src/video/SDL_yuv_sw.c
author Sam Lantinga <slouken@libsdl.org>
Sat, 18 May 2002 11:30:38 +0000
changeset 366 c94b390687d2
parent 297 f6ffac90895c
child 739 22dbf364c017
permissions -rw-r--r--
Avoid a conflict in the definition of int32
slouken@0
     1
/*
slouken@0
     2
    SDL - Simple DirectMedia Layer
slouken@297
     3
    Copyright (C) 1997, 1998, 1999, 2000, 2001, 2002  Sam Lantinga
slouken@0
     4
slouken@0
     5
    This library is free software; you can redistribute it and/or
slouken@0
     6
    modify it under the terms of the GNU Library General Public
slouken@0
     7
    License as published by the Free Software Foundation; either
slouken@0
     8
    version 2 of the License, or (at your option) any later version.
slouken@0
     9
slouken@0
    10
    This library is distributed in the hope that it will be useful,
slouken@0
    11
    but WITHOUT ANY WARRANTY; without even the implied warranty of
slouken@0
    12
    MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the GNU
slouken@0
    13
    Library General Public License for more details.
slouken@0
    14
slouken@0
    15
    You should have received a copy of the GNU Library General Public
slouken@0
    16
    License along with this library; if not, write to the Free
slouken@0
    17
    Foundation, Inc., 59 Temple Place, Suite 330, Boston, MA  02111-1307  USA
slouken@0
    18
slouken@0
    19
    Sam Lantinga
slouken@252
    20
    slouken@libsdl.org
slouken@0
    21
*/
slouken@0
    22
slouken@0
    23
#ifdef SAVE_RCSID
slouken@0
    24
static char rcsid =
slouken@0
    25
 "@(#) $Id$";
slouken@0
    26
#endif
slouken@0
    27
slouken@0
    28
/* This is the software implementation of the YUV video overlay support */
slouken@0
    29
slouken@0
    30
/* This code was derived from code carrying the following copyright notices:
slouken@0
    31
slouken@0
    32
 * Copyright (c) 1995 The Regents of the University of California.
slouken@0
    33
 * All rights reserved.
slouken@0
    34
 * 
slouken@0
    35
 * Permission to use, copy, modify, and distribute this software and its
slouken@0
    36
 * documentation for any purpose, without fee, and without written agreement is
slouken@0
    37
 * hereby granted, provided that the above copyright notice and the following
slouken@0
    38
 * two paragraphs appear in all copies of this software.
slouken@0
    39
 * 
slouken@0
    40
 * IN NO EVENT SHALL THE UNIVERSITY OF CALIFORNIA BE LIABLE TO ANY PARTY FOR
slouken@0
    41
 * DIRECT, INDIRECT, SPECIAL, INCIDENTAL, OR CONSEQUENTIAL DAMAGES ARISING OUT
slouken@0
    42
 * OF THE USE OF THIS SOFTWARE AND ITS DOCUMENTATION, EVEN IF THE UNIVERSITY OF
slouken@0
    43
 * CALIFORNIA HAS BEEN ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
slouken@0
    44
 * 
slouken@0
    45
 * THE UNIVERSITY OF CALIFORNIA SPECIFICALLY DISCLAIMS ANY WARRANTIES,
slouken@0
    46
 * INCLUDING, BUT NOT LIMITED TO, THE IMPLIED WARRANTIES OF MERCHANTABILITY
slouken@0
    47
 * AND FITNESS FOR A PARTICULAR PURPOSE.  THE SOFTWARE PROVIDED HEREUNDER IS
slouken@0
    48
 * ON AN "AS IS" BASIS, AND THE UNIVERSITY OF CALIFORNIA HAS NO OBLIGATION TO
slouken@0
    49
 * PROVIDE MAINTENANCE, SUPPORT, UPDATES, ENHANCEMENTS, OR MODIFICATIONS.
slouken@0
    50
slouken@0
    51
 * Copyright (c) 1995 Erik Corry
slouken@0
    52
 * All rights reserved.
slouken@0
    53
 * 
slouken@0
    54
 * Permission to use, copy, modify, and distribute this software and its
slouken@0
    55
 * documentation for any purpose, without fee, and without written agreement is
slouken@0
    56
 * hereby granted, provided that the above copyright notice and the following
slouken@0
    57
 * two paragraphs appear in all copies of this software.
slouken@0
    58
 * 
slouken@0
    59
 * IN NO EVENT SHALL ERIK CORRY BE LIABLE TO ANY PARTY FOR DIRECT, INDIRECT,
slouken@0
    60
 * SPECIAL, INCIDENTAL, OR CONSEQUENTIAL DAMAGES ARISING OUT OF THE USE OF
slouken@0
    61
 * THIS SOFTWARE AND ITS DOCUMENTATION, EVEN IF ERIK CORRY HAS BEEN ADVISED
slouken@0
    62
 * OF THE POSSIBILITY OF SUCH DAMAGE.
slouken@0
    63
 * 
slouken@0
    64
 * ERIK CORRY SPECIFICALLY DISCLAIMS ANY WARRANTIES, INCLUDING, BUT NOT
slouken@0
    65
 * LIMITED TO, THE IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A
slouken@0
    66
 * PARTICULAR PURPOSE.  THE SOFTWARE PROVIDED HEREUNDER IS ON AN "AS IS"
slouken@0
    67
 * BASIS, AND ERIK CORRY HAS NO OBLIGATION TO PROVIDE MAINTENANCE, SUPPORT,
slouken@0
    68
 * UPDATES, ENHANCEMENTS, OR MODIFICATIONS.
slouken@0
    69
slouken@0
    70
 * Portions of this software Copyright (c) 1995 Brown University.
slouken@0
    71
 * All rights reserved.
slouken@0
    72
 * 
slouken@0
    73
 * Permission to use, copy, modify, and distribute this software and its
slouken@0
    74
 * documentation for any purpose, without fee, and without written agreement
slouken@0
    75
 * is hereby granted, provided that the above copyright notice and the
slouken@0
    76
 * following two paragraphs appear in all copies of this software.
slouken@0
    77
 * 
slouken@0
    78
 * IN NO EVENT SHALL BROWN UNIVERSITY BE LIABLE TO ANY PARTY FOR
slouken@0
    79
 * DIRECT, INDIRECT, SPECIAL, INCIDENTAL, OR CONSEQUENTIAL DAMAGES ARISING OUT
slouken@0
    80
 * OF THE USE OF THIS SOFTWARE AND ITS DOCUMENTATION, EVEN IF BROWN
slouken@0
    81
 * UNIVERSITY HAS BEEN ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
slouken@0
    82
 * 
slouken@0
    83
 * BROWN UNIVERSITY SPECIFICALLY DISCLAIMS ANY WARRANTIES, INCLUDING, BUT NOT
slouken@0
    84
 * LIMITED TO, THE IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A
slouken@0
    85
 * PARTICULAR PURPOSE.  THE SOFTWARE PROVIDED HEREUNDER IS ON AN "AS IS"
slouken@0
    86
 * BASIS, AND BROWN UNIVERSITY HAS NO OBLIGATION TO PROVIDE MAINTENANCE,
slouken@0
    87
 * SUPPORT, UPDATES, ENHANCEMENTS, OR MODIFICATIONS.
slouken@0
    88
 */
slouken@0
    89
slouken@0
    90
#include <stdlib.h>
slouken@0
    91
#include <string.h>
slouken@0
    92
slouken@0
    93
#include "SDL_error.h"
slouken@0
    94
#include "SDL_video.h"
slouken@0
    95
#include "SDL_stretch_c.h"
slouken@0
    96
#include "SDL_yuvfuncs.h"
slouken@0
    97
#include "SDL_yuv_sw_c.h"
slouken@0
    98
slouken@0
    99
/* Function to check the CPU flags */
slouken@0
   100
#define MMX_CPU		0x800000
slouken@0
   101
#ifdef USE_ASMBLIT
slouken@0
   102
#define CPU_Flags()	Hermes_X86_CPU()
slouken@0
   103
#else
slouken@0
   104
#define CPU_Flags()	0L
slouken@0
   105
#endif
slouken@0
   106
slouken@0
   107
#ifdef USE_ASMBLIT
slouken@0
   108
#define X86_ASSEMBLER
slouken@0
   109
#define HermesConverterInterface	void
slouken@0
   110
#define HermesClearInterface		void
slouken@0
   111
#define STACKCALL
slouken@0
   112
slouken@0
   113
#include "HeadX86.h"
slouken@0
   114
#endif
slouken@0
   115
slouken@0
   116
/* The functions used to manipulate software video overlays */
slouken@0
   117
static struct private_yuvhwfuncs sw_yuvfuncs = {
slouken@0
   118
	SDL_LockYUV_SW,
slouken@0
   119
	SDL_UnlockYUV_SW,
slouken@0
   120
	SDL_DisplayYUV_SW,
slouken@0
   121
	SDL_FreeYUV_SW
slouken@0
   122
};
slouken@0
   123
slouken@0
   124
/* RGB conversion lookup tables */
slouken@0
   125
struct private_yuvhwdata {
slouken@0
   126
	SDL_Surface *stretch;
slouken@0
   127
	SDL_Surface *display;
slouken@0
   128
	Uint8 *pixels;
slouken@0
   129
	int *colortab;
slouken@0
   130
	Uint32 *rgb_2_pix;
slouken@0
   131
	void (*Display1X)(int *colortab, Uint32 *rgb_2_pix,
slouken@0
   132
                          unsigned char *lum, unsigned char *cr,
slouken@0
   133
                          unsigned char *cb, unsigned char *out,
slouken@0
   134
                          int rows, int cols, int mod );
slouken@0
   135
	void (*Display2X)(int *colortab, Uint32 *rgb_2_pix,
slouken@0
   136
	                  unsigned char *lum, unsigned char *cr,
slouken@0
   137
                          unsigned char *cb, unsigned char *out,
slouken@0
   138
                          int rows, int cols, int mod );
slouken@0
   139
slouken@0
   140
	/* These are just so we don't have to allocate them separately */
slouken@0
   141
	Uint16 pitches[3];
slouken@0
   142
	Uint8 *planes[3];
slouken@0
   143
};
slouken@0
   144
slouken@0
   145
slouken@0
   146
/* The colorspace conversion functions */
slouken@0
   147
slouken@0
   148
extern void Color565DitherYV12MMX1X( int *colortab, Uint32 *rgb_2_pix,
slouken@0
   149
                                     unsigned char *lum, unsigned char *cr,
slouken@0
   150
                                     unsigned char *cb, unsigned char *out,
slouken@0
   151
                                     int rows, int cols, int mod );
slouken@0
   152
extern void ColorRGBDitherYV12MMX1X( int *colortab, Uint32 *rgb_2_pix,
slouken@0
   153
                                     unsigned char *lum, unsigned char *cr,
slouken@0
   154
                                     unsigned char *cb, unsigned char *out,
slouken@0
   155
                                     int rows, int cols, int mod );
slouken@0
   156
slouken@0
   157
static void Color16DitherYV12Mod1X( int *colortab, Uint32 *rgb_2_pix,
slouken@0
   158
                                    unsigned char *lum, unsigned char *cr,
slouken@0
   159
                                    unsigned char *cb, unsigned char *out,
slouken@0
   160
                                    int rows, int cols, int mod )
slouken@0
   161
{
slouken@0
   162
    unsigned short* row1;
slouken@0
   163
    unsigned short* row2;
slouken@0
   164
    unsigned char* lum2;
slouken@0
   165
    int x, y;
slouken@0
   166
    int cr_r;
slouken@0
   167
    int crb_g;
slouken@0
   168
    int cb_b;
slouken@0
   169
    int cols_2 = cols / 2;
slouken@0
   170
slouken@0
   171
    row1 = (unsigned short*) out;
slouken@0
   172
    row2 = row1 + cols + mod;
slouken@0
   173
    lum2 = lum + cols;
slouken@0
   174
slouken@0
   175
    mod += cols + mod;
slouken@0
   176
slouken@0
   177
    y = rows / 2;
slouken@0
   178
    while( y-- )
slouken@0
   179
    {
slouken@0
   180
        x = cols_2;
slouken@0
   181
        while( x-- )
slouken@0
   182
        {
slouken@0
   183
            register int L;
slouken@0
   184
slouken@0
   185
            cr_r   = 0*768+256 + colortab[ *cr + 0*256 ];
slouken@0
   186
            crb_g  = 1*768+256 + colortab[ *cr + 1*256 ]
slouken@0
   187
                               + colortab[ *cb + 2*256 ];
slouken@0
   188
            cb_b   = 2*768+256 + colortab[ *cb + 3*256 ];
slouken@0
   189
            ++cr; ++cb;
slouken@0
   190
slouken@0
   191
            L = *lum++;
slouken@0
   192
            *row1++ = (rgb_2_pix[ L + cr_r ] |
slouken@0
   193
                       rgb_2_pix[ L + crb_g ] |
slouken@0
   194
                       rgb_2_pix[ L + cb_b ]);
slouken@0
   195
slouken@0
   196
            L = *lum++;
slouken@0
   197
            *row1++ = (rgb_2_pix[ L + cr_r ] |
slouken@0
   198
                       rgb_2_pix[ L + crb_g ] |
slouken@0
   199
                       rgb_2_pix[ L + cb_b ]);
slouken@0
   200
slouken@0
   201
slouken@0
   202
            /* Now, do second row.  */
slouken@0
   203
slouken@0
   204
            L = *lum2++;
slouken@0
   205
            *row2++ = (rgb_2_pix[ L + cr_r ] |
slouken@0
   206
                       rgb_2_pix[ L + crb_g ] |
slouken@0
   207
                       rgb_2_pix[ L + cb_b ]);
slouken@0
   208
slouken@0
   209
            L = *lum2++;
slouken@0
   210
            *row2++ = (rgb_2_pix[ L + cr_r ] |
slouken@0
   211
                       rgb_2_pix[ L + crb_g ] |
slouken@0
   212
                       rgb_2_pix[ L + cb_b ]);
slouken@0
   213
        }
slouken@0
   214
slouken@0
   215
        /*
slouken@0
   216
         * These values are at the start of the next line, (due
slouken@0
   217
         * to the ++'s above),but they need to be at the start
slouken@0
   218
         * of the line after that.
slouken@0
   219
         */
slouken@0
   220
        lum  += cols;
slouken@0
   221
        lum2 += cols;
slouken@0
   222
        row1 += mod;
slouken@0
   223
        row2 += mod;
slouken@0
   224
    }
slouken@0
   225
}
slouken@0
   226
slouken@0
   227
static void Color24DitherYV12Mod1X( int *colortab, Uint32 *rgb_2_pix,
slouken@0
   228
                                    unsigned char *lum, unsigned char *cr,
slouken@0
   229
                                    unsigned char *cb, unsigned char *out,
slouken@0
   230
                                    int rows, int cols, int mod )
slouken@0
   231
{
slouken@0
   232
    unsigned int value;
slouken@0
   233
    unsigned char* row1;
slouken@0
   234
    unsigned char* row2;
slouken@0
   235
    unsigned char* lum2;
slouken@0
   236
    int x, y;
slouken@0
   237
    int cr_r;
slouken@0
   238
    int crb_g;
slouken@0
   239
    int cb_b;
slouken@0
   240
    int cols_2 = cols / 2;
slouken@0
   241
slouken@0
   242
    row1 = out;
slouken@0
   243
    row2 = row1 + cols*3 + mod*3;
slouken@0
   244
    lum2 = lum + cols;
slouken@0
   245
slouken@0
   246
    mod += cols + mod;
slouken@0
   247
    mod *= 3;
slouken@0
   248
slouken@0
   249
    y = rows / 2;
slouken@0
   250
    while( y-- )
slouken@0
   251
    {
slouken@0
   252
        x = cols_2;
slouken@0
   253
        while( x-- )
slouken@0
   254
        {
slouken@0
   255
            register int L;
slouken@0
   256
slouken@0
   257
            cr_r   = 0*768+256 + colortab[ *cr + 0*256 ];
slouken@0
   258
            crb_g  = 1*768+256 + colortab[ *cr + 1*256 ]
slouken@0
   259
                               + colortab[ *cb + 2*256 ];
slouken@0
   260
            cb_b   = 2*768+256 + colortab[ *cb + 3*256 ];
slouken@0
   261
            ++cr; ++cb;
slouken@0
   262
slouken@0
   263
            L = *lum++;
slouken@0
   264
            value = (rgb_2_pix[ L + cr_r ] |
slouken@0
   265
                     rgb_2_pix[ L + crb_g ] |
slouken@0
   266
                     rgb_2_pix[ L + cb_b ]);
slouken@0
   267
            *row1++ = (value      ) & 0xFF;
slouken@0
   268
            *row1++ = (value >>  8) & 0xFF;
slouken@0
   269
            *row1++ = (value >> 16) & 0xFF;
slouken@0
   270
slouken@0
   271
            L = *lum++;
slouken@0
   272
            value = (rgb_2_pix[ L + cr_r ] |
slouken@0
   273
                     rgb_2_pix[ L + crb_g ] |
slouken@0
   274
                     rgb_2_pix[ L + cb_b ]);
slouken@0
   275
            *row1++ = (value      ) & 0xFF;
slouken@0
   276
            *row1++ = (value >>  8) & 0xFF;
slouken@0
   277
            *row1++ = (value >> 16) & 0xFF;
slouken@0
   278
slouken@0
   279
slouken@0
   280
            /* Now, do second row.  */
slouken@0
   281
slouken@0
   282
            L = *lum2++;
slouken@0
   283
            value = (rgb_2_pix[ L + cr_r ] |
slouken@0
   284
                     rgb_2_pix[ L + crb_g ] |
slouken@0
   285
                     rgb_2_pix[ L + cb_b ]);
slouken@0
   286
            *row2++ = (value      ) & 0xFF;
slouken@0
   287
            *row2++ = (value >>  8) & 0xFF;
slouken@0
   288
            *row2++ = (value >> 16) & 0xFF;
slouken@0
   289
slouken@0
   290
            L = *lum2++;
slouken@0
   291
            value = (rgb_2_pix[ L + cr_r ] |
slouken@0
   292
                     rgb_2_pix[ L + crb_g ] |
slouken@0
   293
                     rgb_2_pix[ L + cb_b ]);
slouken@0
   294
            *row2++ = (value      ) & 0xFF;
slouken@0
   295
            *row2++ = (value >>  8) & 0xFF;
slouken@0
   296
            *row2++ = (value >> 16) & 0xFF;
slouken@0
   297
        }
slouken@0
   298
slouken@0
   299
        /*
slouken@0
   300
         * These values are at the start of the next line, (due
slouken@0
   301
         * to the ++'s above),but they need to be at the start
slouken@0
   302
         * of the line after that.
slouken@0
   303
         */
slouken@0
   304
        lum  += cols;
slouken@0
   305
        lum2 += cols;
slouken@0
   306
        row1 += mod;
slouken@0
   307
        row2 += mod;
slouken@0
   308
    }
slouken@0
   309
}
slouken@0
   310
slouken@0
   311
static void Color32DitherYV12Mod1X( int *colortab, Uint32 *rgb_2_pix,
slouken@0
   312
                                    unsigned char *lum, unsigned char *cr,
slouken@0
   313
                                    unsigned char *cb, unsigned char *out,
slouken@0
   314
                                    int rows, int cols, int mod )
slouken@0
   315
{
slouken@0
   316
    unsigned int* row1;
slouken@0
   317
    unsigned int* row2;
slouken@0
   318
    unsigned char* lum2;
slouken@0
   319
    int x, y;
slouken@0
   320
    int cr_r;
slouken@0
   321
    int crb_g;
slouken@0
   322
    int cb_b;
slouken@0
   323
    int cols_2 = cols / 2;
slouken@0
   324
slouken@0
   325
    row1 = (unsigned int*) out;
slouken@0
   326
    row2 = row1 + cols + mod;
slouken@0
   327
    lum2 = lum + cols;
slouken@0
   328
slouken@0
   329
    mod += cols + mod;
slouken@0
   330
slouken@0
   331
    y = rows / 2;
slouken@0
   332
    while( y-- )
slouken@0
   333
    {
slouken@0
   334
        x = cols_2;
slouken@0
   335
        while( x-- )
slouken@0
   336
        {
slouken@0
   337
            register int L;
slouken@0
   338
slouken@0
   339
            cr_r   = 0*768+256 + colortab[ *cr + 0*256 ];
slouken@0
   340
            crb_g  = 1*768+256 + colortab[ *cr + 1*256 ]
slouken@0
   341
                               + colortab[ *cb + 2*256 ];
slouken@0
   342
            cb_b   = 2*768+256 + colortab[ *cb + 3*256 ];
slouken@0
   343
            ++cr; ++cb;
slouken@0
   344
slouken@0
   345
            L = *lum++;
slouken@0
   346
            *row1++ = (rgb_2_pix[ L + cr_r ] |
slouken@0
   347
                       rgb_2_pix[ L + crb_g ] |
slouken@0
   348
                       rgb_2_pix[ L + cb_b ]);
slouken@0
   349
slouken@0
   350
            L = *lum++;
slouken@0
   351
            *row1++ = (rgb_2_pix[ L + cr_r ] |
slouken@0
   352
                       rgb_2_pix[ L + crb_g ] |
slouken@0
   353
                       rgb_2_pix[ L + cb_b ]);
slouken@0
   354
slouken@0
   355
slouken@0
   356
            /* Now, do second row.  */
slouken@0
   357
slouken@0
   358
            L = *lum2++;
slouken@0
   359
            *row2++ = (rgb_2_pix[ L + cr_r ] |
slouken@0
   360
                       rgb_2_pix[ L + crb_g ] |
slouken@0
   361
                       rgb_2_pix[ L + cb_b ]);
slouken@0
   362
slouken@0
   363
            L = *lum2++;
slouken@0
   364
            *row2++ = (rgb_2_pix[ L + cr_r ] |
slouken@0
   365
                       rgb_2_pix[ L + crb_g ] |
slouken@0
   366
                       rgb_2_pix[ L + cb_b ]);
slouken@0
   367
        }
slouken@0
   368
slouken@0
   369
        /*
slouken@0
   370
         * These values are at the start of the next line, (due
slouken@0
   371
         * to the ++'s above),but they need to be at the start
slouken@0
   372
         * of the line after that.
slouken@0
   373
         */
slouken@0
   374
        lum  += cols;
slouken@0
   375
        lum2 += cols;
slouken@0
   376
        row1 += mod;
slouken@0
   377
        row2 += mod;
slouken@0
   378
    }
slouken@0
   379
}
slouken@0
   380
slouken@0
   381
/*
slouken@0
   382
 * In this function I make use of a nasty trick. The tables have the lower
slouken@0
   383
 * 16 bits replicated in the upper 16. This means I can write ints and get
slouken@0
   384
 * the horisontal doubling for free (almost).
slouken@0
   385
 */
slouken@0
   386
static void Color16DitherYV12Mod2X( int *colortab, Uint32 *rgb_2_pix,
slouken@0
   387
                                    unsigned char *lum, unsigned char *cr,
slouken@0
   388
                                    unsigned char *cb, unsigned char *out,
slouken@0
   389
                                    int rows, int cols, int mod )
slouken@0
   390
{
slouken@0
   391
    unsigned int* row1 = (unsigned int*) out;
slouken@0
   392
    const int next_row = cols+(mod/2);
slouken@0
   393
    unsigned int* row2 = row1 + 2*next_row;
slouken@0
   394
    unsigned char* lum2;
slouken@0
   395
    int x, y;
slouken@0
   396
    int cr_r;
slouken@0
   397
    int crb_g;
slouken@0
   398
    int cb_b;
slouken@0
   399
    int cols_2 = cols / 2;
slouken@0
   400
slouken@0
   401
    lum2 = lum + cols;
slouken@0
   402
slouken@0
   403
    mod = (next_row * 3) + (mod/2);
slouken@0
   404
slouken@0
   405
    y = rows / 2;
slouken@0
   406
    while( y-- )
slouken@0
   407
    {
slouken@0
   408
        x = cols_2;
slouken@0
   409
        while( x-- )
slouken@0
   410
        {
slouken@0
   411
            register int L;
slouken@0
   412
slouken@0
   413
            cr_r   = 0*768+256 + colortab[ *cr + 0*256 ];
slouken@0
   414
            crb_g  = 1*768+256 + colortab[ *cr + 1*256 ]
slouken@0
   415
                               + colortab[ *cb + 2*256 ];
slouken@0
   416
            cb_b   = 2*768+256 + colortab[ *cb + 3*256 ];
slouken@0
   417
            ++cr; ++cb;
slouken@0
   418
slouken@0
   419
            L = *lum++;
slouken@0
   420
            row1[0] = row1[next_row] = (rgb_2_pix[ L + cr_r ] |
slouken@0
   421
                                        rgb_2_pix[ L + crb_g ] |
slouken@0
   422
                                        rgb_2_pix[ L + cb_b ]);
slouken@0
   423
            row1++;
slouken@0
   424
slouken@0
   425
            L = *lum++;
slouken@0
   426
            row1[0] = row1[next_row] = (rgb_2_pix[ L + cr_r ] |
slouken@0
   427
                                        rgb_2_pix[ L + crb_g ] |
slouken@0
   428
                                        rgb_2_pix[ L + cb_b ]);
slouken@0
   429
            row1++;
slouken@0
   430
slouken@0
   431
slouken@0
   432
            /* Now, do second row. */
slouken@0
   433
slouken@0
   434
            L = *lum2++;
slouken@0
   435
            row2[0] = row2[next_row] = (rgb_2_pix[ L + cr_r ] |
slouken@0
   436
                                        rgb_2_pix[ L + crb_g ] |
slouken@0
   437
                                        rgb_2_pix[ L + cb_b ]);
slouken@0
   438
            row2++;
slouken@0
   439
slouken@0
   440
            L = *lum2++;
slouken@0
   441
            row2[0] = row2[next_row] = (rgb_2_pix[ L + cr_r ] |
slouken@0
   442
                                        rgb_2_pix[ L + crb_g ] |
slouken@0
   443
                                        rgb_2_pix[ L + cb_b ]);
slouken@0
   444
            row2++;
slouken@0
   445
        }
slouken@0
   446
slouken@0
   447
        /*
slouken@0
   448
         * These values are at the start of the next line, (due
slouken@0
   449
         * to the ++'s above),but they need to be at the start
slouken@0
   450
         * of the line after that.
slouken@0
   451
         */
slouken@0
   452
        lum  += cols;
slouken@0
   453
        lum2 += cols;
slouken@0
   454
        row1 += mod;
slouken@0
   455
        row2 += mod;
slouken@0
   456
    }
slouken@0
   457
}
slouken@0
   458
slouken@0
   459
static void Color24DitherYV12Mod2X( int *colortab, Uint32 *rgb_2_pix,
slouken@0
   460
                                    unsigned char *lum, unsigned char *cr,
slouken@0
   461
                                    unsigned char *cb, unsigned char *out,
slouken@0
   462
                                    int rows, int cols, int mod )
slouken@0
   463
{
slouken@0
   464
    unsigned int value;
slouken@0
   465
    unsigned char* row1 = out;
slouken@0
   466
    const int next_row = (cols*2 + mod) * 3;
slouken@0
   467
    unsigned char* row2 = row1 + 2*next_row;
slouken@0
   468
    unsigned char* lum2;
slouken@0
   469
    int x, y;
slouken@0
   470
    int cr_r;
slouken@0
   471
    int crb_g;
slouken@0
   472
    int cb_b;
slouken@0
   473
    int cols_2 = cols / 2;
slouken@0
   474
slouken@0
   475
    lum2 = lum + cols;
slouken@0
   476
slouken@0
   477
    mod = next_row*3 + mod*3;
slouken@0
   478
slouken@0
   479
    y = rows / 2;
slouken@0
   480
    while( y-- )
slouken@0
   481
    {
slouken@0
   482
        x = cols_2;
slouken@0
   483
        while( x-- )
slouken@0
   484
        {
slouken@0
   485
            register int L;
slouken@0
   486
slouken@0
   487
            cr_r   = 0*768+256 + colortab[ *cr + 0*256 ];
slouken@0
   488
            crb_g  = 1*768+256 + colortab[ *cr + 1*256 ]
slouken@0
   489
                               + colortab[ *cb + 2*256 ];
slouken@0
   490
            cb_b   = 2*768+256 + colortab[ *cb + 3*256 ];
slouken@0
   491
            ++cr; ++cb;
slouken@0
   492
slouken@0
   493
            L = *lum++;
slouken@0
   494
            value = (rgb_2_pix[ L + cr_r ] |
slouken@0
   495
                     rgb_2_pix[ L + crb_g ] |
slouken@0
   496
                     rgb_2_pix[ L + cb_b ]);
slouken@0
   497
            row1[0+0] = row1[3+0] = row1[next_row+0] = row1[next_row+3+0] =
slouken@0
   498
                     (value      ) & 0xFF;
slouken@0
   499
            row1[0+1] = row1[3+1] = row1[next_row+1] = row1[next_row+3+1] =
slouken@0
   500
                     (value >>  8) & 0xFF;
slouken@0
   501
            row1[0+2] = row1[3+2] = row1[next_row+2] = row1[next_row+3+2] =
slouken@0
   502
                     (value >> 16) & 0xFF;
slouken@0
   503
            row1 += 2*3;
slouken@0
   504
slouken@0
   505
            L = *lum++;
slouken@0
   506
            value = (rgb_2_pix[ L + cr_r ] |
slouken@0
   507
                     rgb_2_pix[ L + crb_g ] |
slouken@0
   508
                     rgb_2_pix[ L + cb_b ]);
slouken@0
   509
            row1[0+0] = row1[3+0] = row1[next_row+0] = row1[next_row+3+0] =
slouken@0
   510
                     (value      ) & 0xFF;
slouken@0
   511
            row1[0+1] = row1[3+1] = row1[next_row+1] = row1[next_row+3+1] =
slouken@0
   512
                     (value >>  8) & 0xFF;
slouken@0
   513
            row1[0+2] = row1[3+2] = row1[next_row+2] = row1[next_row+3+2] =
slouken@0
   514
                     (value >> 16) & 0xFF;
slouken@0
   515
            row1 += 2*3;
slouken@0
   516
slouken@0
   517
slouken@0
   518
            /* Now, do second row. */
slouken@0
   519
slouken@0
   520
            L = *lum2++;
slouken@0
   521
            value = (rgb_2_pix[ L + cr_r ] |
slouken@0
   522
                     rgb_2_pix[ L + crb_g ] |
slouken@0
   523
                     rgb_2_pix[ L + cb_b ]);
slouken@0
   524
            row2[0+0] = row2[3+0] = row2[next_row+0] = row2[next_row+3+0] =
slouken@0
   525
                     (value      ) & 0xFF;
slouken@0
   526
            row2[0+1] = row2[3+1] = row2[next_row+1] = row2[next_row+3+1] =
slouken@0
   527
                     (value >>  8) & 0xFF;
slouken@0
   528
            row2[0+2] = row2[3+2] = row2[next_row+2] = row2[next_row+3+2] =
slouken@0
   529
                     (value >> 16) & 0xFF;
slouken@0
   530
            row2 += 2*3;
slouken@0
   531
slouken@0
   532
            L = *lum2++;
slouken@0
   533
            value = (rgb_2_pix[ L + cr_r ] |
slouken@0
   534
                     rgb_2_pix[ L + crb_g ] |
slouken@0
   535
                     rgb_2_pix[ L + cb_b ]);
slouken@0
   536
            row2[0+0] = row2[3+0] = row2[next_row+0] = row2[next_row+3+0] =
slouken@0
   537
                     (value      ) & 0xFF;
slouken@0
   538
            row2[0+1] = row2[3+1] = row2[next_row+1] = row2[next_row+3+1] =
slouken@0
   539
                     (value >>  8) & 0xFF;
slouken@0
   540
            row2[0+2] = row2[3+2] = row2[next_row+2] = row2[next_row+3+2] =
slouken@0
   541
                     (value >> 16) & 0xFF;
slouken@0
   542
            row2 += 2*3;
slouken@0
   543
        }
slouken@0
   544
slouken@0
   545
        /*
slouken@0
   546
         * These values are at the start of the next line, (due
slouken@0
   547
         * to the ++'s above),but they need to be at the start
slouken@0
   548
         * of the line after that.
slouken@0
   549
         */
slouken@0
   550
        lum  += cols;
slouken@0
   551
        lum2 += cols;
slouken@0
   552
        row1 += mod;
slouken@0
   553
        row2 += mod;
slouken@0
   554
    }
slouken@0
   555
}
slouken@0
   556
slouken@0
   557
static void Color32DitherYV12Mod2X( int *colortab, Uint32 *rgb_2_pix,
slouken@0
   558
                                    unsigned char *lum, unsigned char *cr,
slouken@0
   559
                                    unsigned char *cb, unsigned char *out,
slouken@0
   560
                                    int rows, int cols, int mod )
slouken@0
   561
{
slouken@0
   562
    unsigned int* row1 = (unsigned int*) out;
slouken@0
   563
    const int next_row = cols*2+mod;
slouken@0
   564
    unsigned int* row2 = row1 + 2*next_row;
slouken@0
   565
    unsigned char* lum2;
slouken@0
   566
    int x, y;
slouken@0
   567
    int cr_r;
slouken@0
   568
    int crb_g;
slouken@0
   569
    int cb_b;
slouken@0
   570
    int cols_2 = cols / 2;
slouken@0
   571
slouken@0
   572
    lum2 = lum + cols;
slouken@0
   573
slouken@0
   574
    mod = (next_row * 3) + mod;
slouken@0
   575
slouken@0
   576
    y = rows / 2;
slouken@0
   577
    while( y-- )
slouken@0
   578
    {
slouken@0
   579
        x = cols_2;
slouken@0
   580
        while( x-- )
slouken@0
   581
        {
slouken@0
   582
            register int L;
slouken@0
   583
slouken@0
   584
            cr_r   = 0*768+256 + colortab[ *cr + 0*256 ];
slouken@0
   585
            crb_g  = 1*768+256 + colortab[ *cr + 1*256 ]
slouken@0
   586
                               + colortab[ *cb + 2*256 ];
slouken@0
   587
            cb_b   = 2*768+256 + colortab[ *cb + 3*256 ];
slouken@0
   588
            ++cr; ++cb;
slouken@0
   589
slouken@0
   590
            L = *lum++;
slouken@0
   591
            row1[0] = row1[1] = row1[next_row] = row1[next_row+1] =
slouken@0
   592
                                       (rgb_2_pix[ L + cr_r ] |
slouken@0
   593
                                        rgb_2_pix[ L + crb_g ] |
slouken@0
   594
                                        rgb_2_pix[ L + cb_b ]);
slouken@0
   595
            row1 += 2;
slouken@0
   596
slouken@0
   597
            L = *lum++;
slouken@0
   598
            row1[0] = row1[1] = row1[next_row] = row1[next_row+1] =
slouken@0
   599
                                       (rgb_2_pix[ L + cr_r ] |
slouken@0
   600
                                        rgb_2_pix[ L + crb_g ] |
slouken@0
   601
                                        rgb_2_pix[ L + cb_b ]);
slouken@0
   602
            row1 += 2;
slouken@0
   603
slouken@0
   604
slouken@0
   605
            /* Now, do second row. */
slouken@0
   606
slouken@0
   607
            L = *lum2++;
slouken@0
   608
            row2[0] = row2[1] = row2[next_row] = row2[next_row+1] =
slouken@0
   609
                                       (rgb_2_pix[ L + cr_r ] |
slouken@0
   610
                                        rgb_2_pix[ L + crb_g ] |
slouken@0
   611
                                        rgb_2_pix[ L + cb_b ]);
slouken@0
   612
            row2 += 2;
slouken@0
   613
slouken@0
   614
            L = *lum2++;
slouken@0
   615
            row2[0] = row2[1] = row2[next_row] = row2[next_row+1] =
slouken@0
   616
                                       (rgb_2_pix[ L + cr_r ] |
slouken@0
   617
                                        rgb_2_pix[ L + crb_g ] |
slouken@0
   618
                                        rgb_2_pix[ L + cb_b ]);
slouken@0
   619
            row2 += 2;
slouken@0
   620
        }
slouken@0
   621
slouken@0
   622
        /*
slouken@0
   623
         * These values are at the start of the next line, (due
slouken@0
   624
         * to the ++'s above),but they need to be at the start
slouken@0
   625
         * of the line after that.
slouken@0
   626
         */
slouken@0
   627
        lum  += cols;
slouken@0
   628
        lum2 += cols;
slouken@0
   629
        row1 += mod;
slouken@0
   630
        row2 += mod;
slouken@0
   631
    }
slouken@0
   632
}
slouken@0
   633
slouken@0
   634
static void Color16DitherYUY2Mod1X( int *colortab, Uint32 *rgb_2_pix,
slouken@0
   635
                                    unsigned char *lum, unsigned char *cr,
slouken@0
   636
                                    unsigned char *cb, unsigned char *out,
slouken@0
   637
                                    int rows, int cols, int mod )
slouken@0
   638
{
slouken@0
   639
    unsigned short* row;
slouken@0
   640
    int x, y;
slouken@0
   641
    int cr_r;
slouken@0
   642
    int crb_g;
slouken@0
   643
    int cb_b;
slouken@0
   644
    int cols_2 = cols / 2;
slouken@0
   645
slouken@0
   646
    row = (unsigned short*) out;
slouken@0
   647
slouken@0
   648
    y = rows;
slouken@0
   649
    while( y-- )
slouken@0
   650
    {
slouken@0
   651
        x = cols_2;
slouken@0
   652
        while( x-- )
slouken@0
   653
        {
slouken@0
   654
            register int L;
slouken@0
   655
slouken@0
   656
            cr_r   = 0*768+256 + colortab[ *cr + 0*256 ];
slouken@0
   657
            crb_g  = 1*768+256 + colortab[ *cr + 1*256 ]
slouken@0
   658
                               + colortab[ *cb + 2*256 ];
slouken@0
   659
            cb_b   = 2*768+256 + colortab[ *cb + 3*256 ];
slouken@0
   660
            cr += 4; cb += 4;
slouken@0
   661
slouken@0
   662
            L = *lum; lum += 2;
slouken@0
   663
            *row++ = (rgb_2_pix[ L + cr_r ] |
slouken@0
   664
                       rgb_2_pix[ L + crb_g ] |
slouken@0
   665
                       rgb_2_pix[ L + cb_b ]);
slouken@0
   666
slouken@0
   667
            L = *lum; lum += 2;
slouken@0
   668
            *row++ = (rgb_2_pix[ L + cr_r ] |
slouken@0
   669
                       rgb_2_pix[ L + crb_g ] |
slouken@0
   670
                       rgb_2_pix[ L + cb_b ]);
slouken@0
   671
slouken@0
   672
        }
slouken@0
   673
slouken@0
   674
        row += mod;
slouken@0
   675
    }
slouken@0
   676
}
slouken@0
   677
slouken@0
   678
static void Color24DitherYUY2Mod1X( int *colortab, Uint32 *rgb_2_pix,
slouken@0
   679
                                    unsigned char *lum, unsigned char *cr,
slouken@0
   680
                                    unsigned char *cb, unsigned char *out,
slouken@0
   681
                                    int rows, int cols, int mod )
slouken@0
   682
{
slouken@0
   683
    unsigned int value;
slouken@0
   684
    unsigned char* row;
slouken@0
   685
    int x, y;
slouken@0
   686
    int cr_r;
slouken@0
   687
    int crb_g;
slouken@0
   688
    int cb_b;
slouken@0
   689
    int cols_2 = cols / 2;
slouken@0
   690
slouken@0
   691
    row = (unsigned char*) out;
slouken@0
   692
    mod *= 3;
slouken@0
   693
    y = rows;
slouken@0
   694
    while( y-- )
slouken@0
   695
    {
slouken@0
   696
        x = cols_2;
slouken@0
   697
        while( x-- )
slouken@0
   698
        {
slouken@0
   699
            register int L;
slouken@0
   700
slouken@0
   701
            cr_r   = 0*768+256 + colortab[ *cr + 0*256 ];
slouken@0
   702
            crb_g  = 1*768+256 + colortab[ *cr + 1*256 ]
slouken@0
   703
                               + colortab[ *cb + 2*256 ];
slouken@0
   704
            cb_b   = 2*768+256 + colortab[ *cb + 3*256 ];
slouken@0
   705
            cr += 4; cb += 4;
slouken@0
   706
slouken@0
   707
            L = *lum; lum += 2;
slouken@0
   708
            value = (rgb_2_pix[ L + cr_r ] |
slouken@0
   709
                     rgb_2_pix[ L + crb_g ] |
slouken@0
   710
                     rgb_2_pix[ L + cb_b ]);
slouken@0
   711
            *row++ = (value      ) & 0xFF;
slouken@0
   712
            *row++ = (value >>  8) & 0xFF;
slouken@0
   713
            *row++ = (value >> 16) & 0xFF;
slouken@0
   714
slouken@0
   715
            L = *lum; lum += 2;
slouken@0
   716
            value = (rgb_2_pix[ L + cr_r ] |
slouken@0
   717
                     rgb_2_pix[ L + crb_g ] |
slouken@0
   718
                     rgb_2_pix[ L + cb_b ]);
slouken@0
   719
            *row++ = (value      ) & 0xFF;
slouken@0
   720
            *row++ = (value >>  8) & 0xFF;
slouken@0
   721
            *row++ = (value >> 16) & 0xFF;
slouken@0
   722
slouken@0
   723
        }
slouken@0
   724
        row += mod;
slouken@0
   725
    }
slouken@0
   726
}
slouken@0
   727
slouken@0
   728
static void Color32DitherYUY2Mod1X( int *colortab, Uint32 *rgb_2_pix,
slouken@0
   729
                                    unsigned char *lum, unsigned char *cr,
slouken@0
   730
                                    unsigned char *cb, unsigned char *out,
slouken@0
   731
                                    int rows, int cols, int mod )
slouken@0
   732
{
slouken@0
   733
    unsigned int* row;
slouken@0
   734
    int x, y;
slouken@0
   735
    int cr_r;
slouken@0
   736
    int crb_g;
slouken@0
   737
    int cb_b;
slouken@0
   738
    int cols_2 = cols / 2;
slouken@0
   739
slouken@0
   740
    row = (unsigned int*) out;
slouken@0
   741
    y = rows;
slouken@0
   742
    while( y-- )
slouken@0
   743
    {
slouken@0
   744
        x = cols_2;
slouken@0
   745
        while( x-- )
slouken@0
   746
        {
slouken@0
   747
            register int L;
slouken@0
   748
slouken@0
   749
            cr_r   = 0*768+256 + colortab[ *cr + 0*256 ];
slouken@0
   750
            crb_g  = 1*768+256 + colortab[ *cr + 1*256 ]
slouken@0
   751
                               + colortab[ *cb + 2*256 ];
slouken@0
   752
            cb_b   = 2*768+256 + colortab[ *cb + 3*256 ];
slouken@0
   753
            cr += 4; cb += 4;
slouken@0
   754
slouken@0
   755
            L = *lum; lum += 2;
slouken@0
   756
            *row++ = (rgb_2_pix[ L + cr_r ] |
slouken@0
   757
                       rgb_2_pix[ L + crb_g ] |
slouken@0
   758
                       rgb_2_pix[ L + cb_b ]);
slouken@0
   759
slouken@0
   760
            L = *lum; lum += 2;
slouken@0
   761
            *row++ = (rgb_2_pix[ L + cr_r ] |
slouken@0
   762
                       rgb_2_pix[ L + crb_g ] |
slouken@0
   763
                       rgb_2_pix[ L + cb_b ]);
slouken@0
   764
slouken@0
   765
slouken@0
   766
        }
slouken@0
   767
        row += mod;
slouken@0
   768
    }
slouken@0
   769
}
slouken@0
   770
slouken@0
   771
/*
slouken@0
   772
 * In this function I make use of a nasty trick. The tables have the lower
slouken@0
   773
 * 16 bits replicated in the upper 16. This means I can write ints and get
slouken@0
   774
 * the horisontal doubling for free (almost).
slouken@0
   775
 */
slouken@0
   776
static void Color16DitherYUY2Mod2X( int *colortab, Uint32 *rgb_2_pix,
slouken@0
   777
                                    unsigned char *lum, unsigned char *cr,
slouken@0
   778
                                    unsigned char *cb, unsigned char *out,
slouken@0
   779
                                    int rows, int cols, int mod )
slouken@0
   780
{
slouken@0
   781
    unsigned int* row = (unsigned int*) out;
slouken@0
   782
    const int next_row = cols+(mod/2);
slouken@0
   783
    int x, y;
slouken@0
   784
    int cr_r;
slouken@0
   785
    int crb_g;
slouken@0
   786
    int cb_b;
slouken@0
   787
    int cols_2 = cols / 2;
slouken@0
   788
slouken@0
   789
    y = rows;
slouken@0
   790
    while( y-- )
slouken@0
   791
    {
slouken@0
   792
        x = cols_2;
slouken@0
   793
        while( x-- )
slouken@0
   794
        {
slouken@0
   795
            register int L;
slouken@0
   796
slouken@0
   797
            cr_r   = 0*768+256 + colortab[ *cr + 0*256 ];
slouken@0
   798
            crb_g  = 1*768+256 + colortab[ *cr + 1*256 ]
slouken@0
   799
                               + colortab[ *cb + 2*256 ];
slouken@0
   800
            cb_b   = 2*768+256 + colortab[ *cb + 3*256 ];
slouken@0
   801
            cr += 4; cb += 4;
slouken@0
   802
slouken@0
   803
            L = *lum; lum += 2;
slouken@0
   804
            row[0] = row[next_row] = (rgb_2_pix[ L + cr_r ] |
slouken@0
   805
                                        rgb_2_pix[ L + crb_g ] |
slouken@0
   806
                                        rgb_2_pix[ L + cb_b ]);
slouken@0
   807
            row++;
slouken@0
   808
slouken@0
   809
            L = *lum; lum += 2;
slouken@0
   810
            row[0] = row[next_row] = (rgb_2_pix[ L + cr_r ] |
slouken@0
   811
                                        rgb_2_pix[ L + crb_g ] |
slouken@0
   812
                                        rgb_2_pix[ L + cb_b ]);
slouken@0
   813
            row++;
slouken@0
   814
slouken@0
   815
        }
slouken@0
   816
        row += next_row;
slouken@0
   817
    }
slouken@0
   818
}
slouken@0
   819
slouken@0
   820
static void Color24DitherYUY2Mod2X( int *colortab, Uint32 *rgb_2_pix,
slouken@0
   821
                                    unsigned char *lum, unsigned char *cr,
slouken@0
   822
                                    unsigned char *cb, unsigned char *out,
slouken@0
   823
                                    int rows, int cols, int mod )
slouken@0
   824
{
slouken@0
   825
    unsigned int value;
slouken@0
   826
    unsigned char* row = out;
slouken@0
   827
    const int next_row = (cols*2 + mod) * 3;
slouken@0
   828
    int x, y;
slouken@0
   829
    int cr_r;
slouken@0
   830
    int crb_g;
slouken@0
   831
    int cb_b;
slouken@0
   832
    int cols_2 = cols / 2;
slouken@0
   833
    y = rows;
slouken@0
   834
    while( y-- )
slouken@0
   835
    {
slouken@0
   836
        x = cols_2;
slouken@0
   837
        while( x-- )
slouken@0
   838
        {
slouken@0
   839
            register int L;
slouken@0
   840
slouken@0
   841
            cr_r   = 0*768+256 + colortab[ *cr + 0*256 ];
slouken@0
   842
            crb_g  = 1*768+256 + colortab[ *cr + 1*256 ]
slouken@0
   843
                               + colortab[ *cb + 2*256 ];
slouken@0
   844
            cb_b   = 2*768+256 + colortab[ *cb + 3*256 ];
slouken@0
   845
            cr += 4; cb += 4;
slouken@0
   846
slouken@0
   847
            L = *lum; lum += 2;
slouken@0
   848
            value = (rgb_2_pix[ L + cr_r ] |
slouken@0
   849
                     rgb_2_pix[ L + crb_g ] |
slouken@0
   850
                     rgb_2_pix[ L + cb_b ]);
slouken@0
   851
            row[0+0] = row[3+0] = row[next_row+0] = row[next_row+3+0] =
slouken@0
   852
                     (value      ) & 0xFF;
slouken@0
   853
            row[0+1] = row[3+1] = row[next_row+1] = row[next_row+3+1] =
slouken@0
   854
                     (value >>  8) & 0xFF;
slouken@0
   855
            row[0+2] = row[3+2] = row[next_row+2] = row[next_row+3+2] =
slouken@0
   856
                     (value >> 16) & 0xFF;
slouken@0
   857
            row += 2*3;
slouken@0
   858
slouken@0
   859
            L = *lum; lum += 2;
slouken@0
   860
            value = (rgb_2_pix[ L + cr_r ] |
slouken@0
   861
                     rgb_2_pix[ L + crb_g ] |
slouken@0
   862
                     rgb_2_pix[ L + cb_b ]);
slouken@0
   863
            row[0+0] = row[3+0] = row[next_row+0] = row[next_row+3+0] =
slouken@0
   864
                     (value      ) & 0xFF;
slouken@0
   865
            row[0+1] = row[3+1] = row[next_row+1] = row[next_row+3+1] =
slouken@0
   866
                     (value >>  8) & 0xFF;
slouken@0
   867
            row[0+2] = row[3+2] = row[next_row+2] = row[next_row+3+2] =
slouken@0
   868
                     (value >> 16) & 0xFF;
slouken@0
   869
            row += 2*3;
slouken@0
   870
slouken@0
   871
        }
slouken@0
   872
        row += next_row;
slouken@0
   873
    }
slouken@0
   874
}
slouken@0
   875
slouken@0
   876
static void Color32DitherYUY2Mod2X( int *colortab, Uint32 *rgb_2_pix,
slouken@0
   877
                                    unsigned char *lum, unsigned char *cr,
slouken@0
   878
                                    unsigned char *cb, unsigned char *out,
slouken@0
   879
                                    int rows, int cols, int mod )
slouken@0
   880
{
slouken@0
   881
    unsigned int* row = (unsigned int*) out;
slouken@0
   882
    const int next_row = cols*2+mod;
slouken@0
   883
    int x, y;
slouken@0
   884
    int cr_r;
slouken@0
   885
    int crb_g;
slouken@0
   886
    int cb_b;
slouken@0
   887
    int cols_2 = cols / 2;
slouken@0
   888
    mod+=mod;
slouken@0
   889
    y = rows;
slouken@0
   890
    while( y-- )
slouken@0
   891
    {
slouken@0
   892
        x = cols_2;
slouken@0
   893
        while( x-- )
slouken@0
   894
        {
slouken@0
   895
            register int L;
slouken@0
   896
slouken@0
   897
            cr_r   = 0*768+256 + colortab[ *cr + 0*256 ];
slouken@0
   898
            crb_g  = 1*768+256 + colortab[ *cr + 1*256 ]
slouken@0
   899
                               + colortab[ *cb + 2*256 ];
slouken@0
   900
            cb_b   = 2*768+256 + colortab[ *cb + 3*256 ];
slouken@0
   901
            cr += 4; cb += 4;
slouken@0
   902
slouken@0
   903
            L = *lum; lum += 2;
slouken@0
   904
            row[0] = row[1] = row[next_row] = row[next_row+1] =
slouken@0
   905
                                       (rgb_2_pix[ L + cr_r ] |
slouken@0
   906
                                        rgb_2_pix[ L + crb_g ] |
slouken@0
   907
                                        rgb_2_pix[ L + cb_b ]);
slouken@0
   908
            row += 2;
slouken@0
   909
slouken@0
   910
            L = *lum; lum += 2;
slouken@0
   911
            row[0] = row[1] = row[next_row] = row[next_row+1] =
slouken@0
   912
                                       (rgb_2_pix[ L + cr_r ] |
slouken@0
   913
                                        rgb_2_pix[ L + crb_g ] |
slouken@0
   914
                                        rgb_2_pix[ L + cb_b ]);
slouken@0
   915
            row += 2;
slouken@0
   916
slouken@0
   917
slouken@0
   918
        }
slouken@0
   919
slouken@0
   920
        row += next_row;
slouken@0
   921
    }
slouken@0
   922
}
slouken@0
   923
slouken@0
   924
/*
slouken@0
   925
 * How many 1 bits are there in the Uint32.
slouken@0
   926
 * Low performance, do not call often.
slouken@0
   927
 */
slouken@0
   928
static int number_of_bits_set( Uint32 a )
slouken@0
   929
{
slouken@0
   930
    if(!a) return 0;
slouken@0
   931
    if(a & 1) return 1 + number_of_bits_set(a >> 1);
slouken@0
   932
    return(number_of_bits_set(a >> 1));
slouken@0
   933
}
slouken@0
   934
slouken@0
   935
/*
slouken@0
   936
 * How many 0 bits are there at least significant end of Uint32.
slouken@0
   937
 * Low performance, do not call often.
slouken@0
   938
 */
slouken@0
   939
static int free_bits_at_bottom( Uint32 a )
slouken@0
   940
{
slouken@0
   941
      /* assume char is 8 bits */
slouken@0
   942
    if(!a) return sizeof(Uint32) * 8;
slouken@0
   943
    if(((Sint32)a) & 1l) return 0;
slouken@0
   944
    return 1 + free_bits_at_bottom ( a >> 1);
slouken@0
   945
}
slouken@0
   946
slouken@0
   947
slouken@0
   948
SDL_Overlay *SDL_CreateYUV_SW(_THIS, int width, int height, Uint32 format, SDL_Surface *display)
slouken@0
   949
{
slouken@0
   950
	SDL_Overlay *overlay;
slouken@0
   951
	struct private_yuvhwdata *swdata;
slouken@0
   952
	int *Cr_r_tab;
slouken@0
   953
	int *Cr_g_tab;
slouken@0
   954
	int *Cb_g_tab;
slouken@0
   955
	int *Cb_b_tab;
slouken@0
   956
	Uint32 *r_2_pix_alloc;
slouken@0
   957
	Uint32 *g_2_pix_alloc;
slouken@0
   958
	Uint32 *b_2_pix_alloc;
slouken@0
   959
	int i, cpu_mmx;
slouken@0
   960
	int CR, CB;
slouken@0
   961
	Uint32 Rmask, Gmask, Bmask;
slouken@0
   962
slouken@0
   963
	/* Only RGB packed pixel conversion supported */
slouken@0
   964
	if ( (display->format->BytesPerPixel != 2) &&
slouken@0
   965
	     (display->format->BytesPerPixel != 3) &&
slouken@0
   966
	     (display->format->BytesPerPixel != 4) ) {
slouken@0
   967
		SDL_SetError("Can't use YUV data on non 16/24/32 bit surfaces");
slouken@0
   968
		return(NULL);
slouken@0
   969
	}
slouken@0
   970
slouken@0
   971
	/* Verify that we support the format */
slouken@0
   972
	switch (format) {
slouken@0
   973
	    case SDL_YV12_OVERLAY:
slouken@0
   974
	    case SDL_IYUV_OVERLAY:
slouken@0
   975
	    case SDL_YUY2_OVERLAY:
slouken@0
   976
	    case SDL_UYVY_OVERLAY:
slouken@0
   977
	    case SDL_YVYU_OVERLAY:
slouken@0
   978
		break;
slouken@0
   979
	    default:
slouken@0
   980
		SDL_SetError("Unsupported YUV format");
slouken@0
   981
		return(NULL);
slouken@0
   982
	}
slouken@0
   983
slouken@0
   984
	/* Create the overlay structure */
slouken@0
   985
	overlay = (SDL_Overlay *)malloc(sizeof *overlay);
slouken@0
   986
	if ( overlay == NULL ) {
slouken@0
   987
		SDL_OutOfMemory();
slouken@0
   988
		return(NULL);
slouken@0
   989
	}
slouken@0
   990
	memset(overlay, 0, (sizeof *overlay));
slouken@0
   991
slouken@0
   992
	/* Fill in the basic members */
slouken@0
   993
	overlay->format = format;
slouken@0
   994
	overlay->w = width;
slouken@0
   995
	overlay->h = height;
slouken@0
   996
slouken@0
   997
	/* Set up the YUV surface function structure */
slouken@0
   998
	overlay->hwfuncs = &sw_yuvfuncs;
slouken@0
   999
slouken@0
  1000
	/* Create the pixel data and lookup tables */
slouken@0
  1001
	swdata = (struct private_yuvhwdata *)malloc(sizeof *swdata);
slouken@0
  1002
	overlay->hwdata = swdata;
slouken@0
  1003
	if ( swdata == NULL ) {
slouken@0
  1004
		SDL_OutOfMemory();
slouken@0
  1005
		SDL_FreeYUVOverlay(overlay);
slouken@0
  1006
		return(NULL);
slouken@0
  1007
	}
slouken@0
  1008
	swdata->stretch = NULL;
slouken@0
  1009
	swdata->display = display;
slouken@0
  1010
	swdata->pixels = (Uint8 *) malloc(width*height*2);
slouken@0
  1011
	swdata->colortab = (int *)malloc(4*256*sizeof(int));
slouken@0
  1012
	Cr_r_tab = &swdata->colortab[0*256];
slouken@0
  1013
	Cr_g_tab = &swdata->colortab[1*256];
slouken@0
  1014
	Cb_g_tab = &swdata->colortab[2*256];
slouken@0
  1015
	Cb_b_tab = &swdata->colortab[3*256];
slouken@0
  1016
	swdata->rgb_2_pix = (Uint32 *)malloc(3*768*sizeof(Uint32));
slouken@0
  1017
	r_2_pix_alloc = &swdata->rgb_2_pix[0*768];
slouken@0
  1018
	g_2_pix_alloc = &swdata->rgb_2_pix[1*768];
slouken@0
  1019
	b_2_pix_alloc = &swdata->rgb_2_pix[2*768];
slouken@0
  1020
	if ( ! swdata->pixels || ! swdata->colortab || ! swdata->rgb_2_pix ) {
slouken@0
  1021
		SDL_OutOfMemory();
slouken@0
  1022
		SDL_FreeYUVOverlay(overlay);
slouken@0
  1023
		return(NULL);
slouken@0
  1024
	}
slouken@0
  1025
slouken@0
  1026
	/* Generate the tables for the display surface */
slouken@0
  1027
	for (i=0; i<256; i++) {
slouken@0
  1028
		/* Gamma correction (luminescence table) and chroma correction
slouken@0
  1029
		   would be done here.  See the Berkeley mpeg_play sources.
slouken@0
  1030
		*/
slouken@0
  1031
		CB = CR = (i-128);
slouken@0
  1032
		Cr_r_tab[i] = (int) ( (0.419/0.299) * CR);
slouken@0
  1033
		Cr_g_tab[i] = (int) (-(0.299/0.419) * CR);
slouken@0
  1034
		Cb_g_tab[i] = (int) (-(0.114/0.331) * CB); 
slouken@0
  1035
		Cb_b_tab[i] = (int) ( (0.587/0.331) * CB);
slouken@0
  1036
	}
slouken@0
  1037
slouken@0
  1038
	/* 
slouken@0
  1039
	 * Set up entries 0-255 in rgb-to-pixel value tables.
slouken@0
  1040
	 */
slouken@0
  1041
	Rmask = display->format->Rmask;
slouken@0
  1042
	Gmask = display->format->Gmask;
slouken@0
  1043
	Bmask = display->format->Bmask;
slouken@0
  1044
	for ( i=0; i<256; ++i ) {
slouken@0
  1045
		r_2_pix_alloc[i+256] = i >> (8 - number_of_bits_set(Rmask));
slouken@0
  1046
		r_2_pix_alloc[i+256] <<= free_bits_at_bottom(Rmask);
slouken@0
  1047
		g_2_pix_alloc[i+256] = i >> (8 - number_of_bits_set(Gmask));
slouken@0
  1048
		g_2_pix_alloc[i+256] <<= free_bits_at_bottom(Gmask);
slouken@0
  1049
		b_2_pix_alloc[i+256] = i >> (8 - number_of_bits_set(Bmask));
slouken@0
  1050
		b_2_pix_alloc[i+256] <<= free_bits_at_bottom(Bmask);
slouken@0
  1051
	}
slouken@0
  1052
slouken@0
  1053
	/*
slouken@0
  1054
	 * If we have 16-bit output depth, then we double the value
slouken@0
  1055
	 * in the top word. This means that we can write out both
slouken@0
  1056
	 * pixels in the pixel doubling mode with one op. It is 
slouken@0
  1057
	 * harmless in the normal case as storing a 32-bit value
slouken@0
  1058
	 * through a short pointer will lose the top bits anyway.
slouken@0
  1059
	 */
slouken@0
  1060
	if( display->format->BytesPerPixel == 2 ) {
slouken@0
  1061
		for ( i=0; i<256; ++i ) {
slouken@0
  1062
			r_2_pix_alloc[i+256] |= (r_2_pix_alloc[i+256]) << 16;
slouken@0
  1063
			g_2_pix_alloc[i+256] |= (g_2_pix_alloc[i+256]) << 16;
slouken@0
  1064
			b_2_pix_alloc[i+256] |= (b_2_pix_alloc[i+256]) << 16;
slouken@0
  1065
		}
slouken@0
  1066
	}
slouken@0
  1067
slouken@0
  1068
	/*
slouken@0
  1069
	 * Spread out the values we have to the rest of the array so that
slouken@0
  1070
	 * we do not need to check for overflow.
slouken@0
  1071
	 */
slouken@0
  1072
	for ( i=0; i<256; ++i ) {
slouken@0
  1073
		r_2_pix_alloc[i] = r_2_pix_alloc[256];
slouken@0
  1074
		r_2_pix_alloc[i+512] = r_2_pix_alloc[511];
slouken@0
  1075
		g_2_pix_alloc[i] = g_2_pix_alloc[256];
slouken@0
  1076
		g_2_pix_alloc[i+512] = g_2_pix_alloc[511];
slouken@0
  1077
		b_2_pix_alloc[i] = b_2_pix_alloc[256];
slouken@0
  1078
		b_2_pix_alloc[i+512] = b_2_pix_alloc[511];
slouken@0
  1079
	}
slouken@0
  1080
slouken@0
  1081
	/* You have chosen wisely... */
slouken@0
  1082
	switch (format) {
slouken@0
  1083
	    case SDL_YV12_OVERLAY:
slouken@0
  1084
	    case SDL_IYUV_OVERLAY:
slouken@0
  1085
		cpu_mmx = CPU_Flags() & MMX_CPU;
slouken@0
  1086
		if ( display->format->BytesPerPixel == 2 ) {
slouken@0
  1087
#if defined(i386) && defined(__GNUC__) && defined(USE_ASMBLIT)
slouken@0
  1088
			/* inline assembly functions */
slouken@0
  1089
			if ( cpu_mmx && (Rmask == 0xF800) &&
slouken@0
  1090
			                (Gmask == 0x07E0) &&
slouken@0
  1091
				        (Bmask == 0x001F) &&
slouken@0
  1092
			                (width & 15) == 0) {
slouken@0
  1093
/*printf("Using MMX 16-bit 565 dither\n");*/
slouken@0
  1094
				swdata->Display1X = Color565DitherYV12MMX1X;
slouken@0
  1095
			} else {
slouken@0
  1096
/*printf("Using C 16-bit dither\n");*/
slouken@0
  1097
				swdata->Display1X = Color16DitherYV12Mod1X;
slouken@0
  1098
			}
slouken@0
  1099
#else
slouken@0
  1100
			swdata->Display1X = Color16DitherYV12Mod1X;
slouken@0
  1101
#endif
slouken@0
  1102
			swdata->Display2X = Color16DitherYV12Mod2X;
slouken@0
  1103
		}
slouken@0
  1104
		if ( display->format->BytesPerPixel == 3 ) {
slouken@0
  1105
			swdata->Display1X = Color24DitherYV12Mod1X;
slouken@0
  1106
			swdata->Display2X = Color24DitherYV12Mod2X;
slouken@0
  1107
		}
slouken@0
  1108
		if ( display->format->BytesPerPixel == 4 ) {
slouken@0
  1109
#if defined(i386) && defined(__GNUC__) && defined(USE_ASMBLIT)
slouken@0
  1110
			/* inline assembly functions */
slouken@0
  1111
			if ( cpu_mmx && (Rmask == 0x00FF0000) &&
slouken@0
  1112
			                (Gmask == 0x0000FF00) &&
slouken@0
  1113
				        (Bmask == 0x000000FF) && 
slouken@0
  1114
			                (width & 15) == 0) {
slouken@0
  1115
/*printf("Using MMX 32-bit dither\n");*/
slouken@0
  1116
				swdata->Display1X = ColorRGBDitherYV12MMX1X;
slouken@0
  1117
			} else {
slouken@0
  1118
/*printf("Using C 32-bit dither\n");*/
slouken@0
  1119
				swdata->Display1X = Color32DitherYV12Mod1X;
slouken@0
  1120
			}
slouken@0
  1121
#else
slouken@0
  1122
			swdata->Display1X = Color32DitherYV12Mod1X;
slouken@0
  1123
#endif
slouken@0
  1124
			swdata->Display2X = Color32DitherYV12Mod2X;
slouken@0
  1125
		}
slouken@0
  1126
		break;
slouken@0
  1127
	    case SDL_YUY2_OVERLAY:
slouken@0
  1128
	    case SDL_UYVY_OVERLAY:
slouken@0
  1129
	    case SDL_YVYU_OVERLAY:
slouken@0
  1130
		if ( display->format->BytesPerPixel == 2 ) {
slouken@0
  1131
			swdata->Display1X = Color16DitherYUY2Mod1X;
slouken@0
  1132
			swdata->Display2X = Color16DitherYUY2Mod2X;
slouken@0
  1133
		}
slouken@0
  1134
		if ( display->format->BytesPerPixel == 3 ) {
slouken@0
  1135
			swdata->Display1X = Color24DitherYUY2Mod1X;
slouken@0
  1136
			swdata->Display2X = Color24DitherYUY2Mod2X;
slouken@0
  1137
		}
slouken@0
  1138
		if ( display->format->BytesPerPixel == 4 ) {
slouken@0
  1139
			swdata->Display1X = Color32DitherYUY2Mod1X;
slouken@0
  1140
			swdata->Display2X = Color32DitherYUY2Mod2X;
slouken@0
  1141
		}
slouken@0
  1142
		break;
slouken@0
  1143
	    default:
slouken@0
  1144
		/* We should never get here (caught above) */
slouken@0
  1145
		break;
slouken@0
  1146
	}
slouken@0
  1147
slouken@0
  1148
	/* Find the pitch and offset values for the overlay */
slouken@0
  1149
	overlay->pitches = swdata->pitches;
slouken@0
  1150
	overlay->pixels = swdata->planes;
slouken@0
  1151
	switch (format) {
slouken@0
  1152
	    case SDL_YV12_OVERLAY:
slouken@0
  1153
	    case SDL_IYUV_OVERLAY:
slouken@0
  1154
		overlay->pitches[0] = overlay->w;
slouken@0
  1155
		overlay->pitches[1] = overlay->pitches[0] / 2;
slouken@0
  1156
		overlay->pitches[2] = overlay->pitches[0] / 2;
slouken@0
  1157
	        overlay->pixels[0] = swdata->pixels;
slouken@0
  1158
	        overlay->pixels[1] = overlay->pixels[0] +
slouken@0
  1159
		                     overlay->pitches[0] * overlay->h;
slouken@0
  1160
	        overlay->pixels[2] = overlay->pixels[1] +
slouken@0
  1161
		                     overlay->pitches[1] * overlay->h / 2;
slouken@0
  1162
		overlay->planes = 3;
slouken@0
  1163
		break;
slouken@0
  1164
	    case SDL_YUY2_OVERLAY:
slouken@0
  1165
	    case SDL_UYVY_OVERLAY:
slouken@0
  1166
	    case SDL_YVYU_OVERLAY:
slouken@0
  1167
		overlay->pitches[0] = overlay->w*2;
slouken@0
  1168
	        overlay->pixels[0] = swdata->pixels;
slouken@0
  1169
		overlay->planes = 1;
slouken@0
  1170
		break;
slouken@0
  1171
	    default:
slouken@0
  1172
		/* We should never get here (caught above) */
slouken@0
  1173
		break;
slouken@0
  1174
	}
slouken@0
  1175
slouken@0
  1176
	/* We're all done.. */
slouken@0
  1177
	return(overlay);
slouken@0
  1178
}
slouken@0
  1179
slouken@0
  1180
int SDL_LockYUV_SW(_THIS, SDL_Overlay *overlay)
slouken@0
  1181
{
slouken@0
  1182
	return(0);
slouken@0
  1183
}
slouken@0
  1184
slouken@0
  1185
void SDL_UnlockYUV_SW(_THIS, SDL_Overlay *overlay)
slouken@0
  1186
{
slouken@0
  1187
	return;
slouken@0
  1188
}
slouken@0
  1189
slouken@0
  1190
int SDL_DisplayYUV_SW(_THIS, SDL_Overlay *overlay, SDL_Rect *dstrect)
slouken@0
  1191
{
slouken@0
  1192
	struct private_yuvhwdata *swdata;
slouken@0
  1193
	SDL_Surface *stretch;
slouken@0
  1194
	SDL_Surface *display;
slouken@0
  1195
	int scale_2x;
slouken@0
  1196
	Uint8 *lum, *Cr, *Cb;
slouken@0
  1197
	Uint8 *dst;
slouken@0
  1198
	int mod;
slouken@0
  1199
slouken@0
  1200
	swdata = overlay->hwdata;
slouken@0
  1201
	scale_2x = 0;
slouken@0
  1202
	stretch = 0;
slouken@0
  1203
	if ( (overlay->w != dstrect->w) || (overlay->h != dstrect->h) ) {
slouken@0
  1204
		if ( (dstrect->w == 2*overlay->w) &&
slouken@0
  1205
		     (dstrect->h == 2*overlay->h) ) {
slouken@0
  1206
			scale_2x = 1;
slouken@0
  1207
		} else {
slouken@0
  1208
			if ( ! swdata->stretch ) {
slouken@0
  1209
				display = swdata->display;
slouken@0
  1210
				swdata->stretch = SDL_CreateRGBSurface(
slouken@0
  1211
					SDL_SWSURFACE,
slouken@0
  1212
					overlay->w, overlay->h,
slouken@0
  1213
					display->format->BitsPerPixel,
slouken@0
  1214
					display->format->Rmask,
slouken@0
  1215
					display->format->Gmask,
slouken@0
  1216
					display->format->Bmask, 0);
slouken@0
  1217
				if ( ! swdata->stretch ) {
slouken@0
  1218
					return(-1);
slouken@0
  1219
				}
slouken@0
  1220
			}
slouken@0
  1221
			stretch = swdata->stretch;
slouken@0
  1222
		}
slouken@0
  1223
	}
slouken@0
  1224
slouken@0
  1225
	if ( stretch ) {
slouken@0
  1226
		display = stretch;
slouken@0
  1227
	} else {
slouken@0
  1228
		display = swdata->display;
slouken@0
  1229
	}
slouken@0
  1230
	switch (overlay->format) {
slouken@0
  1231
	    case SDL_YV12_OVERLAY:
slouken@0
  1232
		lum = overlay->pixels[0];
slouken@0
  1233
		Cr =  overlay->pixels[1];
slouken@0
  1234
		Cb =  overlay->pixels[2];
slouken@0
  1235
		break;
slouken@0
  1236
	    case SDL_IYUV_OVERLAY:
slouken@0
  1237
		lum = overlay->pixels[0];
slouken@0
  1238
		Cr =  overlay->pixels[2];
slouken@0
  1239
		Cb =  overlay->pixels[1];
slouken@0
  1240
		break;
slouken@0
  1241
	    case SDL_YUY2_OVERLAY:
slouken@0
  1242
		lum = overlay->pixels[0];
slouken@0
  1243
		Cr = lum + 3;
slouken@0
  1244
		Cb = lum + 1;
slouken@0
  1245
		break;
slouken@0
  1246
	    case SDL_UYVY_OVERLAY:
slouken@0
  1247
		lum = overlay->pixels[0]+1;
slouken@0
  1248
		Cr = lum + 1;
slouken@0
  1249
		Cb = lum - 1;
slouken@0
  1250
		break;
slouken@0
  1251
	    case SDL_YVYU_OVERLAY:
slouken@0
  1252
		lum = overlay->pixels[0];
slouken@0
  1253
		Cr = lum + 1;
slouken@0
  1254
		Cb = lum + 3;
slouken@0
  1255
		break;
slouken@0
  1256
	    default:
slouken@292
  1257
		SDL_SetError("Unsupported YUV format in blit");
slouken@0
  1258
		return(-1);
slouken@0
  1259
	}
slouken@0
  1260
	if ( SDL_MUSTLOCK(display) ) {
slouken@0
  1261
        	if ( SDL_LockSurface(display) < 0 ) {
slouken@0
  1262
			return(-1);
slouken@0
  1263
		}
slouken@0
  1264
	}
slouken@0
  1265
	if ( stretch ) {
slouken@0
  1266
		dst = (Uint8 *)stretch->pixels;
slouken@0
  1267
	} else {
slouken@0
  1268
		dst = (Uint8 *)display->pixels
slouken@0
  1269
			+ dstrect->x * display->format->BytesPerPixel
slouken@0
  1270
			+ dstrect->y * display->pitch;
slouken@0
  1271
	}
slouken@0
  1272
	mod = (display->pitch / display->format->BytesPerPixel);
slouken@0
  1273
slouken@0
  1274
	if ( scale_2x ) {
slouken@0
  1275
		mod -= (overlay->w * 2);
slouken@0
  1276
		swdata->Display2X(swdata->colortab, swdata->rgb_2_pix,
slouken@0
  1277
		                  lum, Cr, Cb, dst, overlay->h, overlay->w,mod);
slouken@0
  1278
	} else {
slouken@0
  1279
		mod -= overlay->w;
slouken@0
  1280
		swdata->Display1X(swdata->colortab, swdata->rgb_2_pix,
slouken@0
  1281
		                  lum, Cr, Cb, dst, overlay->h, overlay->w,mod);
slouken@0
  1282
	}
slouken@0
  1283
	if ( SDL_MUSTLOCK(display) ) {
slouken@0
  1284
		SDL_UnlockSurface(display);
slouken@0
  1285
	}
slouken@0
  1286
	if ( stretch ) {
slouken@0
  1287
		display = swdata->display;
slouken@0
  1288
		SDL_SoftStretch(stretch, NULL, display, dstrect);
slouken@0
  1289
	}
slouken@0
  1290
	SDL_UpdateRects(display, 1, dstrect);
slouken@0
  1291
slouken@0
  1292
	return(0);
slouken@0
  1293
}
slouken@0
  1294
slouken@0
  1295
void SDL_FreeYUV_SW(_THIS, SDL_Overlay *overlay)
slouken@0
  1296
{
slouken@0
  1297
	struct private_yuvhwdata *swdata;
slouken@0
  1298
slouken@0
  1299
	swdata = overlay->hwdata;
slouken@0
  1300
	if ( swdata ) {
slouken@9
  1301
		if ( swdata->stretch ) {
slouken@9
  1302
			SDL_FreeSurface(swdata->stretch);
slouken@9
  1303
		}
slouken@0
  1304
		if ( swdata->pixels ) {
slouken@0
  1305
			free(swdata->pixels);
slouken@0
  1306
		}
slouken@0
  1307
		if ( swdata->colortab ) {
slouken@0
  1308
			free(swdata->colortab);
slouken@0
  1309
		}
slouken@0
  1310
		if ( swdata->rgb_2_pix ) {
slouken@0
  1311
			free(swdata->rgb_2_pix);
slouken@0
  1312
		}
slouken@0
  1313
		free(swdata);
slouken@0
  1314
	}
slouken@0
  1315
}