[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 1.2, Sat Mar 16 15:52:50 2002 UTC revision 1.33.2.17, Mon Dec 9 10:47:05 2002 UTC
# Line 1  Line 1 
1  /**************************************************************************  /*****************************************************************************
2   *   *
3   *      XVID MPEG-4 VIDEO CODEC   *      XVID MPEG-4 VIDEO CODEC
4   *      native api   *  - Native API implementation  -
5   *   *
6   *      This program is an implementation of a part of one or more MPEG-4   *      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   *      Video tools as specified in ISO/IEC 14496-2 standard.  Those intending
# Line 24  Line 24 
24   *   *
25   *      You should have received a copy of the GNU General Public License   *      You should have received a copy of the GNU General Public License
26   *      along with this program; if not, write to the Free Software   *      along with this program; if not, write to the Free Software
27   *      Foundation, Inc., 675 Mass Ave, Cambridge, MA 02139, USA.   *  Foundation, Inc., 59 Temple Place, Suite 330, Boston, MA  02111-1307 USA
28   *   *
29   *************************************************************************/   ****************************************************************************/
30    
31  /**************************************************************************  /*****************************************************************************
32   *   *
33   *      History:   *  History
34   *   *
35   *  22.12.2001  API change: added xvid_init() - Isibaar   *      - 23.06.2002    added XVID_CPU_CHKONLY
36   *      16.12.2001      inital version; (c)2001 peter ross <pross@cs.rmit.edu.au>   *  - 17.03.2002        Added interpolate8x8_halfpel_hv_xmm
37     *  - 22.12.2001  API change: added xvid_init() - Isibaar
38     *  - 16.12.2001        inital version; (c)2001 peter ross <pross@cs.rmit.edu.au>
39   *   *
40   *************************************************************************/   *  $Id$
41     *
42     ****************************************************************************/
43    
44  #include "xvid.h"  #include "xvid.h"
45  #include "decoder.h"  #include "decoder.h"
# Line 46  Line 49 
49  #include "dct/fdct.h"  #include "dct/fdct.h"
50  #include "image/colorspace.h"  #include "image/colorspace.h"
51  #include "image/interpolate8x8.h"  #include "image/interpolate8x8.h"
52    #include "image/reduced.h"
53  #include "utils/mem_transfer.h"  #include "utils/mem_transfer.h"
54    #include "utils/mbfunctions.h"
55  #include "quant/quant_h263.h"  #include "quant/quant_h263.h"
56  #include "quant/quant_mpeg4.h"  #include "quant/quant_mpeg4.h"
57    #include "motion/motion.h"
58  #include "motion/sad.h"  #include "motion/sad.h"
59  #include "utils/emms.h"  #include "utils/emms.h"
60  #include "utils/timer.h"  #include "utils/timer.h"
61    #include "bitstream/mbcoding.h"
62    
63    #if defined(ARCH_X86) && defined(EXPERIMENTAL_SSE2_CODE)
64    
65    #ifdef WIN32
66    #include <windows.h>
67    #else
68    #include <signal.h>
69    #include <setjmp.h>
70    #endif
71    
72    
73    #ifndef WIN32
74    
75    static jmp_buf mark;
76    
77    static void
78    sigill_handler(int signal)
79    {
80       longjmp(mark, 1);
81    }
82    #endif
83    
84    
85    /*
86    calls the funcptr, and returns whether SIGILL (illegal instruction) was signalled
87    return values:
88    -1 : could not determine
89    0  : SIGILL was *not* signalled
90    1  : SIGILL was signalled
91    */
92    
93    int
94    sigill_check(void (*func)())
95    {
96    #ifdef WIN32
97            _try {
98                    func();
99            }
100            _except(EXCEPTION_EXECUTE_HANDLER) {
101    
102                    if (_exception_code() == STATUS_ILLEGAL_INSTRUCTION)
103                            return 1;
104            }
105            return 0;
106    #else
107        void * old_handler;
108        int jmpret;
109    
110    
111        old_handler = signal(SIGILL, sigill_handler);
112        if (old_handler == SIG_ERR)
113        {
114            return -1;
115        }
116    
117        jmpret = setjmp(mark);
118        if (jmpret == 0)
119        {
120            func();
121        }
122    
123        signal(SIGILL, old_handler);
124    
125        return jmpret;
126    #endif
127    }
128    #endif
129    
130    /*****************************************************************************
131     * XviD Init Entry point
132     *
133     * Well this function initialize all internal function pointers according
134     * to the CPU features forced by the library client or autodetected (depending
135     * on the XVID_CPU_FORCE flag). It also initializes vlc coding tables and all
136     * image colorspace transformation tables.
137     *
138     * Returned value : XVID_ERR_OK
139     *                  + API_VERSION in the input XVID_INIT_PARAM structure
140     *                  + core build  "   "    "       "               "
141     *
142     ****************************************************************************/
143    
144    
145  int xvid_init(void *handle, int opt, void *param1, void *param2)  static
146    int xvid_init_init(XVID_INIT_PARAM * init_param)
147  {  {
148          int cpu_flags;          int cpu_flags;
         XVID_INIT_PARAM *init_param;  
149    
150          init_param = (XVID_INIT_PARAM *) param1;          /* Inform the client the API version */
151            init_param->api_version = API_VERSION;
152    
153            /* Inform the client the core build - unused because we're still alpha */
154            init_param->core_build = 1000;
155    
156            /* Do we have to force CPU features  ? */
157            if ((init_param->cpu_flags & XVID_CPU_FORCE)) {
158    
         // force specific cpu settings?  
         if((init_param->cpu_flags & XVID_CPU_FORCE) > 0)  
159                  cpu_flags = init_param->cpu_flags;                  cpu_flags = init_param->cpu_flags;
         else {  
160    
161  #ifdef ARCH_X86          } else {
162    
163                  cpu_flags = check_cpu_features();                  cpu_flags = check_cpu_features();
164  #else  
165                  cpu_flags = 0;  #if defined(ARCH_X86) && defined(EXPERIMENTAL_SSE2_CODE)
166                    if ((cpu_flags & XVID_CPU_SSE) && sigill_check(sse_os_trigger))
167                            cpu_flags &= ~XVID_CPU_SSE;
168    
169                    if ((cpu_flags & XVID_CPU_SSE2) && sigill_check(sse2_os_trigger))
170                            cpu_flags &= ~XVID_CPU_SSE2;
171  #endif  #endif
172            }
173    
174            if ((init_param->cpu_flags & XVID_CPU_CHKONLY))
175            {
176                  init_param->cpu_flags = cpu_flags;                  init_param->cpu_flags = cpu_flags;
177                    return XVID_ERR_OK;
178          }          }
179    
180          // initialize the function pointers          init_param->cpu_flags = cpu_flags;
181    
182    
183            /* Initialize the function pointers */
184          idct_int32_init();          idct_int32_init();
185            init_vlc_tables();
186    
187            /* Fixed Point Forward/Inverse DCT transformations */
188          fdct = fdct_int32;          fdct = fdct_int32;
189          idct = idct_int32;          idct = idct_int32;
190    
191            /* Only needed on PPC Altivec archs */
192            sadInit = 0;
193    
194            /* Restore FPU context : emms_c is a nop functions */
195          emms = emms_c;          emms = emms_c;
196    
197            /* Quantization functions */
198          quant_intra = quant_intra_c;          quant_intra = quant_intra_c;
199          dequant_intra = dequant_intra_c;          dequant_intra = dequant_intra_c;
200          quant_inter = quant_inter_c;          quant_inter = quant_inter_c;
# Line 91  Line 205 
205          quant4_inter = quant4_inter_c;          quant4_inter = quant4_inter_c;
206          dequant4_inter = dequant4_inter_c;          dequant4_inter = dequant4_inter_c;
207    
208            /* Block transfer related functions */
209          transfer_8to16copy = transfer_8to16copy_c;          transfer_8to16copy = transfer_8to16copy_c;
210          transfer_16to8copy = transfer_16to8copy_c;          transfer_16to8copy = transfer_16to8copy_c;
211          transfer_8to16sub = transfer_8to16sub_c;          transfer_8to16sub = transfer_8to16sub_c;
212            transfer_8to16sub2 = transfer_8to16sub2_c;
213          transfer_16to8add = transfer_16to8add_c;          transfer_16to8add = transfer_16to8add_c;
214          transfer8x8_copy = transfer8x8_copy_c;          transfer8x8_copy = transfer8x8_copy_c;
215    
216            /* Interlacing functions */
217            MBFieldTest = MBFieldTest_c;
218    
219            /* Image interpolation related functions */
220          interpolate8x8_halfpel_h = interpolate8x8_halfpel_h_c;          interpolate8x8_halfpel_h = interpolate8x8_halfpel_h_c;
221          interpolate8x8_halfpel_v = interpolate8x8_halfpel_v_c;          interpolate8x8_halfpel_v = interpolate8x8_halfpel_v_c;
222          interpolate8x8_halfpel_hv = interpolate8x8_halfpel_hv_c;          interpolate8x8_halfpel_hv = interpolate8x8_halfpel_hv_c;
223    
224            interpolate16x16_lowpass_h = interpolate16x16_lowpass_h_c;
225            interpolate16x16_lowpass_v = interpolate16x16_lowpass_v_c;
226            interpolate16x16_lowpass_hv = interpolate16x16_lowpass_hv_c;
227    
228            interpolate8x8_lowpass_h = interpolate8x8_lowpass_h_c;
229            interpolate8x8_lowpass_v = interpolate8x8_lowpass_v_c;
230            interpolate8x8_lowpass_hv = interpolate8x8_lowpass_hv_c;
231    
232            interpolate8x8_6tap_lowpass_h = interpolate8x8_6tap_lowpass_h_c;
233            interpolate8x8_6tap_lowpass_v = interpolate8x8_6tap_lowpass_v_c;
234    
235            interpolate8x8_avg2 = interpolate8x8_avg2_c;
236            interpolate8x8_avg4 = interpolate8x8_avg4_c;
237    
238            /* reduced resoltuion */
239    
240            copy_upsampled_8x8_16to8 = xvid_Copy_Upsampled_8x8_16To8_C;
241            add_upsampled_8x8_16to8 = xvid_Add_Upsampled_8x8_16To8_C;
242    #ifdef ARCH_X86
243            vfilter_31 = xvid_VFilter_31_x86;
244            hfilter_31 = xvid_HFilter_31_x86;
245    #else
246            vfilter_31 = xvid_VFilter_31_C;
247            hfilter_31 = xvid_HFilter_31_C;
248    #endif
249            filter_18x18_to_8x8 = xvid_Filter_18x18_To_8x8_C;
250            filter_diff_18x18_to_8x8 = xvid_Filter_Diff_18x18_To_8x8_C;
251    
252            /* Initialize internal colorspace transformation tables */
253          colorspace_init();          colorspace_init();
254    
255            /* All colorspace transformation functions User Format->YV12 */
256            yv12_to_yv12    = yv12_to_yv12_c;
257          rgb555_to_yv12 = rgb555_to_yv12_c;          rgb555_to_yv12 = rgb555_to_yv12_c;
258          rgb565_to_yv12 = rgb565_to_yv12_c;          rgb565_to_yv12 = rgb565_to_yv12_c;
259          rgb24_to_yv12 = rgb24_to_yv12_c;          bgr_to_yv12     = bgr_to_yv12_c;
260          rgb32_to_yv12 = rgb32_to_yv12_c;          bgra_to_yv12    = bgra_to_yv12_c;
261          yuv_to_yv12 = yuv_to_yv12_c;          abgr_to_yv12    = abgr_to_yv12_c;
262            rgba_to_yv12    = rgba_to_yv12_c;
263          yuyv_to_yv12 = yuyv_to_yv12_c;          yuyv_to_yv12 = yuyv_to_yv12_c;
264          uyvy_to_yv12 = uyvy_to_yv12_c;          uyvy_to_yv12 = uyvy_to_yv12_c;
265    
266            rgb555i_to_yv12 = rgb555i_to_yv12_c;
267            rgb565i_to_yv12 = rgb565i_to_yv12_c;
268            bgri_to_yv12    = bgri_to_yv12_c;
269            bgrai_to_yv12   = bgrai_to_yv12_c;
270            abgri_to_yv12   = abgri_to_yv12_c;
271            rgbai_to_yv12   = rgbai_to_yv12_c;
272            yuyvi_to_yv12   = yuyvi_to_yv12_c;
273            uyvyi_to_yv12   = uyvyi_to_yv12_c;
274    
275    
276            /* All colorspace transformation functions YV12->User format */
277          yv12_to_rgb555 = yv12_to_rgb555_c;          yv12_to_rgb555 = yv12_to_rgb555_c;
278          yv12_to_rgb565 = yv12_to_rgb565_c;          yv12_to_rgb565 = yv12_to_rgb565_c;
279          yv12_to_rgb24 = yv12_to_rgb24_c;          yv12_to_bgr     = yv12_to_bgr_c;
280          yv12_to_rgb32 = yv12_to_rgb32_c;          yv12_to_bgra    = yv12_to_bgra_c;
281          yv12_to_yuv = yv12_to_yuv_c;          yv12_to_abgr    = yv12_to_abgr_c;
282            yv12_to_rgba    = yv12_to_rgba_c;
283          yv12_to_yuyv = yv12_to_yuyv_c;          yv12_to_yuyv = yv12_to_yuyv_c;
284          yv12_to_uyvy = yv12_to_uyvy_c;          yv12_to_uyvy = yv12_to_uyvy_c;
285    
286            yv12_to_rgb555i = yv12_to_rgb555i_c;
287            yv12_to_rgb565i = yv12_to_rgb565i_c;
288            yv12_to_bgri    = yv12_to_bgri_c;
289            yv12_to_bgrai   = yv12_to_bgrai_c;
290            yv12_to_abgri   = yv12_to_abgri_c;
291            yv12_to_rgbai   = yv12_to_rgbai_c;
292            yv12_to_yuyvi   = yv12_to_yuyvi_c;
293            yv12_to_uyvyi   = yv12_to_uyvyi_c;
294    
295            /* Functions used in motion estimation algorithms */
296          calc_cbp = calc_cbp_c;          calc_cbp = calc_cbp_c;
297          sad16 = sad16_c;          sad16 = sad16_c;
298          sad8 = sad8_c;          sad8 = sad8_c;
299            sad16bi  = sad16bi_c;
300            sad8bi   = sad8bi_c;
301          dev16 = dev16_c;          dev16 = dev16_c;
302            sad16v   = sad16v_c;
303    
304    //      Halfpel8_Refine = Halfpel8_Refine_c;
305    
306  #ifdef ARCH_X86  #ifdef ARCH_X86
307    
308            if ((cpu_flags & XVID_CPU_MMX) || (cpu_flags & XVID_CPU_MMXEXT) ||
309                    (cpu_flags & XVID_CPU_3DNOW) || (cpu_flags & XVID_CPU_3DNOWEXT) ||
310                    (cpu_flags & XVID_CPU_SSE) || (cpu_flags & XVID_CPU_SSE2))
311            {
312                    /* Restore FPU context : emms_c is a nop functions */
313                    emms = emms_mmx;
314            }
315    
316          if((cpu_flags & XVID_CPU_MMX) > 0) {          if((cpu_flags & XVID_CPU_MMX) > 0) {
317    
318                    /* Forward and Inverse Discrete Cosine Transformation functions */
319                  fdct = fdct_mmx;                  fdct = fdct_mmx;
320                  idct = idct_mmx;                  idct = idct_mmx;
321    
322                  emms = emms_mmx;                  /* Quantization related functions */
   
323                  quant_intra = quant_intra_mmx;                  quant_intra = quant_intra_mmx;
324                  dequant_intra = dequant_intra_mmx;                  dequant_intra = dequant_intra_mmx;
325                  quant_inter = quant_inter_mmx;                  quant_inter = quant_inter_mmx;
# Line 141  Line 330 
330                  quant4_inter = quant4_inter_mmx;                  quant4_inter = quant4_inter_mmx;
331                  dequant4_inter = dequant4_inter_mmx;                  dequant4_inter = dequant4_inter_mmx;
332    
333                    /* Block related functions */
334                  transfer_8to16copy = transfer_8to16copy_mmx;                  transfer_8to16copy = transfer_8to16copy_mmx;
335                  transfer_16to8copy = transfer_16to8copy_mmx;                  transfer_16to8copy = transfer_16to8copy_mmx;
336                  transfer_8to16sub = transfer_8to16sub_mmx;                  transfer_8to16sub = transfer_8to16sub_mmx;
337                    transfer_8to16sub2 = transfer_8to16sub2_mmx;
338                  transfer_16to8add = transfer_16to8add_mmx;                  transfer_16to8add = transfer_16to8add_mmx;
339                  transfer8x8_copy = transfer8x8_copy_mmx;                  transfer8x8_copy = transfer8x8_copy_mmx;
340    
341                    /* Interlacing Functions */
342                    MBFieldTest = MBFieldTest_mmx;
343    
344                    /* Image Interpolation related functions */
345                  interpolate8x8_halfpel_h = interpolate8x8_halfpel_h_mmx;                  interpolate8x8_halfpel_h = interpolate8x8_halfpel_h_mmx;
346                  interpolate8x8_halfpel_v = interpolate8x8_halfpel_v_mmx;                  interpolate8x8_halfpel_v = interpolate8x8_halfpel_v_mmx;
347                  interpolate8x8_halfpel_hv = interpolate8x8_halfpel_hv_mmx;                  interpolate8x8_halfpel_hv = interpolate8x8_halfpel_hv_mmx;
348    
349                  rgb24_to_yv12 = rgb24_to_yv12_mmx;                  interpolate8x8_6tap_lowpass_h = interpolate8x8_6tap_lowpass_h_mmx;
350                  rgb32_to_yv12 = rgb32_to_yv12_mmx;                  interpolate8x8_6tap_lowpass_v = interpolate8x8_6tap_lowpass_v_mmx;
351                  yuv_to_yv12 = yuv_to_yv12_mmx;  
352                    interpolate8x8_avg2 = interpolate8x8_avg2_mmx;
353                    interpolate8x8_avg4 = interpolate8x8_avg4_mmx;
354    
355                    /* reduced resolution */
356                    copy_upsampled_8x8_16to8 = xvid_Copy_Upsampled_8x8_16To8_mmx;
357                    add_upsampled_8x8_16to8 = xvid_Add_Upsampled_8x8_16To8_mmx;
358                    hfilter_31 = xvid_HFilter_31_mmx;
359                    filter_18x18_to_8x8 = xvid_Filter_18x18_To_8x8_mmx;
360                    filter_diff_18x18_to_8x8 = xvid_Filter_Diff_18x18_To_8x8_mmx;
361    
362                    /* image input xxx_to_yv12 related functions */
363                    yv12_to_yv12  = yv12_to_yv12_mmx;
364                    bgr_to_yv12   = bgr_to_yv12_mmx;
365                    bgra_to_yv12  = bgra_to_yv12_mmx;
366                  yuyv_to_yv12 = yuyv_to_yv12_mmx;                  yuyv_to_yv12 = yuyv_to_yv12_mmx;
367                  uyvy_to_yv12 = uyvy_to_yv12_mmx;                  uyvy_to_yv12 = uyvy_to_yv12_mmx;
368    
369                  yv12_to_rgb24 = yv12_to_rgb24_mmx;                  /* image output yv12_to_xxx related functions */
370                  yv12_to_rgb32 = yv12_to_rgb32_mmx;                  yv12_to_bgr   = yv12_to_bgr_mmx;
371                    yv12_to_bgra  = yv12_to_bgra_mmx;
372                  yv12_to_yuyv = yv12_to_yuyv_mmx;                  yv12_to_yuyv = yv12_to_yuyv_mmx;
373                  yv12_to_uyvy = yv12_to_uyvy_mmx;                  yv12_to_uyvy = yv12_to_uyvy_mmx;
374    
375                    yv12_to_yuyvi = yv12_to_yuyvi_mmx;
376                    yv12_to_uyvyi = yv12_to_uyvyi_mmx;
377    
378                    /* Motion estimation related functions */
379                  calc_cbp = calc_cbp_mmx;                  calc_cbp = calc_cbp_mmx;
380                  sad16 = sad16_mmx;                  sad16 = sad16_mmx;
381                  sad8 = sad8_mmx;                  sad8 = sad8_mmx;
382                    sad16bi = sad16bi_mmx;
383                    sad8bi  = sad8bi_mmx;
384                  dev16 = dev16_mmx;                  dev16 = dev16_mmx;
385                    sad16v   = sad16v_mmx;
386    
387          }          }
388    
389            /* these 3dnow functions are faster than mmx, but slower than xmm. */
390            if ((cpu_flags & XVID_CPU_3DNOW) > 0) {
391    
392                    /* ME functions */
393                    sad16bi = sad16bi_3dn;
394                    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) > 0) {
402    
403                    /* Inverse DCT */
404                  idct = idct_xmm;                  idct = idct_xmm;
405    
406                    /* Interpolation */
407                  interpolate8x8_halfpel_h = interpolate8x8_halfpel_h_xmm;                  interpolate8x8_halfpel_h = interpolate8x8_halfpel_h_xmm;
408                  interpolate8x8_halfpel_v = interpolate8x8_halfpel_v_xmm;                  interpolate8x8_halfpel_v = interpolate8x8_halfpel_v_xmm;
409                  yuv_to_yv12 = yuv_to_yv12_xmm;                  interpolate8x8_halfpel_hv = interpolate8x8_halfpel_hv_xmm;
410    
411                    /* reduced resolution */
412                    copy_upsampled_8x8_16to8 = xvid_Copy_Upsampled_8x8_16To8_xmm;
413                    add_upsampled_8x8_16to8 = xvid_Add_Upsampled_8x8_16To8_xmm;
414    
415                    /* Quantization */
416                    dequant_intra = dequant_intra_xmm;
417                    dequant_inter = dequant_inter_xmm;
418    
419                    /* Buffer transfer */
420                    transfer_8to16sub2 = transfer_8to16sub2_xmm;
421    
422                    /* Colorspace transformation */
423                    yv12_to_yv12  = yv12_to_yv12_xmm;
424                    yuyv_to_yv12  = yuyv_to_yv12_xmm;
425                    uyvy_to_yv12  = uyvy_to_yv12_xmm;
426    
427                    /* ME functions */
428                  sad16 = sad16_xmm;                  sad16 = sad16_xmm;
429                  sad8 = sad8_xmm;                  sad8 = sad8_xmm;
430                    sad16bi = sad16bi_xmm;
431                    sad8bi  = sad8bi_xmm;
432                  dev16 = dev16_xmm;                  dev16 = dev16_xmm;
433                    sad16v   = sad16v_xmm;
434          }          }
435    
436          if((cpu_flags & XVID_CPU_3DNOW) > 0) {          if((cpu_flags & XVID_CPU_3DNOW) > 0) {
437    
438                    /* Interpolation */
439                  interpolate8x8_halfpel_h = interpolate8x8_halfpel_h_3dn;                  interpolate8x8_halfpel_h = interpolate8x8_halfpel_h_3dn;
440                  interpolate8x8_halfpel_v = interpolate8x8_halfpel_v_3dn;                  interpolate8x8_halfpel_v = interpolate8x8_halfpel_v_3dn;
441                    interpolate8x8_halfpel_hv = interpolate8x8_halfpel_hv_3dn;
442          }          }
443    
444            if ((cpu_flags & XVID_CPU_SSE2) > 0) {
445    #ifdef EXPERIMENTAL_SSE2_CODE
446    
447                    calc_cbp = calc_cbp_sse2;
448    
449                    /* Quantization */
450                    quant_intra   = quant_intra_sse2;
451                    dequant_intra = dequant_intra_sse2;
452                    quant_inter   = quant_inter_sse2;
453                    dequant_inter = dequant_inter_sse2;
454    
455                    /* ME */
456                    sad16    = sad16_sse2;
457                    dev16    = dev16_sse2;
458    
459                    /* Forward and Inverse DCT */
460                    idct  = idct_sse2;
461                    fdct = fdct_sse2;
462  #endif  #endif
463            }
464    
465          // API version  #endif
         init_param->api_version = API_VERSION;  
466    
467          // something clever has to be done for this  #ifdef ARCH_IA64
468          init_param->core_build = 1000;          if ((cpu_flags & XVID_CPU_IA64) > 0) { //use assembler routines?
469              idct_ia64_init();
470              fdct = fdct_ia64;
471              idct = idct_ia64;   //not yet working, crashes
472              interpolate8x8_halfpel_h = interpolate8x8_halfpel_h_ia64;
473              interpolate8x8_halfpel_v = interpolate8x8_halfpel_v_ia64;
474              interpolate8x8_halfpel_hv = interpolate8x8_halfpel_hv_ia64;
475              sad16 = sad16_ia64;
476              sad16bi = sad16bi_ia64;
477              sad8 = sad8_ia64;
478              dev16 = dev16_ia64;
479    //        Halfpel8_Refine = Halfpel8_Refine_ia64;
480              quant_intra = quant_intra_ia64;
481              dequant_intra = dequant_intra_ia64;
482              quant_inter = quant_inter_ia64;
483              dequant_inter = dequant_inter_ia64;
484              transfer_8to16copy = transfer_8to16copy_ia64;
485              transfer_16to8copy = transfer_16to8copy_ia64;
486              transfer_8to16sub = transfer_8to16sub_ia64;
487              transfer_8to16sub2 = transfer_8to16sub2_ia64;
488              transfer_16to8add = transfer_16to8add_ia64;
489              transfer8x8_copy = transfer8x8_copy_ia64;
490              DEBUG("Using IA-64 assembler routines.\n");
491            }
492    #endif
493    
494    #ifdef ARCH_PPC
495    #ifdef ARCH_PPC_ALTIVEC
496            calc_cbp = calc_cbp_altivec;
497            fdct = fdct_altivec;
498            idct = idct_altivec;
499            sadInit = sadInit_altivec;
500            sad16 = sad16_altivec;
501            sad8 = sad8_altivec;
502            dev16 = dev16_altivec;
503    #else
504            calc_cbp = calc_cbp_ppc;
505    #endif
506    #endif
507    
508            return XVID_ERR_OK;
509    }
510    
511    
512    
513    static int
514    xvid_init_convert(XVID_INIT_CONVERTINFO* convert)
515    {
516            // const int flip1 = (convert->input.colorspace & XVID_CSP_VFLIP) ^ (convert->output.colorspace & XVID_CSP_VFLIP);
517            const int width = convert->width;
518            const int height = convert->height;
519            const int width2 = convert->width/2;
520            const int height2 = convert->height/2;
521            IMAGE img;
522    
523            switch (convert->input.colorspace & ~XVID_CSP_VFLIP)
524            {
525                    case XVID_CSP_YV12 :
526                            img.y = convert->input.y;
527                            img.v = (uint8_t*)convert->input.y + width*height;
528                            img.u = (uint8_t*)convert->input.y + width*height + width2*height2;
529                            image_output(&img, width, height, width,
530                                                    convert->output.y, convert->output.y_stride,
531                                                    convert->output.colorspace, convert->interlacing);
532                            break;
533    
534                    default :
535                            return XVID_ERR_FORMAT;
536            }
537    
538    
539            emms();
540          return XVID_ERR_OK;          return XVID_ERR_OK;
541  }  }
542    
543  int xvid_decore(void * handle, int opt, void * param1, void * param2)  
544    int
545    xvid_init(void *handle,
546                      int opt,
547                      void *param1,
548                      void *param2)
549  {  {
550          switch (opt)          switch (opt)
551          {          {
552                    case XVID_INIT_INIT :
553                            return xvid_init_init((XVID_INIT_PARAM*)param1);
554    
555                    case XVID_INIT_CONVERT :
556                            return xvid_init_convert((XVID_INIT_CONVERTINFO*)param1);
557    
558                    default :
559                            return XVID_ERR_FAIL;
560            }
561    }
562    
563    /*****************************************************************************
564     * XviD Native decoder entry point
565     *
566     * This function is just a wrapper to all the option cases.
567     *
568     * Returned values : XVID_ERR_FAIL when opt is invalid
569     *                   else returns the wrapped function result
570     *
571     ****************************************************************************/
572    
573    int
574    xvid_decore(void *handle,
575                            int opt,
576                            void *param1,
577                            void *param2)
578    {
579            switch (opt) {
580          case XVID_DEC_DECODE :          case XVID_DEC_DECODE :
581          return decoder_decode((DECODER *) handle, (XVID_DEC_FRAME *) param1);                  return decoder_decode((DECODER *) handle, (XVID_DEC_FRAME *) param1, (XVID_DEC_STATS*) param2);
582    
583          case XVID_DEC_CREATE :          case XVID_DEC_CREATE :
584          return decoder_create((XVID_DEC_PARAM *) param1);          return decoder_create((XVID_DEC_PARAM *) param1);
# Line 216  Line 592 
592  }  }
593    
594    
595  int xvid_encore(void * handle, int opt, void * param1, void * param2)  /*****************************************************************************
596  {   * XviD Native encoder entry point
597          switch (opt)   *
598     * This function is just a wrapper to all the option cases.
599     *
600     * Returned values : XVID_ERR_FAIL when opt is invalid
601     *                   else returns the wrapped function result
602     *
603     ****************************************************************************/
604    
605    int
606    xvid_encore(void *handle,
607                            int opt,
608                            void *param1,
609                            void *param2)
610          {          {
611            switch (opt) {
612          case XVID_ENC_ENCODE :          case XVID_ENC_ENCODE :
613          return encoder_encode((Encoder *) handle, (XVID_ENC_FRAME *) param1, (XVID_ENC_STATS *) param2);  
614                    if (((Encoder *) handle)->mbParam.max_bframes >= 0)
615                    return encoder_encode_bframes((Encoder *) handle, (XVID_ENC_FRAME *) param1,
616                                                              (XVID_ENC_STATS *) param2);
617                    else
618                    return encoder_encode((Encoder *) handle, (XVID_ENC_FRAME *) param1,
619                                                              (XVID_ENC_STATS *) param2);
620    
621          case XVID_ENC_CREATE :          case XVID_ENC_CREATE :
622          return encoder_create((XVID_ENC_PARAM *) param1);          return encoder_create((XVID_ENC_PARAM *) param1);

Legend:
Removed from v.1.2  
changed lines
  Added in v.1.33.2.17

No admin address has been configured
ViewVC Help
Powered by ViewVC 1.0.4