Merge branch 'mirror' into vdpau
[FFMpeg-mirror/ffmpeg-vdpau.git] / libavcodec / rv30dsp.c
blob13b218b881f7798d6e44f7088d6f60a1ee1f93d2
1 /*
2 * RV30 decoder motion compensation functions
3 * Copyright (c) 2007 Konstantin Shishkov
5 * This file is part of FFmpeg.
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.
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.
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
22 /**
23 * @file rv30dsp.c
24 * RV30 decoder motion compensation functions
27 #include "avcodec.h"
28 #include "dsputil.h"
30 #define RV30_LOWPASS(OPNAME, OP) \
31 static av_unused void OPNAME ## rv30_tpel8_h_lowpass(uint8_t *dst, uint8_t *src, int dstStride, int srcStride, const int C1, const int C2){\
32 const int h=8;\
33 uint8_t *cm = ff_cropTbl + MAX_NEG_CROP;\
34 int i;\
35 for(i=0; i<h; i++)\
37 OP(dst[0], -(src[-1]+src[2]) + src[0]*C1 + src[1]*C2);\
38 OP(dst[1], -(src[ 0]+src[3]) + src[1]*C1 + src[2]*C2);\
39 OP(dst[2], -(src[ 1]+src[4]) + src[2]*C1 + src[3]*C2);\
40 OP(dst[3], -(src[ 2]+src[5]) + src[3]*C1 + src[4]*C2);\
41 OP(dst[4], -(src[ 3]+src[6]) + src[4]*C1 + src[5]*C2);\
42 OP(dst[5], -(src[ 4]+src[7]) + src[5]*C1 + src[6]*C2);\
43 OP(dst[6], -(src[ 5]+src[8]) + src[6]*C1 + src[7]*C2);\
44 OP(dst[7], -(src[ 6]+src[9]) + src[7]*C1 + src[8]*C2);\
45 dst+=dstStride;\
46 src+=srcStride;\
50 static void OPNAME ## rv30_tpel8_v_lowpass(uint8_t *dst, uint8_t *src, int dstStride, int srcStride, const int C1, const int C2){\
51 const int w=8;\
52 uint8_t *cm = ff_cropTbl + MAX_NEG_CROP;\
53 int i;\
54 for(i=0; i<w; i++)\
56 const int srcA= src[-1*srcStride];\
57 const int src0= src[0 *srcStride];\
58 const int src1= src[1 *srcStride];\
59 const int src2= src[2 *srcStride];\
60 const int src3= src[3 *srcStride];\
61 const int src4= src[4 *srcStride];\
62 const int src5= src[5 *srcStride];\
63 const int src6= src[6 *srcStride];\
64 const int src7= src[7 *srcStride];\
65 const int src8= src[8 *srcStride];\
66 const int src9= src[9 *srcStride];\
67 OP(dst[0*dstStride], -(srcA+src2) + src0*C1 + src1*C2);\
68 OP(dst[1*dstStride], -(src0+src3) + src1*C1 + src2*C2);\
69 OP(dst[2*dstStride], -(src1+src4) + src2*C1 + src3*C2);\
70 OP(dst[3*dstStride], -(src2+src5) + src3*C1 + src4*C2);\
71 OP(dst[4*dstStride], -(src3+src6) + src4*C1 + src5*C2);\
72 OP(dst[5*dstStride], -(src4+src7) + src5*C1 + src6*C2);\
73 OP(dst[6*dstStride], -(src5+src8) + src6*C1 + src7*C2);\
74 OP(dst[7*dstStride], -(src6+src9) + src7*C1 + src8*C2);\
75 dst++;\
76 src++;\
80 static void OPNAME ## rv30_tpel8_h3_lowpass(uint8_t *dst, uint8_t *src, int dstStride, int srcStride){\
81 const int h=8+2;\
82 uint8_t *cm = ff_cropTbl + MAX_NEG_CROP;\
83 int i;\
84 for(i=0; i<h; i++)\
86 OP(dst[0], 6*src[0]+9*src[1]+src[2]);\
87 OP(dst[1], 6*src[1]+9*src[2]+src[3]);\
88 OP(dst[2], 6*src[2]+9*src[3]+src[4]);\
89 OP(dst[3], 6*src[3]+9*src[4]+src[5]);\
90 OP(dst[4], 6*src[4]+9*src[5]+src[6]);\
91 OP(dst[5], 6*src[5]+9*src[6]+src[7]);\
92 OP(dst[6], 6*src[6]+9*src[7]+src[8]);\
93 OP(dst[7], 6*src[7]+9*src[8]+src[9]);\
94 dst+=dstStride;\
95 src+=srcStride;\
99 static void OPNAME ## rv30_tpel8_v3_lowpass(uint8_t *dst, uint8_t *src, int dstStride, int srcStride){\
100 const int w=8;\
101 uint8_t *cm = ff_cropTbl + MAX_NEG_CROP;\
102 int i;\
103 for(i=0; i<w; i++)\
105 const int src0= src[0 *srcStride];\
106 const int src1= src[1 *srcStride];\
107 const int src2= src[2 *srcStride];\
108 const int src3= src[3 *srcStride];\
109 const int src4= src[4 *srcStride];\
110 const int src5= src[5 *srcStride];\
111 const int src6= src[6 *srcStride];\
112 const int src7= src[7 *srcStride];\
113 const int src8= src[8 *srcStride];\
114 const int src9= src[9 *srcStride];\
115 OP(dst[0*dstStride], 6*src0 + 9*src1 + src2);\
116 OP(dst[1*dstStride], 6*src1 + 9*src2 + src3);\
117 OP(dst[2*dstStride], 6*src2 + 9*src3 + src4);\
118 OP(dst[3*dstStride], 6*src3 + 9*src4 + src5);\
119 OP(dst[4*dstStride], 6*src4 + 9*src5 + src6);\
120 OP(dst[5*dstStride], 6*src5 + 9*src6 + src7);\
121 OP(dst[6*dstStride], 6*src6 + 9*src7 + src8);\
122 OP(dst[7*dstStride], 6*src7 + 9*src8 + src9);\
123 dst ++;\
124 src ++;\
128 static void OPNAME ## rv30_tpel8_hv_lowpass(uint8_t *dst, uint8_t *src, int dstStride, int srcStride){\
129 uint8_t half[8*10];\
130 put_rv30_tpel8_h3_lowpass(half, src, 8, srcStride);\
131 OPNAME ## rv30_tpel8_v3_lowpass(dst, half, dstStride, 8);\
134 static void OPNAME ## rv30_tpel16_v_lowpass(uint8_t *dst, uint8_t *src, int dstStride, int srcStride, const int C1, const int C2){\
135 OPNAME ## rv30_tpel8_v_lowpass(dst , src , dstStride, srcStride, C1, C2);\
136 OPNAME ## rv30_tpel8_v_lowpass(dst+8, src+8, dstStride, srcStride, C1, C2);\
137 src += 8*srcStride;\
138 dst += 8*dstStride;\
139 OPNAME ## rv30_tpel8_v_lowpass(dst , src , dstStride, srcStride, C1, C2);\
140 OPNAME ## rv30_tpel8_v_lowpass(dst+8, src+8, dstStride, srcStride, C1, C2);\
143 static void OPNAME ## rv30_tpel16_h_lowpass(uint8_t *dst, uint8_t *src, int dstStride, int srcStride, const int C1, const int C2){\
144 OPNAME ## rv30_tpel8_h_lowpass(dst , src , dstStride, srcStride, C1, C2);\
145 OPNAME ## rv30_tpel8_h_lowpass(dst+8, src+8, dstStride, srcStride, C1, C2);\
146 src += 8*srcStride;\
147 dst += 8*dstStride;\
148 OPNAME ## rv30_tpel8_h_lowpass(dst , src , dstStride, srcStride, C1, C2);\
149 OPNAME ## rv30_tpel8_h_lowpass(dst+8, src+8, dstStride, srcStride, C1, C2);\
152 static void OPNAME ## rv30_tpel16_hv_lowpass(uint8_t *dst, uint8_t *src, int dstStride, int srcStride){\
153 OPNAME ## rv30_tpel8_hv_lowpass(dst , src , dstStride, srcStride);\
154 OPNAME ## rv30_tpel8_hv_lowpass(dst+8, src+8, dstStride, srcStride);\
155 src += 8*srcStride;\
156 dst += 8*dstStride;\
157 OPNAME ## rv30_tpel8_hv_lowpass(dst , src , dstStride, srcStride);\
158 OPNAME ## rv30_tpel8_hv_lowpass(dst+8, src+8, dstStride, srcStride);\
162 #define RV30_MC(OPNAME, SIZE) \
163 static void OPNAME ## rv30_tpel ## SIZE ## _mc10_c(uint8_t *dst, uint8_t *src, int stride){\
164 OPNAME ## rv30_tpel ## SIZE ## _h_lowpass(dst, src, stride, stride, 12, 6);\
167 static void OPNAME ## rv30_tpel ## SIZE ## _mc20_c(uint8_t *dst, uint8_t *src, int stride){\
168 OPNAME ## rv30_tpel ## SIZE ## _h_lowpass(dst, src, stride, stride, 6, 12);\
171 static void OPNAME ## rv30_tpel ## SIZE ## _mc01_c(uint8_t *dst, uint8_t *src, int stride){\
172 OPNAME ## rv30_tpel ## SIZE ## _v_lowpass(dst, src, stride, stride, 12, 6);\
175 static void OPNAME ## rv30_tpel ## SIZE ## _mc02_c(uint8_t *dst, uint8_t *src, int stride){\
176 OPNAME ## rv30_tpel ## SIZE ## _v_lowpass(dst, src, stride, stride, 6, 12);\
179 static void OPNAME ## rv30_tpel ## SIZE ## _mc11_c(uint8_t *dst, uint8_t *src, int stride){\
180 uint8_t half[SIZE*SIZE];\
181 put_rv30_tpel ## SIZE ## _h_lowpass(half, src, SIZE, stride, 12, 6);\
182 OPNAME ## rv30_tpel ## SIZE ## _v_lowpass(dst, src, stride, stride, 12, 6);\
185 static void OPNAME ## rv30_tpel ## SIZE ## _mc12_c(uint8_t *dst, uint8_t *src, int stride){\
186 uint8_t half[SIZE*SIZE];\
187 put_rv30_tpel ## SIZE ## _h_lowpass(half, src, SIZE, stride, 12, 6);\
188 OPNAME ## rv30_tpel ## SIZE ## _v_lowpass(dst, src, stride, stride, 6, 12);\
191 static void OPNAME ## rv30_tpel ## SIZE ## _mc21_c(uint8_t *dst, uint8_t *src, int stride){\
192 uint8_t half[SIZE*SIZE];\
193 put_rv30_tpel ## SIZE ## _h_lowpass(half, src, SIZE, stride, 6, 12);\
194 OPNAME ## rv30_tpel ## SIZE ## _v_lowpass(dst, src, stride, stride, 12, 6);\
197 static void OPNAME ## rv30_tpel ## SIZE ## _mc22_c(uint8_t *dst, uint8_t *src, int stride){\
198 OPNAME ## rv30_tpel ## SIZE ## _hv_lowpass(dst, src, stride, stride);\
202 #define op_avg(a, b) a = (((a)+cm[((b) + 8)>>4]+1)>>1)
203 #define op_put(a, b) a = cm[((b) + 8)>>4]
205 RV30_LOWPASS(put_ , op_put)
206 RV30_LOWPASS(avg_ , op_avg)
207 RV30_MC(put_, 8)
208 RV30_MC(put_, 16)
209 RV30_MC(avg_, 8)
210 RV30_MC(avg_, 16)
212 void ff_rv30dsp_init(DSPContext* c, AVCodecContext *avctx) {
213 c->put_rv30_tpel_pixels_tab[0][ 0] = c->put_h264_qpel_pixels_tab[0][0];
214 c->put_rv30_tpel_pixels_tab[0][ 1] = put_rv30_tpel16_mc10_c;
215 c->put_rv30_tpel_pixels_tab[0][ 2] = put_rv30_tpel16_mc20_c;
216 c->put_rv30_tpel_pixels_tab[0][ 4] = put_rv30_tpel16_mc01_c;
217 c->put_rv30_tpel_pixels_tab[0][ 5] = put_rv30_tpel16_mc11_c;
218 c->put_rv30_tpel_pixels_tab[0][ 6] = put_rv30_tpel16_mc21_c;
219 c->put_rv30_tpel_pixels_tab[0][ 8] = put_rv30_tpel16_mc02_c;
220 c->put_rv30_tpel_pixels_tab[0][ 9] = put_rv30_tpel16_mc12_c;
221 c->put_rv30_tpel_pixels_tab[0][10] = put_rv30_tpel16_mc22_c;
222 c->avg_rv30_tpel_pixels_tab[0][ 0] = c->avg_h264_qpel_pixels_tab[0][0];
223 c->avg_rv30_tpel_pixels_tab[0][ 1] = avg_rv30_tpel16_mc10_c;
224 c->avg_rv30_tpel_pixels_tab[0][ 2] = avg_rv30_tpel16_mc20_c;
225 c->avg_rv30_tpel_pixels_tab[0][ 4] = avg_rv30_tpel16_mc01_c;
226 c->avg_rv30_tpel_pixels_tab[0][ 5] = avg_rv30_tpel16_mc11_c;
227 c->avg_rv30_tpel_pixels_tab[0][ 6] = avg_rv30_tpel16_mc21_c;
228 c->avg_rv30_tpel_pixels_tab[0][ 8] = avg_rv30_tpel16_mc02_c;
229 c->avg_rv30_tpel_pixels_tab[0][ 9] = avg_rv30_tpel16_mc12_c;
230 c->avg_rv30_tpel_pixels_tab[0][10] = avg_rv30_tpel16_mc22_c;
231 c->put_rv30_tpel_pixels_tab[1][ 0] = c->put_h264_qpel_pixels_tab[1][0];
232 c->put_rv30_tpel_pixels_tab[1][ 1] = put_rv30_tpel8_mc10_c;
233 c->put_rv30_tpel_pixels_tab[1][ 2] = put_rv30_tpel8_mc20_c;
234 c->put_rv30_tpel_pixels_tab[1][ 4] = put_rv30_tpel8_mc01_c;
235 c->put_rv30_tpel_pixels_tab[1][ 5] = put_rv30_tpel8_mc11_c;
236 c->put_rv30_tpel_pixels_tab[1][ 6] = put_rv30_tpel8_mc21_c;
237 c->put_rv30_tpel_pixels_tab[1][ 8] = put_rv30_tpel8_mc02_c;
238 c->put_rv30_tpel_pixels_tab[1][ 9] = put_rv30_tpel8_mc12_c;
239 c->put_rv30_tpel_pixels_tab[1][10] = put_rv30_tpel8_mc22_c;
240 c->avg_rv30_tpel_pixels_tab[1][ 0] = c->avg_h264_qpel_pixels_tab[1][0];
241 c->avg_rv30_tpel_pixels_tab[1][ 1] = avg_rv30_tpel8_mc10_c;
242 c->avg_rv30_tpel_pixels_tab[1][ 2] = avg_rv30_tpel8_mc20_c;
243 c->avg_rv30_tpel_pixels_tab[1][ 4] = avg_rv30_tpel8_mc01_c;
244 c->avg_rv30_tpel_pixels_tab[1][ 5] = avg_rv30_tpel8_mc11_c;
245 c->avg_rv30_tpel_pixels_tab[1][ 6] = avg_rv30_tpel8_mc21_c;
246 c->avg_rv30_tpel_pixels_tab[1][ 8] = avg_rv30_tpel8_mc02_c;
247 c->avg_rv30_tpel_pixels_tab[1][ 9] = avg_rv30_tpel8_mc12_c;
248 c->avg_rv30_tpel_pixels_tab[1][10] = avg_rv30_tpel8_mc22_c;