[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 1.33.2.1, Mon Sep 23 20:36:01 2002 UTC revision 1.45.2.23, Thu Dec 11 17:18:29 2003 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) 2001-2003 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 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   *  it under the terms of the GNU General Public License as published by
# Line 26  Line 19 
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   *   *
  ****************************************************************************/  
   
 /*****************************************************************************  
  *  
  *  History  
  *  
  *      - 23.06.2002    added XVID_CPU_CHKONLY  
  *  - 17.03.2002        Added interpolate8x8_halfpel_hv_xmm  
  *  - 22.12.2001  API change: added xvid_init() - Isibaar  
  *  - 16.12.2001        inital version; (c)2001 peter ross <pross@cs.rmit.edu.au>  
  *  
22   *  $Id$   *  $Id$
23   *   *
24   ****************************************************************************/   ****************************************************************************/
25    
26    #include <stdio.h>
27    #include <stdlib.h>
28    #include <string.h>
29    #include <time.h>
30    
31  #include "xvid.h"  #include "xvid.h"
32  #include "decoder.h"  #include "decoder.h"
33  #include "encoder.h"  #include "encoder.h"
# Line 49  Line 36 
36  #include "dct/fdct.h"  #include "dct/fdct.h"
37  #include "image/colorspace.h"  #include "image/colorspace.h"
38  #include "image/interpolate8x8.h"  #include "image/interpolate8x8.h"
39    #include "image/reduced.h"
40  #include "utils/mem_transfer.h"  #include "utils/mem_transfer.h"
41  #include "quant/quant_h263.h"  #include "utils/mbfunctions.h"
42  #include "quant/quant_mpeg4.h"  #include "quant/quant.h"
43  #include "motion/motion.h"  #include "motion/motion.h"
44  #include "motion/sad.h"  #include "motion/sad.h"
45  #include "utils/emms.h"  #include "utils/emms.h"
46  #include "utils/timer.h"  #include "utils/timer.h"
47  #include "bitstream/mbcoding.h"  #include "bitstream/mbcoding.h"
48    #include "image/qpel.h"
49    #include "image/postprocessing.h"
50    
51  #if defined(ARCH_X86) && defined(EXPERIMENTAL_SSE2_CODE)  #if defined(_DEBUG)
52    unsigned int xvid_debug = 0; /* xvid debug mask */
53    #endif
54    
55  #ifdef WIN32  #if defined(ARCH_IS_IA32)
56    #if defined(_MSC_VER)
57  #include <windows.h>  #include <windows.h>
58  #else  #else
59  #include <signal.h>  #include <signal.h>
60  #include <setjmp.h>  #include <setjmp.h>
 #endif  
   
   
 #ifndef WIN32  
61    
62  static jmp_buf mark;  static jmp_buf mark;
63    
# Line 81  Line 70 
70    
71    
72  /*  /*
73  calls the funcptr, and returns whether SIGILL (illegal instruction) was signalled   * Calls the funcptr, and returns whether SIGILL (illegal instruction) was
74  return values:   * signalled
75  -1 : could not determine   *
76  0  : SIGILL was *not* signalled   * Return values:
77  1  : SIGILL was signalled   *  -1 : could not determine
78     *   0 : SIGILL was *not* signalled
79     *   1 : SIGILL was signalled
80  */  */
81    
82  int  int
83  sigill_check(void (*func)())  sigill_check(void (*func)())
84  {  {
85  #ifdef WIN32  #if defined(_MSC_VER)
86          _try {          _try {
87                  func();                  func();
88          }          }
# Line 125  Line 116 
116  }  }
117  #endif  #endif
118    
119    
120    /* detect cpu flags  */
121    static unsigned int
122    detect_cpu_flags()
123    {
124            /* enable native assembly optimizations by default */
125            unsigned int cpu_flags = XVID_CPU_ASM;
126    
127    #if defined(ARCH_IS_IA32)
128            cpu_flags |= check_cpu_features();
129            if ((cpu_flags & XVID_CPU_SSE) && sigill_check(sse_os_trigger))
130                    cpu_flags &= ~XVID_CPU_SSE;
131    
132            if ((cpu_flags & XVID_CPU_SSE2) && sigill_check(sse2_os_trigger))
133                    cpu_flags &= ~XVID_CPU_SSE2;
134    #endif
135    
136    #if defined(ARCH_IS_PPC)
137    #if defined(ARCH_IS_PPC_ALTIVEC)
138            cpu_flags |= XVID_CPU_ALTIVEC;
139    #endif
140    #endif
141    
142            return cpu_flags;
143    }
144    
145    
146  /*****************************************************************************  /*****************************************************************************
147   * XviD Init Entry point   * XviD Init Entry point
148   *   *
# Line 139  Line 157 
157   *   *
158   ****************************************************************************/   ****************************************************************************/
159    
 int  
 xvid_init(void *handle,  
                   int opt,  
                   void *param1,  
                   void *param2)  
 {  
         int cpu_flags;  
         XVID_INIT_PARAM *init_param;  
   
         init_param = (XVID_INIT_PARAM *) param1;  
   
         /* Inform the client the API version */  
         init_param->api_version = API_VERSION;  
   
         /* Inform the client the core build - unused because we're still alpha */  
         init_param->core_build = 1000;  
   
         /* Do we have to force CPU features  ? */  
         if ((init_param->cpu_flags & XVID_CPU_FORCE)) {  
   
                 cpu_flags = init_param->cpu_flags;  
160    
161          } else {  static
162    int xvid_gbl_init(xvid_gbl_init_t * init)
                 cpu_flags = check_cpu_features();  
   
 #if defined(ARCH_X86) && defined(EXPERIMENTAL_SSE2_CODE)  
                 if ((cpu_flags & XVID_CPU_SSE) && sigill_check(sse_os_trigger))  
                         cpu_flags &= ~XVID_CPU_SSE;  
   
                 if ((cpu_flags & XVID_CPU_SSE2) && sigill_check(sse2_os_trigger))  
                         cpu_flags &= ~XVID_CPU_SSE2;  
 #endif  
         }  
   
         if ((init_param->cpu_flags & XVID_CPU_CHKONLY))  
163          {          {
164                  init_param->cpu_flags = cpu_flags;          unsigned int cpu_flags;
                 return XVID_ERR_OK;  
         }  
165    
166          init_param->cpu_flags = cpu_flags;          if (XVID_VERSION_MAJOR(init->version) != 1) /* v1.x.x */
167                    return XVID_ERR_VERSION;
168    
169            cpu_flags = (init->cpu_flags & XVID_CPU_FORCE) ? init->cpu_flags : detect_cpu_flags();
170    
171          /* Initialize the function pointers */          /* Initialize the function pointers */
172          idct_int32_init();          idct_int32_init();
173          init_vlc_tables();          init_vlc_tables();
174            init_postproc();
175    
176          /* Fixed Point Forward/Inverse DCT transformations */          /* Fixed Point Forward/Inverse DCT transformations */
177          fdct = fdct_int32;          fdct = fdct_int32;
# Line 197  Line 183 
183          /* Restore FPU context : emms_c is a nop functions */          /* Restore FPU context : emms_c is a nop functions */
184          emms = emms_c;          emms = emms_c;
185    
186            /* Qpel stuff */
187            xvid_QP_Funcs = &xvid_QP_Funcs_C;
188            xvid_QP_Add_Funcs = &xvid_QP_Add_Funcs_C;
189            xvid_Init_QP();
190    
191          /* Quantization functions */          /* Quantization functions */
192          quant_intra   = quant_intra_c;          quant_h263_intra   = quant_h263_intra_c;
193          dequant_intra = dequant_intra_c;          quant_h263_inter   = quant_h263_inter_c;
194          quant_inter   = quant_inter_c;          dequant_h263_intra = dequant_h263_intra_c;
195          dequant_inter = dequant_inter_c;          dequant_h263_inter = dequant_h263_inter_c;
196    
197          quant4_intra   = quant4_intra_c;          quant_mpeg_intra   = quant_mpeg_intra_c;
198          dequant4_intra = dequant4_intra_c;          quant_mpeg_inter   = quant_mpeg_inter_c;
199          quant4_inter   = quant4_inter_c;          dequant_mpeg_intra = dequant_mpeg_intra_c;
200          dequant4_inter = dequant4_inter_c;          dequant_mpeg_inter = dequant_mpeg_inter_c;
201    
202          /* Block transfer related functions */          /* Block transfer related functions */
203          transfer_8to16copy = transfer_8to16copy_c;          transfer_8to16copy = transfer_8to16copy_c;
204          transfer_16to8copy = transfer_16to8copy_c;          transfer_16to8copy = transfer_16to8copy_c;
205          transfer_8to16sub  = transfer_8to16sub_c;          transfer_8to16sub  = transfer_8to16sub_c;
206            transfer_8to16subro  = transfer_8to16subro_c;
207          transfer_8to16sub2 = transfer_8to16sub2_c;          transfer_8to16sub2 = transfer_8to16sub2_c;
208          transfer_16to8add  = transfer_16to8add_c;          transfer_16to8add  = transfer_16to8add_c;
209          transfer8x8_copy   = transfer8x8_copy_c;          transfer8x8_copy   = transfer8x8_copy_c;
210    
211            /* Interlacing functions */
212            MBFieldTest = MBFieldTest_c;
213    
214          /* Image interpolation related functions */          /* Image interpolation related functions */
215          interpolate8x8_halfpel_h  = interpolate8x8_halfpel_h_c;          interpolate8x8_halfpel_h  = interpolate8x8_halfpel_h_c;
216          interpolate8x8_halfpel_v  = interpolate8x8_halfpel_v_c;          interpolate8x8_halfpel_v  = interpolate8x8_halfpel_v_c;
217          interpolate8x8_halfpel_hv = interpolate8x8_halfpel_hv_c;          interpolate8x8_halfpel_hv = interpolate8x8_halfpel_hv_c;
218    
219            interpolate16x16_lowpass_h = interpolate16x16_lowpass_h_c;
220            interpolate16x16_lowpass_v = interpolate16x16_lowpass_v_c;
221            interpolate16x16_lowpass_hv = interpolate16x16_lowpass_hv_c;
222    
223            interpolate8x8_lowpass_h = interpolate8x8_lowpass_h_c;
224            interpolate8x8_lowpass_v = interpolate8x8_lowpass_v_c;
225            interpolate8x8_lowpass_hv = interpolate8x8_lowpass_hv_c;
226    
227            interpolate8x8_6tap_lowpass_h = interpolate8x8_6tap_lowpass_h_c;
228            interpolate8x8_6tap_lowpass_v = interpolate8x8_6tap_lowpass_v_c;
229    
230            interpolate8x8_avg2 = interpolate8x8_avg2_c;
231            interpolate8x8_avg4 = interpolate8x8_avg4_c;
232    
233            /* reduced resolution */
234            copy_upsampled_8x8_16to8 = xvid_Copy_Upsampled_8x8_16To8_C;
235            add_upsampled_8x8_16to8 = xvid_Add_Upsampled_8x8_16To8_C;
236            vfilter_31 = xvid_VFilter_31_C;
237            hfilter_31 = xvid_HFilter_31_C;
238            filter_18x18_to_8x8 = xvid_Filter_18x18_To_8x8_C;
239            filter_diff_18x18_to_8x8 = xvid_Filter_Diff_18x18_To_8x8_C;
240    
241          /* Initialize internal colorspace transformation tables */          /* Initialize internal colorspace transformation tables */
242          colorspace_init();          colorspace_init();
243    
244          /* All colorspace transformation functions User Format->YV12 */          /* All colorspace transformation functions User Format->YV12 */
245            yv12_to_yv12    = yv12_to_yv12_c;
246          rgb555_to_yv12 = rgb555_to_yv12_c;          rgb555_to_yv12 = rgb555_to_yv12_c;
247          rgb565_to_yv12 = rgb565_to_yv12_c;          rgb565_to_yv12 = rgb565_to_yv12_c;
248          rgb24_to_yv12  = rgb24_to_yv12_c;          bgr_to_yv12     = bgr_to_yv12_c;
249          rgb32_to_yv12  = rgb32_to_yv12_c;          bgra_to_yv12    = bgra_to_yv12_c;
250          yuv_to_yv12    = yuv_to_yv12_c;          abgr_to_yv12    = abgr_to_yv12_c;
251            rgba_to_yv12    = rgba_to_yv12_c;
252          yuyv_to_yv12   = yuyv_to_yv12_c;          yuyv_to_yv12   = yuyv_to_yv12_c;
253          uyvy_to_yv12   = uyvy_to_yv12_c;          uyvy_to_yv12   = uyvy_to_yv12_c;
254    
255            rgb555i_to_yv12 = rgb555i_to_yv12_c;
256            rgb565i_to_yv12 = rgb565i_to_yv12_c;
257            bgri_to_yv12    = bgri_to_yv12_c;
258            bgrai_to_yv12   = bgrai_to_yv12_c;
259            abgri_to_yv12   = abgri_to_yv12_c;
260            rgbai_to_yv12   = rgbai_to_yv12_c;
261            yuyvi_to_yv12   = yuyvi_to_yv12_c;
262            uyvyi_to_yv12   = uyvyi_to_yv12_c;
263    
264    
265          /* All colorspace transformation functions YV12->User format */          /* All colorspace transformation functions YV12->User format */
266          yv12_to_rgb555 = yv12_to_rgb555_c;          yv12_to_rgb555 = yv12_to_rgb555_c;
267          yv12_to_rgb565 = yv12_to_rgb565_c;          yv12_to_rgb565 = yv12_to_rgb565_c;
268          yv12_to_rgb24  = yv12_to_rgb24_c;          yv12_to_bgr     = yv12_to_bgr_c;
269          yv12_to_rgb32  = yv12_to_rgb32_c;          yv12_to_bgra    = yv12_to_bgra_c;
270          yv12_to_yuv    = yv12_to_yuv_c;          yv12_to_abgr    = yv12_to_abgr_c;
271            yv12_to_rgba    = yv12_to_rgba_c;
272          yv12_to_yuyv   = yv12_to_yuyv_c;          yv12_to_yuyv   = yv12_to_yuyv_c;
273          yv12_to_uyvy   = yv12_to_uyvy_c;          yv12_to_uyvy   = yv12_to_uyvy_c;
274    
275            yv12_to_rgb555i = yv12_to_rgb555i_c;
276            yv12_to_rgb565i = yv12_to_rgb565i_c;
277            yv12_to_bgri    = yv12_to_bgri_c;
278            yv12_to_bgrai   = yv12_to_bgrai_c;
279            yv12_to_abgri   = yv12_to_abgri_c;
280            yv12_to_rgbai   = yv12_to_rgbai_c;
281            yv12_to_yuyvi   = yv12_to_yuyvi_c;
282            yv12_to_uyvyi   = yv12_to_uyvyi_c;
283    
284          /* Functions used in motion estimation algorithms */          /* Functions used in motion estimation algorithms */
285          calc_cbp = calc_cbp_c;          calc_cbp = calc_cbp_c;
286          sad16    = sad16_c;          sad16    = sad16_c;
# Line 250  Line 289 
289          sad8bi   = sad8bi_c;          sad8bi   = sad8bi_c;
290          dev16    = dev16_c;          dev16    = dev16_c;
291          sad16v   = sad16v_c;          sad16v   = sad16v_c;
292            sse8_16bit = sse8_16bit_c;
293    
294  //      Halfpel8_Refine = Halfpel8_Refine_c;  #if defined(ARCH_IS_IA32)
295    
296  #ifdef ARCH_X86          if ((cpu_flags & XVID_CPU_ASM)) {
297          if ((cpu_flags & XVID_CPU_MMX) > 0) {                  vfilter_31 = xvid_VFilter_31_x86;
298                    hfilter_31 = xvid_HFilter_31_x86;
299            }
300    
301            if ((cpu_flags & XVID_CPU_MMX) || (cpu_flags & XVID_CPU_MMXEXT) ||
302                    (cpu_flags & XVID_CPU_3DNOW) || (cpu_flags & XVID_CPU_3DNOWEXT) ||
303                    (cpu_flags & XVID_CPU_SSE) || (cpu_flags & XVID_CPU_SSE2))
304            {
305                    /* Restore FPU context : emms_c is a nop functions */
306                    emms = emms_mmx;
307            }
308    
309            if ((cpu_flags & XVID_CPU_MMX)) {
310    
311                  /* Forward and Inverse Discrete Cosine Transformation functions */                  /* Forward and Inverse Discrete Cosine Transformation functions */
312                  fdct = fdct_mmx;                  fdct = fdct_mmx_skal;
313                  idct = idct_mmx;                  idct = idct_mmx;
314    
315                  /* To restore FPU context after mmx use */                  /* Qpel stuff */
316                  emms = emms_mmx;                  xvid_QP_Funcs = &xvid_QP_Funcs_mmx;
317                    xvid_QP_Add_Funcs = &xvid_QP_Add_Funcs_mmx;
318    
319                  /* Quantization related functions */                  /* Quantization related functions */
320                  quant_intra   = quant_intra_mmx;                  quant_h263_intra   = quant_h263_intra_mmx;
321                  dequant_intra = dequant_intra_mmx;                  quant_h263_inter   = quant_h263_inter_mmx;
322                  quant_inter   = quant_inter_mmx;                  dequant_h263_intra = dequant_h263_intra_mmx;
323                  dequant_inter = dequant_inter_mmx;                  dequant_h263_inter = dequant_h263_inter_mmx;
324    
325                  quant4_intra   = quant4_intra_mmx;                  quant_mpeg_intra   = quant_mpeg_intra_mmx;
326                  dequant4_intra = dequant4_intra_mmx;                  quant_mpeg_inter   = quant_mpeg_inter_mmx;
327                  quant4_inter   = quant4_inter_mmx;                  dequant_mpeg_intra = dequant_mpeg_intra_mmx;
328                  dequant4_inter = dequant4_inter_mmx;                  dequant_mpeg_inter = dequant_mpeg_inter_mmx;
329    
330                  /* Block related functions */                  /* Block related functions */
331                  transfer_8to16copy = transfer_8to16copy_mmx;                  transfer_8to16copy = transfer_8to16copy_mmx;
332                  transfer_16to8copy = transfer_16to8copy_mmx;                  transfer_16to8copy = transfer_16to8copy_mmx;
333                  transfer_8to16sub  = transfer_8to16sub_mmx;                  transfer_8to16sub  = transfer_8to16sub_mmx;
334                    transfer_8to16subro  = transfer_8to16subro_mmx;
335                  transfer_8to16sub2 = transfer_8to16sub2_mmx;                  transfer_8to16sub2 = transfer_8to16sub2_mmx;
336                  transfer_16to8add  = transfer_16to8add_mmx;                  transfer_16to8add  = transfer_16to8add_mmx;
337                  transfer8x8_copy   = transfer8x8_copy_mmx;                  transfer8x8_copy   = transfer8x8_copy_mmx;
338    
339                    /* Interlacing Functions */
340                    MBFieldTest = MBFieldTest_mmx;
341    
342                  /* Image Interpolation related functions */                  /* Image Interpolation related functions */
343                  interpolate8x8_halfpel_h  = interpolate8x8_halfpel_h_mmx;                  interpolate8x8_halfpel_h  = interpolate8x8_halfpel_h_mmx;
344                  interpolate8x8_halfpel_v  = interpolate8x8_halfpel_v_mmx;                  interpolate8x8_halfpel_v  = interpolate8x8_halfpel_v_mmx;
345                  interpolate8x8_halfpel_hv = interpolate8x8_halfpel_hv_mmx;                  interpolate8x8_halfpel_hv = interpolate8x8_halfpel_hv_mmx;
346    
347                  /* Image RGB->YV12 related functions */                  interpolate8x8_6tap_lowpass_h = interpolate8x8_6tap_lowpass_h_mmx;
348                  rgb24_to_yv12 = rgb24_to_yv12_mmx;                  interpolate8x8_6tap_lowpass_v = interpolate8x8_6tap_lowpass_v_mmx;
349                  rgb32_to_yv12 = rgb32_to_yv12_mmx;  
350                  yuv_to_yv12   = yuv_to_yv12_mmx;                  interpolate8x8_avg2 = interpolate8x8_avg2_mmx;
351                    interpolate8x8_avg4 = interpolate8x8_avg4_mmx;
352    
353                    /* reduced resolution */
354                    copy_upsampled_8x8_16to8 = xvid_Copy_Upsampled_8x8_16To8_mmx;
355                    add_upsampled_8x8_16to8 = xvid_Add_Upsampled_8x8_16To8_mmx;
356                    hfilter_31 = xvid_HFilter_31_mmx;
357                    filter_18x18_to_8x8 = xvid_Filter_18x18_To_8x8_mmx;
358                    filter_diff_18x18_to_8x8 = xvid_Filter_Diff_18x18_To_8x8_mmx;
359    
360                    /* image input xxx_to_yv12 related functions */
361                    yv12_to_yv12  = yv12_to_yv12_mmx;
362                    bgr_to_yv12   = bgr_to_yv12_mmx;
363                    bgra_to_yv12  = bgra_to_yv12_mmx;
364                  yuyv_to_yv12  = yuyv_to_yv12_mmx;                  yuyv_to_yv12  = yuyv_to_yv12_mmx;
365                  uyvy_to_yv12  = uyvy_to_yv12_mmx;                  uyvy_to_yv12  = uyvy_to_yv12_mmx;
366    
367                  /* Image YV12->RGB related functions */                  /* image output yv12_to_xxx related functions */
368                  yv12_to_rgb24 = yv12_to_rgb24_mmx;                  yv12_to_bgr   = yv12_to_bgr_mmx;
369                  yv12_to_rgb32 = yv12_to_rgb32_mmx;                  yv12_to_bgra  = yv12_to_bgra_mmx;
370                  yv12_to_yuyv  = yv12_to_yuyv_mmx;                  yv12_to_yuyv  = yv12_to_yuyv_mmx;
371                  yv12_to_uyvy  = yv12_to_uyvy_mmx;                  yv12_to_uyvy  = yv12_to_uyvy_mmx;
372    
373                    yv12_to_yuyvi = yv12_to_yuyvi_mmx;
374                    yv12_to_uyvyi = yv12_to_uyvyi_mmx;
375    
376                  /* Motion estimation related functions */                  /* Motion estimation related functions */
377                  calc_cbp = calc_cbp_mmx;                  calc_cbp = calc_cbp_mmx;
378                  sad16    = sad16_mmx;                  sad16    = sad16_mmx;
# Line 308  Line 380 
380                  sad16bi = sad16bi_mmx;                  sad16bi = sad16bi_mmx;
381                  sad8bi  = sad8bi_mmx;                  sad8bi  = sad8bi_mmx;
382                  dev16    = dev16_mmx;                  dev16    = dev16_mmx;
383                    sad16v   = sad16v_mmx;
384                    sse8_16bit = sse8_16bit_mmx;
385          }          }
386    
387          /* these 3dnow functions are faster than mmx, but slower than xmm. */          /* these 3dnow functions are faster than mmx, but slower than xmm. */
388          if ((cpu_flags & XVID_CPU_3DNOW) > 0) {          if ((cpu_flags & XVID_CPU_3DNOW)) {
389    
390                    emms = emms_3dn;
391    
392                  /* ME functions */                  /* ME functions */
393                  sad16bi = sad16bi_3dn;                  sad16bi = sad16bi_3dn;
394                  sad8bi  = sad8bi_3dn;                  sad8bi  = sad8bi_3dn;
395    
396                    yuyv_to_yv12  = yuyv_to_yv12_3dn;
397                    uyvy_to_yv12  = uyvy_to_yv12_3dn;
398          }          }
399    
400    
401          if ((cpu_flags & XVID_CPU_MMXEXT) > 0) {          if ((cpu_flags & XVID_CPU_MMXEXT)) {
402    
403                  /* Inverse DCT */                  /* DCT */
404                    fdct = fdct_xmm_skal;
405                  idct = idct_xmm;                  idct = idct_xmm;
406    
407                  /* Interpolation */                  /* Interpolation */
# Line 330  Line 409 
409                  interpolate8x8_halfpel_v  = interpolate8x8_halfpel_v_xmm;                  interpolate8x8_halfpel_v  = interpolate8x8_halfpel_v_xmm;
410                  interpolate8x8_halfpel_hv = interpolate8x8_halfpel_hv_xmm;                  interpolate8x8_halfpel_hv = interpolate8x8_halfpel_hv_xmm;
411    
412                    /* reduced resolution */
413                    copy_upsampled_8x8_16to8 = xvid_Copy_Upsampled_8x8_16To8_xmm;
414                    add_upsampled_8x8_16to8 = xvid_Add_Upsampled_8x8_16To8_xmm;
415    
416                  /* Quantization */                  /* Quantization */
417                  dequant_intra = dequant_intra_xmm;                  quant_mpeg_intra = quant_mpeg_intra_xmm;
418                  dequant_inter = dequant_inter_xmm;                  quant_mpeg_inter = quant_mpeg_inter_xmm;
419    
420                    dequant_h263_intra = dequant_h263_intra_xmm;
421                    dequant_h263_inter = dequant_h263_inter_xmm;
422    
423                  /* Buffer transfer */                  /* Buffer transfer */
424                  transfer_8to16sub2 = transfer_8to16sub2_xmm;                  transfer_8to16sub2 = transfer_8to16sub2_xmm;
425    
426                  /* Colorspace transformation */                  /* Colorspace transformation */
427                  yuv_to_yv12 = yuv_to_yv12_xmm;                  yv12_to_yv12  = yv12_to_yv12_xmm;
428                    yuyv_to_yv12  = yuyv_to_yv12_xmm;
429                    uyvy_to_yv12  = uyvy_to_yv12_xmm;
430    
431                  /* ME functions */                  /* ME functions */
432                  sad16 = sad16_xmm;                  sad16 = sad16_xmm;
# Line 347  Line 435 
435                  sad8bi  = sad8bi_xmm;                  sad8bi  = sad8bi_xmm;
436                  dev16 = dev16_xmm;                  dev16 = dev16_xmm;
437                  sad16v   = sad16v_xmm;                  sad16v   = sad16v_xmm;
                 fprintf(stderr,"sad16v=XMM\n");  
   
438          }          }
439    
440          if ((cpu_flags & XVID_CPU_3DNOW) > 0) {          if ((cpu_flags & XVID_CPU_3DNOW)) {
441    
442                  /* Interpolation */                  /* Interpolation */
443                  interpolate8x8_halfpel_h = interpolate8x8_halfpel_h_3dn;                  interpolate8x8_halfpel_h = interpolate8x8_halfpel_h_3dn;
# Line 359  Line 445 
445                  interpolate8x8_halfpel_hv = interpolate8x8_halfpel_hv_3dn;                  interpolate8x8_halfpel_hv = interpolate8x8_halfpel_hv_3dn;
446          }          }
447    
448          if ((cpu_flags & XVID_CPU_SSE2) > 0) {          if ((cpu_flags & XVID_CPU_3DNOWEXT)) {
449  #ifdef EXPERIMENTAL_SSE2_CODE  
450                    /* Inverse DCT */
451                    idct =  idct_3dne;
452    
453                    /* Buffer transfer */
454                    transfer_8to16copy =  transfer_8to16copy_3dne;
455                    transfer_16to8copy = transfer_16to8copy_3dne;
456                    transfer_8to16sub =  transfer_8to16sub_3dne;
457                    transfer_8to16subro =  transfer_8to16subro_3dne;
458                    transfer_8to16sub2 =  transfer_8to16sub2_3dne;
459                    transfer_16to8add = transfer_16to8add_3dne;
460                    transfer8x8_copy = transfer8x8_copy_3dne;
461    
462                    /* Quantization */
463                    quant_h263_intra = quant_h263_intra_3dne;
464                    quant_h263_inter = quant_h263_inter_3dne;
465                    dequant_mpeg_intra = dequant_mpeg_intra_3dne;
466                    dequant_mpeg_inter = dequant_mpeg_inter_3dne;
467                    dequant_h263_intra = dequant_h263_intra_3dne;
468                    dequant_h263_inter = dequant_h263_inter_3dne;
469    
470                    /* ME functions */
471                    calc_cbp = calc_cbp_3dne;
472                    sad16 = sad16_3dne;
473                    sad8 = sad8_3dne;
474                    sad16bi = sad16bi_3dne;
475                    sad8bi = sad8bi_3dne;
476                    dev16 = dev16_3dne;
477    
478                    /* Interpolation */
479                    interpolate8x8_halfpel_h = interpolate8x8_halfpel_h_3dne;
480                    interpolate8x8_halfpel_v = interpolate8x8_halfpel_v_3dne;
481                    interpolate8x8_halfpel_hv = interpolate8x8_halfpel_hv_3dne;
482            }
483    
484    #if defined(EXPERIMENTAL_SSE2_CODE) /* mark the whole SSE2 stuff as experimental. At least on
485                                                                               my P4, it crashes... */
486            if ((cpu_flags & XVID_CPU_SSE2)) {
487    
488                  calc_cbp = calc_cbp_sse2;                  calc_cbp = calc_cbp_sse2;
489    
490                  /* Quantization */                  /* Quantization */
491                  quant_intra   = quant_intra_sse2;                  quant_h263_intra   = quant_h263_intra_sse2;
492                  dequant_intra = dequant_intra_sse2;                  quant_h263_inter   = quant_h263_inter_sse2;
493                  quant_inter   = quant_inter_sse2;                  dequant_h263_intra = dequant_h263_intra_sse2;
494                  dequant_inter = dequant_inter_sse2;                  dequant_h263_inter = dequant_h263_inter_sse2;
495    
496                  /* ME */                  /* SAD operators */
497                  sad16    = sad16_sse2;                  sad16    = sad16_sse2;
498                  dev16    = dev16_sse2;                  dev16    = dev16_sse2;
499    
500                  /* Forward and Inverse DCT */                  /* DCT operators */
501                  idct  = idct_sse2;                  fdct = fdct_sse2_skal;
502                  fdct = fdct_sse2;                  idct = idct_sse2_dmitry;
 #endif  
503          }          }
504    #endif
505  #endif  #endif
506    
507  #ifdef ARCH_IA64  #if defined(ARCH_IS_IA64)
508          if ((cpu_flags & XVID_CPU_IA64) > 0) { //use assembler routines?          if ((cpu_flags & XVID_CPU_ASM)) { /* use assembler routines? */
509            idct_ia64_init();            idct_ia64_init();
510            fdct = fdct_ia64;            fdct = fdct_ia64;
511            idct = idct_ia64;   //not yet working, crashes            idct = idct_ia64;   /*not yet working, crashes */
512            interpolate8x8_halfpel_h = interpolate8x8_halfpel_h_ia64;            interpolate8x8_halfpel_h = interpolate8x8_halfpel_h_ia64;
513            interpolate8x8_halfpel_v = interpolate8x8_halfpel_v_ia64;            interpolate8x8_halfpel_v = interpolate8x8_halfpel_v_ia64;
514            interpolate8x8_halfpel_hv = interpolate8x8_halfpel_hv_ia64;            interpolate8x8_halfpel_hv = interpolate8x8_halfpel_hv_ia64;
# Line 394  Line 516 
516            sad16bi = sad16bi_ia64;            sad16bi = sad16bi_ia64;
517            sad8 = sad8_ia64;            sad8 = sad8_ia64;
518            dev16 = dev16_ia64;            dev16 = dev16_ia64;
519  //        Halfpel8_Refine = Halfpel8_Refine_ia64;  /*        Halfpel8_Refine = Halfpel8_Refine_ia64; */
520            quant_intra = quant_intra_ia64;            quant_h263_intra = quant_h263_intra_ia64;
521            dequant_intra = dequant_intra_ia64;            quant_h263_inter = quant_h263_inter_ia64;
522            quant_inter = quant_inter_ia64;            dequant_h263_intra = dequant_h263_intra_ia64;
523            dequant_inter = dequant_inter_ia64;            dequant_h263_inter = dequant_h263_inter_ia64;
524            transfer_8to16copy = transfer_8to16copy_ia64;            transfer_8to16copy = transfer_8to16copy_ia64;
525            transfer_16to8copy = transfer_16to8copy_ia64;            transfer_16to8copy = transfer_16to8copy_ia64;
526            transfer_8to16sub = transfer_8to16sub_ia64;            transfer_8to16sub = transfer_8to16sub_ia64;
527            transfer_8to16sub2 = transfer_8to16sub2_ia64;            transfer_8to16sub2 = transfer_8to16sub2_ia64;
528            transfer_16to8add = transfer_16to8add_ia64;            transfer_16to8add = transfer_16to8add_ia64;
529            transfer8x8_copy = transfer8x8_copy_ia64;            transfer8x8_copy = transfer8x8_copy_ia64;
           DEBUG("Using IA-64 assembler routines.\n");  
530          }          }
531  #endif  #endif
532    
533  #ifdef ARCH_PPC  #if defined(ARCH_IS_PPC)
534  #ifdef ARCH_PPC_ALTIVEC          if ((cpu_flags & XVID_CPU_ASM))
535            {
536                    calc_cbp = calc_cbp_ppc;
537            }
538    
539            if ((cpu_flags & XVID_CPU_ALTIVEC))
540            {
541          calc_cbp = calc_cbp_altivec;          calc_cbp = calc_cbp_altivec;
542          fdct = fdct_altivec;          fdct = fdct_altivec;
543          idct = idct_altivec;          idct = idct_altivec;
# Line 418  Line 545 
545          sad16 = sad16_altivec;          sad16 = sad16_altivec;
546          sad8 = sad8_altivec;          sad8 = sad8_altivec;
547          dev16 = dev16_altivec;          dev16 = dev16_altivec;
548  #else          }
         calc_cbp = calc_cbp_ppc;  
549  #endif  #endif
550    
551    #if defined(_DEBUG)
552        xvid_debug = init->debug;
553  #endif  #endif
554    
555          return XVID_ERR_OK;      return 0;
556    }
557    
558    
559    static int
560    xvid_gbl_info(xvid_gbl_info_t * info)
561    {
562            if (XVID_VERSION_MAJOR(info->version) != 1) /* v1.x.x */
563                    return XVID_ERR_VERSION;
564    
565            info->actual_version = XVID_VERSION;
566            info->build = "xvid-1.0.0";
567            info->cpu_flags = detect_cpu_flags();
568    
569    #if defined(_SMP) && defined(WIN32)
570            info->num_threads = pthread_num_processors_np();;
571    #else
572            info->num_threads = 0;
573    #endif
574    
575            return 0;
576    }
577    
578    
579    static int
580    xvid_gbl_convert(xvid_gbl_convert_t* convert)
581    {
582            int width;
583            int height;
584            int width2;
585            int height2;
586            IMAGE img;
587    
588            if (XVID_VERSION_MAJOR(convert->version) != 1)   /* v1.x.x */
589                  return XVID_ERR_VERSION;
590    
591    #if 0
592            const int flip1 = (convert->input.colorspace & XVID_CSP_VFLIP) ^ (convert->output.colorspace & XVID_CSP_VFLIP);
593    #endif
594            width = convert->width;
595            height = convert->height;
596            width2 = convert->width/2;
597            height2 = convert->height/2;
598    
599            switch (convert->input.csp & ~XVID_CSP_VFLIP)
600            {
601                    case XVID_CSP_YV12 :
602                            img.y = convert->input.plane[0];
603                            img.v = (uint8_t*)convert->input.plane[0] + convert->input.stride[0]*height;
604                            img.u = (uint8_t*)convert->input.plane[0] + convert->input.stride[0]*height + (convert->input.stride[0]/2)*height2;
605                            image_output(&img, width, height, width,
606                                                    (uint8_t**)convert->output.plane, convert->output.stride,
607                                                    convert->output.csp, convert->interlacing);
608                            break;
609    
610                    default :
611                            return XVID_ERR_FORMAT;
612            }
613    
614    
615            emms();
616            return 0;
617    }
618    
619    /*****************************************************************************
620     * XviD Global Entry point
621     *
622     * Well this function initialize all internal function pointers according
623     * to the CPU features forced by the library client or autodetected (depending
624     * on the XVID_CPU_FORCE flag). It also initializes vlc coding tables and all
625     * image colorspace transformation tables.
626     *
627     ****************************************************************************/
628    
629    
630    int
631    xvid_global(void *handle,
632                      int opt,
633                      void *param1,
634                      void *param2)
635    {
636            switch(opt)
637            {
638                    case XVID_GBL_INIT :
639                            return xvid_gbl_init((xvid_gbl_init_t*)param1);
640    
641            case XVID_GBL_INFO :
642                return xvid_gbl_info((xvid_gbl_info_t*)param1);
643    
644                    case XVID_GBL_CONVERT :
645                            return xvid_gbl_convert((xvid_gbl_convert_t*)param1);
646    
647                    default :
648                            return XVID_ERR_FAIL;
649            }
650  }  }
651    
652  /*****************************************************************************  /*****************************************************************************
# Line 443  Line 666 
666                          void *param2)                          void *param2)
667  {  {
668          switch (opt) {          switch (opt) {
         case XVID_DEC_DECODE:  
                 return decoder_decode((DECODER *) handle, (XVID_DEC_FRAME *) param1);  
   
669          case XVID_DEC_CREATE:          case XVID_DEC_CREATE:
670                  return decoder_create((XVID_DEC_PARAM *) param1);                  return decoder_create((xvid_dec_create_t *) param1);
671    
672          case XVID_DEC_DESTROY:          case XVID_DEC_DESTROY:
673                  return decoder_destroy((DECODER *) handle);                  return decoder_destroy((DECODER *) handle);
674    
675            case XVID_DEC_DECODE:
676                    return decoder_decode((DECODER *) handle, (xvid_dec_frame_t *) param1, (xvid_dec_stats_t*) param2);
677    
678          default:          default:
679                  return XVID_ERR_FAIL;                  return XVID_ERR_FAIL;
680          }          }
# Line 476  Line 699 
699  {  {
700          switch (opt) {          switch (opt) {
701          case XVID_ENC_ENCODE:          case XVID_ENC_ENCODE:
702  #ifdef BFRAMES  
703                  if (((Encoder *) handle)->mbParam.max_bframes >= 0)                  return enc_encode((Encoder *) handle,
704                  return encoder_encode_bframes((Encoder *) handle, (XVID_ENC_FRAME *) param1,                                                            (xvid_enc_frame_t *) param1,
705                                                            (XVID_ENC_STATS *) param2);                                                            (xvid_enc_stats_t *) param2);
                 else  
 #endif  
                 return encoder_encode((Encoder *) handle, (XVID_ENC_FRAME *) param1,  
                                                           (XVID_ENC_STATS *) param2);  
706    
707          case XVID_ENC_CREATE:          case XVID_ENC_CREATE:
708                  return encoder_create((XVID_ENC_PARAM *) param1);                  return enc_create((xvid_enc_create_t *) param1);
709    
710          case XVID_ENC_DESTROY:          case XVID_ENC_DESTROY:
711                  return encoder_destroy((Encoder *) handle);                  return enc_destroy((Encoder *) handle);
712    
713          default:          default:
714                  return XVID_ERR_FAIL;                  return XVID_ERR_FAIL;
Legend:
Removed from v.1.33.2.1  
changed lines
  Added in v.1.45.2.23
No admin address has been configured
 ViewVC Help
Powered by ViewVC 1.0.4