[cvs] / xvidcore / src / xvid.c Repository:
ViewVC logotype

Diff of /xvidcore/src/xvid.c

Parent Directory Parent Directory | Revision Log Revision Log | View Patch Patch

revision 1.33.2.19, Mon Dec 30 10:49:17 2002 UTC revision 1.39, Sat Nov 16 23:38:16 2002 UTC
# Line 3  Line 3 
3   *  XVID MPEG-4 VIDEO CODEC   *  XVID MPEG-4 VIDEO CODEC
4   *  - Native API implementation  -   *  - Native API implementation  -
5   *   *
6   *  This program is an implementation of a part of one or more MPEG-4   *  Copyright(C) 2001-2002 Peter Ross <pross@xvid.org>
  *  Video tools as specified in ISO/IEC 14496-2 standard.  Those intending  
  *  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.  
7   *   *
8   *  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
9   *  it under the terms of the GNU General Public License as published by   *
10     *  XviD is free software; you can redistribute it and/or modify it
11     *  under the terms of the GNU General Public License as published by
12   *  the Free Software Foundation ; either version 2 of the License, or   *  the Free Software Foundation ; either version 2 of the License, or
13   *  (at your option) any later version.   *  (at your option) any later version.
14   *   *
# Line 26  Line 21 
21   *  along with this program ; if not, write to the Free Software   *  along with this program ; if not, write to the Free Software
22   *  Foundation, Inc., 59 Temple Place, Suite 330, Boston, MA  02111-1307 USA   *  Foundation, Inc., 59 Temple Place, Suite 330, Boston, MA  02111-1307 USA
23   *   *
24   ****************************************************************************/   *  Under section 8 of the GNU General Public License, the copyright
25     *  holders of XVID explicitly forbid distribution in the following
26  /*****************************************************************************   *  countries:
27   *   *
28   *  History   *    - Japan
29   *   *    - United States of America
30   *      - 23.06.2002    added XVID_CPU_CHKONLY   *
31   *  - 17.03.2002        Added interpolate8x8_halfpel_hv_xmm   *  Linking XviD statically or dynamically with other modules is making a
32   *  - 22.12.2001  API change: added xvid_init() - Isibaar   *  combined work based on XviD.  Thus, the terms and conditions of the
33   *  - 16.12.2001        inital version; (c)2001 peter ross <pross@cs.rmit.edu.au>   *  GNU General Public License cover the whole combination.
34     *
35     *  As a special exception, the copyright holders of XviD give you
36     *  permission to link XviD with independent modules that communicate with
37     *  XviD solely through the VFW1.1 and DShow interfaces, regardless of the
38     *  license terms of these independent modules, and to copy and distribute
39     *  the resulting combined work under terms of your choice, provided that
40     *  every copy of the combined work is accompanied by a complete copy of
41     *  the source code of XviD (the version of XviD used to produce the
42     *  combined work), being distributed under the terms of the GNU General
43     *  Public License plus this exception.  An independent module is a module
44     *  which is not derived from or based on XviD.
45     *
46     *  Note that people who make modified versions of XviD are not obligated
47     *  to grant this special exception for their modified versions; it is
48     *  their choice whether to do so.  The GNU General Public License gives
49     *  permission to release a modified version without this exception; this
50     *  exception also makes it possible to release a modified version which
51     *  carries forward this exception.
52   *   *
53   *  $Id$   *  $Id$
54   *   *
# Line 49  Line 62 
62  #include "dct/fdct.h"  #include "dct/fdct.h"
63  #include "image/colorspace.h"  #include "image/colorspace.h"
64  #include "image/interpolate8x8.h"  #include "image/interpolate8x8.h"
 #include "image/reduced.h"  
65  #include "utils/mem_transfer.h"  #include "utils/mem_transfer.h"
 #include "utils/mbfunctions.h"  
66  #include "quant/quant_h263.h"  #include "quant/quant_h263.h"
67  #include "quant/quant_mpeg4.h"  #include "quant/quant_mpeg4.h"
68  #include "motion/motion.h"  #include "motion/motion.h"
# Line 83  Line 94 
94    
95    
96  /*  /*
97  calls the funcptr, and returns whether SIGILL (illegal instruction) was signalled   * Calls the funcptr, and returns whether SIGILL (illegal instruction) was signalled
98  return values:   * Return values:
99  -1 : could not determine   * -1 : could not determine
100  0  : SIGILL was *not* signalled   * 0  : SIGILL was *not* signalled
101  1  : SIGILL was signalled   * 1  : SIGILL was signalled
102  */  */
103    
104  int  int
# Line 141  Line 152 
152   *   *
153   ****************************************************************************/   ****************************************************************************/
154    
155    int
156  static  xvid_init(void *handle,
157  int xvid_init_init(XVID_INIT_PARAM * init_param)                    int opt,
158                      void *param1,
159                      void *param2)
160  {  {
161          int cpu_flags;          int cpu_flags;
162            XVID_INIT_PARAM *init_param;
163    
164            init_param = (XVID_INIT_PARAM *) param1;
165    
166          /* Inform the client the API version */          /* Inform the client the API version */
167          init_param->api_version = API_VERSION;          init_param->api_version = API_VERSION;
# Line 213  Line 229 
229          transfer_16to8add  = transfer_16to8add_c;          transfer_16to8add  = transfer_16to8add_c;
230          transfer8x8_copy   = transfer8x8_copy_c;          transfer8x8_copy   = transfer8x8_copy_c;
231    
         /* Interlacing functions */  
         MBFieldTest = MBFieldTest_c;  
   
232          /* Image interpolation related functions */          /* Image interpolation related functions */
233          interpolate8x8_halfpel_h  = interpolate8x8_halfpel_h_c;          interpolate8x8_halfpel_h  = interpolate8x8_halfpel_h_c;
234          interpolate8x8_halfpel_v  = interpolate8x8_halfpel_v_c;          interpolate8x8_halfpel_v  = interpolate8x8_halfpel_v_c;
235          interpolate8x8_halfpel_hv = interpolate8x8_halfpel_hv_c;          interpolate8x8_halfpel_hv = interpolate8x8_halfpel_hv_c;
236    
         interpolate16x16_lowpass_h = interpolate16x16_lowpass_h_c;  
         interpolate16x16_lowpass_v = interpolate16x16_lowpass_v_c;  
         interpolate16x16_lowpass_hv = interpolate16x16_lowpass_hv_c;  
   
         interpolate8x8_lowpass_h = interpolate8x8_lowpass_h_c;  
         interpolate8x8_lowpass_v = interpolate8x8_lowpass_v_c;  
         interpolate8x8_lowpass_hv = interpolate8x8_lowpass_hv_c;  
   
         interpolate8x8_6tap_lowpass_h = interpolate8x8_6tap_lowpass_h_c;  
         interpolate8x8_6tap_lowpass_v = interpolate8x8_6tap_lowpass_v_c;  
   
         interpolate8x8_avg2 = interpolate8x8_avg2_c;  
         interpolate8x8_avg4 = interpolate8x8_avg4_c;  
   
         /* reduced resoltuion */  
   
         copy_upsampled_8x8_16to8 = xvid_Copy_Upsampled_8x8_16To8_C;  
         add_upsampled_8x8_16to8 = xvid_Add_Upsampled_8x8_16To8_C;  
 #ifdef ARCH_X86  
         vfilter_31 = xvid_VFilter_31_x86;  
         hfilter_31 = xvid_HFilter_31_x86;  
 #else  
         vfilter_31 = xvid_VFilter_31_C;  
         hfilter_31 = xvid_HFilter_31_C;  
 #endif  
         filter_18x18_to_8x8 = xvid_Filter_18x18_To_8x8_C;  
         filter_diff_18x18_to_8x8 = xvid_Filter_Diff_18x18_To_8x8_C;  
   
237          /* Initialize internal colorspace transformation tables */          /* Initialize internal colorspace transformation tables */
238          colorspace_init();          colorspace_init();
239    
240          /* All colorspace transformation functions User Format->YV12 */          /* All colorspace transformation functions User Format->YV12 */
         yv12_to_yv12    = yv12_to_yv12_c;  
241          rgb555_to_yv12  = rgb555_to_yv12_c;          rgb555_to_yv12  = rgb555_to_yv12_c;
242          rgb565_to_yv12  = rgb565_to_yv12_c;          rgb565_to_yv12  = rgb565_to_yv12_c;
243          bgr_to_yv12     = bgr_to_yv12_c;          rgb24_to_yv12  = rgb24_to_yv12_c;
244          bgra_to_yv12    = bgra_to_yv12_c;          rgb32_to_yv12  = rgb32_to_yv12_c;
245          abgr_to_yv12    = abgr_to_yv12_c;          yuv_to_yv12    = yuv_to_yv12_c;
         rgba_to_yv12    = rgba_to_yv12_c;  
246          yuyv_to_yv12    = yuyv_to_yv12_c;          yuyv_to_yv12    = yuyv_to_yv12_c;
247          uyvy_to_yv12    = uyvy_to_yv12_c;          uyvy_to_yv12    = uyvy_to_yv12_c;
248    
         rgb555i_to_yv12 = rgb555i_to_yv12_c;  
         rgb565i_to_yv12 = rgb565i_to_yv12_c;  
         bgri_to_yv12    = bgri_to_yv12_c;  
         bgrai_to_yv12   = bgrai_to_yv12_c;  
         abgri_to_yv12   = abgri_to_yv12_c;  
         rgbai_to_yv12   = rgbai_to_yv12_c;  
         yuyvi_to_yv12   = yuyvi_to_yv12_c;  
         uyvyi_to_yv12   = uyvyi_to_yv12_c;  
   
   
249          /* All colorspace transformation functions YV12->User format */          /* All colorspace transformation functions YV12->User format */
250          yv12_to_rgb555  = yv12_to_rgb555_c;          yv12_to_rgb555  = yv12_to_rgb555_c;
251          yv12_to_rgb565  = yv12_to_rgb565_c;          yv12_to_rgb565  = yv12_to_rgb565_c;
252          yv12_to_bgr     = yv12_to_bgr_c;          yv12_to_rgb24  = yv12_to_rgb24_c;
253          yv12_to_bgra    = yv12_to_bgra_c;          yv12_to_rgb32  = yv12_to_rgb32_c;
254          yv12_to_abgr    = yv12_to_abgr_c;          yv12_to_yuv    = yv12_to_yuv_c;
         yv12_to_rgba    = yv12_to_rgba_c;  
255          yv12_to_yuyv    = yv12_to_yuyv_c;          yv12_to_yuyv    = yv12_to_yuyv_c;
256          yv12_to_uyvy    = yv12_to_uyvy_c;          yv12_to_uyvy    = yv12_to_uyvy_c;
257    
         yv12_to_rgb555i = yv12_to_rgb555i_c;  
         yv12_to_rgb565i = yv12_to_rgb565i_c;  
         yv12_to_bgri    = yv12_to_bgri_c;  
         yv12_to_bgrai   = yv12_to_bgrai_c;  
         yv12_to_abgri   = yv12_to_abgri_c;  
         yv12_to_rgbai   = yv12_to_rgbai_c;  
         yv12_to_yuyvi   = yv12_to_yuyvi_c;  
         yv12_to_uyvyi   = yv12_to_uyvyi_c;  
   
258          /* Functions used in motion estimation algorithms */          /* Functions used in motion estimation algorithms */
259          calc_cbp = calc_cbp_c;          calc_cbp = calc_cbp_c;
260          sad16    = sad16_c;          sad16    = sad16_c;
# Line 299  Line 262 
262          sad16bi  = sad16bi_c;          sad16bi  = sad16bi_c;
263          sad8bi   = sad8bi_c;          sad8bi   = sad8bi_c;
264          dev16    = dev16_c;          dev16    = dev16_c;
         sad16v   = sad16v_c;  
265    
266  //      Halfpel8_Refine = Halfpel8_Refine_c;          Halfpel8_Refine = Halfpel8_Refine_c;
267    
268  #ifdef ARCH_X86  #ifdef ARCH_X86
   
         if ((cpu_flags & XVID_CPU_MMX) || (cpu_flags & XVID_CPU_MMXEXT) ||  
                 (cpu_flags & XVID_CPU_3DNOW) || (cpu_flags & XVID_CPU_3DNOWEXT) ||  
                 (cpu_flags & XVID_CPU_SSE) || (cpu_flags & XVID_CPU_SSE2))  
         {  
                 /* Restore FPU context : emms_c is a nop functions */  
                 emms = emms_mmx;  
         }  
   
269          if ((cpu_flags & XVID_CPU_MMX) > 0) {          if ((cpu_flags & XVID_CPU_MMX) > 0) {
270    
271                  /* Forward and Inverse Discrete Cosine Transformation functions */                  /* Forward and Inverse Discrete Cosine Transformation functions */
272                  fdct = fdct_mmx;                  fdct = fdct_mmx;
273                  idct = idct_mmx;                  idct = idct_mmx;
274    
275                    /* To restore FPU context after mmx use */
276                    emms = emms_mmx;
277    
278                  /* Quantization related functions */                  /* Quantization related functions */
279                  quant_intra   = quant_intra_mmx;                  quant_intra   = quant_intra_mmx;
280                  dequant_intra = dequant_intra_mmx;                  dequant_intra = dequant_intra_mmx;
# Line 338  Line 294 
294                  transfer_16to8add  = transfer_16to8add_mmx;                  transfer_16to8add  = transfer_16to8add_mmx;
295                  transfer8x8_copy   = transfer8x8_copy_mmx;                  transfer8x8_copy   = transfer8x8_copy_mmx;
296    
                 /* Interlacing Functions */  
                 MBFieldTest = MBFieldTest_mmx;  
297    
298                  /* Image Interpolation related functions */                  /* Image Interpolation related functions */
299                  interpolate8x8_halfpel_h  = interpolate8x8_halfpel_h_mmx;                  interpolate8x8_halfpel_h  = interpolate8x8_halfpel_h_mmx;
300                  interpolate8x8_halfpel_v  = interpolate8x8_halfpel_v_mmx;                  interpolate8x8_halfpel_v  = interpolate8x8_halfpel_v_mmx;
301                  interpolate8x8_halfpel_hv = interpolate8x8_halfpel_hv_mmx;                  interpolate8x8_halfpel_hv = interpolate8x8_halfpel_hv_mmx;
302    
303                  interpolate8x8_6tap_lowpass_h = interpolate8x8_6tap_lowpass_h_mmx;                  /* Image RGB->YV12 related functions */
304                  interpolate8x8_6tap_lowpass_v = interpolate8x8_6tap_lowpass_v_mmx;                  rgb24_to_yv12 = rgb24_to_yv12_mmx;
305                    rgb32_to_yv12 = rgb32_to_yv12_mmx;
306                  interpolate8x8_avg2 = interpolate8x8_avg2_mmx;                  yuv_to_yv12   = yuv_to_yv12_mmx;
                 interpolate8x8_avg4 = interpolate8x8_avg4_mmx;  
   
                 /* reduced resolution */  
                 copy_upsampled_8x8_16to8 = xvid_Copy_Upsampled_8x8_16To8_mmx;  
                 add_upsampled_8x8_16to8 = xvid_Add_Upsampled_8x8_16To8_mmx;  
                 hfilter_31 = xvid_HFilter_31_mmx;  
                 filter_18x18_to_8x8 = xvid_Filter_18x18_To_8x8_mmx;  
                 filter_diff_18x18_to_8x8 = xvid_Filter_Diff_18x18_To_8x8_mmx;  
   
                 /* image input xxx_to_yv12 related functions */  
                 yv12_to_yv12  = yv12_to_yv12_mmx;  
                 bgr_to_yv12   = bgr_to_yv12_mmx;  
                 bgra_to_yv12  = bgra_to_yv12_mmx;  
307                  yuyv_to_yv12  = yuyv_to_yv12_mmx;                  yuyv_to_yv12  = yuyv_to_yv12_mmx;
308                  uyvy_to_yv12  = uyvy_to_yv12_mmx;                  uyvy_to_yv12  = uyvy_to_yv12_mmx;
309    
310                  /* image output yv12_to_xxx related functions */                  /* Image YV12->RGB related functions */
311                  yv12_to_bgr   = yv12_to_bgr_mmx;                  yv12_to_rgb24 = yv12_to_rgb24_mmx;
312                  yv12_to_bgra  = yv12_to_bgra_mmx;                  yv12_to_rgb32 = yv12_to_rgb32_mmx;
313                  yv12_to_yuyv  = yv12_to_yuyv_mmx;                  yv12_to_yuyv  = yv12_to_yuyv_mmx;
314                  yv12_to_uyvy  = yv12_to_uyvy_mmx;                  yv12_to_uyvy  = yv12_to_uyvy_mmx;
315    
                 yv12_to_yuyvi = yv12_to_yuyvi_mmx;  
                 yv12_to_uyvyi = yv12_to_uyvyi_mmx;  
   
316                  /* Motion estimation related functions */                  /* Motion estimation related functions */
317                  calc_cbp = calc_cbp_mmx;                  calc_cbp = calc_cbp_mmx;
318                  sad16    = sad16_mmx;                  sad16    = sad16_mmx;
# Line 382  Line 320 
320                  sad16bi = sad16bi_mmx;                  sad16bi = sad16bi_mmx;
321                  sad8bi  = sad8bi_mmx;                  sad8bi  = sad8bi_mmx;
322                  dev16    = dev16_mmx;                  dev16    = dev16_mmx;
                 sad16v   = sad16v_mmx;  
323    
324          }          }
325    
# Line 392  Line 329 
329                  /* ME functions */                  /* ME functions */
330                  sad16bi = sad16bi_3dn;                  sad16bi = sad16bi_3dn;
331                  sad8bi  = sad8bi_3dn;                  sad8bi  = sad8bi_3dn;
   
                 yuyv_to_yv12  = yuyv_to_yv12_3dn;  
                 uyvy_to_yv12  = uyvy_to_yv12_3dn;  
332          }          }
333    
334    
# Line 408  Line 342 
342                  interpolate8x8_halfpel_v  = interpolate8x8_halfpel_v_xmm;                  interpolate8x8_halfpel_v  = interpolate8x8_halfpel_v_xmm;
343                  interpolate8x8_halfpel_hv = interpolate8x8_halfpel_hv_xmm;                  interpolate8x8_halfpel_hv = interpolate8x8_halfpel_hv_xmm;
344    
                 /* reduced resolution */  
                 copy_upsampled_8x8_16to8 = xvid_Copy_Upsampled_8x8_16To8_xmm;  
                 add_upsampled_8x8_16to8 = xvid_Add_Upsampled_8x8_16To8_xmm;  
   
345                  /* Quantization */                  /* Quantization */
                 quant4_intra = quant4_intra_xmm;  
                 quant4_inter = quant4_inter_xmm;  
   
346                  dequant_intra = dequant_intra_xmm;                  dequant_intra = dequant_intra_xmm;
347                  dequant_inter = dequant_inter_xmm;                  dequant_inter = dequant_inter_xmm;
348    
# Line 423  Line 350 
350                  transfer_8to16sub2 = transfer_8to16sub2_xmm;                  transfer_8to16sub2 = transfer_8to16sub2_xmm;
351    
352                  /* Colorspace transformation */                  /* Colorspace transformation */
353                  yv12_to_yv12  = yv12_to_yv12_xmm;                  yuv_to_yv12 = yuv_to_yv12_xmm;
                 yuyv_to_yv12  = yuyv_to_yv12_xmm;  
                 uyvy_to_yv12  = uyvy_to_yv12_xmm;  
354    
355                  /* ME functions */                  /* ME functions */
356                  sad16 = sad16_xmm;                  sad16 = sad16_xmm;
# Line 433  Line 358 
358                  sad16bi = sad16bi_xmm;                  sad16bi = sad16bi_xmm;
359                  sad8bi  = sad8bi_xmm;                  sad8bi  = sad8bi_xmm;
360                  dev16 = dev16_xmm;                  dev16 = dev16_xmm;
361                  sad16v   = sad16v_xmm;  
362          }          }
363    
364          if ((cpu_flags & XVID_CPU_3DNOW) > 0) {          if ((cpu_flags & XVID_CPU_3DNOW) > 0) {
# Line 444  Line 369 
369                  interpolate8x8_halfpel_hv = interpolate8x8_halfpel_hv_3dn;                  interpolate8x8_halfpel_hv = interpolate8x8_halfpel_hv_3dn;
370          }          }
371    
         if ((cpu_flags & XVID_CPU_3DNOWEXT) > 0) {  
   
                 /* Inverse DCT */  
                 idct =  idct_3dne;  
   
                 /* Buffer transfer */  
                 transfer_8to16copy =  transfer_8to16copy_3dne;  
                 transfer_16to8copy = transfer_16to8copy_3dne;  
                 transfer_8to16sub =  transfer_8to16sub_3dne;  
                 transfer_8to16sub2 =  transfer_8to16sub2_3dne;  
                 transfer_16to8add = transfer_16to8add_3dne;  
                 transfer8x8_copy = transfer8x8_copy_3dne;  
   
                 /* Quantization */  
                 dequant4_intra = dequant4_intra_3dne;  
                 dequant4_inter = dequant4_inter_3dne;  
                 quant_intra = quant_intra_3dne;  
                 quant_inter = quant_inter_3dne;  
                 dequant_intra = dequant_intra_3dne;  
                 dequant_inter = dequant_inter_3dne;  
   
                 /* ME functions */  
                 calc_cbp = calc_cbp_3dne;  
                 sad16 = sad16_3dne;  
                 sad8 = sad8_3dne;  
                 sad16bi = sad16bi_3dne;  
                 sad8bi = sad8bi_3dne;  
                 dev16 = dev16_3dne;  
   
                 /* Interpolation */  
                 interpolate8x8_halfpel_h = interpolate8x8_halfpel_h_3dne;  
                 interpolate8x8_halfpel_v = interpolate8x8_halfpel_v_3dne;  
                 interpolate8x8_halfpel_hv = interpolate8x8_halfpel_hv_3dne;  
         }  
   
   
372          if ((cpu_flags & XVID_CPU_SSE2) > 0) {          if ((cpu_flags & XVID_CPU_SSE2) > 0) {
373  #ifdef EXPERIMENTAL_SSE2_CODE  #ifdef EXPERIMENTAL_SSE2_CODE
374    
# Line 507  Line 396 
396          if ((cpu_flags & XVID_CPU_IA64) > 0) { //use assembler routines?          if ((cpu_flags & XVID_CPU_IA64) > 0) { //use assembler routines?
397            idct_ia64_init();            idct_ia64_init();
398            fdct = fdct_ia64;            fdct = fdct_ia64;
399            idct = idct_ia64;   //not yet working, crashes            idct = idct_ia64;
400            interpolate8x8_halfpel_h = interpolate8x8_halfpel_h_ia64;            interpolate8x8_halfpel_h = interpolate8x8_halfpel_h_ia64;
401            interpolate8x8_halfpel_v = interpolate8x8_halfpel_v_ia64;            interpolate8x8_halfpel_v = interpolate8x8_halfpel_v_ia64;
402            interpolate8x8_halfpel_hv = interpolate8x8_halfpel_hv_ia64;            interpolate8x8_halfpel_hv = interpolate8x8_halfpel_hv_ia64;
# Line 515  Line 404 
404            sad16bi = sad16bi_ia64;            sad16bi = sad16bi_ia64;
405            sad8 = sad8_ia64;            sad8 = sad8_ia64;
406            dev16 = dev16_ia64;            dev16 = dev16_ia64;
407  //        Halfpel8_Refine = Halfpel8_Refine_ia64;            Halfpel8_Refine = Halfpel8_Refine_ia64;
408            quant_intra = quant_intra_ia64;            quant_intra = quant_intra_ia64;
409            dequant_intra = dequant_intra_ia64;            dequant_intra = dequant_intra_ia64;
410            quant_inter = quant_inter_ia64;            quant_inter = quant_inter_ia64;
# Line 526  Line 415 
415            transfer_8to16sub2 = transfer_8to16sub2_ia64;            transfer_8to16sub2 = transfer_8to16sub2_ia64;
416            transfer_16to8add = transfer_16to8add_ia64;            transfer_16to8add = transfer_16to8add_ia64;
417            transfer8x8_copy = transfer8x8_copy_ia64;            transfer8x8_copy = transfer8x8_copy_ia64;
418            DEBUG("Using IA-64 assembler routines.\n");  //        DEBUG("Using IA-64 assembler routines.\n");
419          }          }
420  #endif  #endif
421    
# Line 547  Line 436 
436          return XVID_ERR_OK;          return XVID_ERR_OK;
437  }  }
438    
   
   
 static int  
 xvid_init_convert(XVID_INIT_CONVERTINFO* convert)  
 {  
         // const int flip1 = (convert->input.colorspace & XVID_CSP_VFLIP) ^ (convert->output.colorspace & XVID_CSP_VFLIP);  
         const int width = convert->width;  
         const int height = convert->height;  
         const int width2 = convert->width/2;  
         const int height2 = convert->height/2;  
         IMAGE img;  
   
         switch (convert->input.colorspace & ~XVID_CSP_VFLIP)  
         {  
                 case XVID_CSP_YV12 :  
                         img.y = convert->input.y;  
                         img.v = (uint8_t*)convert->input.y + width*height;  
                         img.u = (uint8_t*)convert->input.y + width*height + width2*height2;  
                         image_output(&img, width, height, width,  
                                                 convert->output.y, convert->output.y_stride,  
                                                 convert->output.colorspace, convert->interlacing);  
                         break;  
   
                 default :  
                         return XVID_ERR_FORMAT;  
         }  
   
   
         emms();  
         return XVID_ERR_OK;  
 }  
   
   
   
 void fill8(uint8_t * block, int size, int value)  
 {  
         int i;  
         for (i = 0; i < size; i++)  
                 block[i] = value;  
 }  
   
 void fill16(int16_t * block, int size, int value)  
 {  
         int i;  
         for (i = 0; i < size; i++)  
                 block[i] = value;  
 }  
   
 #define RANDOM(min,max) min + (rand() % (max-min))  
   
 void random8(uint8_t * block, int size, int min, int max)  
 {  
         int i;  
         for (i = 0; i < size; i++)  
                 block[i] = RANDOM(min,max);  
 }  
   
 void random16(int16_t * block, int size, int min, int max)  
 {  
         int i;  
         for (i = 0; i < size; i++)  
                 block[i] = RANDOM(min,max);  
 }  
   
 int compare16(const int16_t * blockA, const int16_t * blockB, int size)  
 {  
         int i;  
         for (i = 0; i < size; i++)  
                 if (blockA[i] != blockB[i])  
                         return 1;  
   
         return 0;  
 }  
   
   
   
 int test_h263_intra(quanth263_intraFunc * funcA, quanth263_intraFunc * funcB,  
                                                 const char * nameA, const char * nameB,  
                                                 int min, int max)  
 {  
         int q,i;  
         int64_t timeSTART;  
         int64_t timeA = 0;  
         int64_t timeB = 0;  
         DECLARE_ALIGNED_MATRIX(arrayX, 1, 64, int16_t, CACHE_LINE);  
         DECLARE_ALIGNED_MATRIX(arrayA, 1, 64, int16_t, CACHE_LINE);  
         DECLARE_ALIGNED_MATRIX(arrayB, 1, 64, int16_t, CACHE_LINE);  
   
         for (q = 1; q <= 31; q++)       /* quantizer */  
         {  
                 for (i = min; i < max; i++)     /* input coeff */  
                 {  
                         fill16(arrayX, 64, i);  
   
                         timeSTART = read_counter();  
                         funcA(arrayA, arrayX, q, q);  
                         timeA += read_counter() - timeSTART;  
   
                         timeSTART = read_counter();  
                         funcB(arrayB, arrayX, q, q);  
                         timeB += read_counter() - timeSTART;  
   
                         if (compare16(arrayA, arrayB, 64))  
                         {  
                                 printf("%s/%s error: q=%i, i=%i\n", nameA?nameA:"?", nameB?nameB:"?", q, i);  
                                 return 0;  
                         }  
                 }  
         }  
   
         if (nameA) printf("%s:\t%I64i\n", nameA, timeA);  
         if (nameB) printf("%s:\t%I64i\n", nameB, timeB);  
   
         return 0;  
 }  
   
 int test_h263_inter(quanth263_interFunc * funcA, quanth263_interFunc * funcB,  
                                                 const char * nameA, const char * nameB,  
                                                 int min, int max)  
 {  
         int q,i;  
         int64_t timeSTART;  
         int64_t timeA = 0;  
         int64_t timeB = 0;  
         DECLARE_ALIGNED_MATRIX(arrayX, 1, 64, int16_t, CACHE_LINE);  
         DECLARE_ALIGNED_MATRIX(arrayA, 1, 64, int16_t, CACHE_LINE);  
         DECLARE_ALIGNED_MATRIX(arrayB, 1, 64, int16_t, CACHE_LINE);  
   
         for (q = 1; q <= 31; q++)       /* quantizer */  
         {  
                 for (i = min; i < max; i++)     /* input coeff */  
                 {  
                         fill16(arrayX, 64, i);  
   
                         timeSTART = read_counter();  
                         funcA(arrayA, arrayX, q);  
                         timeA += read_counter() - timeSTART;  
   
                         timeSTART = read_counter();  
                         funcB(arrayB, arrayX, q);  
                         timeB += read_counter() - timeSTART;  
   
                         if (compare16(arrayA, arrayB, 64))  
                         {  
                                 printf("%s/%s error: q=%i, i=%i\n", nameA?nameA:"?", nameB?nameB:"?", q, i);  
                                 return 0;  
                         }  
                 }  
         }  
   
         if (nameA) printf("%s:\t%I64i\n", nameA, timeA);  
         if (nameB) printf("%s:\t%I64i\n", nameB, timeB);  
   
         return 0;  
 }  
   
   
   
 int xvid_init_test()  
 {  
         int cpu_flags;  
   
         printf("xvid_init_test\n");  
   
 #if defined(ARCH_X86)  
         cpu_flags = check_cpu_features();  
   
         emms_mmx();  
   
         printf("--- quant intra ---\n");  
         if (cpu_flags & XVID_CPU_MMX)  
                 test_h263_intra(quant_intra_c, quant_intra_mmx, "c", "mmx", -2048, 2047);  
         if (cpu_flags & XVID_CPU_3DNOWEXT)  
                 test_h263_intra(quant_intra_c, quant_intra_3dne, NULL, "3dne", -2048, 2047);  
         if (cpu_flags & XVID_CPU_SSE2)  
                 test_h263_intra(quant_intra_c, quant_intra_sse2, NULL, "sse2", -2048, 2047);  
   
         printf("\n--- quant inter ---\n");  
         if (cpu_flags & XVID_CPU_MMX)  
                 test_h263_inter(quant_inter_c, quant_inter_mmx, "c", "mmx", -2048, 2047);  
         if (cpu_flags & XVID_CPU_3DNOWEXT)  
                 test_h263_inter(quant_inter_c, quant_inter_3dne, NULL, "3dne", -2048, 2047);  
         if (cpu_flags & XVID_CPU_SSE2)  
                 test_h263_inter(quant_inter_c, quant_inter_sse2, NULL, "sse2", -2048, 2047);  
   
         printf("\n--- dequan intra ---\n");  
         if (cpu_flags & XVID_CPU_MMX)  
                 test_h263_intra(dequant_intra_c, dequant_intra_mmx, "c", "mmx", -256, 255);  
         if (cpu_flags & XVID_CPU_MMXEXT)  
                 test_h263_intra(dequant_intra_c, dequant_intra_xmm, NULL, "xmm", -256, 255);  
         if (cpu_flags & XVID_CPU_3DNOWEXT)  
                 test_h263_intra(dequant_intra_c, dequant_intra_3dne, NULL, "3dne", -256, 255);  
         if (cpu_flags & XVID_CPU_SSE2)  
                 test_h263_intra(dequant_intra_c, dequant_intra_sse2, NULL, "sse2", -256, 255);  
   
         printf("\n--- dequant inter ---\n");  
         if (cpu_flags & XVID_CPU_MMX)  
                 test_h263_inter((quanth263_interFunc*)dequant_inter_c,  
                                                 (quanth263_interFunc*)dequant_inter_mmx, "c", "mmx", -256, 255);  
   
         if (cpu_flags & XVID_CPU_MMXEXT)  
                 test_h263_inter((quanth263_interFunc*)dequant_inter_c,  
                                                 (quanth263_interFunc*)dequant_inter_xmm, NULL, "xmm", -256, 255);  
         if (cpu_flags & XVID_CPU_3DNOWEXT)  
                 test_h263_inter((quanth263_interFunc*)dequant_inter_c,  
                                                 (quanth263_interFunc*)dequant_inter_3dne, NULL, "3dne", -256, 255);  
         if (cpu_flags & XVID_CPU_SSE2)  
                 test_h263_inter((quanth263_interFunc*)dequant_inter_c,  
                                                 (quanth263_interFunc*)dequant_inter_sse2, NULL, "sse2", -256, 255);  
   
         printf("\n--- quant4_intra ---\n");  
         if (cpu_flags & XVID_CPU_MMX)  
                 test_h263_intra((quanth263_intraFunc*)quant4_intra_c,  
                                                 (quanth263_intraFunc*)quant4_intra_mmx, "c", "mmx", -2048, 2047);  
         if (cpu_flags & XVID_CPU_MMXEXT)  
                 test_h263_intra((quanth263_intraFunc*)quant4_intra_c,  
                                                 (quanth263_intraFunc*)quant4_intra_xmm, NULL, "xmm", -2048, 2047);  
   
         printf("\n--- quant4_inter ---\n");  
         if (cpu_flags & XVID_CPU_MMX)  
                 test_h263_inter((quanth263_interFunc*)quant4_inter_c,  
                                                 (quanth263_interFunc*)quant4_inter_mmx, "c", "mmx", -2048, 2047);  
         if (cpu_flags & XVID_CPU_MMXEXT)  
                 test_h263_inter((quanth263_interFunc*)quant4_inter_c,  
                                                 (quanth263_interFunc*)quant4_inter_xmm, NULL, "xmm", -2048, 2047);  
   
   
         printf("\n--- dequant4_intra ---\n");  
         if (cpu_flags & XVID_CPU_MMX)  
                 test_h263_intra((quanth263_intraFunc*)dequant4_intra_c,  
                                                 (quanth263_intraFunc*)dequant4_intra_mmx, "c", "mmx", -256, 255);  
         if (cpu_flags & XVID_CPU_3DNOWEXT)  
                 test_h263_intra((quanth263_intraFunc*)dequant4_intra_c,  
                                                 (quanth263_intraFunc*)dequant4_intra_3dne, NULL, "sse2", -256, 255);  
   
         printf("\n--- dequant4_inter ---\n");  
         if (cpu_flags & XVID_CPU_MMX)  
                 test_h263_inter((quanth263_interFunc*)dequant4_inter_c,  
                                                 (quanth263_interFunc*)dequant4_inter_mmx, "c", "mmx", -256, 255);  
         if (cpu_flags & XVID_CPU_3DNOWEXT)  
                 test_h263_inter((quanth263_interFunc*)dequant4_inter_c,  
                                                 (quanth263_interFunc*)dequant4_inter_3dne, NULL, "sse2", -256, 255);  
   
         emms_mmx();  
   
 #endif  
   
         return XVID_ERR_OK;  
 }  
   
   
 int  
 xvid_init(void *handle,  
                   int opt,  
                   void *param1,  
                   void *param2)  
 {  
         switch(opt)  
         {  
                 case XVID_INIT_INIT :  
                         return xvid_init_init((XVID_INIT_PARAM*)param1);  
   
                 case XVID_INIT_CONVERT :  
                         return xvid_init_convert((XVID_INIT_CONVERTINFO*)param1);  
   
                 case XVID_INIT_TEST :  
                         return xvid_init_test();  
   
                 default :  
                         return XVID_ERR_FAIL;  
         }  
 }  
   
439  /*****************************************************************************  /*****************************************************************************
440   * XviD Native decoder entry point   * XviD Native decoder entry point
441   *   *
# Line 838  Line 454 
454  {  {
455          switch (opt) {          switch (opt) {
456          case XVID_DEC_DECODE:          case XVID_DEC_DECODE:
457                  return decoder_decode((DECODER *) handle, (XVID_DEC_FRAME *) param1, (XVID_DEC_STATS*) param2);                  return decoder_decode((DECODER *) handle, (XVID_DEC_FRAME *) param1);
458    
459          case XVID_DEC_CREATE:          case XVID_DEC_CREATE:
460                  return decoder_create((XVID_DEC_PARAM *) param1);                  return decoder_create((XVID_DEC_PARAM *) param1);
# Line 870  Line 486 
486  {  {
487          switch (opt) {          switch (opt) {
488          case XVID_ENC_ENCODE:          case XVID_ENC_ENCODE:
   
                 if (((Encoder *) handle)->mbParam.max_bframes >= 0)  
                 return encoder_encode_bframes((Encoder *) handle, (XVID_ENC_FRAME *) param1,  
                                                           (XVID_ENC_STATS *) param2);  
                 else  
489                  return encoder_encode((Encoder *) handle, (XVID_ENC_FRAME *) param1,                  return encoder_encode((Encoder *) handle, (XVID_ENC_FRAME *) param1,
490                                                            (XVID_ENC_STATS *) param2);                                                            (XVID_ENC_STATS *) param2);
491    
Legend:
Removed from v.1.33.2.19  
changed lines
  Added in v.1.39
No admin address has been configured
 ViewVC Help
Powered by ViewVC 1.0.4