forked from AviSynth/AviSynthPlus
/
convert.cpp
3654 lines (3204 loc) · 184 KB
/
convert.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
// Avisynth v2.5. Copyright 2002-2009 Ben Rudiak-Gould et al.
// http://www.avisynth.org
// This program is free software; you can redistribute it and/or modify
// it under the terms of the GNU General Public License as published by
// the Free Software Foundation; either version 2 of the License, or
// (at your option) any later version.
//
// This program is distributed in the hope that it will be useful,
// but WITHOUT ANY WARRANTY; without even the implied warranty of
// MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
// GNU General Public License for more details.
//
// You should have received a copy of the GNU General Public License
// along with this program; if not, write to the Free Software
// Foundation, Inc., 675 Mass Ave, Cambridge, MA 02139, USA, or visit
// http://www.gnu.org/copyleft/gpl.html .
//
// Linking Avisynth statically or dynamically with other modules is making a
// combined work based on Avisynth. Thus, the terms and conditions of the GNU
// General Public License cover the whole combination.
//
// As a special exception, the copyright holders of Avisynth give you
// permission to link Avisynth with independent modules that communicate with
// Avisynth solely through the interfaces defined in avisynth.h, regardless of the license
// terms of these independent modules, and to copy and distribute the
// resulting combined work under terms of your choice, provided that
// every copy of the combined work is accompanied by a complete copy of
// the source code of Avisynth (the version of Avisynth used to produce the
// combined work), being distributed under the terms of the GNU General
// Public License plus this exception. An independent module is a module
// which is not derived from or based on Avisynth, such as 3rd-party filters,
// import and export plugins, or graphical user interfaces.
#include "convert.h"
#include "convert_planar.h"
#include "convert_rgb.h"
#include "convert_yv12.h"
#include "convert_yuy2.h"
#include <avs/alignment.h>
#include <avs/minmax.h>
#include <avs/config.h>
#include <emmintrin.h>
#include <smmintrin.h> // SSE4.1
#include <tuple>
#include <map>
#include <algorithm>
#include "convert_avx.h"
#include "convert_avx2.h"
/********************************************************************
***** Declare index of new filters for Avisynth's filter engine *****
********************************************************************/
extern const AVSFunction Convert_filters[] = { // matrix can be "rec601", "rec709", "PC.601" or "PC.709" or "rec2020"
{ "ConvertToRGB", BUILTIN_FUNC_PREFIX, "c[matrix]s[interlaced]b[ChromaInPlacement]s[chromaresample]s", ConvertToRGB::Create, (void *)0 },
{ "ConvertToRGB24", BUILTIN_FUNC_PREFIX, "c[matrix]s[interlaced]b[ChromaInPlacement]s[chromaresample]s", ConvertToRGB::Create, (void *)24 },
{ "ConvertToRGB32", BUILTIN_FUNC_PREFIX, "c[matrix]s[interlaced]b[ChromaInPlacement]s[chromaresample]s", ConvertToRGB::Create, (void *)32 },
{ "ConvertToRGB48", BUILTIN_FUNC_PREFIX, "c[matrix]s[interlaced]b[ChromaInPlacement]s[chromaresample]s", ConvertToRGB::Create, (void *)48 },
{ "ConvertToRGB64", BUILTIN_FUNC_PREFIX, "c[matrix]s[interlaced]b[ChromaInPlacement]s[chromaresample]s", ConvertToRGB::Create, (void *)64 },
{ "ConvertToPlanarRGB", BUILTIN_FUNC_PREFIX, "c[matrix]s[interlaced]b[ChromaInPlacement]s[chromaresample]s", ConvertToRGB::Create, (void *)-1 },
{ "ConvertToPlanarRGBA", BUILTIN_FUNC_PREFIX, "c[matrix]s[interlaced]b[ChromaInPlacement]s[chromaresample]s", ConvertToRGB::Create, (void *)-2 },
{ "ConvertToY8", BUILTIN_FUNC_PREFIX, "c[matrix]s", ConvertToY::Create, (void*)0 }, // user_data == 0 -> only 8 bit sources
{ "ConvertToYV12", BUILTIN_FUNC_PREFIX, "c[interlaced]b[matrix]s[ChromaInPlacement]s[chromaresample]s[ChromaOutPlacement]s", ConvertToYV12::Create, (void*)0 },
{ "ConvertToYV24", BUILTIN_FUNC_PREFIX, "c[interlaced]b[matrix]s[ChromaInPlacement]s[chromaresample]s", ConvertToPlanarGeneric::CreateYUV444, (void*)0},
{ "ConvertToYV16", BUILTIN_FUNC_PREFIX, "c[interlaced]b[matrix]s[ChromaInPlacement]s[chromaresample]s", ConvertToPlanarGeneric::CreateYUV422, (void*)0},
{ "ConvertToYV411", BUILTIN_FUNC_PREFIX, "c[interlaced]b[matrix]s[ChromaInPlacement]s[chromaresample]s", ConvertToPlanarGeneric::CreateYV411, (void*)0},
{ "ConvertToYUY2", BUILTIN_FUNC_PREFIX, "c[interlaced]b[matrix]s[ChromaInPlacement]s[chromaresample]s", ConvertToYUY2::Create },
{ "ConvertBackToYUY2", BUILTIN_FUNC_PREFIX, "c[matrix]s", ConvertBackToYUY2::Create },
{ "ConvertToY", BUILTIN_FUNC_PREFIX, "c[matrix]s", ConvertToY::Create, (void*)1 }, // user_data == 1 -> any bit depth sources
{ "ConvertToYUV411", BUILTIN_FUNC_PREFIX, "c[interlaced]b[matrix]s[ChromaInPlacement]s[chromaresample]s", ConvertToPlanarGeneric::CreateYV411, (void*)1}, // alias for ConvertToYV411, 8 bit check later
{ "ConvertToYUV420", BUILTIN_FUNC_PREFIX, "c[interlaced]b[matrix]s[ChromaInPlacement]s[chromaresample]s[ChromaOutPlacement]s", ConvertToPlanarGeneric::CreateYUV420, (void*)1},
{ "ConvertToYUV422", BUILTIN_FUNC_PREFIX, "c[interlaced]b[matrix]s[ChromaInPlacement]s[chromaresample]s", ConvertToPlanarGeneric::CreateYUV422, (void*)1},
{ "ConvertToYUV444", BUILTIN_FUNC_PREFIX, "c[interlaced]b[matrix]s[ChromaInPlacement]s[chromaresample]s", ConvertToPlanarGeneric::CreateYUV444, (void*)1},
{ "ConvertTo8bit", BUILTIN_FUNC_PREFIX, "c[bits]i[truerange]b[dither]f[dither_bits]i[fulls]b[fulld]b", ConvertBits::Create, (void *)8 },
{ "ConvertTo16bit", BUILTIN_FUNC_PREFIX, "c[bits]i[truerange]b[dither]f[dither_bits]i[fulls]b[fulld]b", ConvertBits::Create, (void *)16 },
{ "ConvertToFloat", BUILTIN_FUNC_PREFIX, "c[bits]i[truerange]b[dither]f[dither_bits]i[fulls]b[fulld]b", ConvertBits::Create, (void *)32 },
{ "ConvertBits", BUILTIN_FUNC_PREFIX, "c[bits]i[truerange]b[dither]f[dither_bits]i[fulls]b[fulld]b", ConvertBits::Create, (void *)0 },
{ "AddAlphaPlane", BUILTIN_FUNC_PREFIX, "c[mask].", AddAlphaPlane::Create},
{ "RemoveAlphaPlane", BUILTIN_FUNC_PREFIX, "c", RemoveAlphaPlane::Create},
{ 0 }
};
// for YUY2
static const int crv_rec601 = int(1.596*65536+0.5);
static const int cgv_rec601 = int(0.813*65536+0.5);
static const int cgu_rec601 = int(0.391*65536+0.5);
static const int cbu_rec601 = int(2.018*65536+0.5);
static const int crv_rec709 = int(1.793*65536+0.5);
static const int cgv_rec709 = int(0.533*65536+0.5);
static const int cgu_rec709 = int(0.213*65536+0.5);
static const int cbu_rec709 = int(2.112*65536+0.5);
static const int crv_pc601 = int(1.407*65536+0.5);
static const int cgv_pc601 = int(0.717*65536+0.5);
static const int cgu_pc601 = int(0.345*65536+0.5);
static const int cbu_pc601 = int(1.779*65536+0.5);
static const int crv_pc709 = int(1.581*65536+0.5);
static const int cgv_pc709 = int(0.470*65536+0.5);
static const int cgu_pc709 = int(0.188*65536+0.5);
static const int cbu_pc709 = int(1.863*65536+0.5);
static const int cy_rec = int((255.0/219.0)*65536+0.5);
static const int cy_pc = 65536;
// still YUY2 only
static const int crv_values[4] = { crv_rec601, crv_rec709, crv_pc601, crv_pc709 };
static const int cgv_values[4] = { cgv_rec601, cgv_rec709, cgv_pc601, cgv_pc709 };
static const int cgu_values[4] = { cgu_rec601, cgu_rec709, cgu_pc601, cgu_pc709 };
static const int cbu_values[4] = { cbu_rec601, cbu_rec709, cbu_pc601, cbu_pc709 };
static const int cy_values[4] = { cy_rec, cy_rec, cy_pc, cy_pc};
int getMatrix( const char* matrix, IScriptEnvironment* env) {
if (matrix) {
if (!lstrcmpi(matrix, "rec601"))
return Rec601;
if (!lstrcmpi(matrix, "rec709"))
return Rec709;
if (!lstrcmpi(matrix, "PC.601"))
return PC_601;
if (!lstrcmpi(matrix, "PC.709"))
return PC_709;
if (!lstrcmpi(matrix, "PC601"))
return PC_601;
if (!lstrcmpi(matrix, "PC709"))
return PC_709;
if (!lstrcmpi(matrix, "AVERAGE"))
return AVERAGE;
if (!lstrcmpi(matrix, "rec2020"))
return Rec2020;
env->ThrowError("Convert: Unknown colormatrix");
}
return Rec601; // Default colorspace conversion for AviSynth
}
/****************************************
******* Convert to RGB / RGBA ******
***************************************/
// YUY2 only
ConvertToRGB::ConvertToRGB( PClip _child, bool rgb24, const char* matrix,
IScriptEnvironment* env )
: GenericVideoFilter(_child)
{
theMatrix = Rec601;
// no rec2020 here
if (matrix) {
if (!lstrcmpi(matrix, "rec709"))
theMatrix = Rec709;
else if (!lstrcmpi(matrix, "PC.601"))
theMatrix = PC_601;
else if (!lstrcmpi(matrix, "PC601"))
theMatrix = PC_601;
else if (!lstrcmpi(matrix, "PC.709"))
theMatrix = PC_709;
else if (!lstrcmpi(matrix, "PC709"))
theMatrix = PC_709;
else if (!lstrcmpi(matrix, "rec601"))
theMatrix = Rec601;
else
env->ThrowError("ConvertToRGB: invalid \"matrix\" parameter (must be matrix=\"Rec601\", \"Rec709\", \"PC.601\" or \"PC.709\")");
}
vi.pixel_type = rgb24 ? VideoInfo::CS_BGR24 : VideoInfo::CS_BGR32;
}
#if defined(__SSE2__)
template<int rgb_size>
static AVS_FORCEINLINE __m128i convert_yuy2_to_rgb_sse2_core(const __m128i& src_luma_scaled, const __m128i& src_chroma, const __m128i &alpha,
const __m128i& v128, const __m128i& zero, const __m128i& rounder, const __m128i &ff,
const __m128i& ymul, const __m128i& bmul, const __m128i& gmul, const __m128i& rmul) {
__m128i chroma_scaled = _mm_sub_epi16(src_chroma, v128); //V2-128 | U2-128 | V1-128 | U1-128 | V1-128 | U1-128 | V0-128 | U0-128
__m128i luma_scaled = _mm_madd_epi16(src_luma_scaled, ymul); // (y1-16)*cy | (y0-16)*cy
luma_scaled = _mm_add_epi32(luma_scaled, rounder); // (y1-16)*cy + 8192 | (y0-16)*cy + 8192
__m128i chroma_scaled2 = _mm_shuffle_epi32(chroma_scaled, _MM_SHUFFLE(2, 2, 0, 0)); //V1-128 | U1-128 | V1-128 | U1-128 | V0-128 | U0-128 | V0-128 | U0-128
chroma_scaled = _mm_add_epi16(chroma_scaled, chroma_scaled2); // V0+V1-256 | U0+U1-256 | (V0-128)*2 | (U0-128)*2
__m128i b = _mm_madd_epi16(chroma_scaled, bmul); // 0 + (U0+U1-256)*cbu | 0 + (U0-128)*2*cbu
__m128i g = _mm_madd_epi16(chroma_scaled, gmul); // (V0+V1-256)*cgv + (U0+U1-256)*cgu | (V0-128)*2*cgv + (U0-128)*2*cgu
__m128i r = _mm_madd_epi16(chroma_scaled, rmul); // (V0+V1-256)*crv + 0 | (V0-128)*2*crv + 0
b = _mm_add_epi32(luma_scaled, b);
g = _mm_add_epi32(luma_scaled, g);
r = _mm_add_epi32(luma_scaled, r);
b = _mm_srai_epi32(b, 14); //b3 b3 b3 b3 | b2 b2 b2 b2 | b1 b1 b1 b1 | b0 b0 b0 b0
g = _mm_srai_epi32(g, 14); //g3 g3 g3 g3 | g2 g2 g2 g2 | g1 g1 g1 g1 | g0 g0 g0 g0
r = _mm_srai_epi32(r, 14); //g3 g3 g3 g3 | g2 g2 g2 g2 | g1 g1 g1 g1 | g0 g0 g0 g0
if constexpr(rgb_size == 4)
{
b = _mm_max_epi16(b, zero);
g = _mm_max_epi16(g, zero);
r = _mm_max_epi16(r, zero);
b = _mm_min_epi16(b, ff); //00 00 00 b3 | 00 00 00 b2 | 00 00 00 b1 | 00 00 00 b0
g = _mm_min_epi16(g, ff); //00 00 00 g3 | 00 00 00 g2 | 00 00 00 g1 | 00 00 00 g0
r = _mm_min_epi16(r, ff); //00 00 00 r3 | 00 00 00 r2 | 00 00 00 r1 | 00 00 00 r0
r = _mm_slli_epi32(r, 16);
g = _mm_slli_epi32(g, 8);
__m128i rb = _mm_or_si128(r, b);
__m128i rgb = _mm_or_si128(rb, g);
return _mm_or_si128(rgb, alpha);
}
else
{
__m128i bg = _mm_packs_epi32(b, g); //g3g3 | g2g2 | g1g1 | g0g0 | b3b3 | b2b2 | b1b1 | b0b0
r = _mm_packs_epi32(r, zero); //0000 | 0000 | 0000 | 0000 | r3r3 | r2r2 | r1r1 | r0r0
__m128i br = _mm_unpacklo_epi16(bg, r); //r3r3 | b3b3 | r2r2 | b2b2 | r1r1 | b1b1 | r0r0 | b0b0
g = _mm_unpackhi_epi16(bg, zero); //0000 | g3g3 | 0000 | g2g2 | 0000 | g1g1 | 0000 | g0g0
__m128i rgb_lo = _mm_unpacklo_epi16(br, g); //0000 | r1r1 | g1g1 | b1b1 | 0000 | r0r0 | g0g0 | b0b0
rgb_lo = _mm_shufflelo_epi16(rgb_lo, _MM_SHUFFLE(2, 1, 0, 3)); //0000 | r1r1 | g1g1 | b1b1 | r0r0 | g0g0 | b0b0 | 0000
__m128i rgb_hi = _mm_unpackhi_epi16(br, g); //0000 | r3r3 | g3g3 | b3b3 | 0000 | r2r2 | g2g2 | b2b2
rgb_hi = _mm_shufflelo_epi16(rgb_hi, _MM_SHUFFLE(2, 1, 0, 3)); //0000 | r3r3 | g3g3 | b3b3 | r2r2 | g2g2 | b2b2 | 0000
return _mm_packus_epi16(rgb_lo, rgb_hi); //00 | r3 | g3 | b3 | r2 | g2 | b2 | 00 | 00 | r1 | g1 | b1 | r0 | g0 | b0 | 00
}
}
template<int rgb_size>
static void convert_yuy2_to_rgb_sse2(const BYTE *srcp, BYTE* dstp, int src_pitch, int dst_pitch, int height, int width, int crv, int cgv, int cgu, int cbu, int cy, int tv_scale) {
srcp += height * src_pitch;
int mod4_width = width / 4 * 4;
__m128i tv_scale_vector = _mm_set1_epi16(tv_scale);
__m128i zero = _mm_setzero_si128();
__m128i ymul = _mm_set1_epi16(cy / 4);
__m128i bmul = _mm_set_epi16(0, cbu/8, 0, cbu/8, 0, cbu/8, 0, cbu/8);
__m128i gmul = _mm_set_epi16(-cgv/8, -cgu/8, -cgv/8, -cgu/8, -cgv/8, -cgu/8, -cgv/8, -cgu/8);
__m128i rmul = _mm_set_epi16(crv/8, 0, crv/8, 0, crv/8, 0, crv/8, 0);
__m128i alpha = _mm_set1_epi32(0xFF000000);
__m128i ff = _mm_set1_epi16(0x00FF);
__m128i v128 = _mm_set1_epi16(128);
__m128i rounder = _mm_set1_epi32(1 << 13);
for (int y = 0; y < height; ++y) {
srcp -= src_pitch;
for (int x = 0; x < mod4_width-2; x+= 4) {
__m128i src = _mm_loadu_si128(reinterpret_cast<const __m128i*>(srcp+x*2)); //xx xx xx xx V2 xx U2 xx V1 Y3 U1 Y2 V0 Y1 U0 Y0
__m128i src_luma = _mm_and_si128(src, ff);//0 xx 0 xx 0 xx 0 xx 0 Y3 0 Y2 0 Y1 0 Y0
__m128i src_chroma = _mm_srli_epi16(src, 8); //00 xx 00 xx 00 V2 00 U2 00 V1 00 U1 00 V0 00 U0
src_chroma = _mm_shuffle_epi32(src_chroma, _MM_SHUFFLE(2, 1, 1, 0)); //00 V2 00 U2 00 V1 00 U1 00 V1 00 U1 00 V0 00 U0
__m128i luma_scaled = _mm_sub_epi16(src_luma, tv_scale_vector); //Y3-16 | Y2-16 | Y1-16 | Y0-16
luma_scaled = _mm_unpacklo_epi16(luma_scaled, zero); // 0000 | Y3-16 | 0000 | Y2-16 | 0000 | Y1-16 | 0000 | Y0-16
__m128i rgb = convert_yuy2_to_rgb_sse2_core<rgb_size>(luma_scaled, src_chroma, alpha, v128, zero, rounder, ff, ymul, bmul, gmul, rmul);
if constexpr(rgb_size == 4) {
_mm_store_si128(reinterpret_cast<__m128i*>(dstp+x*4), rgb);
} else {
//input: 00 | r3 | g3 | b3 | r2 | g2 | b2 | 00 | 00 | r1 | g1 | b1 | r0 | g0 | b0 | 00
rgb = _mm_srli_si128(rgb, 1); //00 | 00 | r3 | g3 | b3 | r2 | g2 | b2 || 00 | 00 | r1 | g1 | b1 | r0 | g0 | b0
*reinterpret_cast<int*>(dstp+x*3) = _mm_cvtsi128_si32(rgb);
rgb = _mm_shufflelo_epi16(rgb, _MM_SHUFFLE(2, 3, 3, 3)); //00 | 00 | r3 | g3 | b3 | r2 | g2 | b2 || r1 | g1 | 00 | 00 | 00 | 00 | 00 | 00
rgb = _mm_srli_si128(rgb, 6); //00 | 00 | 00 | 00 | 00 | 00 | 00 | 00 || r3 | g3 | b3 | r2 | g2 | b2 | r1 | g1
_mm_storel_epi64(reinterpret_cast<__m128i*>(dstp+x*3+4), rgb);
}
}
if (mod4_width == width) {
//two pixels left to process
__m128i src = _mm_cvtsi32_si128(*reinterpret_cast<const int*>(srcp+width*2 - 4)); //00 00 00 00 V0 Y1 U0 Y0
__m128i src_luma = _mm_and_si128(src, ff);//0 0 0 0 0 Y1 0 Y0
__m128i src_chroma = _mm_srli_epi16(src, 8); //00 00 00 00 00 V0 00 U0
src_chroma = _mm_shufflelo_epi16(src_chroma, _MM_SHUFFLE(1, 0, 1, 0)); //00 V0 00 U0 00 V0 00 U0
__m128i luma_scaled = _mm_sub_epi16(src_luma, tv_scale_vector); //0 | 0 | Y1-16 | Y0-16
luma_scaled = _mm_unpacklo_epi16(luma_scaled, zero); // 0000 | Y1-16 | 0000 | Y0-16
__m128i rgb = convert_yuy2_to_rgb_sse2_core<rgb_size>(luma_scaled, src_chroma, alpha, v128, zero, rounder, ff, ymul, bmul, gmul, rmul);
if constexpr(rgb_size == 4){
_mm_storel_epi64(reinterpret_cast<__m128i*>(dstp+width*4-8), rgb);
} else {
//input: 00 | 00 | 00 | 00 | 00 | 00 | 00 | 00 | 00 | r1 | g1 | b1 | r0 | g0 | b0 | 00
rgb = _mm_srli_si128(rgb, 1); //00 | 00 | 00 | 00 | 00 | 00 | 00 | 00 || 00 | 00 | r1 | g1 | b1 | r0 | g0 | b0
*reinterpret_cast<int*>(dstp+width*3-6) = _mm_cvtsi128_si32(rgb);
rgb = _mm_srli_si128(rgb, 4); //00 ... r1 | g1
*reinterpret_cast<short*>(dstp+width*3-2) = (short)_mm_cvtsi128_si32(rgb);
}
} else {
//four pixels
__m128i src = _mm_loadl_epi64(reinterpret_cast<const __m128i*>(srcp+width*2-8)); //0 0 0 0 0 0 0 0 V1 Y3 U1 Y2 V0 Y1 U0 Y0
__m128i src_luma = _mm_and_si128(src, ff);//0 0 0 0 0 0 0 0 0 Y3 0 Y2 0 Y1 0 Y0
__m128i src_chroma = _mm_srli_epi16(src, 8); //0 0 0 0 0 0 0 0 0 V1 0 U1 0 V0 0 U0
src_chroma = _mm_shuffle_epi32(src_chroma, _MM_SHUFFLE(1, 1, 1, 0)); //00 V1 00 U1 00 V1 00 U1 00 V1 00 U1 00 V0 00 U0
__m128i luma_scaled = _mm_sub_epi16(src_luma, tv_scale_vector); //Y3-16 | Y2-16 | Y1-16 | Y0-16
luma_scaled = _mm_unpacklo_epi16(luma_scaled, zero); // 0000 | Y3-16 | 0000 | Y2-16 | 0000 | Y1-16 | 0000 | Y0-16
__m128i rgb = convert_yuy2_to_rgb_sse2_core<rgb_size>(luma_scaled, src_chroma, alpha, v128, zero, rounder, ff, ymul, bmul, gmul, rmul);
if constexpr(rgb_size == 4) {
_mm_storeu_si128(reinterpret_cast<__m128i*>(dstp+width*4-16), rgb);
} else {
//input: 00 | r3 | g3 | b3 | r2 | g2 | b2 | 00 | 00 | r1 | g1 | b1 | r0 | g0 | b0 | 00
rgb = _mm_srli_si128(rgb, 1); //00 | 00 | r3 | g3 | b3 | r2 | g2 | b2 || 00 | 00 | r1 | g1 | b1 | r0 | g0 | b0
*reinterpret_cast<int*>(dstp+width*3-12) = _mm_cvtsi128_si32(rgb);
rgb = _mm_shufflelo_epi16(rgb, _MM_SHUFFLE(2, 3, 3, 3)); //00 | 00 | r3 | g3 | b3 | r2 | g2 | b2 || r1 | g1 | 00 | 00 | 00 | 00 | 00 | 00
rgb = _mm_srli_si128(rgb, 6); //00 | 00 | 00 | 00 | 00 | 00 | 00 | 00 || r3 | g3 | b3 | r2 | g2 | b2 | r1 | g1
_mm_storel_epi64(reinterpret_cast<__m128i*>(dstp+width*3-8), rgb);
}
}
dstp += dst_pitch;
}
}
#endif
#ifdef X86_32
template<int rgb_size>
static AVS_FORCEINLINE __m64 convert_yuy2_to_rgb_isse_core(const __m64& src_luma_scaled, const __m64& src_chroma, const __m64 &alpha,
const __m64& v128, const __m64& zero, const __m64& rounder, const __m64 &ff,
const __m64& ymul, const __m64& bmul, const __m64& gmul, const __m64& rmul) {
__m64 chroma_scaled = _mm_sub_pi16(src_chroma, v128); //V1-128 | U1-128 | V0-128 | U0-128
__m64 luma_scaled = _mm_madd_pi16(src_luma_scaled, ymul); // (y1-16)*cy | (y0-16)*cy
luma_scaled = _mm_add_pi32(luma_scaled, rounder); // (y1-16)*cy + 8192 | (y0-16)*cy + 8192
__m64 chroma_scaled2 = _mm_shuffle_pi16(chroma_scaled, _MM_SHUFFLE(1, 0, 1, 0)); // V0-128 | U0-128 | V0-128 | U0-128
chroma_scaled = _mm_add_pi16(chroma_scaled, chroma_scaled2); // V0+V1-256 | U0+U1-256 | (V0-128)*2 | (U0-128)*2
__m64 b = _mm_madd_pi16(chroma_scaled, bmul); // 0 + (U0+U1-256)*cbu | 0 + (U0-128)*2*cbu
__m64 g = _mm_madd_pi16(chroma_scaled, gmul); // (V0+V1-256)*cgv + (U0+U1-256)*cgu | (V0-128)*2*cgv + (U0-128)*2*cgu
__m64 r = _mm_madd_pi16(chroma_scaled, rmul); // (V0+V1-256)*crv + 0 | (V0-128)*2*crv + 0
b = _mm_add_pi32(luma_scaled, b); //b1 | b0
g = _mm_add_pi32(luma_scaled, g); //g1 | g0
r = _mm_add_pi32(luma_scaled, r); //r1 | r0
b = _mm_srai_pi32(b, 14); //BBBB | bbbb
g = _mm_srai_pi32(g, 14); //GGGG | gggg
r = _mm_srai_pi32(r, 14); //RRRR | rrrr
if (rgb_size == 4)
{
b = _mm_max_pi16(b, zero);
g = _mm_max_pi16(g, zero);
r = _mm_max_pi16(r, zero);
b = _mm_min_pi16(b, ff); //00 00 00 b1 | 00 00 00 b0
g = _mm_min_pi16(g, ff); //00 00 00 g1 | 00 00 00 g0
r = _mm_min_pi16(r, ff); //00 00 00 r1 | 00 00 00 r0
r = _mm_slli_pi32(r, 16);
g = _mm_slli_pi32(g, 8);
__m64 rb = _mm_or_si64(r, b);
__m64 rgb = _mm_or_si64(rb, g);
return _mm_or_si64(rgb, alpha);
}
else
{
__m64 bg = _mm_packs_pi32(b, g); //GGGG | gggg | BBBB | bbbb
r = _mm_packs_pi32(r, zero); //0000 | 0000 | RRRR | rrrr
__m64 br = _mm_unpacklo_pi16(bg, r); //RRRR | BBBB | rrrr | bbbb
g = _mm_unpackhi_pi16(bg, zero); //0000 | GGGG | 0000 | gggg
__m64 rgb_lo = _mm_unpacklo_pi16(br, g); //0000 | rrrr | gggg | bbbb
rgb_lo = _mm_slli_si64(rgb_lo, 16); //rrrr | gggg | bbbb | 0000
__m64 rgb_hi = _mm_unpackhi_pi16(br, g); //0000 | RRRR | GGGG | BBBB
return _mm_packs_pu16(rgb_lo, rgb_hi); //00 | RR | GG | BB | rr | gg | bb | 00
}
}
//todo: omg this thing is overcomplicated
template<int rgb_size>
static void convert_yuy2_to_rgb_isse(const BYTE *srcp, BYTE* dstp, int src_pitch, int dst_pitch, int height, int width, int crv, int cgv, int cgu, int cbu, int cy, int tv_scale) {
srcp += height * src_pitch;
int mod4_width = width / 4 * 4;
__m64 tv_scale_vector = _mm_set1_pi16(tv_scale);
__m64 zero = _mm_setzero_si64();
__m64 ymul = _mm_set1_pi16(cy / 4);
__m64 bmul = _mm_set_pi16(0, cbu/8, 0, cbu/8);
__m64 gmul = _mm_set_pi16(-cgv/8, -cgu/8, -cgv/8, -cgu/8);
__m64 rmul = _mm_set_pi16(crv/8, 0, crv/8, 0);
__m64 alpha = _mm_set1_pi32(0xFF000000);
__m64 ff = _mm_set1_pi16(0x00FF);
__m64 v128 = _mm_set1_pi16(128);
__m64 rounder = _mm_set1_pi32(1 << 13);
for (int y = 0; y < height; ++y) {
srcp -= src_pitch;
for (int x = 0; x < mod4_width-2; x+= 4) {
__m64 src = *reinterpret_cast<const __m64*>(srcp+x*2); //V1 Y3 U1 Y2 V0 Y1 U0 Y0
__m64 src2 = _mm_cvtsi32_si64(*reinterpret_cast<const int*>(srcp+x*2+8)); //00 00 00 00 V2 xx U2 xx
__m64 src_luma = _mm_and_si64(src, ff);//0 Y3 0 Y2 0 Y1 0 Y0
__m64 src_chroma = _mm_srli_pi16(src, 8); //00 V1 00 U1 00 V0 00 U0
__m64 src_chroma2 = _mm_srli_pi16(src2, 8); //00 00 00 00 00 V2 00 U2
src_chroma2 = _mm_or_si64(_mm_slli_si64(src_chroma2, 32), _mm_srli_si64(src_chroma, 32)); //00 V2 00 U2 00 V1 00 U1
__m64 luma_scaled = _mm_sub_pi16(src_luma, tv_scale_vector); //Y3-16 | Y2-16 | Y1-16 | Y0-16
__m64 luma_scaled_lo = _mm_unpacklo_pi16(luma_scaled, zero); // 0000 | Y1-16 | 0000 | Y0-16
__m64 luma_scaled_hi = _mm_unpackhi_pi16(luma_scaled, zero);
__m64 rgb_lo = convert_yuy2_to_rgb_isse_core<rgb_size>(luma_scaled_lo, src_chroma, alpha, v128, zero, rounder, ff, ymul, bmul, gmul, rmul);
__m64 rgb_hi = convert_yuy2_to_rgb_isse_core<rgb_size>(luma_scaled_hi, src_chroma2, alpha, v128, zero, rounder, ff, ymul, bmul, gmul, rmul);
if (rgb_size == 4) {
*reinterpret_cast<__m64*>(dstp+x*4) = rgb_lo;
*reinterpret_cast<__m64*>(dstp+x*4+8) = rgb_hi;
} else {
//input: 00 | RR | GG | BB | rr | gg | bb | 00
rgb_lo = _mm_srli_si64(rgb_lo, 8); //00 | 00 | RR | GG | BB | rr | gg | bb
rgb_hi = _mm_slli_si64(rgb_hi, 8); //RR | GG | BB | rr | gg | bb | 00 | 00
*reinterpret_cast<int*>(dstp+x*3) = _mm_cvtsi64_si32(rgb_lo);
rgb_lo = _mm_srli_si64(rgb_lo, 32); //00 | 00 | 00 | 00 | 00 | 00 | RR | GG
rgb_hi = _mm_or_si64(rgb_lo, rgb_hi);
*reinterpret_cast<__m64*>(dstp+x*3+4) = rgb_hi;
}
}
if (mod4_width == width) {
//two pixels left to process
__m64 src = _mm_cvtsi32_si64(*reinterpret_cast<const int*>(srcp+width*2 - 4)); //V1 Y3 U1 Y2 V0 Y1 U0 Y0
__m64 src_luma = _mm_and_si64(src, ff);//0 Y3 0 Y2 0 Y1 0 Y0
__m64 src_chroma = _mm_srli_pi16(src, 8); //00 V1 00 U1 00 V0 00 U0
src_chroma = _mm_shuffle_pi16(src_chroma, _MM_SHUFFLE(1, 0, 1, 0)); //00 V0 00 U0 00 V0 00 U0
__m64 luma_scaled = _mm_sub_pi16(src_luma, tv_scale_vector); //Y3-16 | Y2-16 | Y1-16 | Y0-16
luma_scaled = _mm_unpacklo_pi16(luma_scaled, zero); // 0000 | Y1-16 | 0000 | Y0-16
__m64 rgb = convert_yuy2_to_rgb_isse_core<rgb_size>(luma_scaled, src_chroma, alpha, v128, zero, rounder, ff, ymul, bmul, gmul, rmul);
if (rgb_size == 4){
*reinterpret_cast<__m64*>(dstp+width*4-8) = rgb;
} else {
//input: 00 | RR | GG | BB | rr | gg | bb | 00
rgb = _mm_srli_si64(rgb, 8); //00 | 00 | RR | GG | BB | rr | gg | bb
*reinterpret_cast<int*>(dstp+width*3-6) = _mm_cvtsi64_si32(rgb);
rgb = _mm_srli_si64(rgb, 32); //00 | 00 | 00 | 00 | 00 | 00 | RR | GG
*reinterpret_cast<short*>(dstp+width*3-2) = (short)_mm_cvtsi64_si32(rgb);
}
} else {
//four pixels
__m64 src = *reinterpret_cast<const __m64*>(srcp+width*2 - 8); //V1 Y3 U1 Y2 V0 Y1 U0 Y0
__m64 src_luma = _mm_and_si64(src, ff);//0 Y3 0 Y2 0 Y1 0 Y0
__m64 src_chroma = _mm_srli_pi16(src, 8); //00 V1 00 U1 00 V0 00 U0
__m64 src_chroma2 = _mm_shuffle_pi16(src_chroma, _MM_SHUFFLE(3, 2, 3, 2)); //00 V1 00 U1 00 V1 00 U1
__m64 luma_scaled = _mm_sub_pi16(src_luma, tv_scale_vector); //Y3-16 | Y2-16 | Y1-16 | Y0-16
__m64 luma_scaled_lo = _mm_unpacklo_pi16(luma_scaled, zero); // 0000 | Y1-16 | 0000 | Y0-16
__m64 luma_scaled_hi = _mm_unpackhi_pi16(luma_scaled, zero);
__m64 rgb_lo = convert_yuy2_to_rgb_isse_core<rgb_size>(luma_scaled_lo, src_chroma, alpha, v128, zero, rounder, ff, ymul, bmul, gmul, rmul);
__m64 rgb_hi = convert_yuy2_to_rgb_isse_core<rgb_size>(luma_scaled_hi, src_chroma2, alpha, v128, zero, rounder, ff, ymul, bmul, gmul, rmul);
if (rgb_size == 4) {
*reinterpret_cast<__m64*>(dstp+width*4-16) = rgb_lo;
*reinterpret_cast<__m64*>(dstp+width*4-8) = rgb_hi;
} else {
//input: 00 | RR | GG | BB | rr | gg | bb | 00
rgb_lo = _mm_srli_si64(rgb_lo, 8); //00 | 00 | RR | GG | BB | rr | gg | bb
rgb_hi = _mm_slli_si64(rgb_hi, 8); //RR | GG | BB | rr | gg | bb | 00 | 00
*reinterpret_cast<int*>(dstp+width*3-12) = _mm_cvtsi64_si32(rgb_lo);
rgb_lo = _mm_srli_si64(rgb_lo, 32); //00 | 00 | 00 | 00 | 00 | 00 | RR | GG
rgb_hi = _mm_or_si64(rgb_lo, rgb_hi);
*reinterpret_cast<__m64*>(dstp+width*3-8) = rgb_hi;
}
}
dstp += dst_pitch;
}
_mm_empty();
}
#endif // X86_32
template<int rgb_size>
static void convert_yuy2_to_rgb_c(const BYTE *srcp, BYTE* dstp, int src_pitch, int dst_pitch, int height, int width, int crv, int cgv, int cgu, int cbu, int cy, int tv_scale) {
srcp += height * src_pitch;
for (int y = height; y > 0; --y) {
srcp -= src_pitch;
int x;
for (x = 0; x < width-2; x+=2) {
int scaled_y0 = (srcp[x*2+0] - tv_scale) * cy;
int u0 = srcp[x*2+1]-128;
int v0 = srcp[x*2+3]-128;
int scaled_y1 = (srcp[x*2+2] - tv_scale) * cy;
int u1 = srcp[x*2+5]-128;
int v1 = srcp[x*2+7]-128;
dstp[x*rgb_size + 0] = ScaledPixelClip(scaled_y0 + u0 * cbu); // blue
dstp[x*rgb_size + 1] = ScaledPixelClip(scaled_y0 - u0 * cgu - v0 * cgv); // green
dstp[x*rgb_size + 2] = ScaledPixelClip(scaled_y0 + v0 * crv); // red
dstp[(x+1)*rgb_size + 0] = ScaledPixelClip(scaled_y1 + (u0+u1) * (cbu / 2)); // blue
dstp[(x+1)*rgb_size + 1] = ScaledPixelClip(scaled_y1 - (u0+u1) * (cgu / 2) - (v0+v1) * (cgv/2)); // green
dstp[(x+1)*rgb_size + 2] = ScaledPixelClip(scaled_y1 + (v0+v1) * (crv/2)); // red
if constexpr(rgb_size == 4) {
dstp[x*4+3] = 255;
dstp[x*4+7] = 255;
}
}
int scaled_y0 = (srcp[x*2+0] - tv_scale) * cy;
int scaled_y1 = (srcp[x*2+2] - tv_scale) * cy;
int u = srcp[x*2+1]-128;
int v = srcp[x*2+3]-128;
dstp[x*rgb_size + 0] = ScaledPixelClip(scaled_y0 + u * cbu); // blue
dstp[x*rgb_size + 1] = ScaledPixelClip(scaled_y0 - u * cgu - v * cgv); // green
dstp[x*rgb_size + 2] = ScaledPixelClip(scaled_y0 + v * crv); // red
dstp[(x+1)*rgb_size + 0] = ScaledPixelClip(scaled_y1 + u * cbu); // blue
dstp[(x+1)*rgb_size + 1] = ScaledPixelClip(scaled_y1 - u * cgu - v * cgv); // green
dstp[(x+1)*rgb_size + 2] = ScaledPixelClip(scaled_y1 + v * crv); // red
if constexpr(rgb_size == 4) {
dstp[x*4+3] = 255;
dstp[x*4+7] = 255;
}
dstp += dst_pitch;
}
}
// YUY2 only
PVideoFrame __stdcall ConvertToRGB::GetFrame(int n, IScriptEnvironment* env)
{
PVideoFrame src = child->GetFrame(n, env);
const int src_pitch = src->GetPitch();
const BYTE* srcp = src->GetReadPtr();
PVideoFrame dst = env->NewVideoFrame(vi);
const int dst_pitch = dst->GetPitch();
BYTE* dstp = dst->GetWritePtr();
int tv_scale = theMatrix == Rec601 || theMatrix == Rec709 ? 16 : 0;
#ifdef __SSE2__
if (env->GetCPUFlags() & CPUF_SSE2) {
if (vi.IsRGB32()) {
convert_yuy2_to_rgb_sse2<4>(srcp, dstp, src_pitch, dst_pitch, vi.height, vi.width,
crv_values[theMatrix], cgv_values[theMatrix], cgu_values[theMatrix], cbu_values[theMatrix], cy_values[theMatrix], tv_scale);
} else {
convert_yuy2_to_rgb_sse2<3>(srcp, dstp, src_pitch, dst_pitch, vi.height, vi.width,
crv_values[theMatrix], cgv_values[theMatrix], cgu_values[theMatrix], cbu_values[theMatrix], cy_values[theMatrix], tv_scale);
}
}
else
#endif
#ifdef X86_32
if (env->GetCPUFlags() & CPUF_INTEGER_SSE) {
if (vi.IsRGB32()) {
convert_yuy2_to_rgb_isse<4>(srcp, dstp, src_pitch, dst_pitch, vi.height, vi.width,
crv_values[theMatrix], cgv_values[theMatrix], cgu_values[theMatrix], cbu_values[theMatrix], cy_values[theMatrix], tv_scale);
} else {
convert_yuy2_to_rgb_isse<3>(srcp, dstp, src_pitch, dst_pitch, vi.height, vi.width,
crv_values[theMatrix], cgv_values[theMatrix], cgu_values[theMatrix], cbu_values[theMatrix], cy_values[theMatrix], tv_scale);
}
}
else
#endif
{
if (vi.IsRGB32()) {
convert_yuy2_to_rgb_c<4>(srcp, dstp, src_pitch, dst_pitch, vi.height, vi.width,
crv_values[theMatrix], cgv_values[theMatrix], cgu_values[theMatrix], cbu_values[theMatrix], cy_values[theMatrix], tv_scale);
} else {
convert_yuy2_to_rgb_c<3>(srcp, dstp, src_pitch, dst_pitch, vi.height, vi.width,
crv_values[theMatrix], cgv_values[theMatrix], cgu_values[theMatrix], cbu_values[theMatrix], cy_values[theMatrix], tv_scale);
}
}
return dst;
}
// general for all colorspaces
// however class is constructed only for YUY2 input
AVSValue __cdecl ConvertToRGB::Create(AVSValue args, void* user_data, IScriptEnvironment* env)
{
const bool haveOpts = args[3].Defined() || args[4].Defined();
PClip clip = args[0].AsClip();
const char* const matrix = args[1].AsString(0);
VideoInfo vi = clip->GetVideoInfo();
// common Create for all CreateRGB24/32/48/64/Planar(RGBP:-1, RGPAP:-2) using user_data
int target_rgbtype = (int)reinterpret_cast<intptr_t>(user_data);
// -1,-2: Planar RGB(A)
// 0: not specified (leave if input is packed RGB, convert to rgb32/64 input colorspace dependent)
// 24,32,48,64: RGB24/32/48/64
// planar YUV-like
if (vi.IsPlanar() && (vi.IsYUV() || vi.IsYUVA())) {
bool needConvertFinalBitdepth = false;
int finalBitdepth = -1;
AVSValue new_args[5] = { clip, args[2], args[1], args[3], args[4] };
// conversion to planar or packed RGB is always from 444
clip = ConvertToPlanarGeneric::CreateYUV444(AVSValue(new_args, 5), (void *)1, env).AsClip(); // (void *)1: not restricted to 8 bits
if ((target_rgbtype == 24 || target_rgbtype == 32)) {
if (vi.BitsPerComponent() != 8) {
needConvertFinalBitdepth = true;
finalBitdepth = 8;
target_rgbtype = (target_rgbtype == 24) ? -1 : -2; // planar rgb intermediate
}
}
else if ((target_rgbtype == 48 || target_rgbtype == 64)) {
if (vi.BitsPerComponent() != 16) {
needConvertFinalBitdepth = true;
finalBitdepth = 16;
target_rgbtype = (target_rgbtype == 48) ? -1 : -2; // planar rgb intermediate
}
}
else if(target_rgbtype==0 && vi.ComponentSize()==4)
env->ThrowError("ConvertToRGB: conversion is allowed only from 8 or 16 bit colorspaces");
int rgbtype_param;
bool reallyConvert = true;
switch (target_rgbtype)
{
case -1: case -2:
rgbtype_param = target_rgbtype; break; // planar RGB(A)
case 0:
rgbtype_param = vi.ComponentSize() == 1 ? 4 : 8; break; // input bitdepth adaptive
case 24:
rgbtype_param = 3; break; // RGB24
case 32:
rgbtype_param = 4; break; // RGB32
case 48: {
// instead of C code of YUV444P16->RGB48
// we convert to PlanarRGB then to RGB48 (both is fast)
AVSValue new_args2[5] = { clip, args[1], args[2], args[3], args[4] };
clip = ConvertToRGB::Create(AVSValue(new_args2, 5), (void *)-1, env).AsClip();
vi = clip->GetVideoInfo();
reallyConvert = false;
rgbtype_param = 6; // old option RGB48 target, slow C
}
break; // RGB48
case 64: {
// instead of C code of YUV(A)444P16->RGB64
// we convert to PlanarRGB(A) then to RGB64 (both is fast)
AVSValue new_args2[5] = { clip, args[1], args[2], args[3], args[4] };
clip = ConvertToRGB::Create(AVSValue(new_args2, 5), vi.IsYUVA() ? (void *)-2 : (void *)-1, env).AsClip();
vi = clip->GetVideoInfo();
reallyConvert = false;
rgbtype_param = 8; // old option RGB64 target, slow C
}
break; // RGB64
}
if (reallyConvert) {
clip = new ConvertYUV444ToRGB(clip, getMatrix(matrix, env), rgbtype_param, env);
if (needConvertFinalBitdepth) {
// from any planar rgb(a) -> rgb24/32/48/64
clip = new ConvertBits(clip, -1 /*dither_type*/, finalBitdepth /*target_bitdepth*/, true /*assume_truerange*/, true /*fulls*/, true /*fulld*/, 8 /*n/a dither_bitdepth*/, env);
vi = clip->GetVideoInfo();
// source here is always a 8/16bit planar RGB(A), finally it has to be converted to RGB24/32/48/64
const bool isRGBA = target_rgbtype == -2;
clip = new PlanarRGBtoPackedRGB(clip, isRGBA);
vi = clip->GetVideoInfo();
}
return clip;
}
}
if (haveOpts)
env->ThrowError("ConvertToRGB: ChromaPlacement and ChromaResample options are not supported.");
// planar RGB-like source
if (vi.IsPlanarRGB() || vi.IsPlanarRGBA())
{
if (target_rgbtype < 0) // planar to planar
{
if (vi.IsPlanarRGB()) {
if (target_rgbtype == -1)
return clip;
// prgb->prgba create with default alpha
return new AddAlphaPlane(clip, nullptr, 0.0f, false, env);
}
// planar rgba source
if (target_rgbtype == -2)
return clip;
return new RemoveAlphaPlane(clip, env);
}
// planar to packed 24/32/48/64
bool needConvertFinalBitdepth = false;
int finalBitdepth = -1;
if (target_rgbtype == 24 || target_rgbtype == 32) {
if (vi.BitsPerComponent() != 8) {
needConvertFinalBitdepth = true;
finalBitdepth = 8;
}
}
else if (target_rgbtype == 48 || target_rgbtype == 64) {
if (vi.BitsPerComponent() != 16) {
needConvertFinalBitdepth = true;
finalBitdepth = 16;
}
}
if (needConvertFinalBitdepth) {
// from any bitdepth planar rgb(a) -> 8/16 bits
clip = new ConvertBits(clip, -1 /*dither_type*/, finalBitdepth /*target_bitdepth*/, true /*assume_truerange*/, true /*fulls*/, true /*fulld*/, 8 /*n/a dither_bitdepth*/, env);
vi = clip->GetVideoInfo();
}
return new PlanarRGBtoPackedRGB(clip, (target_rgbtype == 32 || target_rgbtype == 64));
}
// YUY2
if (vi.IsYUV()) // at this point IsYUV means YUY2 (non-planar)
{
if (target_rgbtype == 48 || target_rgbtype == 64)
env->ThrowError("ConvertToRGB: conversion from YUY2 is allowed only to 8 bits");
if (target_rgbtype < 0) {
// rgb32 intermediate is faster
clip = new ConvertToRGB(clip, false, matrix, env); // YUY2->RGB32
return new PackedRGBtoPlanarRGB(clip, true, target_rgbtype == -2);
}
else
return new ConvertToRGB(clip, target_rgbtype == 24, matrix, env);
}
// conversions from packed RGB
if (target_rgbtype == 24 || target_rgbtype == 32) {
if (vi.ComponentSize() != 1) {
// 64->32, 48->24
clip = new ConvertBits(clip, -1 /*dither_type*/, 8 /*target_bitdepth*/, true /*assume_truerange*/, true /*fulls*/, true /*fulld*/, 8 /*n/a dither_bitdepth*/, env);
vi = clip->GetVideoInfo(); // new format
}
}
else if (target_rgbtype == 48 || target_rgbtype == 64) {
if (vi.ComponentSize() != 2) {
// 32->64, 24->48
clip = new ConvertBits(clip, -1 /*dither_type*/, 16 /*target_bitdepth*/, true /*assume_truerange*/, true /*fulls*/, true /*fulld*/, 8 /*n/a dither_bitdepth*/, env);
vi = clip->GetVideoInfo(); // new format
}
}
if(target_rgbtype==32 || target_rgbtype==64)
if (vi.IsRGB24() || vi.IsRGB48())
return new RGBtoRGBA(clip); // 24->32 or 48->64
if(target_rgbtype==24 || target_rgbtype==48)
if (vi.IsRGB32() || vi.IsRGB64())
return new RGBAtoRGB(clip); // 32->24 or 64->48
// <0: target is planar RGB(A)
if (target_rgbtype < 0) {
// RGB24/32/48/64 ->
const bool isSrcRGBA = vi.IsRGB32() || vi.IsRGB64();
const bool isTargetRGBA = target_rgbtype == -2;
return new PackedRGBtoPlanarRGB(clip, isSrcRGBA, isTargetRGBA);
}
return clip;
}
/**********************************
******* Convert to YV12 ******
*********************************/
// for YUY2->YV12 only
// all other sources use ConvertToPlanarGeneric
ConvertToYV12::ConvertToYV12(PClip _child, bool _interlaced, IScriptEnvironment* env)
: GenericVideoFilter(_child),
interlaced(_interlaced)
{
if (vi.width & 1)
env->ThrowError("ConvertToYV12: Image width must be multiple of 2");
if (interlaced && (vi.height & 3))
env->ThrowError("ConvertToYV12: Interlaced image height must be multiple of 4");
if ((!interlaced) && (vi.height & 1))
env->ThrowError("ConvertToYV12: Image height must be multiple of 2");
if (!vi.IsYUY2())
env->ThrowError("ConvertToYV12: Source must be YUY2.");
vi.pixel_type = VideoInfo::CS_YV12;
if ((env->GetCPUFlags() & CPUF_MMX) == 0)
env->ThrowError("ConvertToYV12: YV12 support require a MMX capable processor.");
}
PVideoFrame __stdcall ConvertToYV12::GetFrame(int n, IScriptEnvironment* env) {
PVideoFrame src = child->GetFrame(n, env);
PVideoFrame dst = env->NewVideoFrame(vi);
if (interlaced) {
if ((env->GetCPUFlags() & CPUF_SSE2) && IsPtrAligned(src->GetReadPtr(), 16))
{
convert_yuy2_to_yv12_interlaced_sse2(src->GetReadPtr(), src->GetRowSize(), src->GetPitch(),
dst->GetWritePtr(PLANAR_Y), dst->GetWritePtr(PLANAR_U), dst->GetWritePtr(PLANAR_V),
dst->GetPitch(PLANAR_Y), dst->GetPitch(PLANAR_U), src->GetHeight());
}
else
#ifdef X86_32
if ((env->GetCPUFlags() & CPUF_INTEGER_SSE))
{
convert_yuy2_to_yv12_interlaced_isse(src->GetReadPtr(), src->GetRowSize(), src->GetPitch(),
dst->GetWritePtr(PLANAR_Y), dst->GetWritePtr(PLANAR_U), dst->GetWritePtr(PLANAR_V),
dst->GetPitch(PLANAR_Y), dst->GetPitch(PLANAR_U), src->GetHeight());
}
else
#endif
{
convert_yuy2_to_yv12_interlaced_c(src->GetReadPtr(), src->GetRowSize(), src->GetPitch(),
dst->GetWritePtr(PLANAR_Y), dst->GetWritePtr(PLANAR_U), dst->GetWritePtr(PLANAR_V),
dst->GetPitch(PLANAR_Y), dst->GetPitch(PLANAR_U), src->GetHeight());
}
}
else
{
if ((env->GetCPUFlags() & CPUF_SSE2) && IsPtrAligned(src->GetReadPtr(), 16))
{
convert_yuy2_to_yv12_progressive_sse2(src->GetReadPtr(), src->GetRowSize(), src->GetPitch(),
dst->GetWritePtr(PLANAR_Y), dst->GetWritePtr(PLANAR_U), dst->GetWritePtr(PLANAR_V),
dst->GetPitch(PLANAR_Y), dst->GetPitch(PLANAR_U), src->GetHeight());
}
else
#ifdef X86_32
if ((env->GetCPUFlags() & CPUF_INTEGER_SSE))
{
convert_yuy2_to_yv12_progressive_isse(src->GetReadPtr(), src->GetRowSize(), src->GetPitch(),
dst->GetWritePtr(PLANAR_Y), dst->GetWritePtr(PLANAR_U), dst->GetWritePtr(PLANAR_V),
dst->GetPitch(PLANAR_Y), dst->GetPitch(PLANAR_U), src->GetHeight());
}
else
#endif
{
convert_yuy2_to_yv12_progressive_c(src->GetReadPtr(), src->GetRowSize(), src->GetPitch(),
dst->GetWritePtr(PLANAR_Y), dst->GetWritePtr(PLANAR_U), dst->GetWritePtr(PLANAR_V),
dst->GetPitch(PLANAR_Y), dst->GetPitch(PLANAR_U), src->GetHeight());
}
}
return dst;
}
AVSValue __cdecl ConvertToYV12::Create(AVSValue args, void* user_data, IScriptEnvironment* env)
{
PClip clip = args[0].AsClip();
const VideoInfo& vi = clip->GetVideoInfo();
bool only_8bit = reinterpret_cast<intptr_t>(user_data) == 0;
if (only_8bit && vi.BitsPerComponent() != 8)
env->ThrowError("ConvertToYV12: only 8 bit sources allowed");
if (vi.IsYUY2() && !args[3].Defined() && !args[4].Defined() && !args[5].Defined()) // User has not requested options, do it fast!
return new ConvertToYV12(clip,args[1].AsBool(false),env);
return ConvertToPlanarGeneric::CreateYUV420(args, NULL,env);
}
/**********************************
****** Bitdepth conversions *****
**********************************/
// 10->8
// repeated 4x for sse size 16
static const struct dither2x2_t
{
const BYTE data[4] = {
0, 2,
3, 1,
};
// cycle: 2
alignas(16) const BYTE data_sse2[2 * 16] = {
0, 2, 0, 2, 0, 2, 0, 2, 0, 2, 0, 2, 0, 2, 0, 2,
3, 1, 3, 1, 3, 1, 3, 1, 3, 1, 3, 1, 3, 1, 3, 1
};
dither2x2_t() {};
} dither2x2;
// 12->8
static const struct dither4x4_t
{
const BYTE data[16] = {
0, 8, 2, 10,
12, 4, 14, 6,
3, 11, 1, 9,
15, 7, 13, 5
};
// cycle: 4
alignas(16) const BYTE data_sse2[4 * 16] = {
0, 8, 2, 10, 0, 8, 2, 10, 0, 8, 2, 10, 0, 8, 2, 10,
12, 4, 14, 6, 12, 4, 14, 6, 12, 4, 14, 6, 12, 4, 14, 6,
3, 11, 1, 9, 3, 11, 1, 9, 3, 11, 1, 9, 3, 11, 1, 9,
15, 7, 13, 5, 15, 7, 13, 5, 15, 7, 13, 5, 15, 7, 13, 5
};
dither4x4_t() {};
} dither4x4;
// 14->8
static const struct dither8x8_t
{
const BYTE data[8][8] = {
{ 0, 32, 8, 40, 2, 34, 10, 42}, /* 8x8 Bayer ordered dithering */
{48, 16, 56, 24, 50, 18, 58, 26}, /* pattern. Each input pixel */
{12, 44, 4, 36, 14, 46, 6, 38}, /* is scaled to the 0..63 range */
{60, 28, 52, 20, 62, 30, 54, 22}, /* before looking in this table */
{ 3, 35, 11, 43, 1, 33, 9, 41}, /* to determine the action. */
{51, 19, 59, 27, 49, 17, 57, 25},
{15, 47, 7, 39, 13, 45, 5, 37},
{63, 31, 55, 23, 61, 29, 53, 21}
};
// cycle: 8
alignas(16) const BYTE data_sse2[8][16] = {
{ 0, 32, 8, 40, 2, 34, 10, 42, 0, 32, 8, 40, 2, 34, 10, 42 }, /* 8x8 Bayer ordered dithering */
{ 48, 16, 56, 24, 50, 18, 58, 26, 48, 16, 56, 24, 50, 18, 58, 26 }, /* pattern. Each input pixel */
{ 12, 44, 4, 36, 14, 46, 6, 38, 12, 44, 4, 36, 14, 46, 6, 38 }, /* is scaled to the 0..63 range */
{ 60, 28, 52, 20, 62, 30, 54, 22, 60, 28, 52, 20, 62, 30, 54, 22 }, /* before looking in this table */
{ 3, 35, 11, 43, 1, 33, 9, 41, 3, 35, 11, 43, 1, 33, 9, 41 }, /* to determine the action. */
{ 51, 19, 59, 27, 49, 17, 57, 25, 51, 19, 59, 27, 49, 17, 57, 25 },
{ 15, 47, 7, 39, 13, 45, 5, 37, 15, 47, 7, 39, 13, 45, 5, 37 },
{ 63, 31, 55, 23, 61, 29, 53, 21, 63, 31, 55, 23, 61, 29, 53, 21 }
};
dither8x8_t() {};
} dither8x8;
// 16->8
static const struct dither16x16_t
{
// cycle: 16x
alignas(16) const BYTE data[16][16] = {
{ 0,192, 48,240, 12,204, 60,252, 3,195, 51,243, 15,207, 63,255 },
{ 128, 64,176,112,140, 76,188,124,131, 67,179,115,143, 79,191,127 },
{ 32,224, 16,208, 44,236, 28,220, 35,227, 19,211, 47,239, 31,223 },
{ 160, 96,144, 80,172,108,156, 92,163, 99,147, 83,175,111,159, 95 },
{ 8,200, 56,248, 4,196, 52,244, 11,203, 59,251, 7,199, 55,247 },
{ 136, 72,184,120,132, 68,180,116,139, 75,187,123,135, 71,183,119 },
{ 40,232, 24,216, 36,228, 20,212, 43,235, 27,219, 39,231, 23,215 },
{ 168,104,152, 88,164,100,148, 84,171,107,155, 91,167,103,151, 87 },
{ 2,194, 50,242, 14,206, 62,254, 1,193, 49,241, 13,205, 61,253 },
{ 130, 66,178,114,142, 78,190,126,129, 65,177,113,141, 77,189,125 },
{ 34,226, 18,210, 46,238, 30,222, 33,225, 17,209, 45,237, 29,221 },
{ 162, 98,146, 82,174,110,158, 94,161, 97,145, 81,173,109,157, 93 },
{ 10,202, 58,250, 6,198, 54,246, 9,201, 57,249, 5,197, 53,245 },
{ 138, 74,186,122,134, 70,182,118,137, 73,185,121,133, 69,181,117 },
{ 42,234, 26,218, 38,230, 22,214, 41,233, 25,217, 37,229, 21,213 },
{ 170,106,154, 90,166,102,150, 86,169,105,153, 89,165,101,149, 85 }
};
dither16x16_t() {};
} dither16x16;
template<uint8_t sourcebits, int dither_mode, int TARGET_DITHER_BITDEPTH, int rgb_step>
static void convert_rgb_uint16_to_8_c(const BYTE *srcp, BYTE *dstp, int src_rowsize, int src_height, int src_pitch, int dst_pitch)
{
const uint16_t *srcp0 = reinterpret_cast<const uint16_t *>(srcp);
src_pitch = src_pitch / sizeof(uint16_t);
int src_width = src_rowsize / sizeof(uint16_t);
int _y = 0; // for ordered dither
const int TARGET_BITDEPTH = 8; // here is constant (uint8_t target)
// for test, make it 2,4,6,8. sourcebits-TARGET_DITHER_BITDEPTH cannot exceed 8 bit
// const int TARGET_DITHER_BITDEPTH = 2;
const int max_pixel_value = (1 << TARGET_BITDEPTH) - 1;
const int max_pixel_value_dithered = (1 << TARGET_DITHER_BITDEPTH) - 1;
// precheck ensures:
// TARGET_BITDEPTH >= TARGET_DITHER_BITDEPTH
// sourcebits - TARGET_DITHER_BITDEPTH <= 8
// sourcebits - TARGET_DITHER_BITDEPTH is even (later we can use PRESHIFT)
const int DITHER_BIT_DIFF = (sourcebits - TARGET_DITHER_BITDEPTH); // 2, 4, 6, 8
const int PRESHIFT = DITHER_BIT_DIFF & 1; // 0 or 1: correction for odd bit differences (not used here but generality)
const int DITHER_ORDER = (DITHER_BIT_DIFF + PRESHIFT) / 2;
const int DITHER_SIZE = 1 << DITHER_ORDER; // 9,10=2 11,12=4 13,14=8 15,16=16
const int MASK = DITHER_SIZE - 1;
// 10->8: 0x01 (2x2)
// 11->8: 0x03 (4x4)
// 12->8: 0x03 (4x4)
// 14->8: 0x07 (8x8)
// 16->8: 0x0F (16x16)
const BYTE *matrix;