forked from mypaint/libmypaint
-
Notifications
You must be signed in to change notification settings - Fork 0
/
brushmodes.c
626 lines (545 loc) · 22.6 KB
/
brushmodes.c
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
/* libmypaint - The MyPaint Brush Library
* Copyright (C) 2007-2014 Martin Renold <[email protected]> et. al
*
* Permission to use, copy, modify, and/or distribute this software for any
* purpose with or without fee is hereby granted, provided that the above
* copyright notice and this permission notice appear in all copies.
*
* THE SOFTWARE IS PROVIDED "AS IS" AND THE AUTHOR DISCLAIMS ALL WARRANTIES
* WITH REGARD TO THIS SOFTWARE INCLUDING ALL IMPLIED WARRANTIES OF
* MERCHANTABILITY AND FITNESS. IN NO EVENT SHALL THE AUTHOR BE LIABLE FOR
* ANY SPECIAL, DIRECT, INDIRECT, OR CONSEQUENTIAL DAMAGES OR ANY DAMAGES
* WHATSOEVER RESULTING FROM LOSS OF USE, DATA OR PROFITS, WHETHER IN AN
* ACTION OF CONTRACT, NEGLIGENCE OR OTHER TORTIOUS ACTION, ARISING OUT OF
* OR IN CONNECTION WITH THE USE OR PERFORMANCE OF THIS SOFTWARE.
*/
#include "config.h"
#include <stdlib.h>
#include <assert.h>
#include <math.h>
#include "fastapprox/fastpow.h"
#include "brushmodes.h"
#include "helpers.h"
// parameters to those methods:
//
// rgba: A pointer to 16bit rgba data with premultiplied alpha.
// The range of each components is limited from 0 to 2^15.
//
// mask: Contains the dab shape, that is, the intensity of the dab at
// each pixel. Usually rendering is done for one tile at a
// time. The mask is LRE encoded to jump quickly over regions
// that are not affected by the dab.
//
// opacity: overall strength of the blending mode. Has the same
// influence on the dab as the values inside the mask.
// We are manipulating pixels with premultiplied alpha directly.
// This is an "over" operation (opa = topAlpha).
// In the formula below, topColor is assumed to be premultiplied.
//
// opa_a < opa_b >
// resultAlpha = topAlpha + (1.0 - topAlpha) * bottomAlpha
// resultColor = topColor + (1.0 - topAlpha) * bottomColor
//
void draw_dab_pixels_BlendMode_Normal (uint16_t * mask,
uint16_t * rgba,
uint16_t color_r,
uint16_t color_g,
uint16_t color_b,
uint16_t opacity) {
while (1) {
for (; mask[0]; mask++, rgba+=4) {
uint32_t opa_a = mask[0]*(uint32_t)opacity/(1<<15); // topAlpha
uint32_t opa_b = (1<<15)-opa_a; // bottomAlpha
rgba[3] = opa_a + opa_b * rgba[3] / (1<<15);
rgba[0] = (opa_a*color_r + opa_b*rgba[0])/(1<<15);
rgba[1] = (opa_a*color_g + opa_b*rgba[1])/(1<<15);
rgba[2] = (opa_a*color_b + opa_b*rgba[2])/(1<<15);
}
if (!mask[1]) break;
rgba += mask[1];
mask += 2;
}
};
void draw_dab_pixels_BlendMode_Normal_Paint (uint16_t * mask,
uint16_t * rgba,
uint16_t color_r,
uint16_t color_g,
uint16_t color_b,
uint16_t opacity) {
// convert top to spectral. Already straight color
float spectral_a[10] = {0};
rgb_to_spectral((float)color_r / (1 << 15), (float)color_g / (1 << 15), (float)color_b / (1 << 15), spectral_a);
// pigment-mode does not like very low opacity, probably due to rounding
// errors with int->float->int round-trip. Once we convert to pure
// float engine this might be fixed. For now enforce a minimum opacity:
opacity = MAX(opacity, 150);
while (1) {
for (; mask[0]; mask++, rgba+=4) {
uint32_t opa_a = mask[0]*(uint32_t)opacity/(1<<15); // topAlpha
uint32_t opa_b = (1<<15)-opa_a; // bottomAlpha
// optimization- if background has 0 alpha we can just do normal additive
// blending since there is nothing to mix with.
if (rgba[3] <= 0) {
rgba[3] = opa_a + opa_b * rgba[3] / (1<<15);
rgba[0] = (opa_a*color_r + opa_b*rgba[0])/(1<<15);
rgba[1] = (opa_a*color_g + opa_b*rgba[1])/(1<<15);
rgba[2] = (opa_a*color_b + opa_b*rgba[2])/(1<<15);
continue;
}
//alpha-weighted ratio for WGM (sums to 1.0)
float fac_a = (float)opa_a / (opa_a + opa_b * rgba[3] / (1<<15));
float fac_b = 1.0 - fac_a;
//convert bottom to spectral. Un-premult alpha to obtain reflectance
//color noise is not a problem since low alpha also implies low weight
float spectral_b[10] = {0};
rgb_to_spectral((float)rgba[0] / rgba[3], (float)rgba[1] / rgba[3], (float)rgba[2] / rgba[3], spectral_b);
// mix to the two spectral reflectances using WGM
float spectral_result[10] = {0};
for (int i=0; i<10; i++) {
spectral_result[i] = fastpow(spectral_a[i], fac_a) * fastpow(spectral_b[i], fac_b);
}
// convert back to RGB and premultiply alpha
float rgb_result[3] = {0};
spectral_to_rgb(spectral_result, rgb_result);
rgba[3] = opa_a + opa_b * rgba[3] / (1<<15);
for (int i=0; i<3; i++) {
rgba[i] =(rgb_result[i] * rgba[3]) + 0.5;
}
}
if (!mask[1]) break;
rgba += mask[1];
mask += 2;
}
};
//Posterize. Basically exactly like GIMP's posterize
//reduces colors by adjustable amount (posterize_num).
//posterize the canvas, then blend that via opacity
//does not affect alpha
void draw_dab_pixels_BlendMode_Posterize (uint16_t * mask,
uint16_t * rgba,
uint16_t opacity,
uint16_t posterize_num) {
while (1) {
for (; mask[0]; mask++, rgba+=4) {
float r = (float)rgba[0] / (1<<15);
float g = (float)rgba[1] / (1<<15);
float b = (float)rgba[2] / (1<<15);
uint32_t post_r = (1<<15) * ROUND(r * posterize_num) / posterize_num;
uint32_t post_g = (1<<15) * ROUND(g * posterize_num) / posterize_num;
uint32_t post_b = (1<<15) * ROUND(b * posterize_num) / posterize_num;
uint32_t opa_a = mask[0]*(uint32_t)opacity/(1<<15); // topAlpha
uint32_t opa_b = (1<<15)-opa_a; // bottomAlpha
rgba[0] = (opa_a*post_r + opa_b*rgba[0])/(1<<15);
rgba[1] = (opa_a*post_g + opa_b*rgba[1])/(1<<15);
rgba[2] = (opa_a*post_b + opa_b*rgba[2])/(1<<15);
}
if (!mask[1]) break;
rgba += mask[1];
mask += 2;
}
};
// Colorize: apply the source hue and saturation, retaining the target
// brightness. Same thing as in the PDF spec addendum, and upcoming SVG
// compositing drafts. Colorize should be used at either 1.0 or 0.0, values in
// between probably aren't very useful. This blend mode retains the target
// alpha, and any pure whites and blacks in the target layer.
#define MAX3(a, b, c) ((a)>(b)?MAX((a),(c)):MAX((b),(c)))
#define MIN3(a, b, c) ((a)<(b)?MIN((a),(c)):MIN((b),(c)))
// For consistency, these are the values used by MyPaint's Color and
// Luminosity layer blend modes, which in turn are defined by
// http://dvcs.w3.org/hg/FXTF/rawfile/tip/compositing/index.html.
// Same as ITU Rec. BT.601 (SDTV) rounded to 2 decimal places.
static const float LUMA_RED_COEFF = 0.2126 * (1<<15);
static const float LUMA_GREEN_COEFF = 0.7152 * (1<<15);
static const float LUMA_BLUE_COEFF = 0.0722 * (1<<15);
// See also http://en.wikipedia.org/wiki/YCbCr
/* Returns the sRGB luminance of an RGB triple, expressed as scaled ints. */
#define LUMA(r,g,b) \
((r)*LUMA_RED_COEFF + (g)*LUMA_GREEN_COEFF + (b)*LUMA_BLUE_COEFF)
/*
* Sets the output RGB triple's luminance to that of the input, retaining its
* colour. Inputs and outputs are scaled ints having factor 2**-15, and must
* not store premultiplied alpha.
*/
inline static void
set_rgb16_lum_from_rgb16(const uint16_t topr,
const uint16_t topg,
const uint16_t topb,
uint16_t *botr,
uint16_t *botg,
uint16_t *botb)
{
// Spec: SetLum()
// Colours potentially can go out of band to both sides, hence the
// temporary representation inflation.
const uint16_t botlum = LUMA(*botr, *botg, *botb) / (1<<15);
const uint16_t toplum = LUMA(topr, topg, topb) / (1<<15);
const int16_t diff = botlum - toplum;
int32_t r = topr + diff;
int32_t g = topg + diff;
int32_t b = topb + diff;
// Spec: ClipColor()
// Clip out of band values
int32_t lum = LUMA(r, g, b) / (1<<15);
int32_t cmin = MIN3(r, g, b);
int32_t cmax = MAX3(r, g, b);
if (cmin < 0) {
r = lum + (((r - lum) * lum) / (lum - cmin));
g = lum + (((g - lum) * lum) / (lum - cmin));
b = lum + (((b - lum) * lum) / (lum - cmin));
}
if (cmax > (1<<15)) {
r = lum + (((r - lum) * ((1<<15)-lum)) / (cmax - lum));
g = lum + (((g - lum) * ((1<<15)-lum)) / (cmax - lum));
b = lum + (((b - lum) * ((1<<15)-lum)) / (cmax - lum));
}
#ifdef HEAVY_DEBUG
assert((0 <= r) && (r <= (1<<15)));
assert((0 <= g) && (g <= (1<<15)));
assert((0 <= b) && (b <= (1<<15)));
#endif
*botr = r;
*botg = g;
*botb = b;
}
// The method is an implementation of that described in the official Adobe "PDF
// Blend Modes: Addendum" document, dated January 23, 2006; specifically it's
// the "Color" nonseparable blend mode. We do however use different
// coefficients for the Luma value.
void
draw_dab_pixels_BlendMode_Color (uint16_t *mask,
uint16_t *rgba, // b=bottom, premult
uint16_t color_r, // }
uint16_t color_g, // }-- a=top, !premult
uint16_t color_b, // }
uint16_t opacity)
{
while (1) {
for (; mask[0]; mask++, rgba+=4) {
// De-premult
uint16_t r, g, b;
const uint16_t a = rgba[3];
r = g = b = 0;
if (rgba[3] != 0) {
r = ((1<<15)*((uint32_t)rgba[0])) / a;
g = ((1<<15)*((uint32_t)rgba[1])) / a;
b = ((1<<15)*((uint32_t)rgba[2])) / a;
}
// Apply luminance
set_rgb16_lum_from_rgb16(color_r, color_g, color_b, &r, &g, &b);
// Re-premult
r = ((uint32_t) r) * a / (1<<15);
g = ((uint32_t) g) * a / (1<<15);
b = ((uint32_t) b) * a / (1<<15);
// And combine as normal.
uint32_t opa_a = mask[0] * opacity / (1<<15); // topAlpha
uint32_t opa_b = (1<<15) - opa_a; // bottomAlpha
rgba[0] = (opa_a*r + opa_b*rgba[0])/(1<<15);
rgba[1] = (opa_a*g + opa_b*rgba[1])/(1<<15);
rgba[2] = (opa_a*b + opa_b*rgba[2])/(1<<15);
}
if (!mask[1]) break;
rgba += mask[1];
mask += 2;
}
};
// This blend mode is used for smudging and erasing. Smudging
// allows to "drag" around transparency as if it was a color. When
// smuding over a region that is 60% opaque the result will stay 60%
// opaque (color_a=0.6). For normal erasing color_a is set to 0.0
// and color_r/g/b will be ignored. This function can also do normal
// blending (color_a=1.0).
//
void draw_dab_pixels_BlendMode_Normal_and_Eraser (uint16_t * mask,
uint16_t * rgba,
uint16_t color_r,
uint16_t color_g,
uint16_t color_b,
uint16_t color_a,
uint16_t opacity) {
while (1) {
for (; mask[0]; mask++, rgba+=4) {
uint32_t opa_a = mask[0]*(uint32_t)opacity/(1<<15); // topAlpha
uint32_t opa_b = (1<<15)-opa_a; // bottomAlpha
opa_a = opa_a * color_a / (1<<15);
rgba[3] = opa_a + opa_b * rgba[3] / (1<<15);
rgba[0] = (opa_a*color_r + opa_b*rgba[0])/(1<<15);
rgba[1] = (opa_a*color_g + opa_b*rgba[1])/(1<<15);
rgba[2] = (opa_a*color_b + opa_b*rgba[2])/(1<<15);
}
if (!mask[1]) break;
rgba += mask[1];
mask += 2;
}
};
// Fast sigmoid-like function with constant offsets, used to get a
// fairly smooth transition between additive and spectral blending.
float spectral_blend_factor(float x) {
const float ver_fac = 1.65; // vertical compression factor
const float hor_fac = 8.0f; // horizontal compression factor
const float hor_offs = 3.0f; // horizontal offset (slightly left of center)
const float b = x * hor_fac - hor_offs;
return 0.5 + b / (1 + fabsf(b) * ver_fac);
}
void draw_dab_pixels_BlendMode_Normal_and_Eraser_Paint (uint16_t * mask,
uint16_t * rgba,
uint16_t color_r,
uint16_t color_g,
uint16_t color_b,
uint16_t color_a,
uint16_t opacity) {
// Convert input color to spectral, it is not premultiplied
float spectral_a[10] = {0};
rgb_to_spectral(
(float)color_r / (1<<15),
(float)color_g / (1<<15),
(float)color_b / (1<<15),
spectral_a
);
while (1) {
for (; mask[0]; mask++, rgba+=4) {
const uint32_t opa_a = mask[0]*(uint32_t)opacity/(1<<15); // topAlpha
const uint32_t opa_b = (1<<15)-opa_a; // bottomAlpha
const uint32_t opa_a2 = opa_a * color_a / (1<<15); // erase-adjusted alpha
const uint32_t opa_out = opa_a2 + opa_b * rgba[3] / (1<<15);
uint32_t rgb[3] = {0, 0, 0};
// Spectral blending does not handle low transparency well, so we try to patch that
// up by using mostly additive mixing for lower canvas alphas, gradually moving to
// full spectral blending at mostly opaque pixels.
//
// This does not solve all problems with low opacity, and it creates some new ones
// when mixing bright low-opacity colors into dark low-opacity colors, but the new
// artifacts are not as tough to deal with as the old dark-fringe artifacts.
float spectral_factor = CLAMP(spectral_blend_factor((float)rgba[3] / (1<<15)), 0.0f, 1.0f);
float additive_factor = 1.0 - spectral_factor;
if (additive_factor) {
rgb[0] = (opa_a2 * color_r + opa_b * rgba[0]) / (1 << 15);
rgb[1] = (opa_a2 * color_g + opa_b * rgba[1]) / (1 << 15);
rgb[2] = (opa_a2 * color_b + opa_b * rgba[2]) / (1 << 15);
}
if (spectral_factor && rgba[3] != 0) {
// Convert straightened tile pixel color to a spectral
float spectral_b[10] = {0};
rgb_to_spectral(
(float)rgba[0] / rgba[3],
(float)rgba[1] / rgba[3],
(float)rgba[2] / rgba[3],
spectral_b
);
float fac_a = (float)opa_a / (opa_a + opa_b * rgba[3] / (1 << 15));
fac_a *= (float)color_a / (1 << 15);
float fac_b = 1.0 - fac_a;
// Mix input and tile pixel colors using WGM
float spectral_result[10] = {0};
for (int i = 0; i < 10; i++) {
spectral_result[i] =
fastpow(spectral_a[i], fac_a) * fastpow(spectral_b[i], fac_b);
}
// Convert back to RGB
float rgb_result[3] = {0};
spectral_to_rgb(spectral_result, rgb_result);
for (int i = 0; i < 3; i++) {
rgb[i] = (additive_factor * rgb[i]) + (spectral_factor * rgb_result[i] * opa_out);
}
}
rgba[3] = opa_out;
for (int i = 0; i < 3; i++) {
rgba[i] = rgb[i];
}
}
if (!mask[1]) break;
rgba += mask[1];
mask += 2;
}
};
// This is BlendMode_Normal with locked alpha channel.
//
void draw_dab_pixels_BlendMode_LockAlpha (uint16_t * mask,
uint16_t * rgba,
uint16_t color_r,
uint16_t color_g,
uint16_t color_b,
uint16_t opacity) {
while (1) {
for (; mask[0]; mask++, rgba+=4) {
uint32_t opa_a = mask[0]*(uint32_t)opacity/(1<<15); // topAlpha
uint32_t opa_b = (1<<15)-opa_a; // bottomAlpha
opa_a *= rgba[3];
opa_a /= (1<<15);
rgba[0] = (opa_a*color_r + opa_b*rgba[0])/(1<<15);
rgba[1] = (opa_a*color_g + opa_b*rgba[1])/(1<<15);
rgba[2] = (opa_a*color_b + opa_b*rgba[2])/(1<<15);
}
if (!mask[1]) break;
rgba += mask[1];
mask += 2;
}
};
void draw_dab_pixels_BlendMode_LockAlpha_Paint (uint16_t * mask,
uint16_t * rgba,
uint16_t color_r,
uint16_t color_g,
uint16_t color_b,
uint16_t opacity) {
// convert top to spectral. Already straight color
float spectral_a[10] = {0};
rgb_to_spectral((float)color_r / (1<<15), (float)color_g / (1<<15), (float)color_b / (1<<15), spectral_a);
opacity = MAX(opacity, 150);
while (1) {
for (; mask[0]; mask++, rgba+=4) {
uint32_t opa_a = mask[0]*(uint32_t)opacity/(1<<15); // topAlpha
uint32_t opa_b = (1<<15)-opa_a; // bottomAlpha
opa_a *= rgba[3];
opa_a /= (1<<15);
if (rgba[3] <= 0) {
rgba[0] = (opa_a*color_r + opa_b*rgba[0])/(1<<15);
rgba[1] = (opa_a*color_g + opa_b*rgba[1])/(1<<15);
rgba[2] = (opa_a*color_b + opa_b*rgba[2])/(1<<15);
continue;
}
float fac_a = (float)opa_a / (opa_a + opa_b * rgba[3] / (1<<15));
float fac_b = 1.0 - fac_a;
float spectral_b[10] = {0};
rgb_to_spectral((float)rgba[0] / rgba[3], (float)rgba[1] / rgba[3], (float)rgba[2] / rgba[3], spectral_b);
// mix to the two spectral colors using WGM
float spectral_result[10] = {0};
for (int i=0; i<10; i++) {
spectral_result[i] = fastpow(spectral_a[i], fac_a) * fastpow(spectral_b[i], fac_b);
}
// convert back to RGB
float rgb_result[3] = {0};
spectral_to_rgb(spectral_result, rgb_result);
for (int i=0; i<3; i++) {
rgba[i] =(rgb_result[i] * rgba[3]) + 0.5;
}
}
if (!mask[1]) break;
rgba += mask[1];
mask += 2;
}
};
void get_color_pixels_legacy (
uint16_t * mask,
uint16_t * rgba,
float * sum_weight,
float * sum_r,
float * sum_g,
float * sum_b,
float * sum_a
)
{
// The sum of a 64x64 tile fits into a 32 bit integer, but the sum
// of an arbitrary number of tiles may not fit. We assume that we
// are processing a single tile at a time, so we can use integers.
// But for the result we need floats.
uint32_t weight = 0;
uint32_t r = 0;
uint32_t g = 0;
uint32_t b = 0;
uint32_t a = 0;
while (1) {
for (; mask[0]; mask++, rgba+=4) {
uint32_t opa = mask[0];
weight += opa;
r += opa*rgba[0]/(1<<15);
g += opa*rgba[1]/(1<<15);
b += opa*rgba[2]/(1<<15);
a += opa*rgba[3]/(1<<15);
}
if (!mask[1]) break;
rgba += mask[1];
mask += 2;
}
// convert integer to float outside the performance critical loop
*sum_weight += weight;
*sum_r += r;
*sum_g += g;
*sum_b += b;
*sum_a += a;
};
// Sum up the color/alpha components inside the masked region.
// Called by get_color().
//
// The sample interval guarantees that every n pixels are sampled in
// the provided mask segment.
// Setting the interval to 1 means that all pixels will be sampled,
// but note that this may result in large rounding errors.
//
// The sample rate is the probability of any pixel being sampled,
// with the exception of the guaranteed ones. Range: 0.0..1.0.
// The random sample rate can be set to 0, in which case no random
// sampling will occur.
void get_color_pixels_accumulate (uint16_t * mask,
uint16_t * rgba,
float * sum_weight,
float * sum_r,
float * sum_g,
float * sum_b,
float * sum_a,
float paint,
uint16_t sample_interval,
float random_sample_rate
) {
// Fall back to legacy sampling if using static 0 paint setting
// Indicated by passing a negative paint factor (normal range 0..1)
if (paint < 0.0) {
get_color_pixels_legacy(mask, rgba, sum_weight, sum_r, sum_g, sum_b, sum_a);
return;
}
// Sample the canvas as additive and subtractive
// According to paint parameter
// Average the results normally
// Only sample a partially random subset of pixels
float avg_spectral[10] = {0};
float avg_rgb[3] = {*sum_r, *sum_g, *sum_b};
if (paint > 0.0f) {
rgb_to_spectral(*sum_r, *sum_g, *sum_b, avg_spectral);
}
// Rolling counter determining which pixels to sample
// This sampling _is_ biased (but hopefully not too bad).
// Ideally, the selection of pixels to be sampled should
// be determined before this function is called.
uint16_t interval_counter = 0;
const int random_sample_threshold = (int)(random_sample_rate * RAND_MAX);
while (1) {
for (; mask[0]; mask++, rgba+=4) {
// Sample every n pixels, and a percentage of the rest.
// At least one pixel (the first) will always be sampled.
if (interval_counter == 0 || rand() < random_sample_threshold) {
float a = (float)mask[0] * rgba[3] / (1 << 30);
float alpha_sums = a + *sum_a;
*sum_weight += (float)mask[0] / (1 << 15);
float fac_a, fac_b;
fac_a = fac_b = 1.0f;
if (alpha_sums > 0.0f) {
fac_a = a / alpha_sums;
fac_b = 1.0 - fac_a;
}
if (paint > 0.0f && rgba[3] > 0) {
float spectral[10] = {0};
rgb_to_spectral((float)rgba[0] / rgba[3], (float)rgba[1] / rgba[3], (float)rgba[2] / rgba[3], spectral);
for (int i = 0; i < 10; i++) {
avg_spectral[i] = fastpow(spectral[i], fac_a) * fastpow(avg_spectral[i], fac_b);
}
}
if (paint < 1.0f && rgba[3] > 0) {
for (int i = 0; i < 3; i++) {
avg_rgb[i] = (float)rgba[i] * fac_a / rgba[3] + (float)avg_rgb[i] * fac_b;
}
}
*sum_a += a;
}
interval_counter = (interval_counter + 1) % sample_interval;
}
if (!mask[1]) break;
rgba += mask[1];
mask += 2;
}
// Convert the spectral average to rgb and write the result
// back weighted with the rgb average.
float spec_rgb[3] = {0};
spectral_to_rgb(avg_spectral, spec_rgb);
*sum_r = spec_rgb[0] * paint + (1.0 - paint) * avg_rgb[0];
*sum_g = spec_rgb[1] * paint + (1.0 - paint) * avg_rgb[1];
*sum_b = spec_rgb[2] * paint + (1.0 - paint) * avg_rgb[2];
};