Parent Directory
|
Revision Log
Revision 1.6 - (view) (download)
1 : | chl | 1.6 | /***************************************************************************** |
2 : | * | ||
3 : | * XVID MPEG-4 VIDEO CODEC | ||
4 : | * - 8x8 block-based halfpel interpolation - headers | ||
5 : | * | ||
6 : | * Copyright(C) 2002 Peter Ross <pross@xvid.org> | ||
7 : | * | ||
8 : | * This program is an implementation of a part of one or more MPEG-4 | ||
9 : | * Video tools as specified in ISO/IEC 14496-2 standard. Those intending | ||
10 : | * to use this software module in hardware or software products are | ||
11 : | * advised that its use may infringe existing patents or copyrights, and | ||
12 : | * any such use would be at such party's own risk. The original | ||
13 : | * developer of this software module and his/her company, and subsequent | ||
14 : | * editors and their companies, will have no liability for use of this | ||
15 : | * software or modifications or derivatives thereof. | ||
16 : | * | ||
17 : | * This program is free software; you can redistribute it and/or modify | ||
18 : | * it under the terms of the GNU General Public License as published by | ||
19 : | * the Free Software Foundation; either version 2 of the License, or | ||
20 : | * (at your option) any later version. | ||
21 : | * | ||
22 : | * This program is distributed in the hope that it will be useful, | ||
23 : | * but WITHOUT ANY WARRANTY; without even the implied warranty of | ||
24 : | * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the | ||
25 : | * GNU General Public License for more details. | ||
26 : | * | ||
27 : | * You should have received a copy of the GNU General Public License | ||
28 : | * along with this program; if not, write to the Free Software | ||
29 : | * Foundation, Inc., 59 Temple Place, Suite 330, Boston, MA 02111-1307 USA | ||
30 : | * | ||
31 : | ****************************************************************************/ | ||
32 : | |||
33 : | Isibaar | 1.1 | #include "../utils/mem_transfer.h" |
34 : | |||
35 : | edgomez | 1.3 | typedef void (INTERPOLATE8X8) (uint8_t * const dst, |
36 : | const uint8_t * const src, | ||
37 : | const uint32_t stride, | ||
38 : | const uint32_t rounding); | ||
39 : | typedef INTERPOLATE8X8 *INTERPOLATE8X8_PTR; | ||
40 : | Isibaar | 1.1 | |
41 : | extern INTERPOLATE8X8_PTR interpolate8x8_halfpel_h; | ||
42 : | extern INTERPOLATE8X8_PTR interpolate8x8_halfpel_v; | ||
43 : | extern INTERPOLATE8X8_PTR interpolate8x8_halfpel_hv; | ||
44 : | |||
45 : | INTERPOLATE8X8 interpolate8x8_halfpel_h_c; | ||
46 : | INTERPOLATE8X8 interpolate8x8_halfpel_v_c; | ||
47 : | INTERPOLATE8X8 interpolate8x8_halfpel_hv_c; | ||
48 : | |||
49 : | INTERPOLATE8X8 interpolate8x8_halfpel_h_mmx; | ||
50 : | INTERPOLATE8X8 interpolate8x8_halfpel_v_mmx; | ||
51 : | INTERPOLATE8X8 interpolate8x8_halfpel_hv_mmx; | ||
52 : | |||
53 : | INTERPOLATE8X8 interpolate8x8_halfpel_h_xmm; | ||
54 : | INTERPOLATE8X8 interpolate8x8_halfpel_v_xmm; | ||
55 : | INTERPOLATE8X8 interpolate8x8_halfpel_hv_xmm; | ||
56 : | |||
57 : | INTERPOLATE8X8 interpolate8x8_halfpel_h_3dn; | ||
58 : | INTERPOLATE8X8 interpolate8x8_halfpel_v_3dn; | ||
59 : | INTERPOLATE8X8 interpolate8x8_halfpel_hv_3dn; | ||
60 : | |||
61 : | Isibaar | 1.4 | INTERPOLATE8X8 interpolate8x8_halfpel_h_ia64; |
62 : | INTERPOLATE8X8 interpolate8x8_halfpel_v_ia64; | ||
63 : | INTERPOLATE8X8 interpolate8x8_halfpel_hv_ia64; | ||
64 : | |||
65 : | Isibaar | 1.5 | void interpolate8x8_lowpass_h(uint8_t *dst, uint8_t *src, int32_t dst_stride, int32_t src_stride, int32_t rounding); |
66 : | void interpolate8x8_lowpass_v(uint8_t *dst, uint8_t *src, int32_t dst_stride, int32_t src_stride, int32_t rounding); | ||
67 : | void interpolate8x8_lowpass_hv(uint8_t *dst1, uint8_t *dst2, uint8_t *src, int32_t dst1_stride, int32_t dst2_stride, int32_t src_stride, int32_t rounding); | ||
68 : | void interpolate8x8_bilinear2(uint8_t *dst, uint8_t *src1, uint8_t *src2, int32_t dst_stride, int32_t src_stride, int32_t rounding); | ||
69 : | void interpolate8x8_bilinear4(uint8_t *dst, uint8_t *src1, uint8_t *src2, uint8_t *src3, uint8_t *src4, int32_t stride, int32_t rounding); | ||
70 : | |||
71 : | void interpolate8x8_c(uint8_t * const dst, | ||
72 : | const uint8_t * const src, | ||
73 : | const uint32_t x, | ||
74 : | const uint32_t y, | ||
75 : | const uint32_t stride); | ||
76 : | |||
77 : | edgomez | 1.3 | static __inline void |
78 : | interpolate8x8_switch(uint8_t * const cur, | ||
79 : | const uint8_t * const refn, | ||
80 : | const uint32_t x, | ||
81 : | const uint32_t y, | ||
82 : | const int32_t dx, | ||
83 : | const int dy, | ||
84 : | const uint32_t stride, | ||
85 : | const uint32_t rounding) | ||
86 : | Isibaar | 1.1 | { |
87 : | int32_t ddx, ddy; | ||
88 : | |||
89 : | edgomez | 1.3 | switch (((dx & 1) << 1) + (dy & 1)) // ((dx%2)?2:0)+((dy%2)?1:0) |
90 : | { | ||
91 : | case 0: | ||
92 : | ddx = dx / 2; | ||
93 : | ddy = dy / 2; | ||
94 : | transfer8x8_copy(cur + y * stride + x, | ||
95 : | Isibaar | 1.4 | refn + (int)((y + ddy) * stride + x + ddx), stride); |
96 : | Isibaar | 1.1 | break; |
97 : | |||
98 : | edgomez | 1.3 | case 1: |
99 : | ddx = dx / 2; | ||
100 : | ddy = (dy - 1) / 2; | ||
101 : | interpolate8x8_halfpel_v(cur + y * stride + x, | ||
102 : | Isibaar | 1.4 | refn + (int)((y + ddy) * stride + x + ddx), stride, |
103 : | edgomez | 1.3 | rounding); |
104 : | Isibaar | 1.1 | break; |
105 : | |||
106 : | edgomez | 1.3 | case 2: |
107 : | ddx = (dx - 1) / 2; | ||
108 : | ddy = dy / 2; | ||
109 : | interpolate8x8_halfpel_h(cur + y * stride + x, | ||
110 : | Isibaar | 1.4 | refn + (int)((y + ddy) * stride + x + ddx), stride, |
111 : | edgomez | 1.3 | rounding); |
112 : | Isibaar | 1.1 | break; |
113 : | |||
114 : | edgomez | 1.3 | default: |
115 : | ddx = (dx - 1) / 2; | ||
116 : | ddy = (dy - 1) / 2; | ||
117 : | interpolate8x8_halfpel_hv(cur + y * stride + x, | ||
118 : | Isibaar | 1.4 | refn + (int)((y + ddy) * stride + x + ddx), stride, |
119 : | edgomez | 1.3 | rounding); |
120 : | Isibaar | 1.1 | break; |
121 : | edgomez | 1.3 | } |
122 : | Isibaar | 1.1 | } |
123 : | chenm001 | 1.2 | |
124 : | |||
125 : | Isibaar | 1.5 | static __inline void interpolate8x8_quarterpel(uint8_t * const cur, |
126 : | uint8_t * const refn, | ||
127 : | const uint32_t x, const uint32_t y, | ||
128 : | const int32_t dx, const int dy, | ||
129 : | const uint32_t stride, | ||
130 : | const uint32_t rounding) | ||
131 : | { | ||
132 : | const int32_t xRef = x*4 + dx; | ||
133 : | const int32_t yRef = y*4 + dy; | ||
134 : | |||
135 : | uint8_t *src, *dst; | ||
136 : | int32_t x_int, y_int, x_frac, y_frac; | ||
137 : | |||
138 : | uint8_t halfpel_h[72]; | ||
139 : | uint8_t halfpel_v[64]; | ||
140 : | uint8_t halfpel_hv[64]; | ||
141 : | |||
142 : | x_int = xRef/4; | ||
143 : | if (xRef < 0 && xRef % 4) | ||
144 : | x_int--; | ||
145 : | |||
146 : | x_frac = xRef - (4*x_int); | ||
147 : | |||
148 : | y_int = yRef/4; | ||
149 : | if (yRef < 0 && yRef % 4) | ||
150 : | y_int--; | ||
151 : | |||
152 : | y_frac = yRef - (4*y_int); | ||
153 : | |||
154 : | src = refn + y_int * stride + x_int; | ||
155 : | dst = cur + y * stride + x; | ||
156 : | |||
157 : | switch((y_frac << 2) | (x_frac)) { | ||
158 : | |||
159 : | case 0: | ||
160 : | transfer8x8_copy(dst, src, stride); | ||
161 : | break; | ||
162 : | |||
163 : | case 1: | ||
164 : | interpolate8x8_lowpass_h(halfpel_h, src, 8, stride, rounding); | ||
165 : | interpolate8x8_bilinear2(dst, src, halfpel_h, stride, stride, rounding); | ||
166 : | break; | ||
167 : | |||
168 : | case 2: | ||
169 : | interpolate8x8_lowpass_h(dst, src, stride, stride, rounding); | ||
170 : | break; | ||
171 : | |||
172 : | case 3: | ||
173 : | interpolate8x8_lowpass_h(halfpel_h, src, 8, stride, rounding); | ||
174 : | interpolate8x8_bilinear2(dst, src+1, halfpel_h, stride, stride, rounding); | ||
175 : | break; | ||
176 : | |||
177 : | case 4: | ||
178 : | interpolate8x8_lowpass_v(halfpel_v, src, 8, stride, rounding); | ||
179 : | interpolate8x8_bilinear2(dst, src, halfpel_v, stride, stride, rounding); | ||
180 : | break; | ||
181 : | |||
182 : | case 5: | ||
183 : | interpolate8x8_lowpass_v(halfpel_v, src, 8, stride, rounding); | ||
184 : | interpolate8x8_lowpass_hv(halfpel_hv, halfpel_h, src, 8, 8, stride, rounding); | ||
185 : | interpolate8x8_bilinear4(dst, src, halfpel_h, halfpel_v, halfpel_hv, stride, rounding); | ||
186 : | break; | ||
187 : | |||
188 : | case 6: | ||
189 : | interpolate8x8_lowpass_hv(halfpel_hv, halfpel_h, src, 8, 8, stride, rounding); | ||
190 : | interpolate8x8_bilinear2(dst, halfpel_h, halfpel_hv, stride, 8, 1-rounding); | ||
191 : | break; | ||
192 : | |||
193 : | case 7: | ||
194 : | interpolate8x8_lowpass_v(halfpel_v, src+1, 8, stride, 16-rounding); | ||
195 : | interpolate8x8_lowpass_hv(halfpel_hv, halfpel_h, src, 8, 8, stride, rounding); | ||
196 : | interpolate8x8_bilinear4(dst, src+1, halfpel_h, halfpel_v, halfpel_hv, stride, rounding); | ||
197 : | break; | ||
198 : | |||
199 : | case 8: | ||
200 : | interpolate8x8_lowpass_v(dst, src, stride, stride, rounding); | ||
201 : | break; | ||
202 : | |||
203 : | case 9: | ||
204 : | interpolate8x8_lowpass_v(halfpel_v, src, 8, stride, 16-rounding); | ||
205 : | interpolate8x8_lowpass_hv(halfpel_hv, halfpel_h, src, 8, 8, stride, rounding); | ||
206 : | interpolate8x8_bilinear2(dst, halfpel_v, halfpel_hv, stride, 8, rounding); | ||
207 : | break; | ||
208 : | |||
209 : | case 10: | ||
210 : | interpolate8x8_lowpass_hv(dst, halfpel_h, src, stride, 8, stride, rounding); | ||
211 : | break; | ||
212 : | |||
213 : | case 11: | ||
214 : | interpolate8x8_lowpass_v(halfpel_v, src+1, 8, stride, 16-rounding); | ||
215 : | interpolate8x8_lowpass_hv(halfpel_hv, halfpel_h, src, 8, 8, stride, rounding); | ||
216 : | interpolate8x8_bilinear2(dst, halfpel_v, halfpel_hv, stride, 8, rounding); | ||
217 : | break; | ||
218 : | |||
219 : | case 12: | ||
220 : | interpolate8x8_lowpass_v(halfpel_v, src, 8, stride, rounding); | ||
221 : | interpolate8x8_bilinear2(dst, src+stride, halfpel_v, stride, stride, rounding); | ||
222 : | break; | ||
223 : | |||
224 : | case 13: | ||
225 : | interpolate8x8_lowpass_v(halfpel_v, src, 8, stride, rounding); | ||
226 : | interpolate8x8_lowpass_hv(halfpel_hv, halfpel_h, src, 8, 8, stride, rounding); | ||
227 : | interpolate8x8_bilinear4(dst, src+stride, halfpel_h+8, halfpel_v, halfpel_hv, stride, rounding); | ||
228 : | break; | ||
229 : | |||
230 : | case 14: | ||
231 : | interpolate8x8_lowpass_hv(halfpel_hv, halfpel_h, src, 8, 8, stride, rounding); | ||
232 : | interpolate8x8_bilinear2(dst, halfpel_h+8, halfpel_hv, stride, 8, rounding); | ||
233 : | break; | ||
234 : | |||
235 : | case 15: | ||
236 : | interpolate8x8_lowpass_v(halfpel_v, src+1, 8, stride, rounding); | ||
237 : | interpolate8x8_lowpass_hv(halfpel_hv, halfpel_h, src, 8, 8, stride, rounding); | ||
238 : | interpolate8x8_bilinear4(dst, src+stride+1, halfpel_h+8, halfpel_v, halfpel_hv, stride, rounding); | ||
239 : | break; | ||
240 : | } | ||
241 : | chl | 1.6 | } |
No admin address has been configured | ViewVC Help |
Powered by ViewVC 1.0.4 |