pixon 0.8.0

SIMD-dispatched color conversion for the FFmpeg AVPixelFormat space: one-call decode of validated source frames to RGB / RGBA / Luma / HSV, fused resampling while converting, colorimetry-honest matrices and chroma siting, no_std-friendly.
Documentation
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
1001
1002
1003
1004
1005
1006
1007
1008
1009
1010
1011
1012
1013
1014
1015
1016
1017
1018
1019
1020
1021
1022
1023
1024
1025
1026
1027
1028
1029
1030
1031
1032
1033
1034
1035
1036
1037
1038
1039
1040
1041
1042
1043
1044
1045
//! V410 (Tier 5 packed YUV 4:4:4 10-bit, one u32 word per pixel)
//! dispatchers (Ship 12a).
//!
//! Six entries: `v410_to_{rgb,rgba}_row` (u8) and the matching
//! `_u16` variants for native-depth output, plus
//! `v410_to_luma_row` / `v410_to_luma_u16_row` for direct luma
//! extraction. Routes through the standard `cfg_select!` per-arch
//! block; `use_simd = false` forces scalar.
//!
//! V410 is 4:4:4 (no chroma subsampling): each u32 word encodes one
//! complete pixel as 10-bit U / Y / V packed into bits [9:0] / [19:10]
//! / [29:20] with 2-bit padding at the top. Buffer length is `width`
//! u32 elements — no even-width restriction, no width×2 scaling.
//!
//! `be_input = true` selects the big-endian wire variant: each u32 word
//! is byte-swapped before unpacking, matching QuickTime-style BE V410
//! streams.

#[cfg(any(
  target_arch = "aarch64",
  target_arch = "x86_64",
  target_arch = "wasm32"
))]
use crate::row::arch;
#[cfg(target_arch = "aarch64")]
use crate::row::neon_available;
#[cfg(target_arch = "wasm32")]
use crate::row::simd128_available;
#[cfg(target_arch = "x86_64")]
use crate::row::{avx2_available, avx512_available, sse41_available};
use crate::{
  KernelMatrix,
  row::{rgb_row_bytes, rgb_row_elems, rgba_row_bytes, rgba_row_elems, scalar},
};

/// Converts one row of V410 to packed RGB (u8). See
/// [`scalar::v410_to_rgb_or_rgba_row`] for word layout / numerical
/// contract. `use_simd = false` forces scalar. `be_input = true` selects
/// the big-endian wire variant.
#[cfg_attr(not(tarpaulin), inline(always))]
pub fn v410_to_rgb_row(
  packed: &[u32],
  rgb_out: &mut [u8],
  width: usize,
  matrix: KernelMatrix,
  full_range: bool,
  use_simd: bool,
  be_input: bool,
) {
  assert!(packed.len() >= width, "packed row too short");
  assert!(
    rgb_out.len() >= rgb_row_bytes(width),
    "rgb_out row too short"
  );

  if use_simd {
    cfg_select! {
      target_arch = "aarch64" => {
        if neon_available() {
          // SAFETY: NEON verified at runtime.
          if be_input {
            unsafe {
              arch::neon::v410_to_rgb_or_rgba_row::<false, true>(
                packed, rgb_out, width, matrix, full_range,
              );
            }
          } else {
            unsafe {
              arch::neon::v410_to_rgb_or_rgba_row::<false, false>(
                packed, rgb_out, width, matrix, full_range,
              );
            }
          }
          return;
        }
      }
      target_arch = "x86_64" => {
        if avx512_available() {
          // SAFETY: AVX-512BW verified.
          if be_input {
            unsafe {
              arch::x86_avx512::v410_to_rgb_or_rgba_row::<false, true>(
                packed, rgb_out, width, matrix, full_range,
              );
            }
          } else {
            unsafe {
              arch::x86_avx512::v410_to_rgb_or_rgba_row::<false, false>(
                packed, rgb_out, width, matrix, full_range,
              );
            }
          }
          return;
        }
        if avx2_available() {
          // SAFETY: AVX2 verified.
          if be_input {
            unsafe {
              arch::x86_avx2::v410_to_rgb_or_rgba_row::<false, true>(
                packed, rgb_out, width, matrix, full_range,
              );
            }
          } else {
            unsafe {
              arch::x86_avx2::v410_to_rgb_or_rgba_row::<false, false>(
                packed, rgb_out, width, matrix, full_range,
              );
            }
          }
          return;
        }
        if sse41_available() {
          // SAFETY: SSE4.1 verified.
          if be_input {
            unsafe {
              arch::x86_sse41::v410_to_rgb_or_rgba_row::<false, true>(
                packed, rgb_out, width, matrix, full_range,
              );
            }
          } else {
            unsafe {
              arch::x86_sse41::v410_to_rgb_or_rgba_row::<false, false>(
                packed, rgb_out, width, matrix, full_range,
              );
            }
          }
          return;
        }
      }
      target_arch = "wasm32" => {
        if simd128_available() {
          // SAFETY: simd128 compile-time verified.
          if be_input {
            unsafe {
              arch::wasm_simd128::v410_to_rgb_or_rgba_row::<false, true>(
                packed, rgb_out, width, matrix, full_range,
              );
            }
          } else {
            unsafe {
              arch::wasm_simd128::v410_to_rgb_or_rgba_row::<false, false>(
                packed, rgb_out, width, matrix, full_range,
              );
            }
          }
          return;
        }
      }
      _ => {}
    }
  }

  if be_input {
    scalar::v410_to_rgb_or_rgba_row::<false, true>(packed, rgb_out, width, matrix, full_range);
  } else {
    scalar::v410_to_rgb_or_rgba_row::<false, false>(packed, rgb_out, width, matrix, full_range);
  }
}

/// Converts one row of V410 to packed RGBA (u8) with `α = 0xFF`.
/// `be_input = true` selects the big-endian wire variant.
#[cfg_attr(not(tarpaulin), inline(always))]
pub fn v410_to_rgba_row(
  packed: &[u32],
  rgba_out: &mut [u8],
  width: usize,
  matrix: KernelMatrix,
  full_range: bool,
  use_simd: bool,
  be_input: bool,
) {
  assert!(packed.len() >= width, "packed row too short");
  assert!(
    rgba_out.len() >= rgba_row_bytes(width),
    "rgba_out row too short"
  );

  if use_simd {
    cfg_select! {
      target_arch = "aarch64" => {
        if neon_available() {
          // SAFETY: NEON verified.
          if be_input {
            unsafe {
              arch::neon::v410_to_rgb_or_rgba_row::<true, true>(
                packed, rgba_out, width, matrix, full_range,
              );
            }
          } else {
            unsafe {
              arch::neon::v410_to_rgb_or_rgba_row::<true, false>(
                packed, rgba_out, width, matrix, full_range,
              );
            }
          }
          return;
        }
      }
      target_arch = "x86_64" => {
        if avx512_available() {
          // SAFETY: AVX-512BW verified.
          if be_input {
            unsafe {
              arch::x86_avx512::v410_to_rgb_or_rgba_row::<true, true>(
                packed, rgba_out, width, matrix, full_range,
              );
            }
          } else {
            unsafe {
              arch::x86_avx512::v410_to_rgb_or_rgba_row::<true, false>(
                packed, rgba_out, width, matrix, full_range,
              );
            }
          }
          return;
        }
        if avx2_available() {
          // SAFETY: AVX2 verified.
          if be_input {
            unsafe {
              arch::x86_avx2::v410_to_rgb_or_rgba_row::<true, true>(
                packed, rgba_out, width, matrix, full_range,
              );
            }
          } else {
            unsafe {
              arch::x86_avx2::v410_to_rgb_or_rgba_row::<true, false>(
                packed, rgba_out, width, matrix, full_range,
              );
            }
          }
          return;
        }
        if sse41_available() {
          // SAFETY: SSE4.1 verified.
          if be_input {
            unsafe {
              arch::x86_sse41::v410_to_rgb_or_rgba_row::<true, true>(
                packed, rgba_out, width, matrix, full_range,
              );
            }
          } else {
            unsafe {
              arch::x86_sse41::v410_to_rgb_or_rgba_row::<true, false>(
                packed, rgba_out, width, matrix, full_range,
              );
            }
          }
          return;
        }
      }
      target_arch = "wasm32" => {
        if simd128_available() {
          // SAFETY: simd128 compile-time verified.
          if be_input {
            unsafe {
              arch::wasm_simd128::v410_to_rgb_or_rgba_row::<true, true>(
                packed, rgba_out, width, matrix, full_range,
              );
            }
          } else {
            unsafe {
              arch::wasm_simd128::v410_to_rgb_or_rgba_row::<true, false>(
                packed, rgba_out, width, matrix, full_range,
              );
            }
          }
          return;
        }
      }
      _ => {}
    }
  }

  if be_input {
    scalar::v410_to_rgb_or_rgba_row::<true, true>(packed, rgba_out, width, matrix, full_range);
  } else {
    scalar::v410_to_rgb_or_rgba_row::<true, false>(packed, rgba_out, width, matrix, full_range);
  }
}

/// Converts one row of V410 to packed `u16` RGB at native 10-bit
/// depth (low-bit-packed, `[0, 1023]`). `be_input = true` selects
/// the big-endian wire variant.
#[cfg_attr(not(tarpaulin), inline(always))]
pub fn v410_to_rgb_u16_row(
  packed: &[u32],
  rgb_out: &mut [u16],
  width: usize,
  matrix: KernelMatrix,
  full_range: bool,
  use_simd: bool,
  be_input: bool,
) {
  assert!(packed.len() >= width, "packed row too short");
  assert!(
    rgb_out.len() >= rgb_row_elems(width),
    "rgb_out row too short"
  );

  if use_simd {
    cfg_select! {
      target_arch = "aarch64" => {
        if neon_available() {
          // SAFETY: NEON verified.
          if be_input {
            unsafe {
              arch::neon::v410_to_rgb_u16_or_rgba_u16_row::<false, true>(
                packed, rgb_out, width, matrix, full_range,
              );
            }
          } else {
            unsafe {
              arch::neon::v410_to_rgb_u16_or_rgba_u16_row::<false, false>(
                packed, rgb_out, width, matrix, full_range,
              );
            }
          }
          return;
        }
      }
      target_arch = "x86_64" => {
        if avx512_available() {
          // SAFETY: AVX-512BW verified.
          if be_input {
            unsafe {
              arch::x86_avx512::v410_to_rgb_u16_or_rgba_u16_row::<false, true>(
                packed, rgb_out, width, matrix, full_range,
              );
            }
          } else {
            unsafe {
              arch::x86_avx512::v410_to_rgb_u16_or_rgba_u16_row::<false, false>(
                packed, rgb_out, width, matrix, full_range,
              );
            }
          }
          return;
        }
        if avx2_available() {
          // SAFETY: AVX2 verified.
          if be_input {
            unsafe {
              arch::x86_avx2::v410_to_rgb_u16_or_rgba_u16_row::<false, true>(
                packed, rgb_out, width, matrix, full_range,
              );
            }
          } else {
            unsafe {
              arch::x86_avx2::v410_to_rgb_u16_or_rgba_u16_row::<false, false>(
                packed, rgb_out, width, matrix, full_range,
              );
            }
          }
          return;
        }
        if sse41_available() {
          // SAFETY: SSE4.1 verified.
          if be_input {
            unsafe {
              arch::x86_sse41::v410_to_rgb_u16_or_rgba_u16_row::<false, true>(
                packed, rgb_out, width, matrix, full_range,
              );
            }
          } else {
            unsafe {
              arch::x86_sse41::v410_to_rgb_u16_or_rgba_u16_row::<false, false>(
                packed, rgb_out, width, matrix, full_range,
              );
            }
          }
          return;
        }
      }
      target_arch = "wasm32" => {
        if simd128_available() {
          // SAFETY: simd128 compile-time verified.
          if be_input {
            unsafe {
              arch::wasm_simd128::v410_to_rgb_u16_or_rgba_u16_row::<false, true>(
                packed, rgb_out, width, matrix, full_range,
              );
            }
          } else {
            unsafe {
              arch::wasm_simd128::v410_to_rgb_u16_or_rgba_u16_row::<false, false>(
                packed, rgb_out, width, matrix, full_range,
              );
            }
          }
          return;
        }
      }
      _ => {}
    }
  }

  if be_input {
    scalar::v410_to_rgb_u16_or_rgba_u16_row::<false, true>(
      packed, rgb_out, width, matrix, full_range,
    );
  } else {
    scalar::v410_to_rgb_u16_or_rgba_u16_row::<false, false>(
      packed, rgb_out, width, matrix, full_range,
    );
  }
}

/// Converts one row of V410 to packed `u16` RGBA at native 10-bit
/// depth with `α = 1023` (10-bit opaque maximum). `be_input = true`
/// selects the big-endian wire variant.
#[cfg_attr(not(tarpaulin), inline(always))]
pub fn v410_to_rgba_u16_row(
  packed: &[u32],
  rgba_out: &mut [u16],
  width: usize,
  matrix: KernelMatrix,
  full_range: bool,
  use_simd: bool,
  be_input: bool,
) {
  assert!(packed.len() >= width, "packed row too short");
  assert!(
    rgba_out.len() >= rgba_row_elems(width),
    "rgba_out row too short"
  );

  if use_simd {
    cfg_select! {
      target_arch = "aarch64" => {
        if neon_available() {
          // SAFETY: NEON verified.
          if be_input {
            unsafe {
              arch::neon::v410_to_rgb_u16_or_rgba_u16_row::<true, true>(
                packed, rgba_out, width, matrix, full_range,
              );
            }
          } else {
            unsafe {
              arch::neon::v410_to_rgb_u16_or_rgba_u16_row::<true, false>(
                packed, rgba_out, width, matrix, full_range,
              );
            }
          }
          return;
        }
      }
      target_arch = "x86_64" => {
        if avx512_available() {
          // SAFETY: AVX-512BW verified.
          if be_input {
            unsafe {
              arch::x86_avx512::v410_to_rgb_u16_or_rgba_u16_row::<true, true>(
                packed, rgba_out, width, matrix, full_range,
              );
            }
          } else {
            unsafe {
              arch::x86_avx512::v410_to_rgb_u16_or_rgba_u16_row::<true, false>(
                packed, rgba_out, width, matrix, full_range,
              );
            }
          }
          return;
        }
        if avx2_available() {
          // SAFETY: AVX2 verified.
          if be_input {
            unsafe {
              arch::x86_avx2::v410_to_rgb_u16_or_rgba_u16_row::<true, true>(
                packed, rgba_out, width, matrix, full_range,
              );
            }
          } else {
            unsafe {
              arch::x86_avx2::v410_to_rgb_u16_or_rgba_u16_row::<true, false>(
                packed, rgba_out, width, matrix, full_range,
              );
            }
          }
          return;
        }
        if sse41_available() {
          // SAFETY: SSE4.1 verified.
          if be_input {
            unsafe {
              arch::x86_sse41::v410_to_rgb_u16_or_rgba_u16_row::<true, true>(
                packed, rgba_out, width, matrix, full_range,
              );
            }
          } else {
            unsafe {
              arch::x86_sse41::v410_to_rgb_u16_or_rgba_u16_row::<true, false>(
                packed, rgba_out, width, matrix, full_range,
              );
            }
          }
          return;
        }
      }
      target_arch = "wasm32" => {
        if simd128_available() {
          // SAFETY: simd128 compile-time verified.
          if be_input {
            unsafe {
              arch::wasm_simd128::v410_to_rgb_u16_or_rgba_u16_row::<true, true>(
                packed, rgba_out, width, matrix, full_range,
              );
            }
          } else {
            unsafe {
              arch::wasm_simd128::v410_to_rgb_u16_or_rgba_u16_row::<true, false>(
                packed, rgba_out, width, matrix, full_range,
              );
            }
          }
          return;
        }
      }
      _ => {}
    }
  }

  if be_input {
    scalar::v410_to_rgb_u16_or_rgba_u16_row::<true, true>(
      packed, rgba_out, width, matrix, full_range,
    );
  } else {
    scalar::v410_to_rgb_u16_or_rgba_u16_row::<true, false>(
      packed, rgba_out, width, matrix, full_range,
    );
  }
}

/// Extracts one row of 8-bit luma from a packed V410 buffer.
/// Y values are downshifted from 10-bit to 8-bit via `>> 2`.
/// `be_input = true` selects the big-endian wire variant.
#[cfg_attr(not(tarpaulin), inline(always))]
pub fn v410_to_luma_row(
  packed: &[u32],
  luma_out: &mut [u8],
  width: usize,
  use_simd: bool,
  be_input: bool,
) {
  assert!(packed.len() >= width, "packed row too short");
  assert!(luma_out.len() >= width, "luma_out row too short");

  if use_simd {
    cfg_select! {
      target_arch = "aarch64" => {
        if neon_available() {
          // SAFETY: NEON verified.
          if be_input {
            unsafe {
              arch::neon::v410_to_luma_row::<true>(packed, luma_out, width);
            }
          } else {
            unsafe {
              arch::neon::v410_to_luma_row::<false>(packed, luma_out, width);
            }
          }
          return;
        }
      }
      target_arch = "x86_64" => {
        if avx512_available() {
          // SAFETY: AVX-512BW verified.
          if be_input {
            unsafe {
              arch::x86_avx512::v410_to_luma_row::<true>(packed, luma_out, width);
            }
          } else {
            unsafe {
              arch::x86_avx512::v410_to_luma_row::<false>(packed, luma_out, width);
            }
          }
          return;
        }
        if avx2_available() {
          // SAFETY: AVX2 verified.
          if be_input {
            unsafe {
              arch::x86_avx2::v410_to_luma_row::<true>(packed, luma_out, width);
            }
          } else {
            unsafe {
              arch::x86_avx2::v410_to_luma_row::<false>(packed, luma_out, width);
            }
          }
          return;
        }
        if sse41_available() {
          // SAFETY: SSE4.1 verified.
          if be_input {
            unsafe {
              arch::x86_sse41::v410_to_luma_row::<true>(packed, luma_out, width);
            }
          } else {
            unsafe {
              arch::x86_sse41::v410_to_luma_row::<false>(packed, luma_out, width);
            }
          }
          return;
        }
      }
      target_arch = "wasm32" => {
        if simd128_available() {
          // SAFETY: simd128 compile-time verified.
          if be_input {
            unsafe {
              arch::wasm_simd128::v410_to_luma_row::<true>(packed, luma_out, width);
            }
          } else {
            unsafe {
              arch::wasm_simd128::v410_to_luma_row::<false>(packed, luma_out, width);
            }
          }
          return;
        }
      }
      _ => {}
    }
  }

  if be_input {
    scalar::v410_to_luma_row::<true>(packed, luma_out, width);
  } else {
    scalar::v410_to_luma_row::<false>(packed, luma_out, width);
  }
}

/// Extracts one row of native-depth `u16` luma from a packed V410
/// buffer (low-bit-packed: each `u16` carries the 10-bit Y value in
/// its low 10 bits). `be_input = true` selects the big-endian wire
/// variant.
#[cfg_attr(not(tarpaulin), inline(always))]
pub fn v410_to_luma_u16_row(
  packed: &[u32],
  luma_out: &mut [u16],
  width: usize,
  use_simd: bool,
  be_input: bool,
) {
  assert!(packed.len() >= width, "packed row too short");
  assert!(luma_out.len() >= width, "luma_out row too short");

  if use_simd {
    cfg_select! {
      target_arch = "aarch64" => {
        if neon_available() {
          // SAFETY: NEON verified.
          if be_input {
            unsafe {
              arch::neon::v410_to_luma_u16_row::<true>(packed, luma_out, width);
            }
          } else {
            unsafe {
              arch::neon::v410_to_luma_u16_row::<false>(packed, luma_out, width);
            }
          }
          return;
        }
      }
      target_arch = "x86_64" => {
        if avx512_available() {
          // SAFETY: AVX-512BW verified.
          if be_input {
            unsafe {
              arch::x86_avx512::v410_to_luma_u16_row::<true>(packed, luma_out, width);
            }
          } else {
            unsafe {
              arch::x86_avx512::v410_to_luma_u16_row::<false>(packed, luma_out, width);
            }
          }
          return;
        }
        if avx2_available() {
          // SAFETY: AVX2 verified.
          if be_input {
            unsafe {
              arch::x86_avx2::v410_to_luma_u16_row::<true>(packed, luma_out, width);
            }
          } else {
            unsafe {
              arch::x86_avx2::v410_to_luma_u16_row::<false>(packed, luma_out, width);
            }
          }
          return;
        }
        if sse41_available() {
          // SAFETY: SSE4.1 verified.
          if be_input {
            unsafe {
              arch::x86_sse41::v410_to_luma_u16_row::<true>(packed, luma_out, width);
            }
          } else {
            unsafe {
              arch::x86_sse41::v410_to_luma_u16_row::<false>(packed, luma_out, width);
            }
          }
          return;
        }
      }
      target_arch = "wasm32" => {
        if simd128_available() {
          // SAFETY: simd128 compile-time verified.
          if be_input {
            unsafe {
              arch::wasm_simd128::v410_to_luma_u16_row::<true>(packed, luma_out, width);
            }
          } else {
            unsafe {
              arch::wasm_simd128::v410_to_luma_u16_row::<false>(packed, luma_out, width);
            }
          }
          return;
        }
      }
      _ => {}
    }
  }

  if be_input {
    scalar::v410_to_luma_u16_row::<true>(packed, luma_out, width);
  } else {
    scalar::v410_to_luma_u16_row::<false>(packed, luma_out, width);
  }
}

/// Converts one row of V410 **directly** to planar HSV bytes (OpenCV
/// `cv2.COLOR_RGB2HSV` encoding: `H ∈ [0, 179]`, `S, V ∈ [0, 255]`), without
/// materializing a source-width RGB row. Byte-identical to
/// `rgb_to_hsv_row(v410_to_rgb_row(...))` within the selected tier — the
/// SIMD path stages a fixed 8-bit RGB chunk internally. See
/// `scalar::v410_to_hsv_row` for the reference. `use_simd = false` forces
/// the scalar path; `be_input = true` selects the big-endian wire variant.
#[cfg_attr(not(tarpaulin), inline(always))]
#[allow(clippy::too_many_arguments)]
pub fn v410_to_hsv_row(
  packed: &[u32],
  h_out: &mut [u8],
  s_out: &mut [u8],
  v_out: &mut [u8],
  width: usize,
  matrix: KernelMatrix,
  full_range: bool,
  use_simd: bool,
  be_input: bool,
) {
  assert!(packed.len() >= width, "packed row too short");
  assert!(h_out.len() >= width, "h_out row too short");
  assert!(s_out.len() >= width, "s_out row too short");
  assert!(v_out.len() >= width, "v_out row too short");

  if use_simd {
    cfg_select! {
      target_arch = "aarch64" => {
        if neon_available() {
          // SAFETY: NEON verified at runtime.
          if be_input {
            unsafe {
              arch::neon::v410_to_hsv_row::<true>(
                packed, h_out, s_out, v_out, width, matrix, full_range,
              );
            }
          } else {
            unsafe {
              arch::neon::v410_to_hsv_row::<false>(
                packed, h_out, s_out, v_out, width, matrix, full_range,
              );
            }
          }
          return;
        }
      }
      target_arch = "x86_64" => {
        if avx512_available() {
          // SAFETY: AVX-512BW verified.
          if be_input {
            unsafe {
              arch::x86_avx512::v410_to_hsv_row::<true>(
                packed, h_out, s_out, v_out, width, matrix, full_range,
              );
            }
          } else {
            unsafe {
              arch::x86_avx512::v410_to_hsv_row::<false>(
                packed, h_out, s_out, v_out, width, matrix, full_range,
              );
            }
          }
          return;
        }
        if avx2_available() {
          // SAFETY: AVX2 verified.
          if be_input {
            unsafe {
              arch::x86_avx2::v410_to_hsv_row::<true>(
                packed, h_out, s_out, v_out, width, matrix, full_range,
              );
            }
          } else {
            unsafe {
              arch::x86_avx2::v410_to_hsv_row::<false>(
                packed, h_out, s_out, v_out, width, matrix, full_range,
              );
            }
          }
          return;
        }
        if sse41_available() {
          // SAFETY: SSE4.1 verified.
          if be_input {
            unsafe {
              arch::x86_sse41::v410_to_hsv_row::<true>(
                packed, h_out, s_out, v_out, width, matrix, full_range,
              );
            }
          } else {
            unsafe {
              arch::x86_sse41::v410_to_hsv_row::<false>(
                packed, h_out, s_out, v_out, width, matrix, full_range,
              );
            }
          }
          return;
        }
      }
      target_arch = "wasm32" => {
        if simd128_available() {
          // SAFETY: simd128 compile-time verified.
          if be_input {
            unsafe {
              arch::wasm_simd128::v410_to_hsv_row::<true>(
                packed, h_out, s_out, v_out, width, matrix, full_range,
              );
            }
          } else {
            unsafe {
              arch::wasm_simd128::v410_to_hsv_row::<false>(
                packed, h_out, s_out, v_out, width, matrix, full_range,
              );
            }
          }
          return;
        }
      }
      _ => {}
    }
  }

  if be_input {
    scalar::v410_to_hsv_row::<true>(packed, h_out, s_out, v_out, width, matrix, full_range);
  } else {
    scalar::v410_to_hsv_row::<false>(packed, h_out, s_out, v_out, width, matrix, full_range);
  }
}

#[cfg(all(test, feature = "std"))]
mod tests {
  //! Smoke tests for the public V410 dispatchers. Walker / kernel
  //! correctness lives in the per-arch tests and the scalar reference's
  //! own inline tests; this block verifies the dispatcher correctly
  //! reaches its scalar fallback when SIMD is disabled and panics on
  //! invalid inputs.
  use super::*;

  /// Pack one V410 word (host-native u32) from explicit U / Y / V samples
  /// (10-bit each).
  fn pack_v410(u: u32, y: u32, v: u32) -> u32 {
    debug_assert!(u < 1024 && y < 1024 && v < 1024);
    (v << 20) | (y << 10) | u
  }

  /// Build a `Vec<u32>` V410 row of `width` pixels with `(U, Y, V)` repeated,
  /// in LE-encoded byte form so dispatchers with `be_input = false` recover
  /// the intended logical values on both LE and BE hosts. Any positive width
  /// is valid (4:4:4, no chroma subsampling).
  fn solid_v410(width: usize, u: u32, y: u32, v: u32) -> std::vec::Vec<u32> {
    (0..width)
      .map(|_| u32::from_ne_bytes(pack_v410(u, y, v).to_le_bytes()))
      .collect()
  }

  /// Build a `Vec<u32>` V410 row in BE-encoded byte form so dispatchers with
  /// `be_input = true` recover the intended logical values on both LE and
  /// BE hosts.
  fn solid_v410_be(width: usize, u: u32, y: u32, v: u32) -> std::vec::Vec<u32> {
    (0..width)
      .map(|_| u32::from_ne_bytes(pack_v410(u, y, v).to_be_bytes()))
      .collect()
  }

  #[test]
  #[should_panic(expected = "packed row too short")]
  fn v410_dispatcher_rejects_short_packed() {
    // packed buffer has only 2 elements for width=4 (needs 4).
    let packed = [0u32; 2];
    let mut rgb = [0u8; 4 * 3];
    v410_to_rgb_row(
      &packed,
      &mut rgb,
      4,
      KernelMatrix::Bt709,
      true,
      false,
      false,
    );
  }

  #[test]
  #[should_panic(expected = "rgb_out row too short")]
  fn v410_dispatcher_rejects_short_output() {
    // output buffer has only 2 bytes for width=4 (needs 12).
    let packed = [0u32; 4];
    let mut rgb = [0u8; 2];
    v410_to_rgb_row(
      &packed,
      &mut rgb,
      4,
      KernelMatrix::Bt709,
      true,
      false,
      false,
    );
  }

  #[test]
  fn v410_dispatchers_route_with_simd_false() {
    // Full-range gray (Y=512, U=V=512 at 10-bit). Every dispatcher
    // should reach its scalar fallback when `use_simd = false`,
    // produce the documented gray output, and not panic.
    let buf = solid_v410(8, 512, 512, 512);

    // u8 RGB
    let mut rgb = [0u8; 8 * 3];
    v410_to_rgb_row(&buf, &mut rgb, 8, KernelMatrix::Bt709, true, false, false);
    for px in rgb.chunks(3) {
      assert!(px[0].abs_diff(128) <= 1);
      assert_eq!(px[0], px[1]);
      assert_eq!(px[1], px[2]);
    }

    // u8 RGBA — alpha = 0xFF
    let mut rgba = [0u8; 8 * 4];
    v410_to_rgba_row(&buf, &mut rgba, 8, KernelMatrix::Bt709, true, false, false);
    for px in rgba.chunks(4) {
      assert!(px[0].abs_diff(128) <= 1);
      assert_eq!(px[3], 0xFF);
    }

    // u16 RGB at native 10-bit depth.
    let mut rgb_u16 = [0u16; 8 * 3];
    v410_to_rgb_u16_row(
      &buf,
      &mut rgb_u16,
      8,
      KernelMatrix::Bt709,
      true,
      false,
      false,
    );
    for px in rgb_u16.chunks(3) {
      assert!(px[0].abs_diff(512) <= 2);
      assert_eq!(px[0], px[1]);
      assert_eq!(px[1], px[2]);
    }

    // u16 RGBA — alpha = 1023 (10-bit opaque maximum).
    let mut rgba_u16 = [0u16; 8 * 4];
    v410_to_rgba_u16_row(
      &buf,
      &mut rgba_u16,
      8,
      KernelMatrix::Bt709,
      true,
      false,
      false,
    );
    for px in rgba_u16.chunks(4) {
      assert_eq!(px[3], 1023);
    }

    // u8 luma — Y=512 → 128 after `>> 2`.
    let mut luma = [0u8; 8];
    v410_to_luma_row(&buf, &mut luma, 8, false, false);
    for &y in &luma {
      assert_eq!(y, (512u32 >> 2) as u8);
    }

    // u16 luma — low-packed 10-bit Y value.
    let mut luma_u16 = [0u16; 8];
    v410_to_luma_u16_row(&buf, &mut luma_u16, 8, false, false);
    for &y in &luma_u16 {
      assert_eq!(y, 512);
    }
  }

  #[test]
  fn v410_be_and_le_dispatchers_agree() {
    // BE-encoded data decoded with be_input=true must produce the same
    // output as LE-encoded data decoded with be_input=false.
    let le_buf = solid_v410(8, 512, 512, 512);
    let be_buf = solid_v410_be(8, 512, 512, 512);

    // u8 RGB
    let mut rgb_le = [0u8; 8 * 3];
    let mut rgb_be = [0u8; 8 * 3];
    v410_to_rgb_row(
      &le_buf,
      &mut rgb_le,
      8,
      KernelMatrix::Bt709,
      true,
      false,
      false,
    );
    v410_to_rgb_row(
      &be_buf,
      &mut rgb_be,
      8,
      KernelMatrix::Bt709,
      true,
      false,
      true,
    );
    assert_eq!(
      rgb_le, rgb_be,
      "LE and BE must produce identical RGB output"
    );

    // u8 luma
    let mut luma_le = [0u8; 8];
    let mut luma_be = [0u8; 8];
    v410_to_luma_row(&le_buf, &mut luma_le, 8, false, false);
    v410_to_luma_row(&be_buf, &mut luma_be, 8, false, true);
    assert_eq!(
      luma_le, luma_be,
      "LE and BE must produce identical luma output"
    );
  }
}