| 1 |
/* Libvisual-plugins - Standard plugins for libvisual |
| 2 |
* |
| 3 |
* Copyright (C) 2004, 2005, 2006 Dennis Smit <[email protected]> |
| 4 |
* |
| 5 |
* Authors: Dennis Smit <[email protected]> |
| 6 |
* |
| 7 |
* $Id: morph_alphablend.c,v 1.19 2006/01/27 20:19:18 synap Exp $ |
| 8 |
* |
| 9 |
* This program is free software; you can redistribute it and/or modify |
| 10 |
* it under the terms of the Lesser GNU General Public License as |
| 11 |
* published by the Free Software Foundation; either version 2.1 |
| 12 |
* of the License, or (at your option) any later version. |
| 13 |
* |
| 14 |
* This program is distributed in the hope that it will be useful, |
| 15 |
* but WITHOUT ANY WARRANTY; without even the implied warranty of |
| 16 |
* MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the |
| 17 |
* GNU General Public License for more details. |
| 18 |
* |
| 19 |
* You should have received a copy of the GNU Lesser General Public License |
| 20 |
* along with this program; if not, write to the Free Software |
| 21 |
* Foundation, Inc., 59 Temple Place, Suite 330, Boston, MA 02111-1307 USA |
| 22 |
*/ |
| 23 |
|
| 24 |
#include <stdio.h> |
| 25 |
#include <stdlib.h> |
| 26 |
#include <unistd.h> |
| 27 |
#include <fcntl.h> |
| 28 |
#include <string.h> |
| 29 |
|
| 30 |
#include <libvisual/libvisual.h> |
| 31 |
|
| 32 |
typedef struct { |
| 33 |
uint16_t b:5, g:6, r:5; |
| 34 |
} _color16; |
| 35 |
|
| 36 |
static inline int alpha_blend_buffer (uint8_t *dest, uint8_t *src1, uint8_t *src2, int size, int depth, float alpha); |
| 37 |
|
| 38 |
/* alpha blenders */ |
| 39 |
static inline int alpha_blend_8_c (uint8_t *dest, uint8_t *src1, uint8_t *src2, int size, float alpha); |
| 40 |
static inline int alpha_blend_16_c (uint8_t *dest, uint8_t *src1, uint8_t *src2, int size, float alpha); |
| 41 |
static inline int alpha_blend_24_c (uint8_t *dest, uint8_t *src1, uint8_t *src2, int size, float alpha); |
| 42 |
static inline int alpha_blend_32_c (uint8_t *dest, uint8_t *src1, uint8_t *src2, int size, float alpha); |
| 43 |
|
| 44 |
static inline int alpha_blend_8_mmx (uint8_t *dest, uint8_t *src1, uint8_t *src2, int size, float alpha); |
| 45 |
static inline int alpha_blend_16_mmx (uint8_t *dest, uint8_t *src1, uint8_t *src2, int size, float alpha); |
| 46 |
static inline int alpha_blend_24_mmx (uint8_t *dest, uint8_t *src1, uint8_t *src2, int size, float alpha); |
| 47 |
static inline int alpha_blend_32_mmx (uint8_t *dest, uint8_t *src1, uint8_t *src2, int size, float alpha); |
| 48 |
|
| 49 |
int lv_morph_alpha_init (VisPluginData *plugin); |
| 50 |
int lv_morph_alpha_cleanup (VisPluginData *plugin); |
| 51 |
int lv_morph_alpha_apply (VisPluginData *plugin, float rate, VisAudio *audio, VisVideo *dest, VisVideo *src1, VisVideo *src2); |
| 52 |
|
| 53 |
VISUAL_PLUGIN_API_VERSION_VALIDATOR |
| 54 |
|
| 55 |
const VisPluginInfo *get_plugin_info (int *count) |
| 56 |
{ |
| 57 |
static VisMorphPlugin morph[] = {{ |
| 58 |
.apply = lv_morph_alpha_apply, |
| 59 |
.vidoptions.depth = |
| 60 |
VISUAL_VIDEO_DEPTH_8BIT | |
| 61 |
VISUAL_VIDEO_DEPTH_16BIT | |
| 62 |
VISUAL_VIDEO_DEPTH_24BIT | |
| 63 |
VISUAL_VIDEO_DEPTH_32BIT |
| 64 |
}}; |
| 65 |
|
| 66 |
static VisPluginInfo info[] = {{ |
| 67 |
.type = VISUAL_PLUGIN_TYPE_MORPH, |
| 68 |
|
| 69 |
.plugname = "alphablend", |
| 70 |
.name = "alphablend morph", |
| 71 |
.author = "Dennis Smit <[email protected]>", |
| 72 |
.version = "0.1", |
| 73 |
.about = "An alphablend morph plugin", |
| 74 |
.help = "This morph plugin morphs between two video sources using the alphablend method", |
| 75 |
.license = VISUAL_PLUGIN_LICENSE_LGPL, |
| 76 |
|
| 77 |
.init = lv_morph_alpha_init, |
| 78 |
.cleanup = lv_morph_alpha_cleanup, |
| 79 |
|
| 80 |
.plugin = VISUAL_OBJECT (&morph[0]) |
| 81 |
}}; |
| 82 |
|
| 83 |
*count = sizeof (info) / sizeof (*info); |
| 84 |
|
| 85 |
return info; |
| 86 |
} |
| 87 |
|
| 88 |
int lv_morph_alpha_init (VisPluginData *plugin) |
| 89 |
{ |
| 90 |
return 0; |
| 91 |
} |
| 92 |
|
| 93 |
int lv_morph_alpha_cleanup (VisPluginData *plugin) |
| 94 |
{ |
| 95 |
return 0; |
| 96 |
} |
| 97 |
|
| 98 |
int lv_morph_alpha_apply (VisPluginData *plugin, float rate, VisAudio *audio, VisVideo *dest, VisVideo *src1, VisVideo *src2) |
| 99 |
{ |
| 100 |
visual_log_return_val_if_fail (dest != NULL, -1); |
| 101 |
visual_log_return_val_if_fail (src1 != NULL, -1); |
| 102 |
visual_log_return_val_if_fail (src2 != NULL, -1); |
| 103 |
|
| 104 |
alpha_blend_buffer (visual_video_get_pixels (dest), |
| 105 |
visual_video_get_pixels (src1), |
| 106 |
visual_video_get_pixels (src2), |
| 107 |
visual_video_get_size (dest), dest->depth, rate); |
| 108 |
|
| 109 |
return 0; |
| 110 |
} |
| 111 |
|
| 112 |
static inline int alpha_blend_buffer (uint8_t *dest, uint8_t *src1, uint8_t *src2, int size, int depth, float alpha) |
| 113 |
{ |
| 114 |
|
| 115 |
if (visual_cpu_get_mmx ()) { |
| 116 |
if (depth == VISUAL_VIDEO_DEPTH_8BIT) |
| 117 |
return alpha_blend_8_mmx (dest, src1, src2, size, alpha); |
| 118 |
|
| 119 |
if (depth == VISUAL_VIDEO_DEPTH_16BIT) |
| 120 |
return alpha_blend_16_mmx (dest, src1, src2, size, alpha); |
| 121 |
|
| 122 |
if (depth == VISUAL_VIDEO_DEPTH_24BIT) |
| 123 |
return alpha_blend_24_mmx (dest, src1, src2, size, alpha); |
| 124 |
|
| 125 |
if (depth == VISUAL_VIDEO_DEPTH_32BIT) |
| 126 |
return alpha_blend_32_mmx (dest, src1, src2, size, alpha); |
| 127 |
} else { |
| 128 |
if (depth == VISUAL_VIDEO_DEPTH_8BIT) |
| 129 |
return alpha_blend_8_c (dest, src1, src2, size, alpha); |
| 130 |
|
| 131 |
if (depth == VISUAL_VIDEO_DEPTH_16BIT) |
| 132 |
return alpha_blend_16_c (dest, src1, src2, size, alpha); |
| 133 |
|
| 134 |
if (depth == VISUAL_VIDEO_DEPTH_24BIT) |
| 135 |
return alpha_blend_24_c (dest, src1, src2, size, alpha); |
| 136 |
|
| 137 |
if (depth == VISUAL_VIDEO_DEPTH_32BIT) |
| 138 |
return alpha_blend_32_c (dest, src1, src2, size, alpha); |
| 139 |
|
| 140 |
} |
| 141 |
|
| 142 |
return -1; |
| 143 |
} |
| 144 |
|
| 145 |
/* FIXME TODO blends: c sse mmx |
| 146 |
* 8 x x |
| 147 |
* 16 x |
| 148 |
* 24 x |
| 149 |
* 32 x x |
| 150 |
*/ |
| 151 |
|
| 152 |
static inline int alpha_blend_8_c (uint8_t *dest, uint8_t *src1, uint8_t *src2, int size, float alpha) |
| 153 |
{ |
| 154 |
uint8_t ialpha = (alpha * 255); |
| 155 |
int i; |
| 156 |
|
| 157 |
for (i = 0; i < size; i++) { |
| 158 |
dest[i] = (ialpha * (src2[i] - src1[i])) / 255 + src1[i]; |
| 159 |
} |
| 160 |
|
| 161 |
return 0; |
| 162 |
} |
| 163 |
|
| 164 |
static inline int alpha_blend_16_c (uint8_t *dest, uint8_t *src1, uint8_t *src2, int size, float alpha) |
| 165 |
{ |
| 166 |
uint8_t ialpha = (alpha * 255); |
| 167 |
_color16 *destr = (_color16 *) dest; |
| 168 |
_color16 *src1r = (_color16 *) src1; |
| 169 |
_color16 *src2r = (_color16 *) src2; |
| 170 |
int i; |
| 171 |
|
| 172 |
for (i = 0; i < size / 2; i++) { |
| 173 |
destr[i].r = ((ialpha * (src2r[i].r - src1r[i].r)) / 255 + src1r[i].r); |
| 174 |
destr[i].g = ((ialpha * (src2r[i].g - src1r[i].g)) / 255 + src1r[i].g); |
| 175 |
destr[i].b = ((ialpha * (src2r[i].b - src1r[i].b)) / 255 + src1r[i].b); |
| 176 |
} |
| 177 |
|
| 178 |
return 0; |
| 179 |
} |
| 180 |
|
| 181 |
static inline int alpha_blend_24_c (uint8_t *dest, uint8_t *src1, uint8_t *src2, int size, float alpha) |
| 182 |
{ |
| 183 |
uint8_t ialpha = (alpha * 255); |
| 184 |
int i; |
| 185 |
|
| 186 |
for (i = 0; i < size; i++) { |
| 187 |
dest[i] = (ialpha * (src2[i] - src1[i])) / 255 + src1[i]; |
| 188 |
} |
| 189 |
|
| 190 |
return 0; |
| 191 |
} |
| 192 |
|
| 193 |
static inline int alpha_blend_32_c (uint8_t *dest, uint8_t *src1, uint8_t *src2, int size, float alpha) |
| 194 |
{ |
| 195 |
uint8_t ialpha = (alpha * 255); |
| 196 |
int i; |
| 197 |
|
| 198 |
for (i = 0; i < size; i++) { |
| 199 |
dest[i] = (ialpha * (src2[i] - src1[i])) / 255 + src1[i]; |
| 200 |
} |
| 201 |
|
| 202 |
return 0; |
| 203 |
} |
| 204 |
|
| 205 |
static inline int alpha_blend_8_mmx (uint8_t *dest, uint8_t *src1, uint8_t *src2, int size, float alpha) |
| 206 |
{ |
| 207 |
uint32_t aalpha = (alpha * 255); |
| 208 |
uint32_t ialpha = aalpha; |
| 209 |
int i; |
| 210 |
|
| 211 |
ialpha |= ialpha << 16; |
| 212 |
|
| 213 |
#ifdef VISUAL_ARCH_X86 |
| 214 |
__asm __volatile |
| 215 |
("\n\t pxor %%mm6, %%mm6" |
| 216 |
::); |
| 217 |
|
| 218 |
for (i = size; i > 4; i -= 4) { |
| 219 |
__asm __volatile |
| 220 |
("\n\t movd %[alpha], %%mm3" |
| 221 |
"\n\t movd %[src2], %%mm0" |
| 222 |
"\n\t psllq $32, %%mm3" |
| 223 |
"\n\t movd %[alpha], %%mm2" |
| 224 |
"\n\t movd %[src1], %%mm1" |
| 225 |
"\n\t por %%mm3, %%mm2" |
| 226 |
"\n\t punpcklbw %%mm6, %%mm0" /* interleaving dest */ |
| 227 |
"\n\t punpcklbw %%mm6, %%mm1" /* interleaving source */ |
| 228 |
"\n\t psubsw %%mm1, %%mm0" /* (src - dest) part */ |
| 229 |
"\n\t pmullw %%mm2, %%mm0" /* alpha * (src - dest) */ |
| 230 |
"\n\t psrlw $8, %%mm0" /* / 256 */ |
| 231 |
"\n\t paddb %%mm1, %%mm0" /* + dest */ |
| 232 |
"\n\t packuswb %%mm0, %%mm0" |
| 233 |
"\n\t movd %%mm0, %[dest]" |
| 234 |
: [dest] "=m" (*(dest + i)) |
| 235 |
: [src1] "m" (*(src1 + i)) |
| 236 |
, [src2] "m" (*(src2 + i)) |
| 237 |
, [alpha] "m" (ialpha)); |
| 238 |
} |
| 239 |
|
| 240 |
while (i--) |
| 241 |
dest[i] = (aalpha * (src2[i] - src1[i])) / 255 + src1[i]; |
| 242 |
|
| 243 |
__asm __volatile |
| 244 |
("\n\t emms"); |
| 245 |
#endif |
| 246 |
|
| 247 |
return 0; |
| 248 |
} |
| 249 |
|
| 250 |
static inline int alpha_blend_16_mmx (uint8_t *dest, uint8_t *src1, uint8_t *src2, int size, float alpha) |
| 251 |
{ |
| 252 |
uint8_t ialpha = (alpha * 255); |
| 253 |
_color16 *destr = (_color16 *) dest; |
| 254 |
_color16 *src1r = (_color16 *) src1; |
| 255 |
_color16 *src2r = (_color16 *) src2; |
| 256 |
int i; |
| 257 |
|
| 258 |
for (i = 0; i < size / 2; i++) { |
| 259 |
destr[i].r = ((ialpha * (src2r[i].r - src1r[i].r)) / 255 + src1r[i].r); |
| 260 |
destr[i].g = ((ialpha * (src2r[i].g - src1r[i].g)) / 255 + src1r[i].g); |
| 261 |
destr[i].b = ((ialpha * (src2r[i].b - src1r[i].b)) / 255 + src1r[i].b); |
| 262 |
} |
| 263 |
|
| 264 |
return 0; |
| 265 |
} |
| 266 |
|
| 267 |
static inline int alpha_blend_24_mmx (uint8_t *dest, uint8_t *src1, uint8_t *src2, int size, float alpha) |
| 268 |
{ |
| 269 |
uint8_t ialpha = (alpha * 255); |
| 270 |
int i; |
| 271 |
|
| 272 |
for (i = 0; i < size; i++) { |
| 273 |
dest[i] = (ialpha * (src2[i] - src1[i])) / 255 + src1[i]; |
| 274 |
} |
| 275 |
|
| 276 |
return 0; |
| 277 |
} |
| 278 |
|
| 279 |
static inline int alpha_blend_32_mmx (uint8_t *dest, uint8_t *src1, uint8_t *src2, int size, float alpha) |
| 280 |
{ |
| 281 |
uint32_t ialpha = (alpha * 255); |
| 282 |
int i; |
| 283 |
|
| 284 |
#ifdef VISUAL_ARCH_X86 |
| 285 |
__asm __volatile |
| 286 |
("\n\t pxor %%mm6, %%mm6" |
| 287 |
::); |
| 288 |
|
| 289 |
for (i = 0; i < size; i += 4) { |
| 290 |
__asm __volatile |
| 291 |
("\n\t movd %[src2], %%mm0" |
| 292 |
"\n\t movd %[alpha], %%mm2" |
| 293 |
"\n\t movd %[src1], %%mm1" |
| 294 |
"\n\t psllq $32, %%mm2" |
| 295 |
"\n\t movd %[alpha], %%mm3" |
| 296 |
"\n\t movd %[alpha], %%mm4" |
| 297 |
"\n\t psllq $16, %%mm3" |
| 298 |
"\n\t por %%mm4, %%mm2" |
| 299 |
"\n\t punpcklbw %%mm6, %%mm0" /* interleaving dest */ |
| 300 |
"\n\t por %%mm3, %%mm2" |
| 301 |
"\n\t punpcklbw %%mm6, %%mm1" /* interleaving source */ |
| 302 |
"\n\t psubsw %%mm1, %%mm0" /* (src - dest) part */ |
| 303 |
"\n\t pmullw %%mm2, %%mm0" /* alpha * (src - dest) */ |
| 304 |
"\n\t psrlw $8, %%mm0" /* / 256 */ |
| 305 |
"\n\t paddb %%mm1, %%mm0" /* + dest */ |
| 306 |
"\n\t packuswb %%mm0, %%mm0" |
| 307 |
"\n\t movd %%mm0, %[dest]" |
| 308 |
: [dest] "=m" (*(dest + i)) |
| 309 |
: [src1] "m" (*(src1 + i)) |
| 310 |
, [src2] "m" (*(src2 + i)) |
| 311 |
, [alpha] "m" (ialpha)); |
| 312 |
} |
| 313 |
|
| 314 |
__asm __volatile |
| 315 |
("\n\t emms"); |
| 316 |
#endif |
| 317 |
|
| 318 |
return 0; |
| 319 |
} |
| 320 |
|