1/*
2 * RV30 decoder motion compensation functions
3 * Copyright (c) 2007 Konstantin Shishkov
4 *
5 * This file is part of FFmpeg.
6 *
7 * FFmpeg is free software; you can redistribute it and/or
8 * modify it under the terms of the GNU Lesser General Public
9 * License as published by the Free Software Foundation; either
10 * version 2.1 of the License, or (at your option) any later version.
11 *
12 * FFmpeg is distributed in the hope that it will be useful,
13 * but WITHOUT ANY WARRANTY; without even the implied warranty of
14 * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the GNU
15 * Lesser General Public License for more details.
16 *
17 * You should have received a copy of the GNU Lesser General Public
18 * License along with FFmpeg; if not, write to the Free Software
19 * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA
20 */
21
22/**
23 * @file
24 * RV30 decoder motion compensation functions
25 */
26
27#include "avcodec.h"
28#include "h264chroma.h"
29#include "h264qpel.h"
30#include "mathops.h"
31#include "rv34dsp.h"
32
33#define RV30_LOWPASS(OPNAME, OP) \
34static av_unused void OPNAME ## rv30_tpel8_h_lowpass(uint8_t *dst, uint8_t *src, int dstStride, int srcStride, const int C1, const int C2){\
35    const int h = 8;\
36    const uint8_t *cm = ff_crop_tab + MAX_NEG_CROP;\
37    int i;\
38    for(i = 0; i < h; i++)\
39    {\
40        OP(dst[0], (-(src[-1]+src[2]) + src[0]*C1 + src[1]*C2 + 8)>>4);\
41        OP(dst[1], (-(src[ 0]+src[3]) + src[1]*C1 + src[2]*C2 + 8)>>4);\
42        OP(dst[2], (-(src[ 1]+src[4]) + src[2]*C1 + src[3]*C2 + 8)>>4);\
43        OP(dst[3], (-(src[ 2]+src[5]) + src[3]*C1 + src[4]*C2 + 8)>>4);\
44        OP(dst[4], (-(src[ 3]+src[6]) + src[4]*C1 + src[5]*C2 + 8)>>4);\
45        OP(dst[5], (-(src[ 4]+src[7]) + src[5]*C1 + src[6]*C2 + 8)>>4);\
46        OP(dst[6], (-(src[ 5]+src[8]) + src[6]*C1 + src[7]*C2 + 8)>>4);\
47        OP(dst[7], (-(src[ 6]+src[9]) + src[7]*C1 + src[8]*C2 + 8)>>4);\
48        dst += dstStride;\
49        src += srcStride;\
50    }\
51}\
52\
53static void OPNAME ## rv30_tpel8_v_lowpass(uint8_t *dst, uint8_t *src, int dstStride, int srcStride, const int C1, const int C2){\
54    const int w = 8;\
55    const uint8_t *cm = ff_crop_tab + MAX_NEG_CROP;\
56    int i;\
57    for(i = 0; i < w; i++)\
58    {\
59        const int srcA = src[-1*srcStride];\
60        const int src0 = src[0 *srcStride];\
61        const int src1 = src[1 *srcStride];\
62        const int src2 = src[2 *srcStride];\
63        const int src3 = src[3 *srcStride];\
64        const int src4 = src[4 *srcStride];\
65        const int src5 = src[5 *srcStride];\
66        const int src6 = src[6 *srcStride];\
67        const int src7 = src[7 *srcStride];\
68        const int src8 = src[8 *srcStride];\
69        const int src9 = src[9 *srcStride];\
70        OP(dst[0*dstStride], (-(srcA+src2) + src0*C1 + src1*C2 + 8)>>4);\
71        OP(dst[1*dstStride], (-(src0+src3) + src1*C1 + src2*C2 + 8)>>4);\
72        OP(dst[2*dstStride], (-(src1+src4) + src2*C1 + src3*C2 + 8)>>4);\
73        OP(dst[3*dstStride], (-(src2+src5) + src3*C1 + src4*C2 + 8)>>4);\
74        OP(dst[4*dstStride], (-(src3+src6) + src4*C1 + src5*C2 + 8)>>4);\
75        OP(dst[5*dstStride], (-(src4+src7) + src5*C1 + src6*C2 + 8)>>4);\
76        OP(dst[6*dstStride], (-(src5+src8) + src6*C1 + src7*C2 + 8)>>4);\
77        OP(dst[7*dstStride], (-(src6+src9) + src7*C1 + src8*C2 + 8)>>4);\
78        dst++;\
79        src++;\
80    }\
81}\
82\
83static void OPNAME ## rv30_tpel8_hv_lowpass(uint8_t *dst, uint8_t *src, int dstStride, int srcStride){\
84    const int w = 8;\
85    const int h = 8;\
86    const uint8_t *cm = ff_crop_tab + MAX_NEG_CROP;\
87    int i, j;\
88    for(j = 0; j < h; j++){\
89        for(i = 0; i < w; i++){\
90            OP(dst[i], (\
91                  src[srcStride*-1+i-1]  -12*src[srcStride*-1+i]  -6*src[srcStride*-1+i+1]    +src[srcStride*-1+i+2]+\
92              -12*src[srcStride* 0+i-1] +144*src[srcStride* 0+i] +72*src[srcStride* 0+i+1] -12*src[srcStride* 0+i+2] +\
93               -6*src[srcStride* 1+i-1]  +72*src[srcStride* 1+i] +36*src[srcStride* 1+i+1]  -6*src[srcStride* 1+i+2] +\
94                  src[srcStride* 2+i-1]  -12*src[srcStride* 2+i]  -6*src[srcStride* 2+i+1]    +src[srcStride* 2+i+2] +\
95                  128)>>8);\
96        }\
97        src += srcStride;\
98        dst += dstStride;\
99    }\
100}\
101\
102static void OPNAME ## rv30_tpel8_hhv_lowpass(uint8_t *dst, uint8_t *src, int dstStride, int srcStride){\
103    const int w = 8;\
104    const int h = 8;\
105    const uint8_t *cm = ff_crop_tab + MAX_NEG_CROP;\
106    int i, j;\
107    for(j = 0; j < h; j++){\
108        for(i = 0; i < w; i++){\
109            OP(dst[i], (\
110                  src[srcStride*-1+i-1]  -12*src[srcStride*-1+i+1]  -6*src[srcStride*-1+i]    +src[srcStride*-1+i+2]+\
111              -12*src[srcStride* 0+i-1] +144*src[srcStride* 0+i+1] +72*src[srcStride* 0+i] -12*src[srcStride* 0+i+2]+\
112               -6*src[srcStride* 1+i-1]  +72*src[srcStride* 1+i+1] +36*src[srcStride* 1+i]  -6*src[srcStride* 1+i+2]+\
113                  src[srcStride* 2+i-1]  -12*src[srcStride* 2+i+1]  -6*src[srcStride* 2+i]    +src[srcStride* 2+i+2]+\
114                  128)>>8);\
115        }\
116        src += srcStride;\
117        dst += dstStride;\
118    }\
119}\
120\
121static void OPNAME ## rv30_tpel8_hvv_lowpass(uint8_t *dst, uint8_t *src, int dstStride, int srcStride){\
122    const int w = 8;\
123    const int h = 8;\
124    const uint8_t *cm = ff_crop_tab + MAX_NEG_CROP;\
125    int i, j;\
126    for(j = 0; j < h; j++){\
127        for(i = 0; i < w; i++){\
128            OP(dst[i], (\
129                  src[srcStride*-1+i-1]  -12*src[srcStride*-1+i]  -6*src[srcStride*-1+i+1]    +src[srcStride*-1+i+2]+\
130               -6*src[srcStride* 0+i-1]  +72*src[srcStride* 0+i] +36*src[srcStride* 0+i+1]  -6*src[srcStride* 0+i+2]+\
131              -12*src[srcStride* 1+i-1] +144*src[srcStride* 1+i] +72*src[srcStride* 1+i+1] -12*src[srcStride* 1+i+2]+\
132                  src[srcStride* 2+i-1]  -12*src[srcStride* 2+i]  -6*src[srcStride* 2+i+1]    +src[srcStride* 2+i+2]+\
133                  128)>>8);\
134        }\
135        src += srcStride;\
136        dst += dstStride;\
137    }\
138}\
139\
140static void OPNAME ## rv30_tpel8_hhvv_lowpass(uint8_t *dst, uint8_t *src, int dstStride, int srcStride){\
141    const int w = 8;\
142    const int h = 8;\
143    const uint8_t *cm = ff_crop_tab + MAX_NEG_CROP;\
144    int i, j;\
145    for(j = 0; j < h; j++){\
146        for(i = 0; i < w; i++){\
147            OP(dst[i], (\
148               36*src[i+srcStride*0] +54*src[i+1+srcStride*0] +6*src[i+2+srcStride*0]+\
149               54*src[i+srcStride*1] +81*src[i+1+srcStride*1] +9*src[i+2+srcStride*1]+\
150                6*src[i+srcStride*2] + 9*src[i+1+srcStride*2] +  src[i+2+srcStride*2]+\
151               128)>>8);\
152        }\
153        src += srcStride;\
154        dst += dstStride;\
155    }\
156}\
157\
158static void OPNAME ## rv30_tpel16_v_lowpass(uint8_t *dst, uint8_t *src, int dstStride, int srcStride, const int C1, const int C2){\
159    OPNAME ## rv30_tpel8_v_lowpass(dst  , src  , dstStride, srcStride, C1, C2);\
160    OPNAME ## rv30_tpel8_v_lowpass(dst+8, src+8, dstStride, srcStride, C1, C2);\
161    src += 8*srcStride;\
162    dst += 8*dstStride;\
163    OPNAME ## rv30_tpel8_v_lowpass(dst  , src  , dstStride, srcStride, C1, C2);\
164    OPNAME ## rv30_tpel8_v_lowpass(dst+8, src+8, dstStride, srcStride, C1, C2);\
165}\
166\
167static void OPNAME ## rv30_tpel16_h_lowpass(uint8_t *dst, uint8_t *src, int dstStride, int srcStride, const int C1, const int C2){\
168    OPNAME ## rv30_tpel8_h_lowpass(dst  , src  , dstStride, srcStride, C1, C2);\
169    OPNAME ## rv30_tpel8_h_lowpass(dst+8, src+8, dstStride, srcStride, C1, C2);\
170    src += 8*srcStride;\
171    dst += 8*dstStride;\
172    OPNAME ## rv30_tpel8_h_lowpass(dst  , src  , dstStride, srcStride, C1, C2);\
173    OPNAME ## rv30_tpel8_h_lowpass(dst+8, src+8, dstStride, srcStride, C1, C2);\
174}\
175\
176static void OPNAME ## rv30_tpel16_hv_lowpass(uint8_t *dst, uint8_t *src, int dstStride, int srcStride){\
177    OPNAME ## rv30_tpel8_hv_lowpass(dst  , src  , dstStride, srcStride);\
178    OPNAME ## rv30_tpel8_hv_lowpass(dst+8, src+8, dstStride, srcStride);\
179    src += 8*srcStride;\
180    dst += 8*dstStride;\
181    OPNAME ## rv30_tpel8_hv_lowpass(dst  , src  , dstStride, srcStride);\
182    OPNAME ## rv30_tpel8_hv_lowpass(dst+8, src+8, dstStride, srcStride);\
183}\
184\
185static void OPNAME ## rv30_tpel16_hhv_lowpass(uint8_t *dst, uint8_t *src, int dstStride, int srcStride){\
186    OPNAME ## rv30_tpel8_hhv_lowpass(dst  , src  , dstStride, srcStride);\
187    OPNAME ## rv30_tpel8_hhv_lowpass(dst+8, src+8, dstStride, srcStride);\
188    src += 8*srcStride;\
189    dst += 8*dstStride;\
190    OPNAME ## rv30_tpel8_hhv_lowpass(dst  , src  , dstStride, srcStride);\
191    OPNAME ## rv30_tpel8_hhv_lowpass(dst+8, src+8, dstStride, srcStride);\
192}\
193\
194static void OPNAME ## rv30_tpel16_hvv_lowpass(uint8_t *dst, uint8_t *src, int dstStride, int srcStride){\
195    OPNAME ## rv30_tpel8_hvv_lowpass(dst  , src  , dstStride, srcStride);\
196    OPNAME ## rv30_tpel8_hvv_lowpass(dst+8, src+8, dstStride, srcStride);\
197    src += 8*srcStride;\
198    dst += 8*dstStride;\
199    OPNAME ## rv30_tpel8_hvv_lowpass(dst  , src  , dstStride, srcStride);\
200    OPNAME ## rv30_tpel8_hvv_lowpass(dst+8, src+8, dstStride, srcStride);\
201}\
202\
203static void OPNAME ## rv30_tpel16_hhvv_lowpass(uint8_t *dst, uint8_t *src, int dstStride, int srcStride){\
204    OPNAME ## rv30_tpel8_hhvv_lowpass(dst  , src  , dstStride, srcStride);\
205    OPNAME ## rv30_tpel8_hhvv_lowpass(dst+8, src+8, dstStride, srcStride);\
206    src += 8*srcStride;\
207    dst += 8*dstStride;\
208    OPNAME ## rv30_tpel8_hhvv_lowpass(dst  , src  , dstStride, srcStride);\
209    OPNAME ## rv30_tpel8_hhvv_lowpass(dst+8, src+8, dstStride, srcStride);\
210}\
211\
212
213#define RV30_MC(OPNAME, SIZE) \
214static void OPNAME ## rv30_tpel ## SIZE ## _mc10_c(uint8_t *dst, uint8_t *src, ptrdiff_t stride)\
215{\
216    OPNAME ## rv30_tpel ## SIZE ## _h_lowpass(dst, src, stride, stride, 12, 6);\
217}\
218\
219static void OPNAME ## rv30_tpel ## SIZE ## _mc20_c(uint8_t *dst, uint8_t *src, ptrdiff_t stride)\
220{\
221    OPNAME ## rv30_tpel ## SIZE ## _h_lowpass(dst, src, stride, stride, 6, 12);\
222}\
223\
224static void OPNAME ## rv30_tpel ## SIZE ## _mc01_c(uint8_t *dst, uint8_t *src, ptrdiff_t stride)\
225{\
226    OPNAME ## rv30_tpel ## SIZE ## _v_lowpass(dst, src, stride, stride, 12, 6);\
227}\
228\
229static void OPNAME ## rv30_tpel ## SIZE ## _mc02_c(uint8_t *dst, uint8_t *src, ptrdiff_t stride)\
230{\
231    OPNAME ## rv30_tpel ## SIZE ## _v_lowpass(dst, src, stride, stride, 6, 12);\
232}\
233\
234static void OPNAME ## rv30_tpel ## SIZE ## _mc11_c(uint8_t *dst, uint8_t *src, ptrdiff_t stride)\
235{\
236    OPNAME ## rv30_tpel ## SIZE ## _hv_lowpass(dst, src, stride, stride);\
237}\
238\
239static void OPNAME ## rv30_tpel ## SIZE ## _mc12_c(uint8_t *dst, uint8_t *src, ptrdiff_t stride)\
240{\
241    OPNAME ## rv30_tpel ## SIZE ## _hvv_lowpass(dst, src, stride, stride);\
242}\
243\
244static void OPNAME ## rv30_tpel ## SIZE ## _mc21_c(uint8_t *dst, uint8_t *src, ptrdiff_t stride)\
245{\
246    OPNAME ## rv30_tpel ## SIZE ## _hhv_lowpass(dst, src, stride, stride);\
247}\
248\
249static void OPNAME ## rv30_tpel ## SIZE ## _mc22_c(uint8_t *dst, uint8_t *src, ptrdiff_t stride)\
250{\
251    OPNAME ## rv30_tpel ## SIZE ## _hhvv_lowpass(dst, src, stride, stride);\
252}\
253\
254
255#define op_avg(a, b)  a = (((a)+cm[b]+1)>>1)
256#define op_put(a, b)  a = cm[b]
257
258RV30_LOWPASS(put_       , op_put)
259RV30_LOWPASS(avg_       , op_avg)
260RV30_MC(put_, 8)
261RV30_MC(put_, 16)
262RV30_MC(avg_, 8)
263RV30_MC(avg_, 16)
264
265av_cold void ff_rv30dsp_init(RV34DSPContext *c)
266{
267    H264ChromaContext h264chroma;
268    H264QpelContext qpel;
269
270    ff_rv34dsp_init(c);
271    ff_h264chroma_init(&h264chroma, 8);
272    ff_h264qpel_init(&qpel, 8);
273
274    c->put_pixels_tab[0][ 0] = qpel.put_h264_qpel_pixels_tab[0][0];
275    c->put_pixels_tab[0][ 1] = put_rv30_tpel16_mc10_c;
276    c->put_pixels_tab[0][ 2] = put_rv30_tpel16_mc20_c;
277    c->put_pixels_tab[0][ 4] = put_rv30_tpel16_mc01_c;
278    c->put_pixels_tab[0][ 5] = put_rv30_tpel16_mc11_c;
279    c->put_pixels_tab[0][ 6] = put_rv30_tpel16_mc21_c;
280    c->put_pixels_tab[0][ 8] = put_rv30_tpel16_mc02_c;
281    c->put_pixels_tab[0][ 9] = put_rv30_tpel16_mc12_c;
282    c->put_pixels_tab[0][10] = put_rv30_tpel16_mc22_c;
283    c->avg_pixels_tab[0][ 0] = qpel.avg_h264_qpel_pixels_tab[0][0];
284    c->avg_pixels_tab[0][ 1] = avg_rv30_tpel16_mc10_c;
285    c->avg_pixels_tab[0][ 2] = avg_rv30_tpel16_mc20_c;
286    c->avg_pixels_tab[0][ 4] = avg_rv30_tpel16_mc01_c;
287    c->avg_pixels_tab[0][ 5] = avg_rv30_tpel16_mc11_c;
288    c->avg_pixels_tab[0][ 6] = avg_rv30_tpel16_mc21_c;
289    c->avg_pixels_tab[0][ 8] = avg_rv30_tpel16_mc02_c;
290    c->avg_pixels_tab[0][ 9] = avg_rv30_tpel16_mc12_c;
291    c->avg_pixels_tab[0][10] = avg_rv30_tpel16_mc22_c;
292    c->put_pixels_tab[1][ 0] = qpel.put_h264_qpel_pixels_tab[1][0];
293    c->put_pixels_tab[1][ 1] = put_rv30_tpel8_mc10_c;
294    c->put_pixels_tab[1][ 2] = put_rv30_tpel8_mc20_c;
295    c->put_pixels_tab[1][ 4] = put_rv30_tpel8_mc01_c;
296    c->put_pixels_tab[1][ 5] = put_rv30_tpel8_mc11_c;
297    c->put_pixels_tab[1][ 6] = put_rv30_tpel8_mc21_c;
298    c->put_pixels_tab[1][ 8] = put_rv30_tpel8_mc02_c;
299    c->put_pixels_tab[1][ 9] = put_rv30_tpel8_mc12_c;
300    c->put_pixels_tab[1][10] = put_rv30_tpel8_mc22_c;
301    c->avg_pixels_tab[1][ 0] = qpel.avg_h264_qpel_pixels_tab[1][0];
302    c->avg_pixels_tab[1][ 1] = avg_rv30_tpel8_mc10_c;
303    c->avg_pixels_tab[1][ 2] = avg_rv30_tpel8_mc20_c;
304    c->avg_pixels_tab[1][ 4] = avg_rv30_tpel8_mc01_c;
305    c->avg_pixels_tab[1][ 5] = avg_rv30_tpel8_mc11_c;
306    c->avg_pixels_tab[1][ 6] = avg_rv30_tpel8_mc21_c;
307    c->avg_pixels_tab[1][ 8] = avg_rv30_tpel8_mc02_c;
308    c->avg_pixels_tab[1][ 9] = avg_rv30_tpel8_mc12_c;
309    c->avg_pixels_tab[1][10] = avg_rv30_tpel8_mc22_c;
310
311    c->put_chroma_pixels_tab[0] = h264chroma.put_h264_chroma_pixels_tab[0];
312    c->put_chroma_pixels_tab[1] = h264chroma.put_h264_chroma_pixels_tab[1];
313    c->avg_chroma_pixels_tab[0] = h264chroma.avg_h264_chroma_pixels_tab[0];
314    c->avg_chroma_pixels_tab[1] = h264chroma.avg_h264_chroma_pixels_tab[1];
315}
316