|
637 | |
---|
638 | piDst[x * iDstStep] = xClipY( ((iSum + 32) >> 6) + piSrcTmp[iSrcStride4] ); |
---|
639 | piSrcTmp += iSrcStep; |
---|
640 | } |
---|
641 | piSrc += iSrcStride; |
---|
642 | piDst += iDstStride; |
---|
643 | } |
---|
644 | return; |
---|
645 | } |
---|
646 | |
---|
647 | // ------------------------------------------------------------------------------------------------ |
---|
648 | // DCTIF filters for Chroma |
---|
649 | // ------------------------------------------------------------------------------------------------ |
---|
650 | template<UInt bitDepth> |
---|
651 | __inline Void TRenInterpFilter<bitDepth>::xCTI_Filter2DVerC (Pel* piSrc, Int iSrcStride, Int iWidth, Int iHeight, Int iDstStride, Int*& rpiDst, Int iMV) |
---|
652 | { |
---|
653 | Int* piDst = rpiDst; |
---|
654 | Int iSum; |
---|
655 | Pel* piSrcTmp; |
---|
656 | |
---|
657 | switch (iMV) |
---|
658 | { |
---|
659 | case 1: |
---|
660 | { |
---|
661 | for ( Int y = iHeight; y != 0; y-- ) |
---|
662 | { |
---|
663 | piSrcTmp = &piSrc[ -iSrcStride ]; |
---|
664 | for ( Int x = 0; x < iWidth; x++ ) |
---|
665 | { |
---|
666 | iSum = xCTI_Filter_VP04_C_OCT0( piSrcTmp, iSrcStride ); |
---|
667 | piDst[x ] = iSum; |
---|
668 | piSrcTmp++; |
---|
669 | } |
---|
670 | piSrc += iSrcStride; |
---|
671 | piDst += iDstStride; |
---|
672 | } |
---|
673 | } |
---|
674 | break; |
---|
675 | case 2: |
---|
676 | { |
---|
677 | for ( Int y = iHeight; y != 0; y-- ) |
---|
678 | { |
---|
679 | piSrcTmp = &piSrc[ -iSrcStride ]; |
---|
680 | for ( Int x = 0; x < iWidth; x++ ) |
---|
681 | { |
---|
682 | iSum = xCTI_Filter_VP04_C_QUA0( piSrcTmp, iSrcStride ); |
---|
683 | piDst[x ] = iSum; |
---|
684 | piSrcTmp++; |
---|
685 | } |
---|
686 | piSrc += iSrcStride; |
---|
687 | piDst += iDstStride; |
---|
688 | } |
---|
689 | } |
---|
690 | break; |
---|
691 | case 6: |
---|
692 | { |
---|
693 | for ( Int y = iHeight; y != 0; y-- ) |
---|
694 | { |
---|
695 | piSrcTmp = &piSrc[ -iSrcStride ]; |
---|
696 | for ( Int x = 0; x < iWidth; x++ ) |
---|
697 | { |
---|
698 | iSum = xCTI_Filter_VP04_C_QUA1( piSrcTmp, iSrcStride ); |
---|
699 | piDst[x ] = iSum; |
---|
700 | piSrcTmp++; |
---|
701 | } |
---|
702 | piSrc += iSrcStride; |
---|
703 | piDst += iDstStride; |
---|
704 | } |
---|
705 | } |
---|
706 | break; |
---|
707 | case 3: |
---|
708 | { |
---|
709 | for ( Int y = iHeight; y != 0; y-- ) |
---|
710 | { |
---|
711 | piSrcTmp = &piSrc[ -iSrcStride ]; |
---|
712 | for ( Int x = 0; x < iWidth; x++ ) |
---|
713 | { |
---|
714 | iSum = xCTI_Filter_VP04_C_OCT1( piSrcTmp, iSrcStride ); |
---|
715 | piDst[x ] = iSum; |
---|
716 | piSrcTmp++; |
---|
717 | } |
---|
718 | piSrc += iSrcStride; |
---|
719 | piDst += iDstStride; |
---|
720 | } |
---|
721 | } |
---|
722 | break; |
---|
723 | case 5: |
---|
724 | { |
---|
725 | for ( Int y = iHeight; y != 0; y-- ) |
---|
726 | { |
---|
727 | piSrcTmp = &piSrc[ -iSrcStride ]; |
---|
728 | for ( Int x = 0; x < iWidth; x++ ) |
---|
729 | { |
---|
730 | iSum = xCTI_Filter_VP04_C_OCT2( piSrcTmp, iSrcStride ); |
---|
731 | piDst[x ] = iSum; |
---|
732 | piSrcTmp++; |
---|
733 | } |
---|
734 | piSrc += iSrcStride; |
---|
735 | piDst += iDstStride; |
---|
736 | } |
---|
737 | } |
---|
738 | break; |
---|
739 | case 7: |
---|
740 | { |
---|
741 | for ( Int y = iHeight; y != 0; y-- ) |
---|
742 | { |
---|
743 | piSrcTmp = &piSrc[ -iSrcStride ]; |
---|
744 | for ( Int x = 0; x < iWidth; x++ ) |
---|
745 | { |
---|
746 | iSum = xCTI_Filter_VP04_C_OCT3( piSrcTmp, iSrcStride ); |
---|
747 | piDst[x ] = iSum; |
---|
748 | piSrcTmp++; |
---|
749 | } |
---|
750 | piSrc += iSrcStride; |
---|
751 | piDst += iDstStride; |
---|
752 | } |
---|
753 | } |
---|
754 | break; |
---|
755 | case 4: |
---|
756 | { |
---|
757 | |
---|
758 | for ( Int y = iHeight; y != 0; y-- ) |
---|
759 | { |
---|
760 | piSrcTmp = &piSrc[ -iSrcStride ]; |
---|
761 | for ( Int x = 0; x < iWidth; x++ ) |
---|
762 | { |
---|
763 | iSum = xCTI_Filter_VPS04_C_HAL( piSrcTmp, iSrcStride ); |
---|
764 | piDst[x ] = iSum; |
---|
765 | piSrcTmp++; |
---|
766 | } |
---|
767 | piSrc += iSrcStride; |
---|
768 | piDst += iDstStride; |
---|
769 | } |
---|
770 | } |
---|
771 | break; |
---|
772 | default: |
---|
773 | assert( 0 ); |
---|
774 | } |
---|
775 | return; |
---|
776 | } |
---|
777 | |
---|
778 | template<UInt bitDepth> |
---|
779 | __inline Void TRenInterpFilter<bitDepth>::xCTI_Filter2DHorC(Int* piSrc, Int iSrcStride, Int iWidth, Int iHeight, Int iDstStride, Pel*& rpiDst, Int iMV) |
---|
780 | { |
---|
781 | Pel* piDst = rpiDst; |
---|
782 | Int iSum; |
---|
783 | Int* piSrcTmp; |
---|
784 | |
---|
785 | switch (iMV) |
---|
786 | { |
---|
787 | case 1: |
---|
788 | { |
---|
789 | for ( Int y = iHeight; y != 0; y-- ) |
---|
790 | { |
---|
791 | piSrcTmp = &piSrc[ -1 ]; |
---|
792 | for ( Int x = 0; x < iWidth; x++ ) |
---|
793 | { |
---|
794 | iSum = xCTI_Filter_VI04_C_OCT0( piSrcTmp, 1 ); |
---|
795 | piDst [x ] = xClipC ((iSum + 2048) >> 12 ); |
---|
796 | piSrcTmp++; |
---|
797 | } |
---|
798 | piSrc += iSrcStride; |
---|
799 | piDst += iDstStride; |
---|
800 | } |
---|
801 | } |
---|
802 | break; |
---|
803 | case 2: |
---|
804 | { |
---|
805 | for ( Int y = iHeight; y != 0; y-- ) |
---|
806 | { |
---|
807 | piSrcTmp = &piSrc[ -1 ]; |
---|
808 | for ( Int x = 0; x < iWidth; x++ ) |
---|
809 | { |
---|
810 | iSum = xCTI_Filter_VI04_C_QUA0( piSrcTmp, 1 ); |
---|
811 | piDst [x ] = xClipC ((iSum + 2048) >> 12 ); |
---|
812 | piSrcTmp++; |
---|
813 | } |
---|
814 | piSrc += iSrcStride; |
---|
815 | piDst += iDstStride; |
---|
816 | } |
---|
817 | } |
---|
818 | break; |
---|
819 | case 6: |
---|
820 | { |
---|
821 | for ( Int y = iHeight; y != 0; y-- ) |
---|
822 | { |
---|
823 | piSrcTmp = &piSrc[ -1 ]; |
---|
824 | for ( Int x = 0; x < iWidth; x++ ) |
---|
825 | { |
---|
826 | iSum = xCTI_Filter_VI04_C_QUA1( piSrcTmp, 1 ); |
---|
827 | piDst [x ] = xClipC ((iSum + 2048) >> 12 ); |
---|
828 | piSrcTmp++; |
---|
829 | } |
---|
830 | piSrc += iSrcStride; |
---|
831 | piDst += iDstStride; |
---|
832 | } |
---|
833 | } |
---|
834 | break; |
---|
835 | case 3: |
---|
836 | { |
---|
837 | for ( Int y = iHeight; y != 0; y-- ) |
---|
838 | { |
---|
839 | piSrcTmp = &piSrc[ -1 ]; |
---|
840 | for ( Int x = 0; x < iWidth; x++ ) |
---|
841 | { |
---|
842 | iSum = xCTI_Filter_VI04_C_OCT1( piSrcTmp, 1 ); |
---|
843 | piDst [x ] = xClipC ((iSum + 2048) >> 12 ); |
---|
844 | piSrcTmp++; |
---|
845 | } |
---|
846 | piSrc += iSrcStride; |
---|
847 | piDst += iDstStride; |
---|
848 | } |
---|
849 | } |
---|
850 | break; |
---|
851 | case 5: |
---|
852 | { |
---|
853 | for ( Int y = iHeight; y != 0; y-- ) |
---|
854 | { |
---|
855 | piSrcTmp = &piSrc[ -1 ]; |
---|
856 | for ( Int x = 0; x < iWidth; x++ ) |
---|
857 | { |
---|
858 | iSum = xCTI_Filter_VI04_C_OCT2( piSrcTmp, 1 ); |
---|
859 | piDst [x ] = xClipC ((iSum + 2048) >> 12 ); |
---|
860 | piSrcTmp++; |
---|
861 | } |
---|
862 | piSrc += iSrcStride; |
---|
863 | piDst += iDstStride; |
---|
864 | } |
---|
865 | } |
---|
866 | break; |
---|
867 | case 7: |
---|
868 | { |
---|
869 | for ( Int y = iHeight; y != 0; y-- ) |
---|
870 | { |
---|
871 | piSrcTmp = &piSrc[ -1 ]; |
---|
872 | for ( Int x = 0; x < iWidth; x++ ) |
---|
873 | { |
---|
874 | iSum = xCTI_Filter_VI04_C_OCT3( piSrcTmp, 1 ); |
---|
875 | piDst [x ] = xClipC ((iSum + 2048) >> 12 ); |
---|
876 | piSrcTmp++; |
---|
877 | } |
---|
878 | piSrc += iSrcStride; |
---|
879 | piDst += iDstStride; |
---|
880 | } |
---|
881 | } |
---|
882 | break; |
---|
883 | case 4: |
---|
884 | { |
---|
885 | for ( Int y = iHeight; y != 0; y-- ) |
---|
886 | { |
---|
887 | piSrcTmp = &piSrc[ -1 ]; |
---|
888 | for ( Int x = 0; x < iWidth; x++ ) |
---|
889 | { |
---|
890 | iSum = xCTI_Filter_VIS04_C_HAL( piSrcTmp, 1 ); |
---|
891 | piDst [x ] = xClipC ((iSum + 2048) >> 12 ); |
---|
892 | piSrcTmp++; |
---|
893 | } |
---|
894 | piSrc += iSrcStride; |
---|
895 | piDst += iDstStride; |
---|
896 | } |
---|
897 | } |
---|
898 | break; |
---|
899 | default: |
---|
900 | assert( 0 ); |
---|
901 | } |
---|
902 | |
---|
903 | return; |
---|
904 | } |
---|
905 | |
---|
906 | template<UInt bitDepth> |
---|
907 | __inline Void TRenInterpFilter<bitDepth>::xCTI_Filter1DVerC (Pel* piSrc, Int iSrcStride, Int iWidth, Int iHeight, Int iDstStride, Pel*& rpiDst, Int iMV) |
---|
908 | { |
---|
909 | Pel* piDst = rpiDst; |
---|
910 | Int iSum; |
---|
911 | Pel* piSrcTmp; |
---|
912 | |
---|
913 | switch (iMV) |
---|
914 | { |
---|
915 | case 1: |
---|
916 | { |
---|
917 | for ( Int y = iHeight; y != 0; y-- ) |
---|
918 | { |
---|
919 | piSrcTmp = &piSrc[ -iSrcStride ]; |
---|
920 | for ( Int x = 0; x < iWidth; x++ ) |
---|
921 | { |
---|
922 | iSum = xCTI_Filter_VP04_C_OCT0( piSrcTmp, iSrcStride ); |
---|
923 | piDst[x ] = xClipC ((iSum + 32) >> 6 ); |
---|
924 | piSrcTmp++; |
---|
925 | } |
---|
926 | piSrc += iSrcStride; |
---|
927 | piDst += iDstStride; |
---|
928 | } |
---|
929 | } |
---|
930 | break; |
---|
931 | case 2: |
---|
932 | { |
---|
933 | for ( Int y = iHeight; y != 0; y-- ) |
---|
934 | { |
---|
935 | piSrcTmp = &piSrc[ -iSrcStride ]; |
---|
936 | for ( Int x = 0; x < iWidth; x++ ) |
---|
937 | { |
---|
938 | iSum = xCTI_Filter_VP04_C_QUA0( piSrcTmp, iSrcStride ); |
---|
939 | piDst[x ] = xClipC ((iSum + 32) >> 6 ); |
---|
940 | piSrcTmp++; |
---|
941 | } |
---|
942 | piSrc += iSrcStride; |
---|
943 | piDst += iDstStride; |
---|
944 | } |
---|
945 | } |
---|
946 | break; |
---|
947 | case 6: |
---|
948 | { |
---|
949 | for ( Int y = iHeight; y != 0; y-- ) |
---|
950 | { |
---|
951 | piSrcTmp = &piSrc[ -iSrcStride ]; |
---|
952 | for ( Int x = 0; x < iWidth; x++ ) |
---|
953 | { |
---|
954 | iSum = xCTI_Filter_VP04_C_QUA1( piSrcTmp, iSrcStride ); |
---|
955 | piDst[x ] = xClipC ((iSum + 32) >> 6 ); |
---|
956 | piSrcTmp++; |
---|
957 | } |
---|
958 | piSrc += iSrcStride; |
---|
959 | piDst += iDstStride; |
---|
960 | } |
---|
961 | } |
---|
962 | break; |
---|
963 | case 3: |
---|
964 | { |
---|
965 | for ( Int y = iHeight; y != 0; y-- ) |
---|
966 | { |
---|
967 | piSrcTmp = &piSrc[ -iSrcStride ]; |
---|
968 | for ( Int x = 0; x < iWidth; x++ ) |
---|
969 | { |
---|
970 | iSum = xCTI_Filter_VP04_C_OCT1( piSrcTmp, iSrcStride ); |
---|
971 | piDst[x ] = xClipC ((iSum + 32) >> 6 ); |
---|
972 | piSrcTmp++; |
---|
973 | } |
---|
974 | piSrc += iSrcStride; |
---|
975 | piDst += iDstStride; |
---|
976 | } |
---|
977 | } |
---|
978 | break; |
---|
979 | case 5: |
---|
980 | { |
---|
981 | for ( Int y = iHeight; y != 0; y-- ) |
---|
982 | { |
---|
983 | piSrcTmp = &piSrc[ -iSrcStride ]; |
---|
984 | for ( Int x = 0; x < iWidth; x++ ) |
---|
985 | { |
---|
986 | iSum = xCTI_Filter_VP04_C_OCT2( piSrcTmp, iSrcStride ); |
---|
987 | piDst[x ] = xClipC ((iSum + 32) >> 6 ); |
---|
988 | piSrcTmp++; |
---|
989 | } |
---|
990 | piSrc += iSrcStride; |
---|
991 | piDst += iDstStride; |
---|
992 | } |
---|
993 | } |
---|
994 | break; |
---|
995 | case 7: |
---|
996 | { |
---|
997 | for ( Int y = iHeight; y != 0; y-- ) |
---|
998 | { |
---|
999 | piSrcTmp = &piSrc[ -iSrcStride ]; |
---|
1000 | for ( Int x = 0; x < iWidth; x++ ) |
---|
1001 | { |
---|
1002 | iSum = xCTI_Filter_VP04_C_OCT3( piSrcTmp, iSrcStride ); |
---|
1003 | piDst[x ] = xClipC ((iSum + 32) >> 6 ); |
---|
1004 | piSrcTmp++; |
---|
1005 | } |
---|
1006 | piSrc += iSrcStride; |
---|
1007 | piDst += iDstStride; |
---|
1008 | } |
---|
1009 | } |
---|
1010 | break; |
---|
1011 | case 4: |
---|
1012 | { |
---|
1013 | for ( Int y = iHeight; y != 0; y-- ) |
---|
1014 | { |
---|
1015 | piSrcTmp = &piSrc[-iSrcStride ]; |
---|
1016 | for ( Int x = 0; x < iWidth; x++ ) |
---|
1017 | { |
---|
1018 | iSum = xCTI_Filter_VPS04_C_HAL( piSrcTmp, iSrcStride ); |
---|
1019 | piDst[x ] = xClipC ((iSum + 32) >> 6 ); |
---|
1020 | piSrcTmp++; |
---|
1021 | } |
---|
1022 | piSrc += iSrcStride; |
---|
1023 | piDst += iDstStride; |
---|
1024 | } |
---|
1025 | } |
---|
1026 | break; |
---|
1027 | default: |
---|
1028 | assert( 0 ); |
---|
1029 | } |
---|
1030 | return; |
---|
1031 | } |
---|
1032 | |
---|
1033 | template<UInt bitDepth> |
---|
1034 | __inline Void TRenInterpFilter<bitDepth>::xCTI_Filter1DHorC(Pel* piSrc, Int iSrcStride, Int iWidth, Int iHeight, Int iDstStride, Pel*& rpiDst, Int iMV) |
---|
1035 | { |
---|
1036 | Pel* piDst = rpiDst; |
---|
1037 | Int iSum; |
---|
1038 | Pel* piSrcTmp; |
---|
1039 | |
---|
1040 | switch (iMV) |
---|
1041 | { |
---|
1042 | case 1: |
---|
1043 | { |
---|
1044 | for ( Int y = iHeight; y != 0; y-- ) |
---|
1045 | { |
---|
1046 | piSrcTmp = &piSrc[ -1 ]; |
---|
1047 | for ( Int x = 0; x < iWidth; x++ ) |
---|
1048 | { |
---|
1049 | iSum = xCTI_Filter_VP04_C_OCT0( piSrcTmp, 1 ); |
---|
1050 | piDst[x ] = xClipC ((iSum + 32) >> 6 ); |
---|
1051 | piSrcTmp++; |
---|
1052 | } |
---|
1053 | piSrc += iSrcStride; |
---|
1054 | piDst += iDstStride; |
---|
1055 | } |
---|
1056 | } |
---|
1057 | break; |
---|
1058 | case 2: |
---|
1059 | { |
---|
1060 | for ( Int y = iHeight; y != 0; y-- ) |
---|
1061 | { |
---|
1062 | piSrcTmp = &piSrc[ -1 ]; |
---|
1063 | for ( Int x = 0; x < iWidth; x++ ) |
---|
1064 | { |
---|
1065 | iSum = xCTI_Filter_VP04_C_QUA0( piSrcTmp, 1 ); |
---|
1066 | piDst[x ] = xClipC ((iSum + 32) >> 6 ); |
---|
1067 | piSrcTmp++; |
---|
1068 | } |
---|
1069 | piSrc += iSrcStride; |
---|
1070 | piDst += iDstStride; |
---|
1071 | } |
---|
1072 | } |
---|
1073 | break; |
---|
1074 | case 6: |
---|
1075 | { |
---|
1076 | for ( Int y = iHeight; y != 0; y-- ) |
---|
1077 | { |
---|
1078 | piSrcTmp = &piSrc[ -1 ]; |
---|
1079 | for ( Int x = 0; x < iWidth; x++ ) |
---|
1080 | { |
---|
1081 | iSum = xCTI_Filter_VP04_C_QUA1( piSrcTmp, 1 ); |
---|
1082 | piDst[x ] = xClipC ((iSum + 32) >> 6 ); |
---|
1083 | piSrcTmp++; |
---|
1084 | } |
---|
1085 | piSrc += iSrcStride; |
---|
1086 | piDst += iDstStride; |
---|
1087 | } |
---|
1088 | } |
---|
1089 | break; |
---|
1090 | case 3: |
---|
1091 | { |
---|
1092 | for ( Int y = iHeight; y != 0; y-- ) |
---|
1093 | { |
---|
1094 | piSrcTmp = &piSrc[ -1 ]; |
---|
1095 | for ( Int x = 0; x < iWidth; x++ ) |
---|
1096 | { |
---|
1097 | iSum = xCTI_Filter_VP04_C_OCT1( piSrcTmp, 1 ); |
---|
1098 | piDst[x ] = xClipC ((iSum + 32) >> 6 ); |
---|
1099 | piSrcTmp++; |
---|
1100 | } |
---|
1101 | piSrc += iSrcStride; |
---|
1102 | piDst += iDstStride; |
---|
1103 | } |
---|
1104 | } |
---|
1105 | break; |
---|
1106 | case 5: |
---|
1107 | { |
---|
1108 | for ( Int y = iHeight; y != 0; y-- ) |
---|
1109 | { |
---|
1110 | piSrcTmp = &piSrc[ -1 ]; |
---|
1111 | for ( Int x = 0; x < iWidth; x++ ) |
---|
1112 | { |
---|
1113 | iSum = xCTI_Filter_VP04_C_OCT2( piSrcTmp, 1 ); |
---|
1114 | piDst[x ] = xClipC ((iSum + 32) >> 6 ); |
---|
1115 | piSrcTmp++; |
---|
1116 | } |
---|
1117 | piSrc += iSrcStride; |
---|
1118 | piDst += iDstStride; |
---|
1119 | } |
---|
1120 | } |
---|
1121 | break; |
---|
1122 | case 7: |
---|
1123 | { |
---|
1124 | for ( Int y = iHeight; y != 0; y-- ) |
---|
1125 | { |
---|
1126 | piSrcTmp = &piSrc[ -1 ]; |
---|
1127 | for ( Int x = 0; x < iWidth; x++ ) |
---|
1128 | { |
---|
1129 | iSum = xCTI_Filter_VP04_C_OCT3( piSrcTmp, 1 ); |
---|
1130 | piDst[x ] = xClipC ((iSum + 32) >> 6 ); |
---|
1131 | piSrcTmp++; |
---|
1132 | } |
---|
1133 | piSrc += iSrcStride; |
---|
1134 | piDst += iDstStride; |
---|
1135 | } |
---|
1136 | } |
---|
1137 | break; |
---|
1138 | case 4: |
---|
1139 | { |
---|
1140 | for ( Int y = iHeight; y != 0; y-- ) |
---|
1141 | { |
---|
1142 | piSrcTmp = &piSrc[ -1 ]; |
---|
1143 | for ( Int x = 0; x < iWidth; x++ ) |
---|
1144 | { |
---|
1145 | iSum = xCTI_Filter_VPS04_C_HAL( piSrcTmp, 1 ); |
---|
1146 | piDst[x ] = xClipC ((iSum + 32) >> 6 ); |
---|
1147 | piSrcTmp++; |
---|
1148 | } |
---|
1149 | piSrc += iSrcStride; |
---|
1150 | piDst += iDstStride; |
---|
1151 | } |
---|
1152 | } |
---|
1153 | break; |
---|
1154 | default: |
---|
1155 | assert( 0 ); |
---|
1156 | } |
---|
1157 | return; |
---|
1158 | } |
---|
1159 | |
---|
1160 | template<UInt bitDepth> |
---|
1161 | __inline Int TRenInterpFilter<bitDepth>::xCTI_Filter_VP04_C_OCT0( Pel* pSrc, Int iStride ) |
---|
1162 | {// { -3, 60, 8, -1,} // 1/8 |
---|
1163 | Int iSum, iIdx = 0; |
---|
1164 | |
---|
1165 | Int p0 = pSrc[0]; iIdx+= iStride; |
---|
1166 | Int p1 = pSrc[iIdx]; iIdx+= iStride; |
---|
1167 | Int p2 = pSrc[iIdx]; iIdx+= iStride; |
---|
1168 | Int p3 = pSrc[iIdx]; |
---|
1169 | iSum = (p1<<6) -((p1+p0)<<2) +p0 +(p2<<3) -p3; |
---|
1170 | |
---|
1171 | return iSum; |
---|
1172 | } |
---|
1173 | template<UInt bitDepth> |
---|
1174 | __inline Int TRenInterpFilter<bitDepth>::xCTI_Filter_VI04_C_OCT0( Int* pSrc, Int iStride ) |
---|
1175 | { // { -3, 60, 8, -1,} //1/8 |
---|
1176 | Int iSum, iIdx = 0; |
---|
1177 | |
---|
1178 | Int p0 = pSrc[0]; iIdx+= iStride; |
---|
1179 | Int p1 = pSrc[iIdx]; iIdx+= iStride; |
---|
1180 | Int p2 = pSrc[iIdx]; iIdx+= iStride; |
---|
1181 | Int p3 = pSrc[iIdx]; |
---|
1182 | iSum = (p1<<6) -((p1+p0)<<2) +p0 +(p2<<3) -p3; |
---|
1183 | |
---|
1184 | return iSum; |
---|
1185 | } |
---|
1186 | template<UInt bitDepth> |
---|
1187 | __inline Int TRenInterpFilter<bitDepth>::xCTI_Filter_VP04_C_QUA0( Pel* pSrc, Int iStride ) |
---|
1188 | {// { -4, 54, 16, -2,} // 1/4 |
---|
1189 | Int iSum, iIdx = 0; |
---|
1190 | |
---|
1191 | Int p0 = pSrc[0]; iIdx+= iStride; |
---|
1192 | Int p1 = pSrc[iIdx]; iIdx+= iStride; |
---|
1193 | Int p2 = pSrc[iIdx]; iIdx+= iStride; |
---|
1194 | Int p3 = pSrc[iIdx]; |
---|
1195 | iSum = (p1 << 6) + (p2 << 4) - (p1 << 3) - ( p0 << 2) - ((p1 + p3) << 1); |
---|
1196 | |
---|
1197 | return iSum; |
---|
1198 | } |
---|
1199 | |
---|
1200 | template<UInt bitDepth> |
---|
1201 | __inline Int TRenInterpFilter<bitDepth>::xCTI_Filter_VI04_C_QUA0( Int* pSrc, Int iStride ) |
---|
1202 | { // { -4, 54, 16, -2,} //1/4 |
---|
1203 | Int iSum, iIdx = 0; |
---|
1204 | |
---|
1205 | Int p0 = pSrc[0]; iIdx+= iStride; |
---|
1206 | Int p1 = pSrc[iIdx]; iIdx+= iStride; |
---|
1207 | Int p2 = pSrc[iIdx]; iIdx+= iStride; |
---|
1208 | Int p3 = pSrc[iIdx]; |
---|
1209 | iSum = (p1 << 6) + (p2 << 4) - (p1 << 3) - ( p0 << 2) - ((p1 + p3) << 1); |
---|
1210 | |
---|
1211 | return iSum; |
---|
1212 | } |
---|
1213 | template<UInt bitDepth> |
---|
1214 | __inline Int TRenInterpFilter<bitDepth>::xCTI_Filter_VP04_C_QUA1( Pel* pSrc, Int iStride ) |
---|
1215 | {// { -2, 16, 54, -4,}// 3/4 |
---|
1216 | Int iSum, iIdx = 0; |
---|
1217 | |
---|
1218 | Int p0 = pSrc[0]; iIdx+= iStride; |
---|
1219 | Int p1 = pSrc[iIdx]; iIdx+= iStride; |
---|
1220 | Int p2 = pSrc[iIdx]; iIdx+= iStride; |
---|
1221 | Int p3 = pSrc[iIdx]; |
---|
1222 | iSum = (p2 << 6) + (p1 << 4) - (p2 << 3) - ( p3 << 2) - ((p2 + p0) << 1); |
---|
1223 | |
---|
1224 | return iSum; |
---|
1225 | } |
---|
1226 | |
---|
1227 | template<UInt bitDepth> |
---|
1228 | __inline Int TRenInterpFilter<bitDepth>::xCTI_Filter_VI04_C_QUA1( Int* pSrc, Int iStride ) |
---|
1229 | {// { -2, 16, 54, -4,}// 3/4 |
---|
1230 | Int iSum, iIdx = 0; |
---|
1231 | |
---|
1232 | Int p0 = pSrc[0]; iIdx+= iStride; |
---|
1233 | Int p1 = pSrc[iIdx]; iIdx+= iStride; |
---|
1234 | Int p2 = pSrc[iIdx]; iIdx+= iStride; |
---|
1235 | Int p3 = pSrc[iIdx]; |
---|
1236 | iSum = (p2 << 6) + (p1 << 4) - (p2 << 3) - ( p3 << 2) - ((p2 + p0) << 1); |
---|
1237 | |
---|
1238 | return iSum; |
---|
1239 | } |
---|
1240 | |
---|
1241 | template<UInt bitDepth> |
---|
1242 | __inline Int TRenInterpFilter<bitDepth>::xCTI_Filter_VP04_C_OCT1( Pel* pSrc, Int iStride ) |
---|
1243 | {// { -5, 46, 27, -4,} // 3/8 |
---|
1244 | Int iSum, iIdx = 0; |
---|
1245 | |
---|
1246 | Int p0 = pSrc[0]; iIdx+= iStride; |
---|
1247 | Int p1 = pSrc[iIdx]; iIdx+= iStride; |
---|
1248 | Int p2 = pSrc[iIdx]; iIdx+= iStride; |
---|
1249 | Int p3 = pSrc[iIdx]; |
---|
1250 | Int t = p0 + p2; |
---|
1251 | iSum = ((p1 + p2) << 5) + (p1 << 4) - ( (t + p3) << 2) - ( p1 << 1) - t; |
---|
1252 | |
---|
1253 | return iSum; |
---|
1254 | } |
---|
1255 | |
---|
1256 | template<UInt bitDepth> |
---|
1257 | __inline Int TRenInterpFilter<bitDepth>::xCTI_Filter_VI04_C_OCT1( Int* pSrc, Int iStride ) |
---|
1258 | { // { -5, 46, 27, -4,} //3/8 |
---|
1259 | Int iSum, iIdx = 0; |
---|
1260 | |
---|
1261 | Int p0 = pSrc[0]; iIdx+= iStride; |
---|
1262 | Int p1 = pSrc[iIdx]; iIdx+= iStride; |
---|
1263 | Int p2 = pSrc[iIdx]; iIdx+= iStride; |
---|
1264 | Int p3 = pSrc[iIdx]; |
---|
1265 | Int t = p0 + p2; |
---|
1266 | iSum = ((p1 + p2) << 5) + (p1 << 4) - ( (t + p3) << 2) - ( p1 << 1) - t; |
---|
1267 | |
---|
1268 | return iSum; |
---|
1269 | } |
---|
1270 | |
---|
1271 | template<UInt bitDepth> |
---|
1272 | __inline Int TRenInterpFilter<bitDepth>::xCTI_Filter_VPS04_C_HAL( Pel* pSrc, Int iStride ) |
---|
1273 | { |
---|
1274 | // { -4, 36, 36, -4,}, // 1/2 |
---|
1275 | Int iSum; |
---|
1276 | Int iTemp0 = pSrc[iStride*1]+pSrc[iStride*2]; |
---|
1277 | Int iTemp1 = pSrc[ 0]+pSrc[iStride*3]; |
---|
1278 | |
---|
1279 | iSum = ((iTemp0<<3) + iTemp0 -iTemp1)<<2; |
---|
1280 | |
---|
1281 | return iSum; |
---|
1282 | } |
---|
1283 | |
---|
1284 | template<UInt bitDepth> |
---|
1285 | __inline Int TRenInterpFilter<bitDepth>::xCTI_Filter_VIS04_C_HAL( Int* pSrc, Int iStride ) |
---|
1286 | { |
---|
1287 | // { -4, 36, 36, -4,}, //1/2 |
---|
1288 | Int iSum; |
---|
1289 | Int iTemp0 = pSrc[iStride*1]+pSrc[iStride*2]; |
---|
1290 | Int iTemp1 = pSrc[ 0]+pSrc[iStride*3]; |
---|
1291 | |
---|
1292 | iSum = ((iTemp0<<3) + iTemp0 -iTemp1)<<2; |
---|
1293 | |
---|
1294 | return iSum; |
---|
1295 | } |
---|
1296 | |
---|
1297 | template<UInt bitDepth> |
---|
1298 | __inline Int TRenInterpFilter<bitDepth>::xCTI_Filter_VP04_C_OCT2( Pel* pSrc, Int iStride ) |
---|
1299 | {// { -4, 27, 46, -5,}, // 5/8 |
---|
1300 | Int iSum, iIdx = 0; |
---|
1301 | |
---|
1302 | Int p0 = pSrc[0]; iIdx+= iStride; |
---|
1303 | Int p1 = pSrc[iIdx]; iIdx+= iStride; |
---|
1304 | Int p2 = pSrc[iIdx]; iIdx+= iStride; |
---|
1305 | Int p3 = pSrc[iIdx]; |
---|
1306 | Int t = p1 + p3; |
---|
1307 | iSum = ((p1 + p2) << 5) + (p2 << 4) - ( (t + p0) << 2) - ( p2 << 1) - t; |
---|
1308 | |
---|
1309 | return iSum; |
---|
1310 | } |
---|
1311 | |
---|
1312 | template<UInt bitDepth> |
---|
1313 | __inline Int TRenInterpFilter<bitDepth>::xCTI_Filter_VI04_C_OCT2( Int* pSrc, Int iStride ) |
---|
1314 | { // { -4, 27, 46, -5,}, // 5/8 |
---|
1315 | Int iSum, iIdx = 0; |
---|
1316 | |
---|
1317 | Int p0 = pSrc[0]; iIdx+= iStride; |
---|
1318 | Int p1 = pSrc[iIdx]; iIdx+= iStride; |
---|
1319 | Int p2 = pSrc[iIdx]; iIdx+= iStride; |
---|
1320 | Int p3 = pSrc[iIdx]; |
---|
1321 | Int t = p1 + p3; |
---|
1322 | iSum = ((p1 + p2) << 5) + (p2 << 4) - ( (t + p0) << 2) - ( p2 << 1) - t; |
---|
1323 | |
---|
1324 | return iSum; |
---|
1325 | } |
---|
1326 | |
---|
1327 | template<UInt bitDepth> |
---|
1328 | __inline Int TRenInterpFilter<bitDepth>::xCTI_Filter_VP04_C_OCT3( Pel* pSrc, Int iStride ) |
---|
1329 | {// { -1, 8, 60, -3,} // 7/8 |
---|
1330 | Int iSum, iIdx = 0; |
---|
1331 | |
---|
1332 | Int p0 = pSrc[0]; iIdx+= iStride; |
---|
1333 | Int p1 = pSrc[iIdx]; iIdx+= iStride; |
---|
1334 | Int p2 = pSrc[iIdx]; iIdx+= iStride; |
---|
1335 | Int p3 = pSrc[iIdx]; |
---|
1336 | iSum = (p2<<6) -((p2+p3)<<2) +p3 +(p1<<3) -p0; |
---|
1337 | |
---|
1338 | return iSum; |
---|
1339 | } |
---|
1340 | |
---|
1341 | template<UInt bitDepth> |
---|
1342 | __inline Int TRenInterpFilter<bitDepth>::xCTI_Filter_VI04_C_OCT3( Int* pSrc, Int iStride ) |
---|
1343 | { // { -1, 8, 60, -3,} // 7/8 |
---|
1344 | Int iSum, iIdx = 0; |
---|
1345 | |
---|
1346 | Int p0 = pSrc[0]; iIdx+= iStride; |
---|
1347 | Int p1 = pSrc[iIdx]; iIdx+= iStride; |
---|
1348 | Int p2 = pSrc[iIdx]; iIdx+= iStride; |
---|
1349 | Int p3 = pSrc[iIdx]; |
---|
1350 | iSum = (p2<<6) -((p2+p3)<<2) +p3 +(p1<<3) -p0; |
---|
1351 | |
---|
1352 | return iSum; |
---|
1353 | } |
---|
1354 | |
---|
1355 | #endif // NH_3D |
---|
1356 | #endif // __TRENINTERP__ |
---|