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
1046
1047
1048
1049
1050
1051
1052
1053
1054
1055
1056
1057
1058
1059
1060
1061
1062
1063
1064
1065
1066
1067
1068
1069
1070
1071
1072
1073
1074
1075
1076
1077
1078
1079
1080
1081
1082
1083
1084
1085
1086
1087
1088
1089
1090
1091
1092
1093
1094
1095
1096
1097
1098
1099
1100
1101
1102
1103
1104
1105
1106
1107
1108
1109
1110
1111
1112
1113
1114
1115
1116
1117
1118
1119
1120
1121
1122
1123
1124
1125
1126
1127
1128
1129
1130
1131
1132
1133
1134
1135
1136
1137
1138
1139
1140
1141
1142
1143
1144
1145
1146
1147
1148
1149
1150
1151
1152
1153
1154
1155
1156
1157
1158
1159
1160
1161
1162
1163
1164
1165
1166
1167
1168
1169
1170
1171
1172
1173
1174
1175
1176
1177
1178
1179
1180
1181
1182
1183
1184
1185
1186
1187
1188
1189
1190
1191
1192
1193
1194
1195
1196
1197
1198
1199
1200
1201
1202
1203
1204
1205
1206
1207
1208
1209
1210
1211
1212
1213
1214
1215
1216
1217
1218
1219
1220
1221
1222
1223
1224
1225
1226
1227
1228
1229
1230
1231
1232
1233
1234
1235
1236
1237
1238
1239
1240
1241
1242
1243
1244
1245
1246
1247
1248
1249
1250
1251
1252
1253
1254
1255
1256
1257
1258
1259
1260
1261
1262
1263
1264
1265
1266
1267
1268
1269
1270
1271
1272
1273
1274
1275
1276
1277
1278
1279
1280
1281
1282
1283
1284
1285
1286
1287
1288
1289
1290
1291
1292
1293
1294
1295
1296
1297
1298
1299
1300
1301
1302
1303
1304
1305
1306
1307
1308
1309
1310
1311
1312
1313
1314
1315
1316
1317
1318
1319
1320
1321
1322
1323
1324
1325
1326
1327
1328
1329
1330
1331
1332
1333
1334
1335
1336
1337
1338
1339
1340
1341
1342
1343
1344
1345
1346
1347
1348
1349
1350
1351
1352
1353
1354
1355
1356
1357
//! RFC #238 S8b — chroma-siting-aware 4:4:0 **resample** for `Yuv440p`.
//!
//! 4:4:0 keeps FULL-width chroma, subsampled 2:1 **vertically only**, so the
//! siting reduces to its vertical axis: `Bottom`
//! ([`chroma_440_bottom_sited_v`](super::super::chroma_440_bottom_sited_v),
//! `v = 1`) folds the VERTICAL triangle through the resample (its even output
//! row box-blends the previous chroma row), while every co-sited / horizontal
//! siting keeps the vertical pairing co-sited (`v_phase = 0`, byte-identical to
//! the pre-siting resample):
//!  - the **native fast tier** folds the `v = 1` triangle into the chroma area
//!    weights ([`ResamplePlan::area_chroma_440`]) — one SINGLE-rounding phased
//!    box-average on the half-height grid (no horizontal reconstruction);
//!  - the **encoded row-stage tier** (`with_native(false)`) reconstructs
//!    full-height chroma then bins in RGB.
//!
//! ★ Oracle (native tier): the EXACT code-domain box-average of the UNROUNDED
//! vertical-`v = 1`-reconstructed chroma — a SINGLE rounding — pinned against a
//! YUV-domain oracle. The oracle uses EVEN source heights so the phased-V spans
//! (denominator `2·luma_h`) align with the `sh / 2` chroma rows.

use crate::{
  ChromaLocation, KernelMatrix, PixelSink,
  resample::{AreaResampler, AveragingDomain},
  sinker::MixedSinker,
  source::{Yuv440p, Yuv440pRow, Yuv444p, yuv440p_to, yuv444p_to},
};
use mediaframe::frame::{Yuv440pFrame, Yuv444pFrame};

const M: KernelMatrix = KernelMatrix::Bt601;
const FR: bool = true;

/// Round `a / d` half-up (ties toward `+∞`) — the production
/// `round_div_half_up`, replicated here so the oracle is independent.
fn rdhu(a: u64, d: u64) -> u64 {
  let q = a / d;
  let r = a % d;
  q + u64::from(r >= d - d / 2)
}

/// Exact box-overlap area weights for `src -> out`, mirroring
/// `resample::AxisSpans::area`. Returns per output `(first source cell,
/// overlaps)`.
fn area_weights(src: usize, out: usize) -> Vec<(usize, Vec<u64>)> {
  let (src64, out64) = (src as u64, out as u64);
  (0..out)
    .map(|o| {
      let lo = o as u64 * src64;
      let hi = lo + src64;
      let start = (lo / out64) as usize;
      let mut w = Vec::new();
      let mut i = start as u64;
      loop {
        let clo = i * out64;
        if clo >= hi {
          break;
        }
        let chi = clo + out64;
        let ov = chi.min(hi) - clo.max(lo);
        if ov == 0 {
          break;
        }
        w.push(ov);
        if chi >= hi {
          break;
        }
        i += 1;
      }
      (start, w)
    })
    .collect()
}

/// Co-sited box-average of a full-resolution `sw x sh` u8 plane to `ow x oh`
/// (round-half-up) — the reference for a phase-free plane (luma).
fn bin_cosited(plane: &[u8], sw: usize, sh: usize, ow: usize, oh: usize) -> Vec<u8> {
  let hw = area_weights(sw, ow);
  let vw = area_weights(sh, oh);
  let denom = (sw * sh) as u64;
  let mut out = vec![0u8; ow * oh];
  for (oy, (vs, vwin)) in vw.iter().enumerate() {
    for (ox, (hs, hwin)) in hw.iter().enumerate() {
      let mut s = 0u64;
      for (dy, &vwt) in vwin.iter().enumerate() {
        let mut hsum = 0u64;
        for (dx, &hwt) in hwin.iter().enumerate() {
          hsum += hwt * u64::from(plane[(vs + dy) * sw + hs + dx]);
        }
        s += vwt * hsum;
      }
      out[oy * ow + ox] = rdhu(s, denom) as u8;
    }
  }
  out
}

/// The EXACT bottom-sited (`v = 1`) 4:4:0 chroma oracle for the native tier: the
/// vertical `v = 1` triangle (×2 — even luma row `2i` box-blends chroma rows
/// `{i - 1, i}` with weights `{1, 1}`, odd row `2i + 1` takes chroma row `i` with
/// weight 2, top-edge clamp) kept UNROUNDED over the full-width `sw x sh` grid,
/// box-averaged to `ow x oh` — HORIZONTAL a plain box over `sw` (4:4:0 chroma is
/// full-width, no folded H triangle), VERTICAL over the `sh` luma rows — with a
/// SINGLE round-half-up over `2·sw·sh`. The code-domain twin the folded
/// [`ResamplePlan::area_chroma_440`] realizes when its V axis is
/// [`AxisSpans::area_chroma_phased_v`] (EVEN `sh` only, `ch = sh / 2`).
fn bin_chroma_bottom(c: &[u8], sw: usize, ch: usize, ow: usize, oh: usize) -> Vec<u8> {
  let sh = 2 * ch;
  // ×2 UNROUNDED vertical reconstruction (`sw x sh`), full width (no horizontal
  // triangle for 4:4:0), so the box below applies the one rounding.
  let mut r2 = vec![0u64; sw * sh];
  for r in 0..sh {
    let cr = r / 2;
    let prev = cr.saturating_sub(1);
    for j in 0..sw {
      r2[r * sw + j] = if r & 1 == 0 {
        u64::from(c[prev * sw + j]) + u64::from(c[cr * sw + j]) // even: {1, 1}
      } else {
        2 * u64::from(c[cr * sw + j]) // odd: {2}
      };
    }
  }
  let hw = area_weights(sw, ow);
  let vw = area_weights(sh, oh);
  let denom = (2 * sw * sh) as u64; // ×2 (V triangle) × the box normalization
  let mut out = vec![0u8; ow * oh];
  for (oy, (vs, vwin)) in vw.iter().enumerate() {
    for (ox, (hs, hwin)) in hw.iter().enumerate() {
      let mut s = 0u64;
      for (dy, &vwt) in vwin.iter().enumerate() {
        let mut hsum = 0u64;
        for (dx, &hwt) in hwin.iter().enumerate() {
          hsum += hwt * r2[(vs + dy) * sw + hs + dx];
        }
        s += vwt * hsum;
      }
      out[oy * ow + ox] = rdhu(s, denom) as u8;
    }
  }
  out
}

/// Reconstruct `Yuv440p` chroma to full height (`sw x sh`) for the bottom-sited
/// (`v = 1`) decode — the identity bottom kernel at source width: per luma row
/// the even rows vertically box-blend chroma rows `i - 1` (clamped) and `i`
/// (round-half-up), the odd rows take chroma row `i` straight through. No
/// horizontal reconstruction (full-width chroma). The shared reconstruction step
/// for the reconstruct-then-bin oracles.
fn recon_full_bottom(u: &[u8], v: &[u8], sw: usize, sh: usize) -> (Vec<u8>, Vec<u8>) {
  let vblend = |plane: &[u8], cr: usize, prev: usize| -> Vec<u8> {
    (0..sw)
      .map(|c| {
        let a = u32::from(plane[prev * sw + c]);
        let b = u32::from(plane[cr * sw + c]);
        ((a + b + 1) >> 1) as u8
      })
      .collect::<Vec<u8>>()
  };
  let mut uf = vec![0u8; sw * sh];
  let mut vf = vec![0u8; sw * sh];
  for r in 0..sh {
    let cr = r / 2;
    let (uh, vh) = if r & 1 == 0 {
      let prev = cr.saturating_sub(1);
      (vblend(u, cr, prev), vblend(v, cr, prev))
    } else {
      (
        u[cr * sw..cr * sw + sw].to_vec(),
        v[cr * sw..cr * sw + sw].to_vec(),
      )
    };
    uf[r * sw..r * sw + sw].copy_from_slice(&uh);
    vf[r * sw..r * sw + sw].copy_from_slice(&vh);
  }
  (uf, vf)
}

/// A `Yuv440p` fixture (full-width chroma, `ch = sh / 2`) with a horizontal AND
/// vertical chroma ramp plus varying luma. `sh` must be even.
fn ramp(sw: usize, sh: usize) -> (Vec<u8>, Vec<u8>, Vec<u8>) {
  let ch = sh / 2;
  let mut y = vec![0u8; sw * sh];
  for (i, p) in y.iter_mut().enumerate() {
    *p = (40 + (i as u32 * 3) % 160) as u8;
  }
  let mut u = vec![0u8; sw * ch];
  let mut v = vec![0u8; sw * ch];
  for r in 0..ch {
    for c in 0..sw {
      u[r * sw + c] = (30 + c * 12 + r * 40).min(240) as u8;
      v[r * sw + c] = (230u32.saturating_sub((c * 10 + r * 40) as u32)).max(16) as u8;
    }
  }
  (y, u, v)
}

/// A `Yuv440p` fixture with flat luma and a strong per-ROW chroma step (flat
/// across columns), so the vertical bottom fold is observable in isolation.
/// `sh` must be even.
fn vramp(sw: usize, sh: usize) -> (Vec<u8>, Vec<u8>, Vec<u8>) {
  let ch = sh / 2;
  let y = vec![128u8; sw * sh];
  let mut u = vec![0u8; sw * ch];
  let mut v = vec![0u8; sw * ch];
  for r in 0..ch {
    for c in 0..sw {
      u[r * sw + c] = (20 + r * 40).min(240) as u8;
      v[r * sw + c] = (220u32.saturating_sub((r * 40) as u32)).max(16) as u8;
    }
  }
  (y, u, v)
}

/// A flat-chroma fixture: the vertical blend of a constant is that constant, so
/// `Bottom` must equal co-sited. Luma still varies.
fn flat_chroma(sw: usize, sh: usize) -> (Vec<u8>, Vec<u8>, Vec<u8>) {
  let ch = sh / 2;
  let mut y = vec![0u8; sw * sh];
  for (i, p) in y.iter_mut().enumerate() {
    *p = (40 + (i as u32 * 7) % 170) as u8;
  }
  (y, vec![110u8; sw * ch], vec![140u8; sw * ch])
}

type Outs = (
  Vec<u8>,
  Vec<u8>,
  (Vec<u8>, Vec<u8>, Vec<u8>),
  Vec<u8>,
  Vec<u16>,
);

/// Drive a `Yuv440p` area resample (`sw x sh -> ow x oh`) for the full output
/// set, at `loc` siting and `native` tier.
#[allow(clippy::too_many_arguments)]
fn run(
  y: &[u8],
  u: &[u8],
  v: &[u8],
  sw: usize,
  sh: usize,
  ow: usize,
  oh: usize,
  loc: ChromaLocation,
  native: bool,
  simd: bool,
) -> Outs {
  let mut rgb = vec![0u8; ow * oh * 3];
  let mut rgba = vec![0u8; ow * oh * 4];
  let (mut hh, mut ss, mut vv) = (vec![0u8; ow * oh], vec![0u8; ow * oh], vec![0u8; ow * oh]);
  let mut luma = vec![0u8; ow * oh];
  let mut luma_u16 = vec![0u16; ow * oh];
  {
    let mut sink =
      MixedSinker::<Yuv440p, AreaResampler>::with_resampler(sw, sh, AreaResampler::to(ow, oh))
        .unwrap()
        .with_native(native)
        .with_chroma_location(loc.clone())
        .with_simd(simd)
        .with_rgb(&mut rgb)
        .unwrap()
        .with_rgba(&mut rgba)
        .unwrap()
        .with_hsv(&mut hh, &mut ss, &mut vv)
        .unwrap()
        .with_luma(&mut luma)
        .unwrap()
        .with_luma_u16(&mut luma_u16)
        .unwrap();
    let f = Yuv440pFrame::new(
      y, u, v, sw as u32, sh as u32, sw as u32, sw as u32, sw as u32,
    );
    yuv440p_to(&f, FR, sink.set_kernel_matrix(M)).unwrap();
  }
  (rgb, rgba, (hh, ss, vv), luma, luma_u16)
}

/// The bottom-sited NATIVE oracle: bin Y co-sited and U / V through the exact
/// bottom V-fold chroma oracle to `ow x oh`, then convert ONCE at output width
/// via an identity `Yuv444p` sink — the byte-for-byte ground truth the native
/// tier reproduces for `ChromaLocation::Bottom` (EVEN `sh` only).
#[allow(clippy::too_many_arguments)]
fn bottom_native_oracle(
  y: &[u8],
  u: &[u8],
  v: &[u8],
  sw: usize,
  sh: usize,
  ow: usize,
  oh: usize,
  simd: bool,
) -> Outs {
  let ch = sh / 2;
  let yb = bin_cosited(y, sw, sh, ow, oh);
  let ub = bin_chroma_bottom(u, sw, ch, ow, oh);
  let vb = bin_chroma_bottom(v, sw, ch, ow, oh);
  let mut rgb = vec![0u8; ow * oh * 3];
  let mut rgba = vec![0u8; ow * oh * 4];
  let (mut hh, mut ss, mut vv) = (vec![0u8; ow * oh], vec![0u8; ow * oh], vec![0u8; ow * oh]);
  let mut luma = vec![0u8; ow * oh];
  let mut luma_u16 = vec![0u16; ow * oh];
  {
    let mut sink = MixedSinker::<Yuv444p>::new(ow, oh)
      .with_simd(simd)
      .with_rgb(&mut rgb)
      .unwrap()
      .with_rgba(&mut rgba)
      .unwrap()
      .with_hsv(&mut hh, &mut ss, &mut vv)
      .unwrap()
      .with_luma(&mut luma)
      .unwrap()
      .with_luma_u16(&mut luma_u16)
      .unwrap();
    let f = Yuv444pFrame::new(
      &yb, &ub, &vb, ow as u32, oh as u32, ow as u32, ow as u32, ow as u32,
    );
    yuv444p_to(&f, FR, sink.set_kernel_matrix(M)).unwrap();
  }
  (rgb, rgba, (hh, ss, vv), luma, luma_u16)
}

/// The bottom-sited ENCODED row-stage oracle: reconstruct U / V to full height
/// with the vertical bottom blend ([`recon_full_bottom`]) then run that
/// full-resolution `Yuv444p` frame through a `with_native(false)` RGB-domain
/// resample (convert-each-row-then-bin-RGB, exactly what the `Yuv440p` encoded
/// arm does for `Bottom`).
#[allow(clippy::too_many_arguments)]
fn encoded_oracle_rgb_bottom(
  y: &[u8],
  u: &[u8],
  v: &[u8],
  sw: usize,
  sh: usize,
  ow: usize,
  oh: usize,
  simd: bool,
) -> Vec<u8> {
  let (uf, vf) = recon_full_bottom(u, v, sw, sh);
  let mut rgb = vec![0u8; ow * oh * 3];
  {
    let mut sink =
      MixedSinker::<Yuv444p, AreaResampler>::with_resampler(sw, sh, AreaResampler::to(ow, oh))
        .unwrap()
        .with_native(false)
        .with_simd(simd)
        .with_rgb(&mut rgb)
        .unwrap();
    let f = Yuv444pFrame::new(
      y, &uf, &vf, sw as u32, sh as u32, sw as u32, sw as u32, sw as u32,
    );
    yuv444p_to(&f, FR, sink.set_kernel_matrix(M)).unwrap();
  }
  rgb
}

/// Direct (non-resample) `Yuv440p` `Bottom` decode to RGB — the delay-line
/// kernel path, the already-validated identity Bottom reference the resample
/// path must match at identity dimensions.
fn direct_bottom_rgb(y: &[u8], u: &[u8], v: &[u8], sw: usize, sh: usize, simd: bool) -> Vec<u8> {
  let mut rgb = vec![0u8; sw * sh * 3];
  {
    let mut sink = MixedSinker::<Yuv440p>::new(sw, sh)
      .with_chroma_location(ChromaLocation::Bottom)
      .with_simd(simd)
      .with_rgb(&mut rgb)
      .unwrap();
    let f = Yuv440pFrame::new(
      y, u, v, sw as u32, sh as u32, sw as u32, sw as u32, sw as u32,
    );
    yuv440p_to(&f, FR, sink.set_kernel_matrix(M)).unwrap();
  }
  rgb
}

// ---- co-sited byte-identity (the regression contract) ----------------------

#[test]
#[cfg_attr(
  miri,
  ignore = "SIMD-dispatched row kernels use intrinsics unsupported by Miri"
)]
fn cosited_group_is_byte_identical_across_tiers() {
  // Every co-sited / vertically-central siting must produce the byte-identical
  // pre-siting resample (v_phase 0 && !v_top → neither phased-V plan is built), on
  // BOTH tiers. 4:4:0 has no horizontal phase, so `Center` (v=0.5) stays co-sited;
  // `Top` / `TopLeft` (v=0) now fold the forward triangle and left this group.
  let (y, u, v) = ramp(8, 8);
  for native in [true, false] {
    let base = run(
      &y,
      &u,
      &v,
      8,
      8,
      4,
      4,
      ChromaLocation::Unspecified,
      native,
      true,
    );
    for loc in [
      ChromaLocation::Left,
      ChromaLocation::Center,
      ChromaLocation::other("unassigned-7"),
    ] {
      let got = run(&y, &u, &v, 8, 8, 4, 4, loc.clone(), native, true);
      assert_eq!(got.0, base.0, "rgb {loc:?} native={native}");
      assert_eq!(got.1, base.1, "rgba {loc:?} native={native}");
      assert_eq!(got.2, base.2, "hsv {loc:?} native={native}");
      assert_eq!(got.3, base.3, "luma {loc:?} native={native}");
      assert_eq!(got.4, base.4, "luma_u16 {loc:?} native={native}");
    }
  }
}

#[test]
#[cfg_attr(
  miri,
  ignore = "SIMD-dispatched row kernels use intrinsics unsupported by Miri"
)]
fn bottom_left_resamples_as_bottom_across_tiers() {
  // 4:4:0 has no horizontal phase, so `BottomLeft` (v=1) drives the identical
  // vertical fold as `Bottom` on every tier. The ramp varies vertically, so a
  // `BottomLeft` that stayed co-sited would diverge from the folded `Bottom` —
  // making the equality a genuine regression guard.
  let (y, u, v) = ramp(8, 8);
  for native in [true, false] {
    let bottom = run(&y, &u, &v, 8, 8, 4, 4, ChromaLocation::Bottom, native, true);
    let bl = run(
      &y,
      &u,
      &v,
      8,
      8,
      4,
      4,
      ChromaLocation::BottomLeft,
      native,
      true,
    );
    assert_eq!(bl.0, bottom.0, "rgb BottomLeft==Bottom native={native}");
    assert_eq!(bl.1, bottom.1, "rgba BottomLeft==Bottom native={native}");
    assert_eq!(bl.2, bottom.2, "hsv BottomLeft==Bottom native={native}");
    assert_eq!(bl.3, bottom.3, "luma BottomLeft==Bottom native={native}");
    assert_eq!(
      bl.4, bottom.4,
      "luma_u16 BottomLeft==Bottom native={native}"
    );
  }
}

// ---- bottom native == the exact code-domain oracle -------------------------

#[test]
#[cfg_attr(
  miri,
  ignore = "SIMD-dispatched row kernels use intrinsics unsupported by Miri"
)]
fn bottom_native_equals_code_domain_oracle() {
  // The native tier folds the vertical v=1 triangle into the chroma area
  // weights; its output is the EXACT code-domain box-average of the UNROUNDED
  // V reconstruction (single rounding), for clean 2:1 and fractional ratios
  // (EVEN source height).
  for (sw, sh, ow, oh) in [(8, 8, 4, 4), (8, 8, 5, 3), (12, 8, 4, 4), (16, 8, 6, 5)] {
    let (y, u, v) = ramp(sw, sh);
    let o = bottom_native_oracle(&y, &u, &v, sw, sh, ow, oh, true);
    let n = run(
      &y,
      &u,
      &v,
      sw,
      sh,
      ow,
      oh,
      ChromaLocation::Bottom,
      true,
      true,
    );
    assert_eq!(n.0, o.0, "rgb {sw}x{sh}->{ow}x{oh}");
    assert_eq!(n.1, o.1, "rgba {sw}x{sh}->{ow}x{oh}");
    assert_eq!(n.2, o.2, "hsv {sw}x{sh}->{ow}x{oh}");
    assert_eq!(n.3, o.3, "luma {sw}x{sh}->{ow}x{oh}");
    assert_eq!(n.4, o.4, "luma_u16 {sw}x{sh}->{ow}x{oh}");
  }
}

#[test]
#[cfg_attr(
  miri,
  ignore = "SIMD-dispatched row kernels use intrinsics unsupported by Miri"
)]
fn bottom_native_simd_matches_scalar() {
  // Precomputed integer weights → the SIMD H/V passes are 0-ULP against scalar.
  let (y, u, v) = ramp(8, 8);
  let s = run(&y, &u, &v, 8, 8, 4, 4, ChromaLocation::Bottom, true, false);
  let d = run(&y, &u, &v, 8, 8, 4, 4, ChromaLocation::Bottom, true, true);
  assert_eq!(s.0, d.0, "rgb scalar vs simd");
  assert_eq!(s.2, d.2, "hsv scalar vs simd");
  assert_eq!(s.3, d.3, "luma scalar vs simd");
}

// ---- bottom encoded row-stage / filter / linear == reconstruct-then-bin ----

#[test]
#[cfg_attr(
  miri,
  ignore = "SIMD-dispatched row kernels use intrinsics unsupported by Miri"
)]
fn bottom_encoded_output_equals_rgb_reconstruct_then_bin() {
  // The encoded row-stage tier reconstructs full-height chroma with the vertical
  // bottom blend then bins in RGB — the reconstruct-then-bin oracle.
  for (sw, sh, ow, oh) in [(8, 8, 4, 4), (8, 8, 5, 3), (12, 8, 6, 4)] {
    let (y, u, v) = ramp(sw, sh);
    let oracle = encoded_oracle_rgb_bottom(&y, &u, &v, sw, sh, ow, oh, true);
    let got = run(
      &y,
      &u,
      &v,
      sw,
      sh,
      ow,
      oh,
      ChromaLocation::Bottom,
      false,
      true,
    );
    assert_eq!(got.0, oracle, "rgb {sw}x{sh}->{ow}x{oh}");
  }
}

#[test]
#[cfg_attr(
  miri,
  ignore = "SIMD-dispatched row kernels use intrinsics unsupported by Miri"
)]
fn bottom_filter_equals_reconstruct_then_filter() {
  // The single-kernel filter tier reconstructs full-height chroma with the
  // vertical bottom blend then Triangle-filters it — equal to feeding the same
  // reconstruction through a Yuv444p Triangle filter, and non-vacuously
  // different from the co-sited decode on a vertical ramp.
  use crate::resample::{FilteredResampler, Triangle};
  let (sw, sh, ow, oh) = (8usize, 8usize, 4usize, 4usize);
  let (y, u, v) = vramp(sw, sh);
  let filter_440 = |loc: ChromaLocation| -> Vec<u8> {
    let mut rgb = vec![0u8; ow * oh * 3];
    {
      let mut sink = MixedSinker::<Yuv440p, FilteredResampler<Triangle>>::with_resampler(
        sw,
        sh,
        FilteredResampler::new(ow, oh, Triangle),
      )
      .unwrap()
      .with_chroma_location(loc.clone())
      .with_simd(true)
      .with_rgb(&mut rgb)
      .unwrap();
      let f = Yuv440pFrame::new(
        &y, &u, &v, sw as u32, sh as u32, sw as u32, sw as u32, sw as u32,
      );
      yuv440p_to(&f, FR, sink.set_kernel_matrix(M)).unwrap();
    }
    rgb
  };
  let got = filter_440(ChromaLocation::Bottom);
  let (uf, vf) = recon_full_bottom(&u, &v, sw, sh);
  let mut oracle = vec![0u8; ow * oh * 3];
  {
    let mut sink = MixedSinker::<Yuv444p, FilteredResampler<Triangle>>::with_resampler(
      sw,
      sh,
      FilteredResampler::new(ow, oh, Triangle),
    )
    .unwrap()
    .with_simd(true)
    .with_rgb(&mut oracle)
    .unwrap();
    let f = Yuv444pFrame::new(
      &y, &uf, &vf, sw as u32, sh as u32, sw as u32, sw as u32, sw as u32,
    );
    yuv444p_to(&f, FR, sink.set_kernel_matrix(M)).unwrap();
  }
  assert_eq!(got, oracle, "filter-tier bottom == reconstruct-then-filter");
  assert_ne!(
    got,
    filter_440(ChromaLocation::Left),
    "filter-tier bottom must differ from co-sited on a vertical ramp"
  );
}

#[test]
#[cfg_attr(
  miri,
  ignore = "SIMD-dispatched row kernels use intrinsics unsupported by Miri"
)]
fn bottom_linear_equals_reconstruct_then_linear() {
  // The linear-light tier reconstructs full-height chroma with the vertical
  // bottom blend then resamples in linear light — equal to feeding the same
  // reconstruction through a Yuv444p linear-domain resample.
  let (sw, sh, ow, oh) = (8usize, 8usize, 4usize, 4usize);
  let (y, u, v) = ramp(sw, sh);
  let mut got = vec![0u8; ow * oh * 3];
  {
    let mut sink =
      MixedSinker::<Yuv440p, AreaResampler>::with_resampler(sw, sh, AreaResampler::to(ow, oh))
        .unwrap()
        .with_averaging_domain(AveragingDomain::Linear)
        .with_chroma_location(ChromaLocation::Bottom)
        .with_rgb(&mut got)
        .unwrap();
    let f = Yuv440pFrame::new(
      &y, &u, &v, sw as u32, sh as u32, sw as u32, sw as u32, sw as u32,
    );
    yuv440p_to(&f, FR, sink.set_kernel_matrix(M)).unwrap();
  }
  let (uf, vf) = recon_full_bottom(&u, &v, sw, sh);
  let mut oracle = vec![0u8; ow * oh * 3];
  {
    let mut sink =
      MixedSinker::<Yuv444p, AreaResampler>::with_resampler(sw, sh, AreaResampler::to(ow, oh))
        .unwrap()
        .with_averaging_domain(AveragingDomain::Linear)
        .with_rgb(&mut oracle)
        .unwrap();
    let f = Yuv444pFrame::new(
      &y, &uf, &vf, sw as u32, sh as u32, sw as u32, sw as u32, sw as u32,
    );
    yuv444p_to(&f, FR, sink.set_kernel_matrix(M)).unwrap();
  }
  assert_eq!(got, oracle, "linear-tier bottom == reconstruct-then-linear");
}

// ---- identity resample == direct decode ------------------------------------

#[test]
#[cfg_attr(
  miri,
  ignore = "SIMD-dispatched row kernels use intrinsics unsupported by Miri"
)]
fn bottom_encoded_identity_matches_direct_decode() {
  // At identity dimensions (out == src) the encoded resample tier reconstructs
  // chroma with the SAME bottom kernel as the direct (non-resample) Yuv440p
  // decode and bins with pass-through area weights, so routing Bottom through the
  // resample preserves the decode byte-for-byte.
  let (y, u, v) = vramp(8, 8);
  let res = run(&y, &u, &v, 8, 8, 8, 8, ChromaLocation::Bottom, false, true);
  let direct = direct_bottom_rgb(&y, &u, &v, 8, 8, true);
  assert_eq!(
    res.0, direct,
    "identity encoded resample bottom == direct decode"
  );
}

// ---- non-vacuous + flat-chroma sanity --------------------------------------

#[test]
#[cfg_attr(
  miri,
  ignore = "SIMD-dispatched row kernels use intrinsics unsupported by Miri"
)]
fn bottom_differs_from_cosited_on_vertical_ramp() {
  // The v=1 fold must actually MOVE the chroma: on a purely-vertical chroma ramp
  // Bottom diverges from the co-sited decode on BOTH tiers.
  let (y, u, v) = vramp(8, 8);
  for native in [true, false] {
    let cos = run(&y, &u, &v, 8, 8, 4, 4, ChromaLocation::Left, native, true);
    let bot = run(&y, &u, &v, 8, 8, 4, 4, ChromaLocation::Bottom, native, true);
    assert_ne!(
      bot.0, cos.0,
      "bottom rgb must differ from co-sited (native={native})"
    );
  }
}

#[test]
#[cfg_attr(
  miri,
  ignore = "SIMD-dispatched row kernels use intrinsics unsupported by Miri"
)]
fn bottom_equals_cosited_on_flat_chroma() {
  // On constant chroma the vertical blend is a no-op, so Bottom collapses to the
  // co-sited decode byte-for-byte.
  let (y, u, v) = flat_chroma(8, 8);
  for native in [true, false] {
    let cos = run(&y, &u, &v, 8, 8, 4, 4, ChromaLocation::Left, native, true);
    let bot = run(&y, &u, &v, 8, 8, 4, 4, ChromaLocation::Bottom, native, true);
    assert_eq!(bot.0, cos.0, "flat-chroma bottom rgb (native={native})");
    assert_eq!(bot.2, cos.2, "flat-chroma bottom hsv (native={native})");
  }
}

// ---- IN-SEQUENCE mid-frame phase change is rejected (not silently mixed) -----

/// Drive one Yuv440p resample frame: `begin_frame`, accept row 0 at `loc1`
/// (freezes the vertical phase), flip to `loc2`, then feed the IN-SEQUENCE row 1
/// (still chroma row 0) and return its `process` result.
fn in_sequence_flip_row1<R>(
  mut sink: MixedSinker<'_, Yuv440p, R>,
  y: &[u8],
  u: &[u8],
  v: &[u8],
  loc1: ChromaLocation,
  loc2: ChromaLocation,
) -> Result<(), super::super::MixedSinkerError> {
  sink.set_chroma_location(loc1.clone());
  PixelSink::begin_frame(&mut sink, 8, 8).unwrap();
  let row0 = Yuv440pRow::for_tests(&y[0..8], &u[0..8], &v[0..8], 0, M, FR);
  PixelSink::process(&mut sink, row0).unwrap();
  sink.set_chroma_location(loc2.clone());
  let row1 = Yuv440pRow::for_tests(&y[8..16], &u[0..8], &v[0..8], 1, M, FR);
  PixelSink::process(&mut sink, row1)
}

#[test]
#[cfg_attr(
  miri,
  ignore = "SIMD-dispatched row kernels use intrinsics unsupported by Miri"
)]
fn in_sequence_mid_frame_phase_change_rejected_across_tiers() {
  use super::super::MixedSinkerError;
  let (y, u, v) = ramp(8, 8);
  // Every co-sited ⇆ bottom-vertical flip (both directions) changes the v=1 fold
  // and must reject the in-sequence row 1 with ChromaSitingChanged across all
  // tiers. `Bottom` (h=0.5) and `BottomLeft` (h=0) are both v=1 for 4:4:0.
  for (loc1, loc2) in [
    (ChromaLocation::Left, ChromaLocation::Bottom),
    (ChromaLocation::Bottom, ChromaLocation::Left),
    (ChromaLocation::Center, ChromaLocation::Bottom),
    (ChromaLocation::Bottom, ChromaLocation::Center),
    (ChromaLocation::Left, ChromaLocation::BottomLeft),
    (ChromaLocation::BottomLeft, ChromaLocation::Top),
  ] {
    // Native fast tier.
    let mut rgb = vec![0u8; 4 * 4 * 3];
    let sink = MixedSinker::<Yuv440p, AreaResampler>::with_resampler(8, 8, AreaResampler::to(4, 4))
      .unwrap()
      .with_native(true)
      .with_rgb(&mut rgb)
      .unwrap();
    let err = in_sequence_flip_row1(sink, &y, &u, &v, loc1.clone(), loc2.clone()).unwrap_err();
    assert!(
      matches!(err, MixedSinkerError::ChromaSitingChanged(_)),
      "native {loc1:?}->{loc2:?}: want ChromaSitingChanged, got {err:?}"
    );

    // Encoded row-stage RGB tier.
    let mut rgb = vec![0u8; 4 * 4 * 3];
    let sink = MixedSinker::<Yuv440p, AreaResampler>::with_resampler(8, 8, AreaResampler::to(4, 4))
      .unwrap()
      .with_native(false)
      .with_rgb(&mut rgb)
      .unwrap();
    let err = in_sequence_flip_row1(sink, &y, &u, &v, loc1.clone(), loc2.clone()).unwrap_err();
    assert!(
      matches!(err, MixedSinkerError::ChromaSitingChanged(_)),
      "encoded-rgb {loc1:?}->{loc2:?}: want ChromaSitingChanged, got {err:?}"
    );

    // HSV-only tier (the `HsvDirectPlanarYuv` join).
    let (mut hh, mut ss, mut vv) = (vec![0u8; 4 * 4], vec![0u8; 4 * 4], vec![0u8; 4 * 4]);
    let sink = MixedSinker::<Yuv440p, AreaResampler>::with_resampler(8, 8, AreaResampler::to(4, 4))
      .unwrap()
      .with_native(false)
      .with_hsv(&mut hh, &mut ss, &mut vv)
      .unwrap();
    let err = in_sequence_flip_row1(sink, &y, &u, &v, loc1.clone(), loc2.clone()).unwrap_err();
    assert!(
      matches!(err, MixedSinkerError::ChromaSitingChanged(_)),
      "hsv {loc1:?}->{loc2:?}: want ChromaSitingChanged, got {err:?}"
    );

    // Linear averaging domain.
    let mut rgb = vec![0u8; 4 * 4 * 3];
    let sink = MixedSinker::<Yuv440p, AreaResampler>::with_resampler(8, 8, AreaResampler::to(4, 4))
      .unwrap()
      .with_averaging_domain(AveragingDomain::Linear)
      .with_native(false)
      .with_rgb(&mut rgb)
      .unwrap();
    let err = in_sequence_flip_row1(sink, &y, &u, &v, loc1.clone(), loc2.clone()).unwrap_err();
    assert!(
      matches!(err, MixedSinkerError::ChromaSitingChanged(_)),
      "linear {loc1:?}->{loc2:?}: want ChromaSitingChanged, got {err:?}"
    );

    // Filter tier (single-kernel Triangle FilteredResampler).
    let mut rgb = vec![0u8; 4 * 4 * 3];
    let sink = MixedSinker::<
      Yuv440p,
      crate::resample::FilteredResampler<crate::resample::Triangle>,
    >::with_resampler(
      8,
      8,
      crate::resample::FilteredResampler::new(4, 4, crate::resample::Triangle),
    )
    .unwrap()
    .with_rgb(&mut rgb)
    .unwrap();
    let err = in_sequence_flip_row1(sink, &y, &u, &v, loc1.clone(), loc2.clone()).unwrap_err();
    assert!(
      matches!(err, MixedSinkerError::ChromaSitingChanged(_)),
      "filter {loc1:?}->{loc2:?}: want ChromaSitingChanged, got {err:?}"
    );
  }
}

// ---- cross-frame sink reuse rebuilds the phased join -----------------------

/// Reuse ONE native-tier sink across two frames of the SAME content, siting
/// `loc1` then `loc2`, returning frame 2's RGB.
fn run_reuse_native(
  y: &[u8],
  u: &[u8],
  v: &[u8],
  loc1: ChromaLocation,
  loc2: ChromaLocation,
) -> Vec<u8> {
  let mut rgb = vec![0u8; 4 * 4 * 3];
  {
    let mut sink =
      MixedSinker::<Yuv440p, AreaResampler>::with_resampler(8, 8, AreaResampler::to(4, 4))
        .unwrap()
        .with_native(true)
        .with_rgb(&mut rgb)
        .unwrap();
    let f = Yuv440pFrame::new(y, u, v, 8, 8, 8, 8, 8);
    sink.set_chroma_location(loc1.clone());
    yuv440p_to(&f, FR, sink.set_kernel_matrix(M)).unwrap();
    sink.set_chroma_location(loc2.clone());
    yuv440p_to(&f, FR, sink.set_kernel_matrix(M)).unwrap();
  }
  rgb
}

/// One HSV-only (`with_native(false)` → the `HsvDirectPlanarYuv` join) fresh
/// frame at `loc`. HSV-only takes the direct YUV→HSV binning path, distinct from
/// the RGB-derived HSV a sink with RGB attached would produce, so the reuse
/// oracle must compare against THIS, not an RGB-bearing sink.
fn run_hsv_only(y: &[u8], u: &[u8], v: &[u8], loc: ChromaLocation) -> (Vec<u8>, Vec<u8>, Vec<u8>) {
  let (mut hh, mut ss, mut vv) = (vec![0u8; 4 * 4], vec![0u8; 4 * 4], vec![0u8; 4 * 4]);
  {
    let mut sink =
      MixedSinker::<Yuv440p, AreaResampler>::with_resampler(8, 8, AreaResampler::to(4, 4))
        .unwrap()
        .with_native(false)
        .with_chroma_location(loc.clone())
        .with_hsv(&mut hh, &mut ss, &mut vv)
        .unwrap();
    let f = Yuv440pFrame::new(y, u, v, 8, 8, 8, 8, 8);
    yuv440p_to(&f, FR, sink.set_kernel_matrix(M)).unwrap();
  }
  (hh, ss, vv)
}

/// Reuse ONE HSV-only sink across two frames, siting `loc1` then `loc2`.
fn run_reuse_hsv(
  y: &[u8],
  u: &[u8],
  v: &[u8],
  loc1: ChromaLocation,
  loc2: ChromaLocation,
) -> (Vec<u8>, Vec<u8>, Vec<u8>) {
  let (mut hh, mut ss, mut vv) = (vec![0u8; 4 * 4], vec![0u8; 4 * 4], vec![0u8; 4 * 4]);
  {
    let mut sink =
      MixedSinker::<Yuv440p, AreaResampler>::with_resampler(8, 8, AreaResampler::to(4, 4))
        .unwrap()
        .with_native(false)
        .with_hsv(&mut hh, &mut ss, &mut vv)
        .unwrap();
    let f = Yuv440pFrame::new(y, u, v, 8, 8, 8, 8, 8);
    sink.set_chroma_location(loc1.clone());
    yuv440p_to(&f, FR, sink.set_kernel_matrix(M)).unwrap();
    sink.set_chroma_location(loc2.clone());
    yuv440p_to(&f, FR, sink.set_kernel_matrix(M)).unwrap();
  }
  (hh, ss, vv)
}

#[test]
#[cfg_attr(
  miri,
  ignore = "SIMD-dispatched row kernels use intrinsics unsupported by Miri"
)]
fn native_join_rebuilds_on_siting_change_across_frames() {
  // Reuse one native-tier sink flipping Left ⇆ Bottom (both directions): frame 2
  // must match a FRESH sink for frame 2's siting — no stale-phase carryover.
  let (y, u, v) = vramp(8, 8);
  for (loc1, loc2) in [
    (ChromaLocation::Left, ChromaLocation::Bottom),
    (ChromaLocation::Bottom, ChromaLocation::Left),
  ] {
    let reused = run_reuse_native(&y, &u, &v, loc1.clone(), loc2.clone());
    let fresh = run(&y, &u, &v, 8, 8, 4, 4, loc2.clone(), true, true).0;
    assert_eq!(
      reused, fresh,
      "native reuse {loc1:?}->{loc2:?} must rebuild"
    );
  }
}

#[test]
#[cfg_attr(
  miri,
  ignore = "SIMD-dispatched row kernels use intrinsics unsupported by Miri"
)]
fn hsv_join_rebuilds_on_siting_change_across_frames() {
  // The `HsvDirectPlanarYuv` twin: reuse one HSV-only sink flipping Left ⇆
  // Bottom; frame 2 must match a fresh HSV-only sink for frame 2's siting.
  let (y, u, v) = vramp(8, 8);
  for (loc1, loc2) in [
    (ChromaLocation::Left, ChromaLocation::Bottom),
    (ChromaLocation::Bottom, ChromaLocation::Left),
  ] {
    let reused = run_reuse_hsv(&y, &u, &v, loc1.clone(), loc2.clone());
    let fresh = run_hsv_only(&y, &u, &v, loc2.clone());
    assert_eq!(reused, fresh, "hsv reuse {loc1:?}->{loc2:?} must rebuild");
  }
}

// ============ Top-sited (v = 0) FORWARD one-row delay =======================

/// The code-domain twin the folded [`ResamplePlan::area_chroma_440`] realizes
/// when its V axis is [`AxisSpans::area_chroma_phased_v_top`] — the FORWARD mirror
/// of [`bin_chroma_bottom`]: EVEN luma row `2i` is co-sited on `c[i]` (weight
/// `{2}`), ODD row `2i+1` folds `c[i]` and `c[i+1]` (weight `{1, 1}`, clamped to
/// `c[i]` at the bottom edge), a SINGLE round-half-up over `2·sw·sh`. EVEN `sh`
/// only (`ch = sh / 2`).
fn bin_chroma_top(c: &[u8], sw: usize, ch: usize, ow: usize, oh: usize) -> Vec<u8> {
  let sh = 2 * ch;
  let mut r2 = vec![0u64; sw * sh];
  for r in 0..sh {
    let cr = r / 2;
    let next = (cr + 1).min(ch - 1);
    for j in 0..sw {
      r2[r * sw + j] = if r & 1 == 0 {
        2 * u64::from(c[cr * sw + j]) // even: co-sited {2}
      } else {
        u64::from(c[cr * sw + j]) + u64::from(c[next * sw + j]) // odd: forward {1, 1}
      };
    }
  }
  let hw = area_weights(sw, ow);
  let vw = area_weights(sh, oh);
  let denom = (2 * sw * sh) as u64;
  let mut out = vec![0u8; ow * oh];
  for (oy, (vs, vwin)) in vw.iter().enumerate() {
    for (ox, (hs, hwin)) in hw.iter().enumerate() {
      let mut s = 0u64;
      for (dy, &vwt) in vwin.iter().enumerate() {
        let mut hsum = 0u64;
        for (dx, &hwt) in hwin.iter().enumerate() {
          hsum += hwt * r2[(vs + dy) * sw + hs + dx];
        }
        s += vwt * hsum;
      }
      out[oy * ow + ox] = rdhu(s, denom) as u8;
    }
  }
  out
}

/// Reconstruct `Yuv440p` chroma to full height (`sw x sh`) for the top-sited
/// (`v = 0`) decode — the FORWARD mirror of [`recon_full_bottom`]: even rows take
/// chroma row `i` straight, odd rows forward box-blend `c[i]` and `c[i+1]`
/// (round-half-up, bottom-edge clamp). The shared reconstruction step for the
/// reconstruct-then-bin/filter/linear Top oracles.
fn recon_full_top(u: &[u8], v: &[u8], sw: usize, sh: usize) -> (Vec<u8>, Vec<u8>) {
  let ch = sh.div_ceil(2);
  let vblend = |plane: &[u8], a: usize, b: usize| -> Vec<u8> {
    (0..sw)
      .map(|c| {
        let x = u32::from(plane[a * sw + c]);
        let y = u32::from(plane[b * sw + c]);
        ((x + y + 1) >> 1) as u8
      })
      .collect::<Vec<u8>>()
  };
  let mut uf = vec![0u8; sw * sh];
  let mut vf = vec![0u8; sw * sh];
  for r in 0..sh {
    let cr = r / 2;
    let (uh, vh) = if r & 1 == 0 {
      (
        u[cr * sw..cr * sw + sw].to_vec(),
        v[cr * sw..cr * sw + sw].to_vec(),
      )
    } else {
      let next = (cr + 1).min(ch - 1);
      (vblend(u, cr, next), vblend(v, cr, next))
    };
    uf[r * sw..r * sw + sw].copy_from_slice(&uh);
    vf[r * sw..r * sw + sw].copy_from_slice(&vh);
  }
  (uf, vf)
}

/// The top-sited NATIVE oracle: bin Y co-sited and U / V through the exact forward
/// V-fold chroma oracle to `ow x oh`, then convert ONCE at output width via an
/// identity `Yuv444p` sink (EVEN `sh` only).
#[allow(clippy::too_many_arguments)]
fn top_native_oracle(
  y: &[u8],
  u: &[u8],
  v: &[u8],
  sw: usize,
  sh: usize,
  ow: usize,
  oh: usize,
  simd: bool,
) -> Outs {
  let ch = sh / 2;
  let yb = bin_cosited(y, sw, sh, ow, oh);
  let ub = bin_chroma_top(u, sw, ch, ow, oh);
  let vb = bin_chroma_top(v, sw, ch, ow, oh);
  let mut rgb = vec![0u8; ow * oh * 3];
  let mut rgba = vec![0u8; ow * oh * 4];
  let (mut hh, mut ss, mut vv) = (vec![0u8; ow * oh], vec![0u8; ow * oh], vec![0u8; ow * oh]);
  let mut luma = vec![0u8; ow * oh];
  let mut luma_u16 = vec![0u16; ow * oh];
  {
    let mut sink = MixedSinker::<Yuv444p>::new(ow, oh)
      .with_simd(simd)
      .with_rgb(&mut rgb)
      .unwrap()
      .with_rgba(&mut rgba)
      .unwrap()
      .with_hsv(&mut hh, &mut ss, &mut vv)
      .unwrap()
      .with_luma(&mut luma)
      .unwrap()
      .with_luma_u16(&mut luma_u16)
      .unwrap();
    let f = Yuv444pFrame::new(
      &yb, &ub, &vb, ow as u32, oh as u32, ow as u32, ow as u32, ow as u32,
    );
    yuv444p_to(&f, FR, sink.set_kernel_matrix(M)).unwrap();
  }
  (rgb, rgba, (hh, ss, vv), luma, luma_u16)
}

/// The top-sited ENCODED row-stage oracle: reconstruct U / V to full height with
/// the forward top blend ([`recon_full_top`]) then run that full-resolution
/// `Yuv444p` frame through a `with_native(false)` RGB-domain resample.
#[allow(clippy::too_many_arguments)]
fn encoded_oracle_rgb_top(
  y: &[u8],
  u: &[u8],
  v: &[u8],
  sw: usize,
  sh: usize,
  ow: usize,
  oh: usize,
  simd: bool,
) -> Vec<u8> {
  let (uf, vf) = recon_full_top(u, v, sw, sh);
  let mut rgb = vec![0u8; ow * oh * 3];
  {
    let mut sink =
      MixedSinker::<Yuv444p, AreaResampler>::with_resampler(sw, sh, AreaResampler::to(ow, oh))
        .unwrap()
        .with_native(false)
        .with_simd(simd)
        .with_rgb(&mut rgb)
        .unwrap();
    let f = Yuv444pFrame::new(
      y, &uf, &vf, sw as u32, sh as u32, sw as u32, sw as u32, sw as u32,
    );
    yuv444p_to(&f, FR, sink.set_kernel_matrix(M)).unwrap();
  }
  rgb
}

/// Direct (non-resample) `Yuv440p` `Top` decode to RGB — the FORWARD-delay kernel
/// path (the already-validated identity Top reference).
fn direct_top_rgb(y: &[u8], u: &[u8], v: &[u8], sw: usize, sh: usize, simd: bool) -> Vec<u8> {
  let mut rgb = vec![0u8; sw * sh * 3];
  {
    let mut sink = MixedSinker::<Yuv440p>::new(sw, sh)
      .with_chroma_location(ChromaLocation::Top)
      .with_simd(simd)
      .with_rgb(&mut rgb)
      .unwrap();
    let f = Yuv440pFrame::new(
      y, u, v, sw as u32, sh as u32, sw as u32, sw as u32, sw as u32,
    );
    yuv440p_to(&f, FR, sink.set_kernel_matrix(M)).unwrap();
  }
  rgb
}

#[test]
#[cfg_attr(
  miri,
  ignore = "SIMD-dispatched row kernels use intrinsics unsupported by Miri"
)]
fn top_native_equals_code_domain_oracle() {
  // The native fast tier bins Y co-sited + U / V through the forward V-fold on the
  // half-height grid (single rounding), for clean 2:1 and fractional ratios.
  for (sw, sh, ow, oh) in [(8, 8, 4, 4), (8, 8, 5, 3), (12, 8, 6, 4), (8, 8, 8, 8)] {
    let (y, u, v) = ramp(sw, sh);
    let got = run(&y, &u, &v, sw, sh, ow, oh, ChromaLocation::Top, true, true);
    let oracle = top_native_oracle(&y, &u, &v, sw, sh, ow, oh, true);
    assert_eq!(got.0, oracle.0, "rgb {sw}x{sh}->{ow}x{oh}");
    assert_eq!(got.1, oracle.1, "rgba {sw}x{sh}->{ow}x{oh}");
    assert_eq!(got.2, oracle.2, "hsv {sw}x{sh}->{ow}x{oh}");
    assert_eq!(got.3, oracle.3, "luma {sw}x{sh}->{ow}x{oh}");
    assert_eq!(got.4, oracle.4, "luma_u16 {sw}x{sh}->{ow}x{oh}");
    // `TopLeft` (v=0, no horizontal phase) must decode identically.
    let tl = run(
      &y,
      &u,
      &v,
      sw,
      sh,
      ow,
      oh,
      ChromaLocation::TopLeft,
      true,
      true,
    );
    assert_eq!(tl.0, got.0, "TopLeft==Top rgb {sw}x{sh}->{ow}x{oh}");
  }
}

#[test]
#[cfg_attr(
  miri,
  ignore = "SIMD-dispatched row kernels use intrinsics unsupported by Miri"
)]
fn top_native_simd_matches_scalar() {
  let (y, u, v) = ramp(8, 8);
  let s = run(&y, &u, &v, 8, 8, 4, 4, ChromaLocation::Top, true, false);
  let d = run(&y, &u, &v, 8, 8, 4, 4, ChromaLocation::Top, true, true);
  assert_eq!(
    s, d,
    "top native must be bit-identical across SIMD and scalar"
  );
}

#[test]
#[cfg_attr(
  miri,
  ignore = "SIMD-dispatched row kernels use intrinsics unsupported by Miri"
)]
fn top_encoded_output_equals_rgb_reconstruct_then_bin() {
  // The encoded row-stage tier reconstructs full-height chroma with the forward
  // top blend then bins in RGB — the reconstruct-then-bin oracle.
  for (sw, sh, ow, oh) in [(8, 8, 4, 4), (8, 8, 5, 3), (12, 8, 6, 4)] {
    let (y, u, v) = ramp(sw, sh);
    let oracle = encoded_oracle_rgb_top(&y, &u, &v, sw, sh, ow, oh, true);
    let got = run(&y, &u, &v, sw, sh, ow, oh, ChromaLocation::Top, false, true);
    assert_eq!(got.0, oracle, "rgb {sw}x{sh}->{ow}x{oh}");
  }
}

#[test]
#[cfg_attr(
  miri,
  ignore = "SIMD-dispatched row kernels use intrinsics unsupported by Miri"
)]
fn top_filter_equals_reconstruct_then_filter() {
  // The single-kernel filter tier reconstructs full-height chroma with the forward
  // top blend then Triangle-filters it — equal to feeding the same reconstruction
  // through a Yuv444p Triangle filter, and non-vacuously different from co-sited.
  use crate::resample::{FilteredResampler, Triangle};
  let (sw, sh, ow, oh) = (8usize, 8usize, 4usize, 4usize);
  let (y, u, v) = vramp(sw, sh);
  let filter_440 = |loc: ChromaLocation| -> Vec<u8> {
    let mut rgb = vec![0u8; ow * oh * 3];
    {
      let mut sink = MixedSinker::<Yuv440p, FilteredResampler<Triangle>>::with_resampler(
        sw,
        sh,
        FilteredResampler::new(ow, oh, Triangle),
      )
      .unwrap()
      .with_chroma_location(loc.clone())
      .with_simd(true)
      .with_rgb(&mut rgb)
      .unwrap();
      let f = Yuv440pFrame::new(
        &y, &u, &v, sw as u32, sh as u32, sw as u32, sw as u32, sw as u32,
      );
      yuv440p_to(&f, FR, sink.set_kernel_matrix(M)).unwrap();
    }
    rgb
  };
  let got = filter_440(ChromaLocation::Top);
  let (uf, vf) = recon_full_top(&u, &v, sw, sh);
  let mut oracle = vec![0u8; ow * oh * 3];
  {
    let mut sink = MixedSinker::<Yuv444p, FilteredResampler<Triangle>>::with_resampler(
      sw,
      sh,
      FilteredResampler::new(ow, oh, Triangle),
    )
    .unwrap()
    .with_simd(true)
    .with_rgb(&mut oracle)
    .unwrap();
    let f = Yuv444pFrame::new(
      &y, &uf, &vf, sw as u32, sh as u32, sw as u32, sw as u32, sw as u32,
    );
    yuv444p_to(&f, FR, sink.set_kernel_matrix(M)).unwrap();
  }
  assert_eq!(got, oracle, "filter-tier top == reconstruct-then-filter");
  assert_ne!(
    got,
    filter_440(ChromaLocation::Left),
    "filter-tier top must differ from co-sited on a vertical ramp"
  );
}

#[test]
#[cfg_attr(
  miri,
  ignore = "SIMD-dispatched row kernels use intrinsics unsupported by Miri"
)]
fn top_linear_equals_reconstruct_then_linear() {
  // The linear-light tier reconstructs full-height chroma with the forward top
  // blend then resamples in linear light — equal to feeding the same
  // reconstruction through a Yuv444p linear-domain resample.
  let (sw, sh, ow, oh) = (8usize, 8usize, 4usize, 4usize);
  let (y, u, v) = ramp(sw, sh);
  let mut got = vec![0u8; ow * oh * 3];
  {
    let mut sink =
      MixedSinker::<Yuv440p, AreaResampler>::with_resampler(sw, sh, AreaResampler::to(ow, oh))
        .unwrap()
        .with_averaging_domain(AveragingDomain::Linear)
        .with_chroma_location(ChromaLocation::Top)
        .with_rgb(&mut got)
        .unwrap();
    let f = Yuv440pFrame::new(
      &y, &u, &v, sw as u32, sh as u32, sw as u32, sw as u32, sw as u32,
    );
    yuv440p_to(&f, FR, sink.set_kernel_matrix(M)).unwrap();
  }
  let (uf, vf) = recon_full_top(&u, &v, sw, sh);
  let mut oracle = vec![0u8; ow * oh * 3];
  {
    let mut sink =
      MixedSinker::<Yuv444p, AreaResampler>::with_resampler(sw, sh, AreaResampler::to(ow, oh))
        .unwrap()
        .with_averaging_domain(AveragingDomain::Linear)
        .with_rgb(&mut oracle)
        .unwrap();
    let f = Yuv444pFrame::new(
      &y, &uf, &vf, sw as u32, sh as u32, sw as u32, sw as u32, sw as u32,
    );
    yuv444p_to(&f, FR, sink.set_kernel_matrix(M)).unwrap();
  }
  assert_eq!(got, oracle, "linear-tier top == reconstruct-then-linear");
}

#[test]
#[cfg_attr(
  miri,
  ignore = "SIMD-dispatched row kernels use intrinsics unsupported by Miri"
)]
fn top_encoded_identity_matches_direct_decode() {
  // At identity dimensions the encoded / filter resample tiers reconstruct chroma
  // with the SAME forward top kernel as the direct (non-resample) Yuv440p decode,
  // so routing Top through the resample preserves the decode byte-for-byte —
  // native, row-stage and filter all equal the validated direct Top decode.
  let (y, u, v) = vramp(8, 8);
  let direct = direct_top_rgb(&y, &u, &v, 8, 8, true);
  let row_stage = run(&y, &u, &v, 8, 8, 8, 8, ChromaLocation::Top, false, true);
  assert_eq!(
    row_stage.0, direct,
    "identity encoded resample top == direct decode"
  );
  use crate::resample::{FilteredResampler, Triangle};
  let mut filt = vec![0u8; 8 * 8 * 3];
  {
    let mut sink = MixedSinker::<Yuv440p, FilteredResampler<Triangle>>::with_resampler(
      8,
      8,
      FilteredResampler::new(8, 8, Triangle),
    )
    .unwrap()
    .with_chroma_location(ChromaLocation::Top)
    .with_rgb(&mut filt)
    .unwrap();
    let f = Yuv440pFrame::new(&y, &u, &v, 8, 8, 8, 8, 8);
    yuv440p_to(&f, FR, sink.set_kernel_matrix(M)).unwrap();
  }
  assert_eq!(
    filt, direct,
    "identity filter resample top == direct decode"
  );
}

#[test]
#[cfg_attr(
  miri,
  ignore = "SIMD-dispatched row kernels use intrinsics unsupported by Miri"
)]
fn top_differs_from_bottom_and_cosited_on_vertical_ramp() {
  // The v=0 forward fold must MOVE the chroma: on a vertical ramp Top diverges from
  // co-sited AND from the backward-folded Bottom, on BOTH tiers.
  let (y, u, v) = vramp(8, 8);
  for native in [true, false] {
    let cos = run(&y, &u, &v, 8, 8, 4, 4, ChromaLocation::Left, native, true);
    let bot = run(&y, &u, &v, 8, 8, 4, 4, ChromaLocation::Bottom, native, true);
    let top = run(&y, &u, &v, 8, 8, 4, 4, ChromaLocation::Top, native, true);
    assert_ne!(
      top.0, cos.0,
      "top must differ from co-sited (native={native})"
    );
    assert_ne!(
      top.0, bot.0,
      "top must differ from bottom (native={native})"
    );
  }
}

#[test]
#[cfg_attr(
  miri,
  ignore = "SIMD-dispatched row kernels use intrinsics unsupported by Miri"
)]
fn top_equals_cosited_on_flat_chroma() {
  // On constant chroma the forward vertical blend is a no-op, so Top collapses to
  // the co-sited decode byte-for-byte on both tiers.
  let (y, u, v) = flat_chroma(8, 8);
  for native in [true, false] {
    let cos = run(&y, &u, &v, 8, 8, 4, 4, ChromaLocation::Left, native, true);
    let top = run(&y, &u, &v, 8, 8, 4, 4, ChromaLocation::Top, native, true);
    assert_eq!(top.0, cos.0, "flat-chroma top rgb (native={native})");
    assert_eq!(top.2, cos.2, "flat-chroma top hsv (native={native})");
  }
}

#[test]
#[cfg_attr(
  miri,
  ignore = "SIMD-dispatched row kernels use intrinsics unsupported by Miri"
)]
fn native_join_rebuilds_on_top_siting_change_across_frames() {
  // Reuse one native-tier sink flipping Left ⇆ Top (both directions): frame 2 must
  // match a FRESH sink for frame 2's siting — the Top phase is recorded separately
  // from Bottom in the cached join.
  let (y, u, v) = vramp(8, 8);
  for (loc1, loc2) in [
    (ChromaLocation::Left, ChromaLocation::Top),
    (ChromaLocation::Top, ChromaLocation::Left),
    (ChromaLocation::Bottom, ChromaLocation::Top),
    (ChromaLocation::Top, ChromaLocation::Bottom),
  ] {
    let reused = run_reuse_native(&y, &u, &v, loc1.clone(), loc2.clone());
    let fresh = run(&y, &u, &v, 8, 8, 4, 4, loc2.clone(), true, true).0;
    assert_eq!(
      reused, fresh,
      "native reuse {loc1:?}->{loc2:?} must rebuild"
    );
  }
}