3 |
* XVID MPEG-4 VIDEO CODEC |
* XVID MPEG-4 VIDEO CODEC |
4 |
* - 8bit<->16bit transfer - |
* - 8bit<->16bit transfer - |
5 |
* |
* |
6 |
* Copyright(C) 2001-2002 Peter Ross <pross@xvid.org> |
* Copyright(C) 2001-2003 Peter Ross <pross@xvid.org> |
7 |
* |
* |
8 |
* This file is part of XviD, a free MPEG-4 video encoder/decoder |
* This program is free software ; you can redistribute it and/or modify |
9 |
* |
* it under the terms of the GNU General Public License as published by |
|
* XviD is free software; you can redistribute it and/or modify it |
|
|
* under the terms of the GNU General Public License as published by |
|
10 |
* the Free Software Foundation; either version 2 of the License, or |
* the Free Software Foundation; either version 2 of the License, or |
11 |
* (at your option) any later version. |
* (at your option) any later version. |
12 |
* |
* |
19 |
* along with this program; if not, write to the Free Software |
* along with this program; if not, write to the Free Software |
20 |
* Foundation, Inc., 59 Temple Place, Suite 330, Boston, MA 02111-1307 USA |
* Foundation, Inc., 59 Temple Place, Suite 330, Boston, MA 02111-1307 USA |
21 |
* |
* |
22 |
* Under section 8 of the GNU General Public License, the copyright |
* $Id: mem_transfer.c,v 1.16 2005-09-13 12:12:15 suxen_drol Exp $ |
|
* holders of XVID explicitly forbid distribution in the following |
|
|
* countries: |
|
|
* |
|
|
* - Japan |
|
|
* - United States of America |
|
|
* |
|
|
* Linking XviD statically or dynamically with other modules is making a |
|
|
* combined work based on XviD. Thus, the terms and conditions of the |
|
|
* GNU General Public License cover the whole combination. |
|
|
* |
|
|
* As a special exception, the copyright holders of XviD give you |
|
|
* permission to link XviD with independent modules that communicate with |
|
|
* XviD solely through the VFW1.1 and DShow interfaces, regardless of the |
|
|
* license terms of these independent modules, and to copy and distribute |
|
|
* the resulting combined work under terms of your choice, provided that |
|
|
* every copy of the combined work is accompanied by a complete copy of |
|
|
* the source code of XviD (the version of XviD used to produce the |
|
|
* combined work), being distributed under the terms of the GNU General |
|
|
* Public License plus this exception. An independent module is a module |
|
|
* which is not derived from or based on XviD. |
|
|
* |
|
|
* Note that people who make modified versions of XviD are not obligated |
|
|
* to grant this special exception for their modified versions; it is |
|
|
* their choice whether to do so. The GNU General Public License gives |
|
|
* permission to release a modified version without this exception; this |
|
|
* exception also makes it possible to release a modified version which |
|
|
* carries forward this exception. |
|
|
* |
|
|
* $Id: mem_transfer.c,v 1.8 2002-11-26 23:44:11 edgomez Exp $ |
|
23 |
* |
* |
24 |
****************************************************************************/ |
****************************************************************************/ |
25 |
|
|
32 |
TRANSFER_16TO8COPY_PTR transfer_16to8copy; |
TRANSFER_16TO8COPY_PTR transfer_16to8copy; |
33 |
|
|
34 |
TRANSFER_8TO16SUB_PTR transfer_8to16sub; |
TRANSFER_8TO16SUB_PTR transfer_8to16sub; |
35 |
|
TRANSFER_8TO16SUBRO_PTR transfer_8to16subro; |
36 |
TRANSFER_8TO16SUB2_PTR transfer_8to16sub2; |
TRANSFER_8TO16SUB2_PTR transfer_8to16sub2; |
37 |
|
TRANSFER_8TO16SUB2RO_PTR transfer_8to16sub2ro; |
38 |
TRANSFER_16TO8ADD_PTR transfer_16to8add; |
TRANSFER_16TO8ADD_PTR transfer_16to8add; |
39 |
|
|
40 |
TRANSFER8X8_COPY_PTR transfer8x8_copy; |
TRANSFER8X8_COPY_PTR transfer8x8_copy; |
41 |
|
TRANSFER8X4_COPY_PTR transfer8x4_copy; |
42 |
|
|
43 |
|
#define USE_REFERENCE_C |
44 |
|
|
45 |
/***************************************************************************** |
/***************************************************************************** |
46 |
* |
* |
67 |
const uint8_t * const src, |
const uint8_t * const src, |
68 |
uint32_t stride) |
uint32_t stride) |
69 |
{ |
{ |
70 |
uint32_t i, j; |
int i, j; |
|
|
|
71 |
for (j = 0; j < 8; j++) { |
for (j = 0; j < 8; j++) { |
72 |
for (i = 0; i < 8; i++) { |
for (i = 0; i < 8; i++) { |
73 |
dst[j * 8 + i] = (int16_t) src[j * stride + i]; |
dst[j * 8 + i] = (int16_t) src[j * stride + i]; |
89 |
const int16_t * const src, |
const int16_t * const src, |
90 |
uint32_t stride) |
uint32_t stride) |
91 |
{ |
{ |
92 |
uint32_t i, j; |
int i, j; |
93 |
|
|
94 |
for (j = 0; j < 8; j++) { |
for (j = 0; j < 8; j++) { |
95 |
for (i = 0; i < 8; i++) { |
for (i = 0; i < 8; i++) { |
96 |
|
#ifdef USE_REFERENCE_C |
97 |
int16_t pixel = src[j * 8 + i]; |
int16_t pixel = src[j * 8 + i]; |
98 |
|
|
99 |
if (pixel < 0) { |
if (pixel < 0) { |
102 |
pixel = 255; |
pixel = 255; |
103 |
} |
} |
104 |
dst[j * stride + i] = (uint8_t) pixel; |
dst[j * stride + i] = (uint8_t) pixel; |
105 |
|
#else |
106 |
|
const int16_t pixel = src[j * 8 + i]; |
107 |
|
const uint8_t value = (uint8_t)( (pixel&~255) ? (-pixel)>>(8*sizeof(pixel)-1) : pixel ); |
108 |
|
dst[j*stride + i] = value; |
109 |
|
#endif |
110 |
} |
} |
111 |
} |
} |
112 |
} |
} |
131 |
const uint8_t * ref, |
const uint8_t * ref, |
132 |
const uint32_t stride) |
const uint32_t stride) |
133 |
{ |
{ |
134 |
uint32_t i, j; |
int i, j; |
135 |
|
|
136 |
for (j = 0; j < 8; j++) { |
for (j = 0; j < 8; j++) { |
137 |
for (i = 0; i < 8; i++) { |
for (i = 0; i < 8; i++) { |
138 |
uint8_t c = cur[j * stride + i]; |
const uint8_t c = cur[j * stride + i]; |
139 |
uint8_t r = ref[j * stride + i]; |
const uint8_t r = ref[j * stride + i]; |
140 |
|
|
141 |
cur[j * stride + i] = r; |
cur[j * stride + i] = r; |
142 |
dct[j * 8 + i] = (int16_t) c - (int16_t) r; |
dct[j * 8 + i] = (int16_t) c - (int16_t) r; |
145 |
} |
} |
146 |
|
|
147 |
|
|
148 |
|
void |
149 |
|
transfer_8to16subro_c(int16_t * const dct, |
150 |
|
const uint8_t * const cur, |
151 |
|
const uint8_t * ref, |
152 |
|
const uint32_t stride) |
153 |
|
{ |
154 |
|
int i, j; |
155 |
|
|
156 |
|
for (j = 0; j < 8; j++) { |
157 |
|
for (i = 0; i < 8; i++) { |
158 |
|
const uint8_t c = cur[j * stride + i]; |
159 |
|
const uint8_t r = ref[j * stride + i]; |
160 |
|
dct[j * 8 + i] = (int16_t) c - (int16_t) r; |
161 |
|
} |
162 |
|
} |
163 |
|
} |
164 |
|
|
165 |
|
|
166 |
|
|
167 |
/* |
/* |
168 |
* C - the current buffer |
* C - the current buffer |
169 |
* R1 - the 1st reference buffer |
* R1 - the 1st reference buffer |
174 |
* |
* |
175 |
* R1 (8bit) = R1 |
* R1 (8bit) = R1 |
176 |
* R2 (8bit) = R2 |
* R2 (8bit) = R2 |
177 |
* C (8bit) = C |
* R (temp) = min((R1 + R2)/2, 255) |
178 |
* DCT (16bit)= C - min((R1 + R2)/2, 255) |
* DCT (16bit)= C - R |
179 |
|
* C (8bit) = R |
180 |
*/ |
*/ |
181 |
void |
void |
182 |
transfer_8to16sub2_c(int16_t * const dct, |
transfer_8to16sub2_c(int16_t * const dct, |
189 |
|
|
190 |
for (j = 0; j < 8; j++) { |
for (j = 0; j < 8; j++) { |
191 |
for (i = 0; i < 8; i++) { |
for (i = 0; i < 8; i++) { |
192 |
uint8_t c = cur[j * stride + i]; |
const uint8_t c = cur[j * stride + i]; |
193 |
int r = (ref1[j * stride + i] + ref2[j * stride + i] + 1) / 2; |
const uint8_t r = (ref1[j * stride + i] + ref2[j * stride + i] + 1) >> 1; |
194 |
|
cur[j * stride + i] = r; |
195 |
if (r > 255) { |
dct[j * 8 + i] = (int16_t) c - (int16_t) r; |
196 |
r = 255; |
} |
197 |
} |
} |
198 |
/*cur[j * stride + i] = r; */ |
} |
199 |
|
|
200 |
|
void |
201 |
|
transfer_8to16sub2ro_c(int16_t * const dct, |
202 |
|
const uint8_t * const cur, |
203 |
|
const uint8_t * ref1, |
204 |
|
const uint8_t * ref2, |
205 |
|
const uint32_t stride) |
206 |
|
{ |
207 |
|
uint32_t i, j; |
208 |
|
|
209 |
|
for (j = 0; j < 8; j++) { |
210 |
|
for (i = 0; i < 8; i++) { |
211 |
|
const uint8_t c = cur[j * stride + i]; |
212 |
|
const uint8_t r = (ref1[j * stride + i] + ref2[j * stride + i] + 1) >> 1; |
213 |
dct[j * 8 + i] = (int16_t) c - (int16_t) r; |
dct[j * 8 + i] = (int16_t) c - (int16_t) r; |
214 |
} |
} |
215 |
} |
} |
230 |
const int16_t * const src, |
const int16_t * const src, |
231 |
uint32_t stride) |
uint32_t stride) |
232 |
{ |
{ |
233 |
uint32_t i, j; |
int i, j; |
234 |
|
|
235 |
for (j = 0; j < 8; j++) { |
for (j = 0; j < 8; j++) { |
236 |
for (i = 0; i < 8; i++) { |
for (i = 0; i < 8; i++) { |
237 |
|
#ifdef USE_REFERENCE_C |
238 |
int16_t pixel = (int16_t) dst[j * stride + i] + src[j * 8 + i]; |
int16_t pixel = (int16_t) dst[j * stride + i] + src[j * 8 + i]; |
239 |
|
|
240 |
if (pixel < 0) { |
if (pixel < 0) { |
243 |
pixel = 255; |
pixel = 255; |
244 |
} |
} |
245 |
dst[j * stride + i] = (uint8_t) pixel; |
dst[j * stride + i] = (uint8_t) pixel; |
246 |
|
#else |
247 |
|
const int16_t pixel = (int16_t) dst[j * stride + i] + src[j * 8 + i]; |
248 |
|
const uint8_t value = (uint8_t)( (pixel&~255) ? (-pixel)>>(8*sizeof(pixel)-1) : pixel ); |
249 |
|
dst[j*stride + i] = value; |
250 |
|
#endif |
251 |
|
|
252 |
} |
} |
253 |
} |
} |
254 |
} |
} |
267 |
const uint8_t * const src, |
const uint8_t * const src, |
268 |
const uint32_t stride) |
const uint32_t stride) |
269 |
{ |
{ |
270 |
uint32_t i, j; |
int j, i; |
271 |
|
|
272 |
for (j = 0; j < 8; j++) { |
for (j = 0; j < 8; ++j) { |
273 |
for (i = 0; i < 8; i++) { |
uint8_t *d = dst + j*stride; |
274 |
dst[j * stride + i] = src[j * stride + i]; |
const uint8_t *s = src + j*stride; |
275 |
|
|
276 |
|
for (i = 0; i < 8; ++i) |
277 |
|
{ |
278 |
|
*d++ = *s++; |
279 |
|
} |
280 |
|
} |
281 |
} |
} |
282 |
|
|
283 |
|
/* |
284 |
|
* SRC - the source buffer |
285 |
|
* DST - the destination buffer |
286 |
|
* |
287 |
|
* Then the function does the 8->8 bit transfer and this serie of operations : |
288 |
|
* |
289 |
|
* SRC (8bit) = SRC |
290 |
|
* DST (8bit) = SRC |
291 |
|
*/ |
292 |
|
void |
293 |
|
transfer8x4_copy_c(uint8_t * const dst, |
294 |
|
const uint8_t * const src, |
295 |
|
const uint32_t stride) |
296 |
|
{ |
297 |
|
uint32_t j; |
298 |
|
|
299 |
|
for (j = 0; j < 4; j++) { |
300 |
|
uint32_t *d= (uint32_t*)(dst + j*stride); |
301 |
|
const uint32_t *s = (const uint32_t*)(src + j*stride); |
302 |
|
*(d+0) = *(s+0); |
303 |
|
*(d+1) = *(s+1); |
304 |
} |
} |
305 |
} |
} |