| 1 |
/***************************************************************************** |
| 2 |
* |
| 3 |
* XVID MPEG-4 VIDEO CODEC |
| 4 |
* - Native API implementation - |
| 5 |
* |
| 6 |
* This program is an implementation of a part of one or more MPEG-4 |
| 7 |
* Video tools as specified in ISO/IEC 14496-2 standard. Those intending |
| 8 |
* to use this software module in hardware or software products are |
| 9 |
* advised that its use may infringe existing patents or copyrights, and |
| 10 |
* any such use would be at such party's own risk. The original |
| 11 |
* developer of this software module and his/her company, and subsequent |
| 12 |
* editors and their companies, will have no liability for use of this |
| 13 |
* software or modifications or derivatives thereof. |
| 14 |
* |
| 15 |
* This program is free software ; you can redistribute it and/or modify |
| 16 |
* it under the terms of the GNU General Public License as published by |
| 17 |
* the Free Software Foundation ; either version 2 of the License, or |
| 18 |
* (at your option) any later version. |
| 19 |
* |
| 20 |
* This program is distributed in the hope that it will be useful, |
| 21 |
* but WITHOUT ANY WARRANTY ; without even the implied warranty of |
| 22 |
* MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the |
| 23 |
* GNU General Public License for more details. |
| 24 |
* |
| 25 |
* You should have received a copy of the GNU General Public License |
| 26 |
* along with this program ; if not, write to the Free Software |
| 27 |
* Foundation, Inc., 59 Temple Place, Suite 330, Boston, MA 02111-1307 USA |
| 28 |
* |
| 29 |
****************************************************************************/ |
| 30 |
/***************************************************************************** |
| 31 |
* |
| 32 |
* History |
| 33 |
* |
| 34 |
* - 17.03.2002 Added interpolate8x8_halfpel_hv_xmm |
| 35 |
* - 22.12.2001 API change: added xvid_init() - Isibaar |
| 36 |
* - 16.12.2001 inital version; (c)2001 peter ross <pross@cs.rmit.edu.au> |
| 37 |
* |
| 38 |
* $Id: xvid.c,v 1.17 2002-06-13 21:35:01 edgomez Exp $ |
| 39 |
* |
| 40 |
****************************************************************************/ |
| 41 |
|
| 42 |
#include "xvid.h" |
| 43 |
#include "decoder.h" |
| 44 |
#include "encoder.h" |
| 45 |
#include "bitstream/cbp.h" |
| 46 |
#include "dct/idct.h" |
| 47 |
#include "dct/fdct.h" |
| 48 |
#include "image/colorspace.h" |
| 49 |
#include "image/interpolate8x8.h" |
| 50 |
#include "utils/mem_transfer.h" |
| 51 |
#include "quant/quant_h263.h" |
| 52 |
#include "quant/quant_mpeg4.h" |
| 53 |
#include "motion/sad.h" |
| 54 |
#include "utils/emms.h" |
| 55 |
#include "utils/timer.h" |
| 56 |
#include "bitstream/mbcoding.h" |
| 57 |
|
| 58 |
/***************************************************************************** |
| 59 |
* XviD Init Entry point |
| 60 |
* |
| 61 |
* Well this function initialize all internal function pointers according |
| 62 |
* to the CPU features forced by the library client or autodetected (depending |
| 63 |
* on the XVID_CPU_FORCE flag). It also initializes vlc coding tables and all |
| 64 |
* image colorspace transformation tables. |
| 65 |
* |
| 66 |
* Returned value : XVID_ERR_OK |
| 67 |
* + API_VERSION in the input XVID_INIT_PARAM structure |
| 68 |
* + core build " " " " " |
| 69 |
* |
| 70 |
****************************************************************************/ |
| 71 |
|
| 72 |
int |
| 73 |
xvid_init(void *handle, |
| 74 |
int opt, |
| 75 |
void *param1, |
| 76 |
void *param2) |
| 77 |
{ |
| 78 |
int cpu_flags; |
| 79 |
XVID_INIT_PARAM *init_param; |
| 80 |
|
| 81 |
init_param = (XVID_INIT_PARAM *) param1; |
| 82 |
|
| 83 |
/* Do we have to force CPU features ? */ |
| 84 |
if ((init_param->cpu_flags & XVID_CPU_FORCE) > 0) { |
| 85 |
cpu_flags = init_param->cpu_flags; |
| 86 |
} else { |
| 87 |
|
| 88 |
#ifdef ARCH_X86 |
| 89 |
cpu_flags = check_cpu_features(); |
| 90 |
#else |
| 91 |
cpu_flags = 0; |
| 92 |
#endif |
| 93 |
init_param->cpu_flags = cpu_flags; |
| 94 |
} |
| 95 |
|
| 96 |
/* Initialize the function pointers */ |
| 97 |
idct_int32_init(); |
| 98 |
init_vlc_tables(); |
| 99 |
|
| 100 |
/* Fixed Point Forward/Inverse DCT transformations */ |
| 101 |
fdct = fdct_int32; |
| 102 |
idct = idct_int32; |
| 103 |
|
| 104 |
/* Only needed on PPC Altivec archs */ |
| 105 |
sadInit = 0; |
| 106 |
|
| 107 |
/* Restore FPU context : emms_c is a nop functions */ |
| 108 |
emms = emms_c; |
| 109 |
|
| 110 |
/* Quantization functions */ |
| 111 |
quant_intra = quant_intra_c; |
| 112 |
dequant_intra = dequant_intra_c; |
| 113 |
quant_inter = quant_inter_c; |
| 114 |
dequant_inter = dequant_inter_c; |
| 115 |
|
| 116 |
quant4_intra = quant4_intra_c; |
| 117 |
dequant4_intra = dequant4_intra_c; |
| 118 |
quant4_inter = quant4_inter_c; |
| 119 |
dequant4_inter = dequant4_inter_c; |
| 120 |
|
| 121 |
/* Block transfer related functions */ |
| 122 |
transfer_8to16copy = transfer_8to16copy_c; |
| 123 |
transfer_16to8copy = transfer_16to8copy_c; |
| 124 |
transfer_8to16sub = transfer_8to16sub_c; |
| 125 |
transfer_8to16sub2 = transfer_8to16sub2_c; |
| 126 |
transfer_16to8add = transfer_16to8add_c; |
| 127 |
transfer8x8_copy = transfer8x8_copy_c; |
| 128 |
|
| 129 |
/* Image interpolation related functions */ |
| 130 |
interpolate8x8_halfpel_h = interpolate8x8_halfpel_h_c; |
| 131 |
interpolate8x8_halfpel_v = interpolate8x8_halfpel_v_c; |
| 132 |
interpolate8x8_halfpel_hv = interpolate8x8_halfpel_hv_c; |
| 133 |
|
| 134 |
/* Initialize internal colorspace transformation tables */ |
| 135 |
colorspace_init(); |
| 136 |
|
| 137 |
/* All colorspace transformation functions User Format->YV12 */ |
| 138 |
rgb555_to_yv12 = rgb555_to_yv12_c; |
| 139 |
rgb565_to_yv12 = rgb565_to_yv12_c; |
| 140 |
rgb24_to_yv12 = rgb24_to_yv12_c; |
| 141 |
rgb32_to_yv12 = rgb32_to_yv12_c; |
| 142 |
yuv_to_yv12 = yuv_to_yv12_c; |
| 143 |
yuyv_to_yv12 = yuyv_to_yv12_c; |
| 144 |
uyvy_to_yv12 = uyvy_to_yv12_c; |
| 145 |
|
| 146 |
/* All colorspace transformation functions YV12->User format */ |
| 147 |
yv12_to_rgb555 = yv12_to_rgb555_c; |
| 148 |
yv12_to_rgb565 = yv12_to_rgb565_c; |
| 149 |
yv12_to_rgb24 = yv12_to_rgb24_c; |
| 150 |
yv12_to_rgb32 = yv12_to_rgb32_c; |
| 151 |
yv12_to_yuv = yv12_to_yuv_c; |
| 152 |
yv12_to_yuyv = yv12_to_yuyv_c; |
| 153 |
yv12_to_uyvy = yv12_to_uyvy_c; |
| 154 |
|
| 155 |
/* Functions used in motion estimation algorithms */ |
| 156 |
calc_cbp = calc_cbp_c; |
| 157 |
sad16 = sad16_c; |
| 158 |
sad16bi = sad16bi_c; |
| 159 |
sad8 = sad8_c; |
| 160 |
dev16 = dev16_c; |
| 161 |
|
| 162 |
#ifdef ARCH_X86 |
| 163 |
if ((cpu_flags & XVID_CPU_MMX) > 0) { |
| 164 |
|
| 165 |
/* Forward and Inverse Discrete Cosine Transformation functions */ |
| 166 |
fdct = fdct_mmx; |
| 167 |
idct = idct_mmx; |
| 168 |
|
| 169 |
/* To restore FPU context after mmx use */ |
| 170 |
emms = emms_mmx; |
| 171 |
|
| 172 |
/* Quantization related functions */ |
| 173 |
quant_intra = quant_intra_mmx; |
| 174 |
dequant_intra = dequant_intra_mmx; |
| 175 |
quant_inter = quant_inter_mmx; |
| 176 |
dequant_inter = dequant_inter_mmx; |
| 177 |
|
| 178 |
quant4_intra = quant4_intra_mmx; |
| 179 |
dequant4_intra = dequant4_intra_mmx; |
| 180 |
quant4_inter = quant4_inter_mmx; |
| 181 |
dequant4_inter = dequant4_inter_mmx; |
| 182 |
|
| 183 |
/* Block related functions */ |
| 184 |
transfer_8to16copy = transfer_8to16copy_mmx; |
| 185 |
transfer_16to8copy = transfer_16to8copy_mmx; |
| 186 |
transfer_8to16sub = transfer_8to16sub_mmx; |
| 187 |
transfer_16to8add = transfer_16to8add_mmx; |
| 188 |
transfer8x8_copy = transfer8x8_copy_mmx; |
| 189 |
|
| 190 |
/* Image Interpolation related functions */ |
| 191 |
interpolate8x8_halfpel_h = interpolate8x8_halfpel_h_mmx; |
| 192 |
interpolate8x8_halfpel_v = interpolate8x8_halfpel_v_mmx; |
| 193 |
interpolate8x8_halfpel_hv = interpolate8x8_halfpel_hv_mmx; |
| 194 |
|
| 195 |
/* Image RGB->YV12 related functions */ |
| 196 |
rgb24_to_yv12 = rgb24_to_yv12_mmx; |
| 197 |
rgb32_to_yv12 = rgb32_to_yv12_mmx; |
| 198 |
yuv_to_yv12 = yuv_to_yv12_mmx; |
| 199 |
yuyv_to_yv12 = yuyv_to_yv12_mmx; |
| 200 |
uyvy_to_yv12 = uyvy_to_yv12_mmx; |
| 201 |
|
| 202 |
/* Image YV12->RGB related functions */ |
| 203 |
yv12_to_rgb24 = yv12_to_rgb24_mmx; |
| 204 |
yv12_to_rgb32 = yv12_to_rgb32_mmx; |
| 205 |
yv12_to_yuyv = yv12_to_yuyv_mmx; |
| 206 |
yv12_to_uyvy = yv12_to_uyvy_mmx; |
| 207 |
|
| 208 |
/* Motion estimation related functions */ |
| 209 |
calc_cbp = calc_cbp_mmx; |
| 210 |
sad16 = sad16_mmx; |
| 211 |
sad8 = sad8_mmx; |
| 212 |
dev16 = dev16_mmx; |
| 213 |
|
| 214 |
} |
| 215 |
|
| 216 |
if ((cpu_flags & XVID_CPU_MMXEXT) > 0) { |
| 217 |
|
| 218 |
/* Inverse DCT */ |
| 219 |
idct = idct_xmm; |
| 220 |
|
| 221 |
/* Interpolation */ |
| 222 |
interpolate8x8_halfpel_h = interpolate8x8_halfpel_h_xmm; |
| 223 |
interpolate8x8_halfpel_v = interpolate8x8_halfpel_v_xmm; |
| 224 |
interpolate8x8_halfpel_hv = interpolate8x8_halfpel_hv_xmm; |
| 225 |
|
| 226 |
/* Colorspace transformation */ |
| 227 |
yuv_to_yv12 = yuv_to_yv12_xmm; |
| 228 |
|
| 229 |
/* ME functions */ |
| 230 |
sad16 = sad16_xmm; |
| 231 |
sad8 = sad8_xmm; |
| 232 |
dev16 = dev16_xmm; |
| 233 |
|
| 234 |
} |
| 235 |
|
| 236 |
if ((cpu_flags & XVID_CPU_3DNOW) > 0) { |
| 237 |
|
| 238 |
/* Interpolation */ |
| 239 |
interpolate8x8_halfpel_h = interpolate8x8_halfpel_h_3dn; |
| 240 |
interpolate8x8_halfpel_v = interpolate8x8_halfpel_v_3dn; |
| 241 |
interpolate8x8_halfpel_hv = interpolate8x8_halfpel_hv_3dn; |
| 242 |
} |
| 243 |
|
| 244 |
if ((cpu_flags & XVID_CPU_SSE2) > 0) { |
| 245 |
#ifdef EXPERIMENTAL_SSE2_CODE |
| 246 |
|
| 247 |
/* Quantization */ |
| 248 |
quant_intra = quant_intra_sse2; |
| 249 |
dequant_intra = dequant_intra_sse2; |
| 250 |
quant_inter = quant_inter_sse2; |
| 251 |
dequant_inter = dequant_inter_sse2; |
| 252 |
|
| 253 |
/* ME */ |
| 254 |
calc_cbp = calc_cbp_sse2; |
| 255 |
sad16 = sad16_sse2; |
| 256 |
dev16 = dev16_sse2; |
| 257 |
|
| 258 |
/* Forward and Inverse DCT */ |
| 259 |
idct = idct_sse2; |
| 260 |
fdct = fdct_sse2; |
| 261 |
#endif |
| 262 |
} |
| 263 |
|
| 264 |
#endif |
| 265 |
|
| 266 |
#ifdef ARCH_PPC |
| 267 |
#ifdef ARCH_PPC_ALTIVEC |
| 268 |
calc_cbp = calc_cbp_altivec; |
| 269 |
fdct = fdct_altivec; |
| 270 |
idct = idct_altivec; |
| 271 |
sadInit = sadInit_altivec; |
| 272 |
sad16 = sad16_altivec; |
| 273 |
sad8 = sad8_altivec; |
| 274 |
dev16 = dev16_altivec; |
| 275 |
#else |
| 276 |
calc_cbp = calc_cbp_ppc; |
| 277 |
#endif |
| 278 |
#endif |
| 279 |
|
| 280 |
/* Inform the client the API version */ |
| 281 |
init_param->api_version = API_VERSION; |
| 282 |
|
| 283 |
/* Inform the client the core build - unused because we're still alpha */ |
| 284 |
init_param->core_build = 1000; |
| 285 |
|
| 286 |
return XVID_ERR_OK; |
| 287 |
} |
| 288 |
|
| 289 |
/***************************************************************************** |
| 290 |
* XviD Native decoder entry point |
| 291 |
* |
| 292 |
* This function is just a wrapper to all the option cases. |
| 293 |
* |
| 294 |
* Returned values : XVID_ERR_FAIL when opt is invalid |
| 295 |
* else returns the wrapped function result |
| 296 |
* |
| 297 |
****************************************************************************/ |
| 298 |
|
| 299 |
int |
| 300 |
xvid_decore(void *handle, |
| 301 |
int opt, |
| 302 |
void *param1, |
| 303 |
void *param2) |
| 304 |
{ |
| 305 |
switch (opt) { |
| 306 |
case XVID_DEC_DECODE: |
| 307 |
return decoder_decode((DECODER *) handle, (XVID_DEC_FRAME *) param1); |
| 308 |
|
| 309 |
case XVID_DEC_CREATE: |
| 310 |
return decoder_create((XVID_DEC_PARAM *) param1); |
| 311 |
|
| 312 |
case XVID_DEC_DESTROY: |
| 313 |
return decoder_destroy((DECODER *) handle); |
| 314 |
|
| 315 |
default: |
| 316 |
return XVID_ERR_FAIL; |
| 317 |
} |
| 318 |
} |
| 319 |
|
| 320 |
|
| 321 |
/***************************************************************************** |
| 322 |
* XviD Native encoder entry point |
| 323 |
* |
| 324 |
* This function is just a wrapper to all the option cases. |
| 325 |
* |
| 326 |
* Returned values : XVID_ERR_FAIL when opt is invalid |
| 327 |
* else returns the wrapped function result |
| 328 |
* |
| 329 |
****************************************************************************/ |
| 330 |
|
| 331 |
int |
| 332 |
xvid_encore(void *handle, |
| 333 |
int opt, |
| 334 |
void *param1, |
| 335 |
void *param2) |
| 336 |
{ |
| 337 |
switch (opt) { |
| 338 |
case XVID_ENC_ENCODE: |
| 339 |
return encoder_encode((Encoder *) handle, (XVID_ENC_FRAME *) param1, |
| 340 |
(XVID_ENC_STATS *) param2); |
| 341 |
|
| 342 |
case XVID_ENC_CREATE: |
| 343 |
return encoder_create((XVID_ENC_PARAM *) param1); |
| 344 |
|
| 345 |
case XVID_ENC_DESTROY: |
| 346 |
return encoder_destroy((Encoder *) handle); |
| 347 |
|
| 348 |
default: |
| 349 |
return XVID_ERR_FAIL; |
| 350 |
} |
| 351 |
} |
