-
-
Notifications
You must be signed in to change notification settings - Fork 41
/
BlockData.cpp
1834 lines (1624 loc) · 53 KB
/
BlockData.cpp
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
245
246
247
248
249
250
251
252
253
254
255
256
257
258
259
260
261
262
263
264
265
266
267
268
269
270
271
272
273
274
275
276
277
278
279
280
281
282
283
284
285
286
287
288
289
290
291
292
293
294
295
296
297
298
299
300
301
302
303
304
305
306
307
308
309
310
311
312
313
314
315
316
317
318
319
320
321
322
323
324
325
326
327
328
329
330
331
332
333
334
335
336
337
338
339
340
341
342
343
344
345
346
347
348
349
350
351
352
353
354
355
356
357
358
359
360
361
362
363
364
365
366
367
368
369
370
371
372
373
374
375
376
377
378
379
380
381
382
383
384
385
386
387
388
389
390
391
392
393
394
395
396
397
398
399
400
401
402
403
404
405
406
407
408
409
410
411
412
413
414
415
416
417
418
419
420
421
422
423
424
425
426
427
428
429
430
431
432
433
434
435
436
437
438
439
440
441
442
443
444
445
446
447
448
449
450
451
452
453
454
455
456
457
458
459
460
461
462
463
464
465
466
467
468
469
470
471
472
473
474
475
476
477
478
479
480
481
482
483
484
485
486
487
488
489
490
491
492
493
494
495
496
497
498
499
500
501
502
503
504
505
506
507
508
509
510
511
512
513
514
515
516
517
518
519
520
521
522
523
524
525
526
527
528
529
530
531
532
533
534
535
536
537
538
539
540
541
542
543
544
545
546
547
548
549
550
551
552
553
554
555
556
557
558
559
560
561
562
563
564
565
566
567
568
569
570
571
572
573
574
575
576
577
578
579
580
581
582
583
584
585
586
587
588
589
590
591
592
593
594
595
596
597
598
599
600
601
602
603
604
605
606
607
608
609
610
611
612
613
614
615
616
617
618
619
620
621
622
623
624
625
626
627
628
629
630
631
632
633
634
635
636
637
638
639
640
641
642
643
644
645
646
647
648
649
650
651
652
653
654
655
656
657
658
659
660
661
662
663
664
665
666
667
668
669
670
671
672
673
674
675
676
677
678
679
680
681
682
683
684
685
686
687
688
689
690
691
692
693
694
695
696
697
698
699
700
701
702
703
704
705
706
707
708
709
710
711
712
713
714
715
716
717
718
719
720
721
722
723
724
725
726
727
728
729
730
731
732
733
734
735
736
737
738
739
740
741
742
743
744
745
746
747
748
749
750
751
752
753
754
755
756
757
758
759
760
761
762
763
764
765
766
767
768
769
770
771
772
773
774
775
776
777
778
779
780
781
782
783
784
785
786
787
788
789
790
791
792
793
794
795
796
797
798
799
800
801
802
803
804
805
806
807
808
809
810
811
812
813
814
815
816
817
818
819
820
821
822
823
824
825
826
827
828
829
830
831
832
833
834
835
836
837
838
839
840
841
842
843
844
845
846
847
848
849
850
851
852
853
854
855
856
857
858
859
860
861
862
863
864
865
866
867
868
869
870
871
872
873
874
875
876
877
878
879
880
881
882
883
884
885
886
887
888
889
890
891
892
893
894
895
896
897
898
899
900
901
902
903
904
905
906
907
908
909
910
911
912
913
914
915
916
917
918
919
920
921
922
923
924
925
926
927
928
929
930
931
932
933
934
935
936
937
938
939
940
941
942
943
944
945
946
947
948
949
950
951
952
953
954
955
956
957
958
959
960
961
962
963
964
965
966
967
968
969
970
971
972
973
974
975
976
977
978
979
980
981
982
983
984
985
986
987
988
989
990
991
992
993
994
995
996
997
998
999
1000
#include <assert.h>
#include <string.h>
#include "bcdec.h"
#include "BlockData.hpp"
#include "ColorSpace.hpp"
#include "Debug.hpp"
#include "MipMap.hpp"
#include "mmap.hpp"
#include "ProcessRGB.hpp"
#include "ProcessDxtc.hpp"
#include "Tables.hpp"
#include "TaskDispatch.hpp"
#ifdef __ARM_NEON
# include <arm_neon.h>
#endif
#if defined __SSE4_1__ || defined __AVX2__ || defined _MSC_VER
# ifdef _MSC_VER
# include <intrin.h>
# include <Windows.h>
# define _bswap(x) _byteswap_ulong(x)
# define _bswap64(x) _byteswap_uint64(x)
# else
# include <x86intrin.h>
# endif
#endif
#ifndef _bswap
# define _bswap(x) __builtin_bswap32(x)
# define _bswap64(x) __builtin_bswap64(x)
#endif
static uint8_t table59T58H[8] = { 3,6,11,16,23,32,41,64 };
BlockData::BlockData( const char* fn )
: m_file( fopen( fn, "rb" ) )
{
assert( m_file );
fseek( m_file, 0, SEEK_END );
m_maplen = ftell( m_file );
fseek( m_file, 0, SEEK_SET );
m_data = (uint8_t*)mmap( nullptr, m_maplen, PROT_READ, MAP_SHARED, fileno( m_file ), 0 );
auto data32 = (uint32_t*)m_data;
if( *data32 == 0x03525650 )
{
// PVR
switch( *(data32+2) )
{
case 6:
m_type = Etc1;
break;
case 7:
m_type = Bc1;
break;
case 11:
m_type = Bc3;
break;
case 12:
m_type = Bc4;
break;
case 13:
m_type = Bc5;
break;
case 15:
m_type = Bc7;
break;
case 22:
m_type = Etc2_RGB;
break;
case 23:
m_type = Etc2_RGBA;
break;
case 25:
m_type = Etc2_R11;
break;
case 26:
m_type = Etc2_RG11;
break;
default:
assert( false );
break;
}
m_size.y = *(data32+6);
m_size.x = *(data32+7);
m_dataOffset = 52 + *(data32+12);
}
else if( *data32 == 0x58544BAB )
{
// KTX
switch( *(data32+7) )
{
case 0x9274:
m_type = Etc2_RGB;
break;
case 0x9278:
m_type = Etc2_RGBA;
break;
case 0x9270:
m_type = Etc2_R11;
break;
case 0x9272:
m_type = Etc2_RG11;
break;
default:
assert( false );
break;
}
m_size.x = *(data32+9);
m_size.y = *(data32+10);
m_dataOffset = sizeof( uint32_t ) * 17 + *(data32+15);
}
else if( *data32 == 0x20534444 )
{
// DDS
switch( *(data32+21) )
{
case 0x31545844:
m_type = Bc1;
m_dataOffset = 128;
break;
case 0x35545844:
m_type = Bc3;
m_dataOffset = 128;
break;
case 0x30315844:
m_dataOffset = 148;
// DXGI_FORMAT_BCn
switch( *(data32+32) )
{
case 71:
case 72:
m_type = Bc1;
break;
case 77:
case 78:
m_type = Bc3;
break;
case 80:
m_type = Bc4;
break;
case 83:
m_type = Bc5;
break;
case 98:
case 99:
m_type = Bc7;
break;
default:
assert( false );
break;
};
break;
default:
assert( false );
break;
};
m_size.x = *(data32+4);
m_size.y = *(data32+3);
}
else
{
assert( false );
}
}
static void WritePvrHeader( uint32_t* dst, BlockData::Type type, const v2i& size, int levels )
{
*dst++ = 0x03525650; // version
*dst++ = 0; // flags
switch( type ) // pixelformat[0]
{
case BlockData::Etc1:
*dst++ = 6;
break;
case BlockData::Etc2_RGB:
*dst++ = 22;
break;
case BlockData::Etc2_RGBA:
*dst++ = 23;
break;
case BlockData::Etc2_R11:
*dst++ = 25;
break;
case BlockData::Etc2_RG11:
*dst++ = 26;
break;
case BlockData::Bc1:
*dst++ = 7;
break;
case BlockData::Bc3:
*dst++ = 11;
break;
case BlockData::Bc4:
*dst++ = 12;
break;
case BlockData::Bc5:
*dst++ = 13;
break;
case BlockData::Bc7:
*dst++ = 15;
break;
default:
assert( false );
break;
}
*dst++ = 0; // pixelformat[1]
*dst++ = 0; // colourspace
*dst++ = 0; // channel type
*dst++ = size.y; // height
*dst++ = size.x; // width
*dst++ = 1; // depth
*dst++ = 1; // num surfs
*dst++ = 1; // num faces
*dst++ = levels; // mipmap count
*dst++ = 0; // metadata size
}
static void WriteDdsHeader( uint32_t* dst, BlockData::Type type, const v2i& size, int levels )
{
const uint32_t flags = levels == 1 ? 0x1007 : 0x21007;
uint32_t pitch = size.x * size.y / 2;
if( type == BlockData::Etc2_RGBA || type == BlockData::Bc3 || type == BlockData::Bc5 || type == BlockData::Bc7 || type == BlockData::Etc2_RG11 ) pitch *= 2;
const uint32_t caps = levels == 1 ? 0x1000 : 0x401008;
*dst++ = 0x20534444; // magic
*dst++ = 124; // size
*dst++ = flags; // flags
*dst++ = size.y; // height
*dst++ = size.x; // width
*dst++ = pitch; // pitch
*dst++ = 0; // depth
*dst++ = levels; // mipmap count
memset( dst, 0, 44 );
dst += 11;
*dst++ = 32; // size
*dst++ = 4; // flags
switch( type )
{
case BlockData::Bc1:
memcpy( dst++, "DXT1", 4 );
break;
case BlockData::Bc3:
memcpy( dst++, "DXT5", 4 );
break;
case BlockData::Bc4:
memcpy( dst++, "DX10", 4 );
break;
case BlockData::Bc5:
memcpy( dst++, "DX10", 4 );
break;
case BlockData::Bc7:
memcpy( dst++, "DX10", 4 );
break;
default:
assert( false );
break;
}
memset( dst, 0, 20 );
dst += 5;
*dst++ = caps;
memset( dst, 0, 16 );
dst+= 4;
if( type == BlockData::Bc1 || type == BlockData::Bc3 ) return;
switch( type )
{
case BlockData::Bc4:
*dst++ = 80; // DXGI_FORMAT_BC4_UNORM
break;
case BlockData::Bc5:
*dst++ = 83; // DXGI_FORMAT_BC5_UNORM
break;
case BlockData::Bc7:
*dst++ = 98; // DXGI_FORMAT_BC7_UNORM
break;
default:
assert( false );
break;
}
*dst++ = 3; // DXGI_FORMAT_DIMENSION_TEXTURE2D
*dst++ = 0; // miscFlag
*dst++ = 1; // arraySize
*dst++ = 0; // miscFlags2
}
static uint8_t* OpenForWriting( const char* fn, size_t len, const v2i& size, FILE** f, int levels, BlockData::Type type, BlockData::Format format )
{
*f = fopen( fn, "wb+" );
assert( *f );
fseek( *f, len - 1, SEEK_SET );
const char zero = 0;
fwrite( &zero, 1, 1, *f );
fseek( *f, 0, SEEK_SET );
auto ret = (uint8_t*)mmap( nullptr, len, PROT_WRITE, MAP_SHARED, fileno( *f ), 0 );
auto dst = (uint32_t*)ret;
switch( format )
{
case BlockData::Pvr:
WritePvrHeader( dst, type, size, levels );
break;
case BlockData::Dds:
WriteDdsHeader( dst, type, size, levels );
break;
default:
assert( false );
break;
}
return ret;
}
static int AdjustSizeForMipmaps( const v2i& size, int levels )
{
int len = 0;
v2i current = size;
for( int i=1; i<levels; i++ )
{
assert( current.x != 1 || current.y != 1 );
current.x = std::max( 1, current.x / 2 );
current.y = std::max( 1, current.y / 2 );
len += ( ( current.x + 3 ) & ~3 ) * ( ( current.y + 3 ) & ~3 ) / 2;
}
assert( current.x == 1 && current.y == 1 );
return len;
}
BlockData::BlockData( const char* fn, const v2i& size, bool mipmap, Type type, Format format )
: m_size( size )
, m_dataOffset( 52 )
, m_maplen( m_size.x*m_size.y/2 )
, m_type( type )
{
assert( m_size.x%4 == 0 && m_size.y%4 == 0 );
uint32_t cnt = m_size.x * m_size.y / 16;
DBGPRINT( cnt << " blocks" );
int levels = 1;
if( mipmap )
{
levels = NumberOfMipLevels( size );
DBGPRINT( "Number of mipmaps: " << levels );
m_maplen += AdjustSizeForMipmaps( size, levels );
}
if( type == Etc2_RGBA || type == Bc3 || type == Bc5 || type == Bc7 || type == Etc2_RG11 ) m_maplen *= 2;
switch( format )
{
case Pvr:
m_dataOffset = 52;
break;
case Dds:
m_dataOffset = 128;
if( type == Bc4 || type == Bc5 || type == Bc7 ) m_dataOffset += 20;
break;
default:
assert( false );
break;
}
m_maplen += m_dataOffset;
m_data = OpenForWriting( fn, m_maplen, m_size, &m_file, levels, type, format );
}
BlockData::BlockData( const v2i& size, bool mipmap, Type type )
: m_size( size )
, m_dataOffset( 52 )
, m_file( nullptr )
, m_maplen( m_size.x*m_size.y/2 )
, m_type( type )
{
assert( m_size.x%4 == 0 && m_size.y%4 == 0 );
if( mipmap )
{
const int levels = NumberOfMipLevels( size );
m_maplen += AdjustSizeForMipmaps( size, levels );
}
if( type == Etc2_RGBA || type == Bc3 || type == Bc5 || type == Bc7 || type == Etc2_RG11 ) m_maplen *= 2;
m_maplen += m_dataOffset;
m_data = new uint8_t[m_maplen];
}
BlockData::~BlockData()
{
if( m_file )
{
munmap( m_data, m_maplen );
fclose( m_file );
}
else
{
delete[] m_data;
}
}
void BlockData::Process( const uint32_t* src, uint32_t blocks, size_t offset, size_t width, bool dither, bool useHeuristics )
{
auto dst = ((uint64_t*)( m_data + m_dataOffset )) + offset;
switch( m_type )
{
case Etc1:
if( dither )
{
CompressEtc1RgbDither( src, dst, blocks, width );
}
else
{
CompressEtc1Rgb( src, dst, blocks, width );
}
break;
case Etc2_RGB:
CompressEtc2Rgb( src, dst, blocks, width, useHeuristics );
break;
case Etc2_R11:
CompressEacR( src, dst, blocks, width );
break;
case Etc2_RG11:
dst = ((uint64_t*)( m_data + m_dataOffset )) + offset * 2;
CompressEacRg( src, dst, blocks, width );
break;
case Bc1:
if( dither )
{
CompressBc1Dither( src, dst, blocks, width );
}
else
{
CompressBc1( src, dst, blocks, width );
}
break;
case Bc4:
CompressBc4( src, dst, blocks, width );
break;
case Bc5:
dst = ((uint64_t*)( m_data + m_dataOffset )) + offset * 2;
CompressBc5( src, dst, blocks, width );
break;
default:
assert( false );
break;
}
}
void BlockData::ProcessRGBA( const uint32_t* src, uint32_t blocks, size_t offset, size_t width, bool useHeuristics, const bc7enc_compress_block_params* params )
{
auto dst = ((uint64_t*)( m_data + m_dataOffset )) + offset * 2;
switch( m_type )
{
case Etc2_RGBA:
CompressEtc2Rgba( src, dst, blocks, width, useHeuristics );
break;
case Bc3:
CompressBc3( src, dst, blocks, width );
break;
case Bc7:
CompressBc7( src, dst, blocks, width, params );
break;
default:
assert( false );
break;
}
}
namespace
{
static etcpak_force_inline int32_t expand6(uint32_t value)
{
return (value << 2) | (value >> 4);
}
static etcpak_force_inline int32_t expand7(uint32_t value)
{
return (value << 1) | (value >> 6);
}
static etcpak_force_inline void DecodeT( uint64_t block, uint32_t* dst, uint32_t w )
{
const auto r0 = ( block >> 24 ) & 0x1B;
const auto rh0 = ( r0 >> 3 ) & 0x3;
const auto rl0 = r0 & 0x3;
const auto g0 = ( block >> 20 ) & 0xF;
const auto b0 = ( block >> 16 ) & 0xF;
const auto r1 = ( block >> 12 ) & 0xF;
const auto g1 = ( block >> 8 ) & 0xF;
const auto b1 = ( block >> 4 ) & 0xF;
const auto cr0 = ( ( rh0 << 6 ) | ( rl0 << 4 ) | ( rh0 << 2 ) | rl0);
const auto cg0 = ( g0 << 4 ) | g0;
const auto cb0 = ( b0 << 4 ) | b0;
const auto cr1 = ( r1 << 4 ) | r1;
const auto cg1 = ( g1 << 4 ) | g1;
const auto cb1 = ( b1 << 4 ) | b1;
const auto codeword_hi = ( block >> 2 ) & 0x3;
const auto codeword_lo = block & 0x1;
const auto codeword = ( codeword_hi << 1 ) | codeword_lo;
const auto c2r = clampu8( cr1 + table59T58H[codeword] );
const auto c2g = clampu8( cg1 + table59T58H[codeword] );
const auto c2b = clampu8( cb1 + table59T58H[codeword] );
const auto c3r = clampu8( cr1 - table59T58H[codeword] );
const auto c3g = clampu8( cg1 - table59T58H[codeword] );
const auto c3b = clampu8( cb1 - table59T58H[codeword] );
const uint32_t col_tab[4] = {
uint32_t( cr0 | ( cg0 << 8 ) | ( cb0 << 16 ) | 0xFF000000 ),
uint32_t( c2r | ( c2g << 8 ) | ( c2b << 16 ) | 0xFF000000 ),
uint32_t( cr1 | ( cg1 << 8 ) | ( cb1 << 16 ) | 0xFF000000 ),
uint32_t( c3r | ( c3g << 8 ) | ( c3b << 16 ) | 0xFF000000 )
};
const uint32_t indexes = ( block >> 32 ) & 0xFFFFFFFF;
for( uint8_t j = 0; j < 4; j++ )
{
for( uint8_t i = 0; i < 4; i++ )
{
//2bit indices distributed on two lane 16bit numbers
const uint8_t index = ( ( ( indexes >> ( j + i * 4 + 16 ) ) & 0x1 ) << 1) | ( ( indexes >> ( j + i * 4 ) ) & 0x1);
dst[j * w + i] = col_tab[index];
}
}
}
static etcpak_force_inline void DecodeTAlpha( uint64_t block, uint64_t alpha, uint32_t* dst, uint32_t w )
{
const auto r0 = ( block >> 24 ) & 0x1B;
const auto rh0 = ( r0 >> 3 ) & 0x3;
const auto rl0 = r0 & 0x3;
const auto g0 = ( block >> 20 ) & 0xF;
const auto b0 = ( block >> 16 ) & 0xF;
const auto r1 = ( block >> 12 ) & 0xF;
const auto g1 = ( block >> 8 ) & 0xF;
const auto b1 = ( block >> 4 ) & 0xF;
const auto cr0 = ( ( rh0 << 6 ) | ( rl0 << 4 ) | ( rh0 << 2 ) | rl0);
const auto cg0 = ( g0 << 4 ) | g0;
const auto cb0 = ( b0 << 4 ) | b0;
const auto cr1 = ( r1 << 4 ) | r1;
const auto cg1 = ( g1 << 4 ) | g1;
const auto cb1 = ( b1 << 4 ) | b1;
const auto codeword_hi = ( block >> 2 ) & 0x3;
const auto codeword_lo = block & 0x1;
const auto codeword = (codeword_hi << 1) | codeword_lo;
const int32_t base = alpha >> 56;
const int32_t mul = ( alpha >> 52 ) & 0xF;
const auto tbl = g_alpha[( alpha >> 48 ) & 0xF];
const auto c2r = clampu8( cr1 + table59T58H[codeword] );
const auto c2g = clampu8( cg1 + table59T58H[codeword] );
const auto c2b = clampu8( cb1 + table59T58H[codeword] );
const auto c3r = clampu8( cr1 - table59T58H[codeword] );
const auto c3g = clampu8( cg1 - table59T58H[codeword] );
const auto c3b = clampu8( cb1 - table59T58H[codeword] );
const uint32_t col_tab[4] = {
uint32_t( cr0 | ( cg0 << 8 ) | ( cb0 << 16 ) ),
uint32_t( c2r | ( c2g << 8 ) | ( c2b << 16 ) ),
uint32_t( cr1 | ( cg1 << 8 ) | ( cb1 << 16 ) ),
uint32_t( c3r | ( c3g << 8 ) | ( c3b << 16 ) )
};
const uint32_t indexes = ( block >> 32 ) & 0xFFFFFFFF;
for( uint8_t j = 0; j < 4; j++ )
{
for( uint8_t i = 0; i < 4; i++ )
{
//2bit indices distributed on two lane 16bit numbers
const uint8_t index = ( ( ( indexes >> ( j + i * 4 + 16 ) ) & 0x1 ) << 1 ) | ( ( indexes >> ( j + i * 4 ) ) & 0x1 );
const auto amod = tbl[( alpha >> ( 45 - j * 3 - i * 12 ) ) & 0x7];
const uint32_t a = clampu8( base + amod * mul );
dst[j * w + i] = col_tab[index] | ( a << 24 );
}
}
}
static etcpak_force_inline void DecodeH( uint64_t block, uint32_t* dst, uint32_t w )
{
const uint32_t indexes = ( block >> 32 ) & 0xFFFFFFFF;
const auto r0444 = ( block >> 27 ) & 0xF;
const auto g0444 = ( ( block >> 20 ) & 0x1 ) | ( ( ( block >> 24 ) & 0x7 ) << 1 );
const auto b0444 = ( ( block >> 15 ) & 0x7 ) | ( ( ( block >> 19 ) & 0x1 ) << 3 );
const auto r1444 = ( block >> 11 ) & 0xF;
const auto g1444 = ( block >> 7 ) & 0xF;
const auto b1444 = ( block >> 3 ) & 0xF;
const auto r0 = ( r0444 << 4 ) | r0444;
const auto g0 = ( g0444 << 4 ) | g0444;
const auto b0 = ( b0444 << 4 ) | b0444;
const auto r1 = ( r1444 << 4 ) | r1444;
const auto g1 = ( g1444 << 4 ) | g1444;
const auto b1 = ( b1444 << 4 ) | b1444;
const auto codeword_hi = ( ( block & 0x1 ) << 1 ) | ( ( block & 0x4 ) );
const auto c0 = ( r0444 << 8 ) | ( g0444 << 4 ) | ( b0444 << 0 );
const auto c1 = ( block >> 3 ) & ( ( 1 << 12 ) - 1 );
const auto codeword_lo = ( c0 >= c1 ) ? 1 : 0;
const auto codeword = codeword_hi | codeword_lo;
const uint32_t col_tab[] = {
uint32_t( clampu8( r0 + table59T58H[codeword] ) | ( clampu8( g0 + table59T58H[codeword] ) << 8 ) | ( clampu8( b0 + table59T58H[codeword] ) << 16 ) ),
uint32_t( clampu8( r0 - table59T58H[codeword] ) | ( clampu8( g0 - table59T58H[codeword] ) << 8 ) | ( clampu8( b0 - table59T58H[codeword] ) << 16 ) ),
uint32_t( clampu8( r1 + table59T58H[codeword] ) | ( clampu8( g1 + table59T58H[codeword] ) << 8 ) | ( clampu8( b1 + table59T58H[codeword] ) << 16 ) ),
uint32_t( clampu8( r1 - table59T58H[codeword] ) | ( clampu8( g1 - table59T58H[codeword] ) << 8 ) | ( clampu8( b1 - table59T58H[codeword] ) << 16 ) )
};
for( uint8_t j = 0; j < 4; j++ )
{
for( uint8_t i = 0; i < 4; i++ )
{
const uint8_t index = ( ( ( indexes >> ( j + i * 4 + 16 ) ) & 0x1 ) << 1 ) | ( ( indexes >> ( j + i * 4 ) ) & 0x1 );
dst[j * w + i] = col_tab[index] | 0xFF000000;
}
}
}
static etcpak_force_inline void DecodeHAlpha( uint64_t block, uint64_t alpha, uint32_t* dst, uint32_t w )
{
const uint32_t indexes = ( block >> 32 ) & 0xFFFFFFFF;
const auto r0444 = ( block >> 27 ) & 0xF;
const auto g0444 = ( ( block >> 20 ) & 0x1 ) | ( ( ( block >> 24 ) & 0x7 ) << 1 );
const auto b0444 = ( ( block >> 15 ) & 0x7 ) | ( ( ( block >> 19 ) & 0x1 ) << 3 );
const auto r1444 = ( block >> 11 ) & 0xF;
const auto g1444 = ( block >> 7 ) & 0xF;
const auto b1444 = ( block >> 3 ) & 0xF;
const auto r0 = ( r0444 << 4 ) | r0444;
const auto g0 = ( g0444 << 4 ) | g0444;
const auto b0 = ( b0444 << 4 ) | b0444;
const auto r1 = ( r1444 << 4 ) | r1444;
const auto g1 = ( g1444 << 4 ) | g1444;
const auto b1 = ( b1444 << 4 ) | b1444;
const auto codeword_hi = ( ( block & 0x1 ) << 1 ) | ( ( block & 0x4 ) );
const auto c0 = ( r0444 << 8 ) | ( g0444 << 4 ) | ( b0444 << 0 );
const auto c1 = ( block >> 3 ) & ( ( 1 << 12 ) - 1 );
const auto codeword_lo = ( c0 >= c1 ) ? 1 : 0;
const auto codeword = codeword_hi | codeword_lo;
const int32_t base = alpha >> 56;
const int32_t mul = ( alpha >> 52 ) & 0xF;
const auto tbl = g_alpha[(alpha >> 48) & 0xF];
const uint32_t col_tab[] = {
uint32_t( clampu8( r0 + table59T58H[codeword] ) | ( clampu8( g0 + table59T58H[codeword] ) << 8 ) | ( clampu8( b0 + table59T58H[codeword] ) << 16 ) ),
uint32_t( clampu8( r0 - table59T58H[codeword] ) | ( clampu8( g0 - table59T58H[codeword] ) << 8 ) | ( clampu8( b0 - table59T58H[codeword] ) << 16 ) ),
uint32_t( clampu8( r1 + table59T58H[codeword] ) | ( clampu8( g1 + table59T58H[codeword] ) << 8 ) | ( clampu8( b1 + table59T58H[codeword] ) << 16 ) ),
uint32_t( clampu8( r1 - table59T58H[codeword] ) | ( clampu8( g1 - table59T58H[codeword] ) << 8 ) | ( clampu8( b1 - table59T58H[codeword] ) << 16 ) )
};
for( uint8_t j = 0; j < 4; j++ )
{
for( uint8_t i = 0; i < 4; i++ )
{
const uint8_t index = ( ( ( indexes >> ( j + i * 4 + 16 ) ) & 0x1 ) << 1 ) | ( ( indexes >> ( j + i * 4 ) ) & 0x1 );
const auto amod = tbl[( alpha >> ( 45 - j * 3 - i * 12) ) & 0x7];
const uint32_t a = clampu8( base + amod * mul );
dst[j * w + i] = col_tab[index] | ( a << 24 );
}
}
}
static etcpak_force_inline void DecodePlanar( uint64_t block, uint32_t* dst, uint32_t w )
{
const auto bv = expand6((block >> ( 0 + 32)) & 0x3F);
const auto gv = expand7((block >> ( 6 + 32)) & 0x7F);
const auto rv = expand6((block >> (13 + 32)) & 0x3F);
const auto bh = expand6((block >> (19 + 32)) & 0x3F);
const auto gh = expand7((block >> (25 + 32)) & 0x7F);
const auto rh0 = (block >> (32 - 32)) & 0x01;
const auto rh1 = ((block >> (34 - 32)) & 0x1F) << 1;
const auto rh = expand6(rh0 | rh1);
const auto bo0 = (block >> (39 - 32)) & 0x07;
const auto bo1 = ((block >> (43 - 32)) & 0x3) << 3;
const auto bo2 = ((block >> (48 - 32)) & 0x1) << 5;
const auto bo = expand6(bo0 | bo1 | bo2);
const auto go0 = (block >> (49 - 32)) & 0x3F;
const auto go1 = ((block >> (56 - 32)) & 0x01) << 6;
const auto go = expand7(go0 | go1);
const auto ro = expand6((block >> (57 - 32)) & 0x3F);
#ifdef __ARM_NEON
uint64_t init = uint64_t(uint16_t(rh-ro)) | ( uint64_t(uint16_t(gh-go)) << 16 ) | ( uint64_t(uint16_t(bh-bo)) << 32 );
int16x8_t chco = vreinterpretq_s16_u64( vdupq_n_u64( init ) );
init = uint64_t(uint16_t( (rv-ro) - 4 * (rh-ro) )) | ( uint64_t(uint16_t( (gv-go) - 4 * (gh-go) )) << 16 ) | ( uint64_t(uint16_t( (bv-bo) - 4 * (bh-bo) )) << 32 );
int16x8_t cvco = vreinterpretq_s16_u64( vdupq_n_u64( init ) );
init = uint64_t(4*ro+2) | ( uint64_t(4*go+2) << 16 ) | ( uint64_t(4*bo+2) << 32 ) | ( uint64_t(0xFFF) << 48 );
int16x8_t col = vreinterpretq_s16_u64( vdupq_n_u64( init ) );
for( int j=0; j<4; j++ )
{
for( int i=0; i<4; i++ )
{
uint8x8_t c = vqshrun_n_s16( col, 2 );
vst1_lane_u32( dst+j*w+i, vreinterpret_u32_u8( c ), 0 );
col = vaddq_s16( col, chco );
}
col = vaddq_s16( col, cvco );
}
#elif defined __AVX2__
const auto R0 = 4*ro+2;
const auto G0 = 4*go+2;
const auto B0 = 4*bo+2;
const auto RHO = rh-ro;
const auto GHO = gh-go;
const auto BHO = bh-bo;
__m256i cvco = _mm256_setr_epi16( rv - ro, gv - go, bv - bo, 0, rv - ro, gv - go, bv - bo, 0, rv - ro, gv - go, bv - bo, 0, rv - ro, gv - go, bv - bo, 0 );
__m256i col = _mm256_setr_epi16( R0, G0, B0, 0xFFF, R0+RHO, G0+GHO, B0+BHO, 0xFFF, R0+2*RHO, G0+2*GHO, B0+2*BHO, 0xFFF, R0+3*RHO, G0+3*GHO, B0+3*BHO, 0xFFF );
for( int j=0; j<4; j++ )
{
__m256i c = _mm256_srai_epi16( col, 2 );
__m128i s = _mm_packus_epi16( _mm256_castsi256_si128( c ), _mm256_extracti128_si256( c, 1 ) );
_mm_storeu_si128( (__m128i*)(dst+j*w), s );
col = _mm256_add_epi16( col, cvco );
}
#elif defined __SSE4_1__
__m128i chco = _mm_setr_epi16( rh - ro, gh - go, bh - bo, 0, 0, 0, 0, 0 );
__m128i cvco = _mm_setr_epi16( (rv - ro) - 4 * (rh - ro), (gv - go) - 4 * (gh - go), (bv - bo) - 4 * (bh - bo), 0, 0, 0, 0, 0 );
__m128i col = _mm_setr_epi16( 4*ro+2, 4*go+2, 4*bo+2, 0xFFF, 0, 0, 0, 0 );
for( int j=0; j<4; j++ )
{
for( int i=0; i<4; i++ )
{
__m128i c = _mm_srai_epi16( col, 2 );
__m128i s = _mm_packus_epi16( c, c );
dst[j*w+i] = _mm_cvtsi128_si32( s );
col = _mm_add_epi16( col, chco );
}
col = _mm_add_epi16( col, cvco );
}
#else
for( int j=0; j<4; j++ )
{
for( int i=0; i<4; i++ )
{
const uint32_t r = (i * (rh - ro) + j * (rv - ro) + 4 * ro + 2) >> 2;
const uint32_t g = (i * (gh - go) + j * (gv - go) + 4 * go + 2) >> 2;
const uint32_t b = (i * (bh - bo) + j * (bv - bo) + 4 * bo + 2) >> 2;
if( ( ( r | g | b ) & ~0xFF ) == 0 )
{
dst[j*w+i] = r | ( g << 8 ) | ( b << 16 ) | 0xFF000000;
}
else
{
const auto rc = clampu8( r );
const auto gc = clampu8( g );
const auto bc = clampu8( b );
dst[j*w+i] = rc | ( gc << 8 ) | ( bc << 16 ) | 0xFF000000;
}
}
}
#endif
}
static etcpak_force_inline void DecodePlanarAlpha( uint64_t block, uint64_t alpha, uint32_t* dst, uint32_t w )
{
const auto bv = expand6((block >> ( 0 + 32)) & 0x3F);
const auto gv = expand7((block >> ( 6 + 32)) & 0x7F);
const auto rv = expand6((block >> (13 + 32)) & 0x3F);
const auto bh = expand6((block >> (19 + 32)) & 0x3F);
const auto gh = expand7((block >> (25 + 32)) & 0x7F);
const auto rh0 = (block >> (32 - 32)) & 0x01;
const auto rh1 = ((block >> (34 - 32)) & 0x1F) << 1;
const auto rh = expand6(rh0 | rh1);
const auto bo0 = (block >> (39 - 32)) & 0x07;
const auto bo1 = ((block >> (43 - 32)) & 0x3) << 3;
const auto bo2 = ((block >> (48 - 32)) & 0x1) << 5;
const auto bo = expand6(bo0 | bo1 | bo2);
const auto go0 = (block >> (49 - 32)) & 0x3F;
const auto go1 = ((block >> (56 - 32)) & 0x01) << 6;
const auto go = expand7(go0 | go1);
const auto ro = expand6((block >> (57 - 32)) & 0x3F);
const int32_t base = alpha >> 56;
const int32_t mul = ( alpha >> 52 ) & 0xF;
const auto tbl = g_alpha[( alpha >> 48 ) & 0xF];
#ifdef __ARM_NEON
uint64_t init = uint64_t(uint16_t(rh-ro)) | ( uint64_t(uint16_t(gh-go)) << 16 ) | ( uint64_t(uint16_t(bh-bo)) << 32 );
int16x8_t chco = vreinterpretq_s16_u64( vdupq_n_u64( init ) );
init = uint64_t(uint16_t( (rv-ro) - 4 * (rh-ro) )) | ( uint64_t(uint16_t( (gv-go) - 4 * (gh-go) )) << 16 ) | ( uint64_t(uint16_t( (bv-bo) - 4 * (bh-bo) )) << 32 );
int16x8_t cvco = vreinterpretq_s16_u64( vdupq_n_u64( init ) );
init = uint64_t(4*ro+2) | ( uint64_t(4*go+2) << 16 ) | ( uint64_t(4*bo+2) << 32 );
int16x8_t col = vreinterpretq_s16_u64( vdupq_n_u64( init ) );
for( int j=0; j<4; j++ )
{
for( int i=0; i<4; i++ )
{
const auto amod = tbl[(alpha >> ( 45 - j*3 - i*12 )) & 0x7];
const uint32_t a = clampu8( base + amod * mul );
uint8x8_t c = vqshrun_n_s16( col, 2 );
dst[j*w+i] = vget_lane_u32( vreinterpret_u32_u8( c ), 0 ) | ( a << 24 );
col = vaddq_s16( col, chco );
}
col = vaddq_s16( col, cvco );
}
#elif defined __SSE4_1__
__m128i chco = _mm_setr_epi16( rh - ro, gh - go, bh - bo, 0, 0, 0, 0, 0 );
__m128i cvco = _mm_setr_epi16( (rv - ro) - 4 * (rh - ro), (gv - go) - 4 * (gh - go), (bv - bo) - 4 * (bh - bo), 0, 0, 0, 0, 0 );
__m128i col = _mm_setr_epi16( 4*ro+2, 4*go+2, 4*bo+2, 0, 0, 0, 0, 0 );
for( int j=0; j<4; j++ )
{
for( int i=0; i<4; i++ )
{
const auto amod = tbl[(alpha >> ( 45 - j*3 - i*12 )) & 0x7];
const uint32_t a = clampu8( base + amod * mul );
__m128i c = _mm_srai_epi16( col, 2 );
__m128i s = _mm_packus_epi16( c, c );
dst[j*w+i] = _mm_cvtsi128_si32( s ) | ( a << 24 );
col = _mm_add_epi16( col, chco );
}
col = _mm_add_epi16( col, cvco );
}
#else
for (auto j = 0; j < 4; j++)
{
for (auto i = 0; i < 4; i++)
{
const uint32_t r = (i * (rh - ro) + j * (rv - ro) + 4 * ro + 2) >> 2;
const uint32_t g = (i * (gh - go) + j * (gv - go) + 4 * go + 2) >> 2;
const uint32_t b = (i * (bh - bo) + j * (bv - bo) + 4 * bo + 2) >> 2;
const auto amod = tbl[(alpha >> ( 45 - j*3 - i*12 )) & 0x7];
const uint32_t a = clampu8( base + amod * mul );
if( ( ( r | g | b ) & ~0xFF ) == 0 )
{
dst[j*w+i] = r | ( g << 8 ) | ( b << 16 ) | ( a << 24 );
}
else
{
const auto rc = clampu8( r );
const auto gc = clampu8( g );
const auto bc = clampu8( b );
dst[j*w+i] = rc | ( gc << 8 ) | ( bc << 16 ) | ( a << 24 );
}
}
}
#endif
}
}
BitmapPtr BlockData::Decode()
{
switch( m_type )
{
case Etc1:
case Etc2_RGB:
return DecodeRGB();
case Etc2_RGBA:
return DecodeRGBA();
case Etc2_R11:
return DecodeR();
case Etc2_RG11:
return DecodeRG();
case Bc1:
return DecodeBc1();
case Bc3:
return DecodeBc3();
case Bc4:
return DecodeBc4();
case Bc5:
return DecodeBc5();
case Bc7:
return DecodeBc7();
default:
assert( false );
return nullptr;
}
}
static etcpak_force_inline uint64_t ConvertByteOrder( uint64_t d )
{
uint32_t word[2];
memcpy( word, &d, 8 );
word[0] = _bswap( word[0] );
word[1] = _bswap( word[1] );
memcpy( &d, word, 8 );
return d;
}
static etcpak_force_inline void DecodeRGBPart( uint64_t d, uint32_t* dst, uint32_t w )
{
d = ConvertByteOrder( d );
uint32_t br[2], bg[2], bb[2];
if( d & 0x2 )
{
int32_t dr, dg, db;
uint32_t r0 = ( d & 0xF8000000 ) >> 27;
uint32_t g0 = ( d & 0x00F80000 ) >> 19;
uint32_t b0 = ( d & 0x0000F800 ) >> 11;
dr = ( int32_t(d) << 5 ) >> 29;
dg = ( int32_t(d) << 13 ) >> 29;
db = ( int32_t(d) << 21 ) >> 29;
int32_t r1 = int32_t(r0) + dr;
int32_t g1 = int32_t(g0) + dg;
int32_t b1 = int32_t(b0) + db;
// T mode
if ( (r1 < 0) || (r1 > 31) )
{
DecodeT( d, dst, w );
return;
}
// H mode
if ((g1 < 0) || (g1 > 31))
{
DecodeH( d, dst, w );
return;
}
// P mode
if( (b1 < 0) || (b1 > 31) )
{
DecodePlanar( d, dst, w );
return;
}
br[0] = ( r0 << 3 ) | ( r0 >> 2 );
br[1] = ( r1 << 3 ) | ( r1 >> 2 );
bg[0] = ( g0 << 3 ) | ( g0 >> 2 );
bg[1] = ( g1 << 3 ) | ( g1 >> 2 );
bb[0] = ( b0 << 3 ) | ( b0 >> 2 );
bb[1] = ( b1 << 3 ) | ( b1 >> 2 );
}
else
{
br[0] = ( ( d & 0xF0000000 ) >> 24 ) | ( ( d & 0xF0000000 ) >> 28 );
br[1] = ( ( d & 0x0F000000 ) >> 20 ) | ( ( d & 0x0F000000 ) >> 24 );
bg[0] = ( ( d & 0x00F00000 ) >> 16 ) | ( ( d & 0x00F00000 ) >> 20 );
bg[1] = ( ( d & 0x000F0000 ) >> 12 ) | ( ( d & 0x000F0000 ) >> 16 );
bb[0] = ( ( d & 0x0000F000 ) >> 8 ) | ( ( d & 0x0000F000 ) >> 12 );
bb[1] = ( ( d & 0x00000F00 ) >> 4 ) | ( ( d & 0x00000F00 ) >> 8 );
}
unsigned int tcw[2];
tcw[0] = ( d & 0xE0 ) >> 5;
tcw[1] = ( d & 0x1C ) >> 2;
uint32_t b1 = ( d >> 32 ) & 0xFFFF;
uint32_t b2 = ( d >> 48 );
b1 = ( b1 | ( b1 << 8 ) ) & 0x00FF00FF;
b1 = ( b1 | ( b1 << 4 ) ) & 0x0F0F0F0F;
b1 = ( b1 | ( b1 << 2 ) ) & 0x33333333;
b1 = ( b1 | ( b1 << 1 ) ) & 0x55555555;
b2 = ( b2 | ( b2 << 8 ) ) & 0x00FF00FF;
b2 = ( b2 | ( b2 << 4 ) ) & 0x0F0F0F0F;
b2 = ( b2 | ( b2 << 2 ) ) & 0x33333333;
b2 = ( b2 | ( b2 << 1 ) ) & 0x55555555;
uint32_t idx = b1 | ( b2 << 1 );
if( d & 0x1 )