1 |
/************************************************************************** |
/***************************************************************************** |
2 |
* |
* |
3 |
* XVID MPEG-4 VIDEO CODEC |
* XVID MPEG-4 VIDEO CODEC |
4 |
* sum of absolute difference |
* - SAD calculation module (C part) - |
5 |
* |
* |
6 |
* This program is an implementation of a part of one or more MPEG-4 |
* Copyright(C) 2002 Michael Militzer <isibaar@xvid.org> |
7 |
* Video tools as specified in ISO/IEC 14496-2 standard. Those intending |
* 2002 Peter Ross <pross@xvid.org> |
|
* to use this software module in hardware or software products are |
|
|
* advised that its use may infringe existing patents or copyrights, and |
|
|
* any such use would be at such party's own risk. The original |
|
|
* developer of this software module and his/her company, and subsequent |
|
|
* editors and their companies, will have no liability for use of this |
|
|
* software or modifications or derivatives thereof. |
|
8 |
* |
* |
9 |
* This program is free software; you can redistribute it and/or modify |
* This file is part of XviD, a free MPEG-4 video encoder/decoder |
10 |
* it under the terms of the GNU General Public License as published by |
* |
11 |
|
* XviD is free software; you can redistribute it and/or modify it |
12 |
|
* under the terms of the GNU General Public License as published by |
13 |
* the Free Software Foundation; either version 2 of the License, or |
* the Free Software Foundation; either version 2 of the License, or |
14 |
* (at your option) any later version. |
* (at your option) any later version. |
15 |
* |
* |
20 |
* |
* |
21 |
* You should have received a copy of the GNU General Public License |
* You should have received a copy of the GNU General Public License |
22 |
* along with this program; if not, write to the Free Software |
* along with this program; if not, write to the Free Software |
23 |
* Foundation, Inc., 675 Mass Ave, Cambridge, MA 02139, USA. |
* Foundation, Inc., 59 Temple Place, Suite 330, Boston, MA 02111-1307 USA |
24 |
* |
* |
25 |
*************************************************************************/ |
* Under section 8 of the GNU General Public License, the copyright |
26 |
|
* holders of XVID explicitly forbid distribution in the following |
27 |
/************************************************************************** |
* countries: |
28 |
* |
* |
29 |
* History: |
* - Japan |
30 |
|
* - United States of America |
31 |
* |
* |
32 |
* 14.02.2002 added sad16bi_c() |
* Linking XviD statically or dynamically with other modules is making a |
33 |
* 10.11.2001 initial version; (c)2001 peter ross <pross@cs.rmit.edu.au> |
* combined work based on XviD. Thus, the terms and conditions of the |
34 |
|
* GNU General Public License cover the whole combination. |
35 |
* |
* |
36 |
*************************************************************************/ |
* As a special exception, the copyright holders of XviD give you |
37 |
|
* permission to link XviD with independent modules that communicate with |
38 |
|
* XviD solely through the VFW1.1 and DShow interfaces, regardless of the |
39 |
|
* license terms of these independent modules, and to copy and distribute |
40 |
|
* the resulting combined work under terms of your choice, provided that |
41 |
|
* every copy of the combined work is accompanied by a complete copy of |
42 |
|
* the source code of XviD (the version of XviD used to produce the |
43 |
|
* combined work), being distributed under the terms of the GNU General |
44 |
|
* Public License plus this exception. An independent module is a module |
45 |
|
* which is not derived from or based on XviD. |
46 |
|
* |
47 |
|
* Note that people who make modified versions of XviD are not obligated |
48 |
|
* to grant this special exception for their modified versions; it is |
49 |
|
* their choice whether to do so. The GNU General Public License gives |
50 |
|
* permission to release a modified version without this exception; this |
51 |
|
* exception also makes it possible to release a modified version which |
52 |
|
* carries forward this exception. |
53 |
|
* |
54 |
|
* $Id$ |
55 |
|
* |
56 |
|
****************************************************************************/ |
57 |
|
|
58 |
#include "../portab.h" |
#include "../portab.h" |
59 |
#include "sad.h" |
#include "sad.h" |
61 |
sad16FuncPtr sad16; |
sad16FuncPtr sad16; |
62 |
sad8FuncPtr sad8; |
sad8FuncPtr sad8; |
63 |
sad16biFuncPtr sad16bi; |
sad16biFuncPtr sad16bi; |
64 |
sad8biFuncPtr sad8bi; // not really sad16, but no difference in prototype |
sad8biFuncPtr sad8bi; /* not really sad16, but no difference in prototype */ |
65 |
dev16FuncPtr dev16; |
dev16FuncPtr dev16; |
|
sad16vFuncPtr sad16v; |
|
66 |
|
|
67 |
sadInitFuncPtr sadInit; |
sadInitFuncPtr sadInit; |
68 |
|
|
69 |
#define ABS(X) (((X)>0)?(X):-(X)) |
#define ABS(X) (((X)>0)?(X):-(X)) |
70 |
|
|
71 |
|
#define MRSAD16_CORRFACTOR 8 |
72 |
|
uint32_t |
73 |
|
mrsad16_c(const uint8_t * const cur, |
74 |
|
const uint8_t * const ref, |
75 |
|
const uint32_t stride, |
76 |
|
const uint32_t best_sad) |
77 |
|
{ |
78 |
|
|
79 |
|
uint32_t sad = 0; |
80 |
|
int32_t mean = 0; |
81 |
|
uint32_t i, j; |
82 |
|
uint8_t const *ptr_cur = cur; |
83 |
|
uint8_t const *ptr_ref = ref; |
84 |
|
|
85 |
|
for (j = 0; j < 16; j++) { |
86 |
|
for (i = 0; i < 16; i++) { |
87 |
|
mean += ((int) *(ptr_cur + i) - (int) *(ptr_ref + i)); |
88 |
|
} |
89 |
|
ptr_cur += stride; |
90 |
|
ptr_ref += stride; |
91 |
|
|
92 |
|
} |
93 |
|
mean /= 256; |
94 |
|
|
95 |
|
for (j = 0; j < 16; j++) { |
96 |
|
|
97 |
|
ptr_cur -= stride; |
98 |
|
ptr_ref -= stride; |
99 |
|
|
100 |
|
for (i = 0; i < 16; i++) { |
101 |
|
|
102 |
|
sad += ABS(*(ptr_cur + i) - *(ptr_ref + i) - mean); |
103 |
|
if (sad >= best_sad) { |
104 |
|
return MRSAD16_CORRFACTOR * sad; |
105 |
|
} |
106 |
|
} |
107 |
|
} |
108 |
|
|
109 |
|
return MRSAD16_CORRFACTOR * sad; |
110 |
|
|
111 |
|
} |
112 |
|
|
113 |
|
|
114 |
uint32_t |
uint32_t |
115 |
sad16_c(const uint8_t * const cur, |
sad16_c(const uint8_t * const cur, |
116 |
const uint8_t * const ref, |
const uint8_t * const ref, |
119 |
{ |
{ |
120 |
|
|
121 |
uint32_t sad = 0; |
uint32_t sad = 0; |
122 |
uint32_t j; |
uint32_t i, j; |
123 |
uint8_t const *ptr_cur = cur; |
uint8_t const *ptr_cur = cur; |
124 |
uint8_t const *ptr_ref = ref; |
uint8_t const *ptr_ref = ref; |
125 |
|
|
126 |
for (j = 0; j < 16; j++) { |
for (j = 0; j < 16; j++) { |
|
sad += ABS(ptr_cur[0] - ptr_ref[0]); |
|
|
sad += ABS(ptr_cur[1] - ptr_ref[1]); |
|
|
sad += ABS(ptr_cur[2] - ptr_ref[2]); |
|
|
sad += ABS(ptr_cur[3] - ptr_ref[3]); |
|
|
sad += ABS(ptr_cur[4] - ptr_ref[4]); |
|
|
sad += ABS(ptr_cur[5] - ptr_ref[5]); |
|
|
sad += ABS(ptr_cur[6] - ptr_ref[6]); |
|
|
sad += ABS(ptr_cur[7] - ptr_ref[7]); |
|
|
sad += ABS(ptr_cur[8] - ptr_ref[8]); |
|
|
sad += ABS(ptr_cur[9] - ptr_ref[9]); |
|
|
sad += ABS(ptr_cur[10] - ptr_ref[10]); |
|
|
sad += ABS(ptr_cur[11] - ptr_ref[11]); |
|
|
sad += ABS(ptr_cur[12] - ptr_ref[12]); |
|
|
sad += ABS(ptr_cur[13] - ptr_ref[13]); |
|
|
sad += ABS(ptr_cur[14] - ptr_ref[14]); |
|
|
sad += ABS(ptr_cur[15] - ptr_ref[15]); |
|
127 |
|
|
128 |
if (sad >= best_sad) |
for (i = 0; i < 16; i++) { |
129 |
|
|
130 |
|
sad += ABS(*(ptr_cur + i) - *(ptr_ref + i)); |
131 |
|
|
132 |
|
if (sad >= best_sad) { |
133 |
return sad; |
return sad; |
134 |
|
} |
135 |
|
|
136 |
|
|
137 |
|
} |
138 |
|
|
139 |
ptr_cur += stride; |
ptr_cur += stride; |
140 |
ptr_ref += stride; |
ptr_ref += stride; |
145 |
|
|
146 |
} |
} |
147 |
|
|
148 |
|
|
149 |
|
|
150 |
uint32_t |
uint32_t |
151 |
sad16bi_c(const uint8_t * const cur, |
sad16bi_c(const uint8_t * const cur, |
152 |
const uint8_t * const ref1, |
const uint8_t * const ref1, |
229 |
const uint32_t stride) |
const uint32_t stride) |
230 |
{ |
{ |
231 |
uint32_t sad = 0; |
uint32_t sad = 0; |
232 |
uint32_t j; |
uint32_t i, j; |
233 |
uint8_t const *ptr_cur = cur; |
uint8_t const *ptr_cur = cur; |
234 |
uint8_t const *ptr_ref = ref; |
uint8_t const *ptr_ref = ref; |
235 |
|
|
236 |
for (j = 0; j < 8; j++) { |
for (j = 0; j < 8; j++) { |
237 |
|
|
238 |
sad += ABS(ptr_cur[0] - ptr_ref[0]); |
for (i = 0; i < 8; i++) { |
239 |
sad += ABS(ptr_cur[1] - ptr_ref[1]); |
sad += ABS(*(ptr_cur + i) - *(ptr_ref + i)); |
240 |
sad += ABS(ptr_cur[2] - ptr_ref[2]); |
} |
|
sad += ABS(ptr_cur[3] - ptr_ref[3]); |
|
|
sad += ABS(ptr_cur[4] - ptr_ref[4]); |
|
|
sad += ABS(ptr_cur[5] - ptr_ref[5]); |
|
|
sad += ABS(ptr_cur[6] - ptr_ref[6]); |
|
|
sad += ABS(ptr_cur[7] - ptr_ref[7]); |
|
241 |
|
|
242 |
ptr_cur += stride; |
ptr_cur += stride; |
243 |
ptr_ref += stride; |
ptr_ref += stride; |
248 |
} |
} |
249 |
|
|
250 |
|
|
251 |
|
|
252 |
|
|
253 |
/* average deviation from mean */ |
/* average deviation from mean */ |
254 |
|
|
255 |
uint32_t |
uint32_t |
285 |
|
|
286 |
return dev; |
return dev; |
287 |
} |
} |
|
|
|
|
uint32_t sad16v_c(const uint8_t * const cur, |
|
|
const uint8_t * const ref, |
|
|
const uint32_t stride, |
|
|
int32_t *sad) |
|
|
{ |
|
|
sad[0] = sad8(cur, ref, stride); |
|
|
sad[1] = sad8(cur + 8, ref + 8, stride); |
|
|
sad[2] = sad8(cur + 8*stride, ref + 8*stride, stride); |
|
|
sad[3] = sad8(cur + 8*stride + 8, ref + 8*stride + 8, stride); |
|
|
|
|
|
return sad[0]+sad[1]+sad[2]+sad[3]; |
|
|
} |
|
|
|
|
|
uint32_t sad32v_c(const uint8_t * const cur, |
|
|
const uint8_t * const ref, |
|
|
const uint32_t stride, |
|
|
int32_t *sad) |
|
|
{ |
|
|
sad[0] = sad16(cur, ref, stride, 256*4096); |
|
|
sad[1] = sad16(cur + 8, ref + 8, stride, 256*4096); |
|
|
sad[2] = sad16(cur + 8*stride, ref + 8*stride, stride, 256*4096); |
|
|
sad[3] = sad16(cur + 8*stride + 8, ref + 8*stride + 8, stride, 256*4096); |
|
|
|
|
|
return sad[0]+sad[1]+sad[2]+sad[3]; |
|
|
} |
|
|
|
|
|
|
|
|
|
|
|
#define MRSAD16_CORRFACTOR 8 |
|
|
uint32_t |
|
|
mrsad16_c(const uint8_t * const cur, |
|
|
const uint8_t * const ref, |
|
|
const uint32_t stride, |
|
|
const uint32_t best_sad) |
|
|
{ |
|
|
|
|
|
uint32_t sad = 0; |
|
|
int32_t mean = 0; |
|
|
uint32_t i, j; |
|
|
uint8_t const *ptr_cur = cur; |
|
|
uint8_t const *ptr_ref = ref; |
|
|
|
|
|
for (j = 0; j < 16; j++) { |
|
|
for (i = 0; i < 16; i++) { |
|
|
mean += ((int) *(ptr_cur + i) - (int) *(ptr_ref + i)); |
|
|
} |
|
|
ptr_cur += stride; |
|
|
ptr_ref += stride; |
|
|
|
|
|
} |
|
|
mean /= 256; |
|
|
|
|
|
for (j = 0; j < 16; j++) { |
|
|
|
|
|
ptr_cur -= stride; |
|
|
ptr_ref -= stride; |
|
|
|
|
|
for (i = 0; i < 16; i++) { |
|
|
|
|
|
sad += ABS(*(ptr_cur + i) - *(ptr_ref + i) - mean); |
|
|
if (sad >= best_sad) { |
|
|
return MRSAD16_CORRFACTOR * sad; |
|
|
} |
|
|
} |
|
|
} |
|
|
|
|
|
return MRSAD16_CORRFACTOR * sad; |
|
|
|
|
|
} |
|
|
|
|
|
|
|