1 //---------------------------------------------------------------------------------
2 //
3 // Little Color Management System
4 // Copyright (c) 1998-2013 Marti Maria Saguer
5 //
6 // Permission is hereby granted, free of charge, to any person obtaining
7 // a copy of this software and associated documentation files (the "Software"),
8 // to deal in the Software without restriction, including without limitation
9 // the rights to use, copy, modify, merge, publish, distribute, sublicense,
10 // and/or sell copies of the Software, and to permit persons to whom the Software
11 // is furnished to do so, subject to the following conditions:
12 //
13 // The above copyright notice and this permission notice shall be included in
14 // all copies or substantial portions of the Software.
15 //
16 // THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND,
17 // EXPRESS OR IMPLIED, INCLUDING BUT NOT LIMITED TO
18 // THE WARRANTIES OF MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE AND
19 // NONINFRINGEMENT. IN NO EVENT SHALL THE AUTHORS OR COPYRIGHT HOLDERS BE
20 // LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN AN ACTION
21 // OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN CONNECTION
22 // WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE.
23 //
24 //---------------------------------------------------------------------------------
25 //
26 #include "lcms2_internal.h"
27
28 // Tone curves are powerful constructs that can contain curves specified in diverse ways.
29 // The curve is stored in segments, where each segment can be sampled or specified by parameters.
30 // a 16.bit simplification of the *whole* curve is kept for optimization purposes. For float operation,
31 // each segment is evaluated separately. Plug-ins may be used to define new parametric schemes,
32 // each plug-in may define up to MAX_TYPES_IN_LCMS_PLUGIN functions types. For defining a function,
33 // the plug-in should provide the type id, how many parameters each type has, and a pointer to
34 // a procedure that evaluates the function. In the case of reverse evaluation, the evaluator will
35 // be called with the type id as a negative value, and a sampled version of the reversed curve
36 // will be built.
37
38 // ----------------------------------------------------------------- Implementation
39 // Maxim number of nodes
40 #define MAX_NODES_IN_CURVE 4097
41 #define MINUS_INF (-1E22F)
42 #define PLUS_INF (+1E22F)
43
44 // The list of supported parametric curves
45 typedef struct _cmsParametricCurvesCollection_st {
46
47 int nFunctions; // Number of supported functions in this chunk
48 int FunctionTypes[MAX_TYPES_IN_LCMS_PLUGIN]; // The identification types
49 int ParameterCount[MAX_TYPES_IN_LCMS_PLUGIN]; // Number of parameters for each function
50 cmsParametricCurveEvaluator Evaluator; // The evaluator
51
52 struct _cmsParametricCurvesCollection_st* Next; // Next in list
53
54 } _cmsParametricCurvesCollection;
55
56 // This is the default (built-in) evaluator
57 static cmsFloat64Number DefaultEvalParametricFn(cmsInt32Number Type, const cmsFloat64Number Params[], cmsFloat64Number R);
58
59 // The built-in list
60 static _cmsParametricCurvesCollection DefaultCurves = {
61 9, // # of curve types
62 { 1, 2, 3, 4, 5, 6, 7, 8, 108 }, // Parametric curve ID
63 { 1, 3, 4, 5, 7, 4, 5, 5, 1 }, // Parameters by type
64 DefaultEvalParametricFn, // Evaluator
65 NULL // Next in chain
66 };
67
68 // Duplicates the zone of memory used by the plug-in in the new context
69 static
DupPluginCurvesList(struct _cmsContext_struct * ctx,const struct _cmsContext_struct * src)70 void DupPluginCurvesList(struct _cmsContext_struct* ctx,
71 const struct _cmsContext_struct* src)
72 {
73 _cmsCurvesPluginChunkType newHead = { NULL };
74 _cmsParametricCurvesCollection* entry;
75 _cmsParametricCurvesCollection* Anterior = NULL;
76 _cmsCurvesPluginChunkType* head = (_cmsCurvesPluginChunkType*) src->chunks[CurvesPlugin];
77
78 _cmsAssert(head != NULL);
79
80 // Walk the list copying all nodes
81 for (entry = head->ParametricCurves;
82 entry != NULL;
83 entry = entry ->Next) {
84
85 _cmsParametricCurvesCollection *newEntry = ( _cmsParametricCurvesCollection *) _cmsSubAllocDup(ctx ->MemPool, entry, sizeof(_cmsParametricCurvesCollection));
86
87 if (newEntry == NULL)
88 return;
89
90 // We want to keep the linked list order, so this is a little bit tricky
91 newEntry -> Next = NULL;
92 if (Anterior)
93 Anterior -> Next = newEntry;
94
95 Anterior = newEntry;
96
97 if (newHead.ParametricCurves == NULL)
98 newHead.ParametricCurves = newEntry;
99 }
100
101 ctx ->chunks[CurvesPlugin] = _cmsSubAllocDup(ctx->MemPool, &newHead, sizeof(_cmsCurvesPluginChunkType));
102 }
103
104 // The allocator have to follow the chain
_cmsAllocCurvesPluginChunk(struct _cmsContext_struct * ctx,const struct _cmsContext_struct * src)105 void _cmsAllocCurvesPluginChunk(struct _cmsContext_struct* ctx,
106 const struct _cmsContext_struct* src)
107 {
108 _cmsAssert(ctx != NULL);
109
110 if (src != NULL) {
111
112 // Copy all linked list
113 DupPluginCurvesList(ctx, src);
114 }
115 else {
116 static _cmsCurvesPluginChunkType CurvesPluginChunk = { NULL };
117 ctx ->chunks[CurvesPlugin] = _cmsSubAllocDup(ctx ->MemPool, &CurvesPluginChunk, sizeof(_cmsCurvesPluginChunkType));
118 }
119 }
120
121
122 // The linked list head
123 _cmsCurvesPluginChunkType _cmsCurvesPluginChunk = { NULL };
124
125 // As a way to install new parametric curves
_cmsRegisterParametricCurvesPlugin(cmsContext ContextID,cmsPluginBase * Data)126 cmsBool _cmsRegisterParametricCurvesPlugin(cmsContext ContextID, cmsPluginBase* Data)
127 {
128 _cmsCurvesPluginChunkType* ctx = ( _cmsCurvesPluginChunkType*) _cmsContextGetClientChunk(ContextID, CurvesPlugin);
129 cmsPluginParametricCurves* Plugin = (cmsPluginParametricCurves*) Data;
130 _cmsParametricCurvesCollection* fl;
131
132 if (Data == NULL) {
133
134 ctx -> ParametricCurves = NULL;
135 return TRUE;
136 }
137
138 fl = (_cmsParametricCurvesCollection*) _cmsPluginMalloc(ContextID, sizeof(_cmsParametricCurvesCollection));
139 if (fl == NULL) return FALSE;
140
141 // Copy the parameters
142 fl ->Evaluator = Plugin ->Evaluator;
143 fl ->nFunctions = Plugin ->nFunctions;
144
145 // Make sure no mem overwrites
146 if (fl ->nFunctions > MAX_TYPES_IN_LCMS_PLUGIN)
147 fl ->nFunctions = MAX_TYPES_IN_LCMS_PLUGIN;
148
149 // Copy the data
150 memmove(fl->FunctionTypes, Plugin ->FunctionTypes, fl->nFunctions * sizeof(cmsUInt32Number));
151 memmove(fl->ParameterCount, Plugin ->ParameterCount, fl->nFunctions * sizeof(cmsUInt32Number));
152
153 // Keep linked list
154 fl ->Next = ctx->ParametricCurves;
155 ctx->ParametricCurves = fl;
156
157 // All is ok
158 return TRUE;
159 }
160
161
162 // Search in type list, return position or -1 if not found
163 static
IsInSet(int Type,_cmsParametricCurvesCollection * c)164 int IsInSet(int Type, _cmsParametricCurvesCollection* c)
165 {
166 int i;
167
168 for (i=0; i < c ->nFunctions; i++)
169 if (abs(Type) == c ->FunctionTypes[i]) return i;
170
171 return -1;
172 }
173
174
175 // Search for the collection which contains a specific type
176 static
GetParametricCurveByType(cmsContext ContextID,int Type,int * index)177 _cmsParametricCurvesCollection *GetParametricCurveByType(cmsContext ContextID, int Type, int* index)
178 {
179 _cmsParametricCurvesCollection* c;
180 int Position;
181 _cmsCurvesPluginChunkType* ctx = ( _cmsCurvesPluginChunkType*) _cmsContextGetClientChunk(ContextID, CurvesPlugin);
182
183 for (c = ctx->ParametricCurves; c != NULL; c = c ->Next) {
184
185 Position = IsInSet(Type, c);
186
187 if (Position != -1) {
188 if (index != NULL)
189 *index = Position;
190 return c;
191 }
192 }
193 // If none found, revert for defaults
194 for (c = &DefaultCurves; c != NULL; c = c ->Next) {
195
196 Position = IsInSet(Type, c);
197
198 if (Position != -1) {
199 if (index != NULL)
200 *index = Position;
201 return c;
202 }
203 }
204
205 return NULL;
206 }
207
208 // Low level allocate, which takes care of memory details. nEntries may be zero, and in this case
209 // no optimation curve is computed. nSegments may also be zero in the inverse case, where only the
210 // optimization curve is given. Both features simultaneously is an error
211 static
AllocateToneCurveStruct(cmsContext ContextID,cmsInt32Number nEntries,cmsInt32Number nSegments,const cmsCurveSegment * Segments,const cmsUInt16Number * Values)212 cmsToneCurve* AllocateToneCurveStruct(cmsContext ContextID, cmsInt32Number nEntries,
213 cmsInt32Number nSegments, const cmsCurveSegment* Segments,
214 const cmsUInt16Number* Values)
215 {
216 cmsToneCurve* p;
217 int i;
218
219 // We allow huge tables, which are then restricted for smoothing operations
220 if (nEntries > 65530 || nEntries < 0) {
221 cmsSignalError(ContextID, cmsERROR_RANGE, "Couldn't create tone curve of more than 65530 entries");
222 return NULL;
223 }
224
225 if (nEntries <= 0 && nSegments <= 0) {
226 cmsSignalError(ContextID, cmsERROR_RANGE, "Couldn't create tone curve with zero segments and no table");
227 return NULL;
228 }
229
230 // Allocate all required pointers, etc.
231 p = (cmsToneCurve*) _cmsMallocZero(ContextID, sizeof(cmsToneCurve));
232 if (!p) return NULL;
233
234 // In this case, there are no segments
235 if (nSegments <= 0) {
236 p ->Segments = NULL;
237 p ->Evals = NULL;
238 }
239 else {
240 p ->Segments = (cmsCurveSegment*) _cmsCalloc(ContextID, nSegments, sizeof(cmsCurveSegment));
241 if (p ->Segments == NULL) goto Error;
242
243 p ->Evals = (cmsParametricCurveEvaluator*) _cmsCalloc(ContextID, nSegments, sizeof(cmsParametricCurveEvaluator));
244 if (p ->Evals == NULL) goto Error;
245 }
246
247 p -> nSegments = nSegments;
248
249 // This 16-bit table contains a limited precision representation of the whole curve and is kept for
250 // increasing xput on certain operations.
251 if (nEntries <= 0) {
252 p ->Table16 = NULL;
253 }
254 else {
255 p ->Table16 = (cmsUInt16Number*) _cmsCalloc(ContextID, nEntries, sizeof(cmsUInt16Number));
256 if (p ->Table16 == NULL) goto Error;
257 }
258
259 p -> nEntries = nEntries;
260
261 // Initialize members if requested
262 if (Values != NULL && (nEntries > 0)) {
263
264 for (i=0; i < nEntries; i++)
265 p ->Table16[i] = Values[i];
266 }
267
268 // Initialize the segments stuff. The evaluator for each segment is located and a pointer to it
269 // is placed in advance to maximize performance.
270 if (Segments != NULL && (nSegments > 0)) {
271
272 _cmsParametricCurvesCollection *c;
273
274 p ->SegInterp = (cmsInterpParams**) _cmsCalloc(ContextID, nSegments, sizeof(cmsInterpParams*));
275 if (p ->SegInterp == NULL) goto Error;
276
277 for (i=0; i< nSegments; i++) {
278
279 // Type 0 is a special marker for table-based curves
280 if (Segments[i].Type == 0)
281 p ->SegInterp[i] = _cmsComputeInterpParams(ContextID, Segments[i].nGridPoints, 1, 1, NULL, CMS_LERP_FLAGS_FLOAT);
282
283 memmove(&p ->Segments[i], &Segments[i], sizeof(cmsCurveSegment));
284
285 if (Segments[i].Type == 0 && Segments[i].SampledPoints != NULL)
286 p ->Segments[i].SampledPoints = (cmsFloat32Number*) _cmsDupMem(ContextID, Segments[i].SampledPoints, sizeof(cmsFloat32Number) * Segments[i].nGridPoints);
287 else
288 p ->Segments[i].SampledPoints = NULL;
289
290
291 c = GetParametricCurveByType(ContextID, Segments[i].Type, NULL);
292 if (c != NULL)
293 p ->Evals[i] = c ->Evaluator;
294 }
295 }
296
297 p ->InterpParams = _cmsComputeInterpParams(ContextID, p ->nEntries, 1, 1, p->Table16, CMS_LERP_FLAGS_16BITS);
298 if (p->InterpParams != NULL)
299 return p;
300
301 Error:
302 if (p -> Segments) _cmsFree(ContextID, p ->Segments);
303 if (p -> Evals) _cmsFree(ContextID, p -> Evals);
304 if (p ->Table16) _cmsFree(ContextID, p ->Table16);
305 _cmsFree(ContextID, p);
306 return NULL;
307 }
308
309
310 // Parametric Fn using floating point
311 static
DefaultEvalParametricFn(cmsInt32Number Type,const cmsFloat64Number Params[],cmsFloat64Number R)312 cmsFloat64Number DefaultEvalParametricFn(cmsInt32Number Type, const cmsFloat64Number Params[], cmsFloat64Number R)
313 {
314 cmsFloat64Number e, Val, disc;
315
316 switch (Type) {
317
318 // X = Y ^ Gamma
319 case 1:
320 if (R < 0) {
321
322 if (fabs(Params[0] - 1.0) < MATRIX_DET_TOLERANCE)
323 Val = R;
324 else
325 Val = 0;
326 }
327 else
328 Val = pow(R, Params[0]);
329 break;
330
331 // Type 1 Reversed: X = Y ^1/gamma
332 case -1:
333 if (R < 0) {
334
335 if (fabs(Params[0] - 1.0) < MATRIX_DET_TOLERANCE)
336 Val = R;
337 else
338 Val = 0;
339 }
340 else
341 Val = pow(R, 1/Params[0]);
342 break;
343
344 // CIE 122-1966
345 // Y = (aX + b)^Gamma | X >= -b/a
346 // Y = 0 | else
347 case 2:
348 disc = -Params[2] / Params[1];
349
350 if (R >= disc ) {
351
352 e = Params[1]*R + Params[2];
353
354 if (e > 0)
355 Val = pow(e, Params[0]);
356 else
357 Val = 0;
358 }
359 else
360 Val = 0;
361 break;
362
363 // Type 2 Reversed
364 // X = (Y ^1/g - b) / a
365 case -2:
366 if (R < 0)
367 Val = 0;
368 else
369 Val = (pow(R, 1.0/Params[0]) - Params[2]) / Params[1];
370
371 if (Val < 0)
372 Val = 0;
373 break;
374
375
376 // IEC 61966-3
377 // Y = (aX + b)^Gamma | X <= -b/a
378 // Y = c | else
379 case 3:
380 disc = -Params[2] / Params[1];
381 if (disc < 0)
382 disc = 0;
383
384 if (R >= disc) {
385
386 e = Params[1]*R + Params[2];
387
388 if (e > 0)
389 Val = pow(e, Params[0]) + Params[3];
390 else
391 Val = 0;
392 }
393 else
394 Val = Params[3];
395 break;
396
397
398 // Type 3 reversed
399 // X=((Y-c)^1/g - b)/a | (Y>=c)
400 // X=-b/a | (Y<c)
401 case -3:
402 if (R >= Params[3]) {
403
404 e = R - Params[3];
405
406 if (e > 0)
407 Val = (pow(e, 1/Params[0]) - Params[2]) / Params[1];
408 else
409 Val = 0;
410 }
411 else {
412 Val = -Params[2] / Params[1];
413 }
414 break;
415
416
417 // IEC 61966-2.1 (sRGB)
418 // Y = (aX + b)^Gamma | X >= d
419 // Y = cX | X < d
420 case 4:
421 if (R >= Params[4]) {
422
423 e = Params[1]*R + Params[2];
424
425 if (e > 0)
426 Val = pow(e, Params[0]);
427 else
428 Val = 0;
429 }
430 else
431 Val = R * Params[3];
432 break;
433
434 // Type 4 reversed
435 // X=((Y^1/g-b)/a) | Y >= (ad+b)^g
436 // X=Y/c | Y< (ad+b)^g
437 case -4:
438 e = Params[1] * Params[4] + Params[2];
439 if (e < 0)
440 disc = 0;
441 else
442 disc = pow(e, Params[0]);
443
444 if (R >= disc) {
445
446 Val = (pow(R, 1.0/Params[0]) - Params[2]) / Params[1];
447 }
448 else {
449 Val = R / Params[3];
450 }
451 break;
452
453
454 // Y = (aX + b)^Gamma + e | X >= d
455 // Y = cX + f | X < d
456 case 5:
457 if (R >= Params[4]) {
458
459 e = Params[1]*R + Params[2];
460
461 if (e > 0)
462 Val = pow(e, Params[0]) + Params[5];
463 else
464 Val = Params[5];
465 }
466 else
467 Val = R*Params[3] + Params[6];
468 break;
469
470
471 // Reversed type 5
472 // X=((Y-e)1/g-b)/a | Y >=(ad+b)^g+e), cd+f
473 // X=(Y-f)/c | else
474 case -5:
475
476 disc = Params[3] * Params[4] + Params[6];
477 if (R >= disc) {
478
479 e = R - Params[5];
480 if (e < 0)
481 Val = 0;
482 else
483 Val = (pow(e, 1.0/Params[0]) - Params[2]) / Params[1];
484 }
485 else {
486 Val = (R - Params[6]) / Params[3];
487 }
488 break;
489
490
491 // Types 6,7,8 comes from segmented curves as described in ICCSpecRevision_02_11_06_Float.pdf
492 // Type 6 is basically identical to type 5 without d
493
494 // Y = (a * X + b) ^ Gamma + c
495 case 6:
496 e = Params[1]*R + Params[2];
497
498 if (e < 0)
499 Val = Params[3];
500 else
501 Val = pow(e, Params[0]) + Params[3];
502 break;
503
504 // ((Y - c) ^1/Gamma - b) / a
505 case -6:
506 e = R - Params[3];
507 if (e < 0)
508 Val = 0;
509 else
510 Val = (pow(e, 1.0/Params[0]) - Params[2]) / Params[1];
511 break;
512
513
514 // Y = a * log (b * X^Gamma + c) + d
515 case 7:
516
517 e = Params[2] * pow(R, Params[0]) + Params[3];
518 if (e <= 0)
519 Val = Params[4];
520 else
521 Val = Params[1]*log10(e) + Params[4];
522 break;
523
524 // (Y - d) / a = log(b * X ^Gamma + c)
525 // pow(10, (Y-d) / a) = b * X ^Gamma + c
526 // pow((pow(10, (Y-d) / a) - c) / b, 1/g) = X
527 case -7:
528 Val = pow((pow(10.0, (R-Params[4]) / Params[1]) - Params[3]) / Params[2], 1.0 / Params[0]);
529 break;
530
531
532 //Y = a * b^(c*X+d) + e
533 case 8:
534 Val = (Params[0] * pow(Params[1], Params[2] * R + Params[3]) + Params[4]);
535 break;
536
537
538 // Y = (log((y-e) / a) / log(b) - d ) / c
539 // a=0, b=1, c=2, d=3, e=4,
540 case -8:
541
542 disc = R - Params[4];
543 if (disc < 0) Val = 0;
544 else
545 Val = (log(disc / Params[0]) / log(Params[1]) - Params[3]) / Params[2];
546 break;
547
548 // S-Shaped: (1 - (1-x)^1/g)^1/g
549 case 108:
550 Val = pow(1.0 - pow(1 - R, 1/Params[0]), 1/Params[0]);
551 break;
552
553 // y = (1 - (1-x)^1/g)^1/g
554 // y^g = (1 - (1-x)^1/g)
555 // 1 - y^g = (1-x)^1/g
556 // (1 - y^g)^g = 1 - x
557 // 1 - (1 - y^g)^g
558 case -108:
559 Val = 1 - pow(1 - pow(R, Params[0]), Params[0]);
560 break;
561
562 default:
563 // Unsupported parametric curve. Should never reach here
564 return 0;
565 }
566
567 return Val;
568 }
569
570 // Evaluate a segmented function for a single value. Return -1 if no valid segment found .
571 // If fn type is 0, perform an interpolation on the table
572 static
EvalSegmentedFn(const cmsToneCurve * g,cmsFloat64Number R)573 cmsFloat64Number EvalSegmentedFn(const cmsToneCurve *g, cmsFloat64Number R)
574 {
575 int i;
576
577 for (i = g ->nSegments-1; i >= 0 ; --i) {
578
579 // Check for domain
580 if ((R > g ->Segments[i].x0) && (R <= g ->Segments[i].x1)) {
581
582 // Type == 0 means segment is sampled
583 if (g ->Segments[i].Type == 0) {
584
585 cmsFloat32Number R1 = (cmsFloat32Number) (R - g ->Segments[i].x0) / (g ->Segments[i].x1 - g ->Segments[i].x0);
586 cmsFloat32Number Out;
587
588 // Setup the table (TODO: clean that)
589 g ->SegInterp[i]-> Table = g ->Segments[i].SampledPoints;
590
591 g ->SegInterp[i] -> Interpolation.LerpFloat(&R1, &Out, g ->SegInterp[i]);
592
593 return Out;
594 }
595 else
596 return g ->Evals[i](g->Segments[i].Type, g ->Segments[i].Params, R);
597 }
598 }
599
600 return MINUS_INF;
601 }
602
603 // Access to estimated low-res table
cmsGetToneCurveEstimatedTableEntries(const cmsToneCurve * t)604 cmsUInt32Number CMSEXPORT cmsGetToneCurveEstimatedTableEntries(const cmsToneCurve* t)
605 {
606 _cmsAssert(t != NULL);
607 return t ->nEntries;
608 }
609
cmsGetToneCurveEstimatedTable(const cmsToneCurve * t)610 const cmsUInt16Number* CMSEXPORT cmsGetToneCurveEstimatedTable(const cmsToneCurve* t)
611 {
612 _cmsAssert(t != NULL);
613 return t ->Table16;
614 }
615
616
617 // Create an empty gamma curve, by using tables. This specifies only the limited-precision part, and leaves the
618 // floating point description empty.
cmsBuildTabulatedToneCurve16(cmsContext ContextID,cmsInt32Number nEntries,const cmsUInt16Number Values[])619 cmsToneCurve* CMSEXPORT cmsBuildTabulatedToneCurve16(cmsContext ContextID, cmsInt32Number nEntries, const cmsUInt16Number Values[])
620 {
621 return AllocateToneCurveStruct(ContextID, nEntries, 0, NULL, Values);
622 }
623
624 static
EntriesByGamma(cmsFloat64Number Gamma)625 int EntriesByGamma(cmsFloat64Number Gamma)
626 {
627 if (fabs(Gamma - 1.0) < 0.001) return 2;
628 return 4096;
629 }
630
631
632 // Create a segmented gamma, fill the table
cmsBuildSegmentedToneCurve(cmsContext ContextID,cmsInt32Number nSegments,const cmsCurveSegment Segments[])633 cmsToneCurve* CMSEXPORT cmsBuildSegmentedToneCurve(cmsContext ContextID,
634 cmsInt32Number nSegments, const cmsCurveSegment Segments[])
635 {
636 int i;
637 cmsFloat64Number R, Val;
638 cmsToneCurve* g;
639 int nGridPoints = 4096;
640
641 _cmsAssert(Segments != NULL);
642
643 // Optimizatin for identity curves.
644 if (nSegments == 1 && Segments[0].Type == 1) {
645
646 nGridPoints = EntriesByGamma(Segments[0].Params[0]);
647 }
648
649 g = AllocateToneCurveStruct(ContextID, nGridPoints, nSegments, Segments, NULL);
650 if (g == NULL) return NULL;
651
652 // Once we have the floating point version, we can approximate a 16 bit table of 4096 entries
653 // for performance reasons. This table would normally not be used except on 8/16 bits transforms.
654 for (i=0; i < nGridPoints; i++) {
655
656 R = (cmsFloat64Number) i / (nGridPoints-1);
657
658 Val = EvalSegmentedFn(g, R);
659
660 // Round and saturate
661 g ->Table16[i] = _cmsQuickSaturateWord(Val * 65535.0);
662 }
663
664 return g;
665 }
666
667 // Use a segmented curve to store the floating point table
cmsBuildTabulatedToneCurveFloat(cmsContext ContextID,cmsUInt32Number nEntries,const cmsFloat32Number values[])668 cmsToneCurve* CMSEXPORT cmsBuildTabulatedToneCurveFloat(cmsContext ContextID, cmsUInt32Number nEntries, const cmsFloat32Number values[])
669 {
670 cmsCurveSegment Seg[3];
671
672 // A segmented tone curve should have function segments in the first and last positions
673 // Initialize segmented curve part up to 0 to constant value = samples[0]
674 Seg[0].x0 = MINUS_INF;
675 Seg[0].x1 = 0;
676 Seg[0].Type = 6;
677
678 Seg[0].Params[0] = 1;
679 Seg[0].Params[1] = 0;
680 Seg[0].Params[2] = 0;
681 Seg[0].Params[3] = values[0];
682 Seg[0].Params[4] = 0;
683
684 // From zero to 1
685 Seg[1].x0 = 0;
686 Seg[1].x1 = 1.0;
687 Seg[1].Type = 0;
688
689 Seg[1].nGridPoints = nEntries;
690 Seg[1].SampledPoints = (cmsFloat32Number*) values;
691
692 // Final segment is constant = lastsample
693 Seg[2].x0 = 1.0;
694 Seg[2].x1 = PLUS_INF;
695 Seg[2].Type = 6;
696
697 Seg[2].Params[0] = 1;
698 Seg[2].Params[1] = 0;
699 Seg[2].Params[2] = 0;
700 Seg[2].Params[3] = values[nEntries-1];
701 Seg[2].Params[4] = 0;
702
703
704 return cmsBuildSegmentedToneCurve(ContextID, 3, Seg);
705 }
706
707 // Parametric curves
708 //
709 // Parameters goes as: Curve, a, b, c, d, e, f
710 // Type is the ICC type +1
711 // if type is negative, then the curve is analyticaly inverted
cmsBuildParametricToneCurve(cmsContext ContextID,cmsInt32Number Type,const cmsFloat64Number Params[])712 cmsToneCurve* CMSEXPORT cmsBuildParametricToneCurve(cmsContext ContextID, cmsInt32Number Type, const cmsFloat64Number Params[])
713 {
714 cmsCurveSegment Seg0;
715 int Pos = 0;
716 cmsUInt32Number size;
717 _cmsParametricCurvesCollection* c = GetParametricCurveByType(ContextID, Type, &Pos);
718
719 _cmsAssert(Params != NULL);
720
721 if (c == NULL) {
722 cmsSignalError(ContextID, cmsERROR_UNKNOWN_EXTENSION, "Invalid parametric curve type %d", Type);
723 return NULL;
724 }
725
726 memset(&Seg0, 0, sizeof(Seg0));
727
728 Seg0.x0 = MINUS_INF;
729 Seg0.x1 = PLUS_INF;
730 Seg0.Type = Type;
731
732 size = c->ParameterCount[Pos] * sizeof(cmsFloat64Number);
733 memmove(Seg0.Params, Params, size);
734
735 return cmsBuildSegmentedToneCurve(ContextID, 1, &Seg0);
736 }
737
738
739
740 // Build a gamma table based on gamma constant
cmsBuildGamma(cmsContext ContextID,cmsFloat64Number Gamma)741 cmsToneCurve* CMSEXPORT cmsBuildGamma(cmsContext ContextID, cmsFloat64Number Gamma)
742 {
743 return cmsBuildParametricToneCurve(ContextID, 1, &Gamma);
744 }
745
746
747 // Free all memory taken by the gamma curve
cmsFreeToneCurve(cmsToneCurve * Curve)748 void CMSEXPORT cmsFreeToneCurve(cmsToneCurve* Curve)
749 {
750 cmsContext ContextID;
751
752 // added by Xiaochuan Liu
753 // Curve->InterpParams may be null
754 if (Curve == NULL || Curve->InterpParams == NULL) return;
755
756 ContextID = Curve ->InterpParams->ContextID;
757
758 _cmsFreeInterpParams(Curve ->InterpParams);
759 Curve ->InterpParams = NULL;
760
761 if (Curve -> Table16) {
762 _cmsFree(ContextID, Curve ->Table16);
763 Curve ->Table16 = NULL;
764 }
765
766 if (Curve ->Segments) {
767
768 cmsUInt32Number i;
769
770 for (i=0; i < Curve ->nSegments; i++) {
771
772 if (Curve ->Segments[i].SampledPoints) {
773 _cmsFree(ContextID, Curve ->Segments[i].SampledPoints);
774 Curve ->Segments[i].SampledPoints = NULL;
775 }
776
777 if (Curve ->SegInterp[i] != 0) {
778 _cmsFreeInterpParams(Curve->SegInterp[i]);
779 Curve->SegInterp[i] = NULL;
780 }
781 }
782
783 _cmsFree(ContextID, Curve ->Segments);
784 Curve ->Segments = NULL;
785 _cmsFree(ContextID, Curve ->SegInterp);
786 Curve ->SegInterp = NULL;
787 }
788
789 if (Curve -> Evals) {
790 _cmsFree(ContextID, Curve -> Evals);
791 Curve -> Evals = NULL;
792 }
793
794 if (Curve) {
795 _cmsFree(ContextID, Curve);
796 Curve = NULL;
797 }
798 }
799
800 // Utility function, free 3 gamma tables
cmsFreeToneCurveTriple(cmsToneCurve * Curve[3])801 void CMSEXPORT cmsFreeToneCurveTriple(cmsToneCurve* Curve[3])
802 {
803
804 _cmsAssert(Curve != NULL);
805
806 if (Curve[0] != NULL) cmsFreeToneCurve(Curve[0]);
807 if (Curve[1] != NULL) cmsFreeToneCurve(Curve[1]);
808 if (Curve[2] != NULL) cmsFreeToneCurve(Curve[2]);
809
810 Curve[0] = Curve[1] = Curve[2] = NULL;
811 }
812
813
814 // Duplicate a gamma table
cmsDupToneCurve(const cmsToneCurve * In)815 cmsToneCurve* CMSEXPORT cmsDupToneCurve(const cmsToneCurve* In)
816 {
817 // Xiaochuan Liu
818 // fix openpdf bug(mantis id:0055683, google id:360198)
819 // the function CurveSetElemTypeFree in cmslut.c also needs to check pointer
820 if (In == NULL || In ->InterpParams == NULL || In ->Segments == NULL || In ->Table16 == NULL) return NULL;
821
822 return AllocateToneCurveStruct(In ->InterpParams ->ContextID, In ->nEntries, In ->nSegments, In ->Segments, In ->Table16);
823 }
824
825 // Joins two curves for X and Y. Curves should be monotonic.
826 // We want to get
827 //
828 // y = Y^-1(X(t))
829 //
cmsJoinToneCurve(cmsContext ContextID,const cmsToneCurve * X,const cmsToneCurve * Y,cmsUInt32Number nResultingPoints)830 cmsToneCurve* CMSEXPORT cmsJoinToneCurve(cmsContext ContextID,
831 const cmsToneCurve* X,
832 const cmsToneCurve* Y, cmsUInt32Number nResultingPoints)
833 {
834 cmsToneCurve* out = NULL;
835 cmsToneCurve* Yreversed = NULL;
836 cmsFloat32Number t, x;
837 cmsFloat32Number* Res = NULL;
838 cmsUInt32Number i;
839
840
841 _cmsAssert(X != NULL);
842 _cmsAssert(Y != NULL);
843
844 Yreversed = cmsReverseToneCurveEx(nResultingPoints, Y);
845 if (Yreversed == NULL) goto Error;
846
847 Res = (cmsFloat32Number*) _cmsCalloc(ContextID, nResultingPoints, sizeof(cmsFloat32Number));
848 if (Res == NULL) goto Error;
849
850 //Iterate
851 for (i=0; i < nResultingPoints; i++) {
852
853 t = (cmsFloat32Number) i / (nResultingPoints-1);
854 x = cmsEvalToneCurveFloat(X, t);
855 Res[i] = cmsEvalToneCurveFloat(Yreversed, x);
856 }
857
858 // Allocate space for output
859 out = cmsBuildTabulatedToneCurveFloat(ContextID, nResultingPoints, Res);
860
861 Error:
862
863 if (Res != NULL) _cmsFree(ContextID, Res);
864 if (Yreversed != NULL) cmsFreeToneCurve(Yreversed);
865
866 return out;
867 }
868
869
870
871 // Get the surrounding nodes. This is tricky on non-monotonic tables
872 static
GetInterval(cmsFloat64Number In,const cmsUInt16Number LutTable[],const struct _cms_interp_struc * p)873 int GetInterval(cmsFloat64Number In, const cmsUInt16Number LutTable[], const struct _cms_interp_struc* p)
874 {
875 int i;
876 int y0, y1;
877
878 // A 1 point table is not allowed
879 if (p -> Domain[0] < 1) return -1;
880
881 // Let's see if ascending or descending.
882 if (LutTable[0] < LutTable[p ->Domain[0]]) {
883
884 // Table is overall ascending
885 for (i=p->Domain[0]-1; i >=0; --i) {
886
887 y0 = LutTable[i];
888 y1 = LutTable[i+1];
889
890 if (y0 <= y1) { // Increasing
891 if (In >= y0 && In <= y1) return i;
892 }
893 else
894 if (y1 < y0) { // Decreasing
895 if (In >= y1 && In <= y0) return i;
896 }
897 }
898 }
899 else {
900 // Table is overall descending
901 for (i=0; i < (int) p -> Domain[0]; i++) {
902
903 y0 = LutTable[i];
904 y1 = LutTable[i+1];
905
906 if (y0 <= y1) { // Increasing
907 if (In >= y0 && In <= y1) return i;
908 }
909 else
910 if (y1 < y0) { // Decreasing
911 if (In >= y1 && In <= y0) return i;
912 }
913 }
914 }
915
916 return -1;
917 }
918
919 // Reverse a gamma table
cmsReverseToneCurveEx(cmsInt32Number nResultSamples,const cmsToneCurve * InCurve)920 cmsToneCurve* CMSEXPORT cmsReverseToneCurveEx(cmsInt32Number nResultSamples, const cmsToneCurve* InCurve)
921 {
922 cmsToneCurve *out;
923 cmsFloat64Number a = 0, b = 0, y, x1, y1, x2, y2;
924 int i, j;
925 int Ascending;
926
927 _cmsAssert(InCurve != NULL);
928
929 // Try to reverse it analytically whatever possible
930
931 if (InCurve ->nSegments == 1 && InCurve ->Segments[0].Type > 0 &&
932 /* InCurve -> Segments[0].Type <= 5 */
933 GetParametricCurveByType(InCurve ->InterpParams->ContextID, InCurve ->Segments[0].Type, NULL) != NULL) {
934
935 return cmsBuildParametricToneCurve(InCurve ->InterpParams->ContextID,
936 -(InCurve -> Segments[0].Type),
937 InCurve -> Segments[0].Params);
938 }
939
940 // Nope, reverse the table.
941 out = cmsBuildTabulatedToneCurve16(InCurve ->InterpParams->ContextID, nResultSamples, NULL);
942 if (out == NULL)
943 return NULL;
944
945 // We want to know if this is an ascending or descending table
946 Ascending = !cmsIsToneCurveDescending(InCurve);
947
948 // Iterate across Y axis
949 for (i=0; i < nResultSamples; i++) {
950
951 y = (cmsFloat64Number) i * 65535.0 / (nResultSamples - 1);
952
953 // Find interval in which y is within.
954 j = GetInterval(y, InCurve->Table16, InCurve->InterpParams);
955 if (j >= 0) {
956
957
958 // Get limits of interval
959 x1 = InCurve ->Table16[j];
960 x2 = InCurve ->Table16[j+1];
961
962 y1 = (cmsFloat64Number) (j * 65535.0) / (InCurve ->nEntries - 1);
963 y2 = (cmsFloat64Number) ((j+1) * 65535.0 ) / (InCurve ->nEntries - 1);
964
965 // If collapsed, then use any
966 if (x1 == x2) {
967
968 out ->Table16[i] = _cmsQuickSaturateWord(Ascending ? y2 : y1);
969 continue;
970
971 } else {
972
973 // Interpolate
974 a = (y2 - y1) / (x2 - x1);
975 b = y2 - a * x2;
976 }
977 }
978
979 out ->Table16[i] = _cmsQuickSaturateWord(a* y + b);
980 }
981
982
983 return out;
984 }
985
986 // Reverse a gamma table
cmsReverseToneCurve(const cmsToneCurve * InGamma)987 cmsToneCurve* CMSEXPORT cmsReverseToneCurve(const cmsToneCurve* InGamma)
988 {
989 _cmsAssert(InGamma != NULL);
990
991 return cmsReverseToneCurveEx(4096, InGamma);
992 }
993
994 // From: Eilers, P.H.C. (1994) Smoothing and interpolation with finite
995 // differences. in: Graphic Gems IV, Heckbert, P.S. (ed.), Academic press.
996 //
997 // Smoothing and interpolation with second differences.
998 //
999 // Input: weights (w), data (y): vector from 1 to m.
1000 // Input: smoothing parameter (lambda), length (m).
1001 // Output: smoothed vector (z): vector from 1 to m.
1002
1003 static
smooth2(cmsContext ContextID,cmsFloat32Number w[],cmsFloat32Number y[],cmsFloat32Number z[],cmsFloat32Number lambda,int m)1004 cmsBool smooth2(cmsContext ContextID, cmsFloat32Number w[], cmsFloat32Number y[], cmsFloat32Number z[], cmsFloat32Number lambda, int m)
1005 {
1006 int i, i1, i2;
1007 cmsFloat32Number *c, *d, *e;
1008 cmsBool st;
1009
1010
1011 c = (cmsFloat32Number*) _cmsCalloc(ContextID, MAX_NODES_IN_CURVE, sizeof(cmsFloat32Number));
1012 d = (cmsFloat32Number*) _cmsCalloc(ContextID, MAX_NODES_IN_CURVE, sizeof(cmsFloat32Number));
1013 e = (cmsFloat32Number*) _cmsCalloc(ContextID, MAX_NODES_IN_CURVE, sizeof(cmsFloat32Number));
1014
1015 if (c != NULL && d != NULL && e != NULL) {
1016
1017
1018 d[1] = w[1] + lambda;
1019 c[1] = -2 * lambda / d[1];
1020 e[1] = lambda /d[1];
1021 z[1] = w[1] * y[1];
1022 d[2] = w[2] + 5 * lambda - d[1] * c[1] * c[1];
1023 c[2] = (-4 * lambda - d[1] * c[1] * e[1]) / d[2];
1024 e[2] = lambda / d[2];
1025 z[2] = w[2] * y[2] - c[1] * z[1];
1026
1027 for (i = 3; i < m - 1; i++) {
1028 i1 = i - 1; i2 = i - 2;
1029 d[i]= w[i] + 6 * lambda - c[i1] * c[i1] * d[i1] - e[i2] * e[i2] * d[i2];
1030 c[i] = (-4 * lambda -d[i1] * c[i1] * e[i1])/ d[i];
1031 e[i] = lambda / d[i];
1032 z[i] = w[i] * y[i] - c[i1] * z[i1] - e[i2] * z[i2];
1033 }
1034
1035 i1 = m - 2; i2 = m - 3;
1036
1037 d[m - 1] = w[m - 1] + 5 * lambda -c[i1] * c[i1] * d[i1] - e[i2] * e[i2] * d[i2];
1038 c[m - 1] = (-2 * lambda - d[i1] * c[i1] * e[i1]) / d[m - 1];
1039 z[m - 1] = w[m - 1] * y[m - 1] - c[i1] * z[i1] - e[i2] * z[i2];
1040 i1 = m - 1; i2 = m - 2;
1041
1042 d[m] = w[m] + lambda - c[i1] * c[i1] * d[i1] - e[i2] * e[i2] * d[i2];
1043 z[m] = (w[m] * y[m] - c[i1] * z[i1] - e[i2] * z[i2]) / d[m];
1044 z[m - 1] = z[m - 1] / d[m - 1] - c[m - 1] * z[m];
1045
1046 for (i = m - 2; 1<= i; i--)
1047 z[i] = z[i] / d[i] - c[i] * z[i + 1] - e[i] * z[i + 2];
1048
1049 st = TRUE;
1050 }
1051 else st = FALSE;
1052
1053 if (c != NULL) _cmsFree(ContextID, c);
1054 if (d != NULL) _cmsFree(ContextID, d);
1055 if (e != NULL) _cmsFree(ContextID, e);
1056
1057 return st;
1058 }
1059
1060 // Smooths a curve sampled at regular intervals.
cmsSmoothToneCurve(cmsToneCurve * Tab,cmsFloat64Number lambda)1061 cmsBool CMSEXPORT cmsSmoothToneCurve(cmsToneCurve* Tab, cmsFloat64Number lambda)
1062 {
1063 cmsFloat32Number w[MAX_NODES_IN_CURVE], y[MAX_NODES_IN_CURVE], z[MAX_NODES_IN_CURVE];
1064 int i, nItems, Zeros, Poles;
1065
1066 if (Tab == NULL) return FALSE;
1067
1068 if (cmsIsToneCurveLinear(Tab)) return TRUE; // Nothing to do
1069
1070 nItems = Tab -> nEntries;
1071
1072 if (nItems >= MAX_NODES_IN_CURVE) {
1073 cmsSignalError(Tab ->InterpParams->ContextID, cmsERROR_RANGE, "cmsSmoothToneCurve: too many points.");
1074 return FALSE;
1075 }
1076
1077 memset(w, 0, nItems * sizeof(cmsFloat32Number));
1078 memset(y, 0, nItems * sizeof(cmsFloat32Number));
1079 memset(z, 0, nItems * sizeof(cmsFloat32Number));
1080
1081 for (i=0; i < nItems; i++)
1082 {
1083 y[i+1] = (cmsFloat32Number) Tab -> Table16[i];
1084 w[i+1] = 1.0;
1085 }
1086
1087 if (!smooth2(Tab ->InterpParams->ContextID, w, y, z, (cmsFloat32Number) lambda, nItems)) return FALSE;
1088
1089 // Do some reality - checking...
1090 Zeros = Poles = 0;
1091 for (i=nItems; i > 1; --i) {
1092
1093 if (z[i] == 0.) Zeros++;
1094 if (z[i] >= 65535.) Poles++;
1095 if (z[i] < z[i-1]) {
1096 cmsSignalError(Tab ->InterpParams->ContextID, cmsERROR_RANGE, "cmsSmoothToneCurve: Non-Monotonic.");
1097 return FALSE;
1098 }
1099 }
1100
1101 if (Zeros > (nItems / 3)) {
1102 cmsSignalError(Tab ->InterpParams->ContextID, cmsERROR_RANGE, "cmsSmoothToneCurve: Degenerated, mostly zeros.");
1103 return FALSE;
1104 }
1105 if (Poles > (nItems / 3)) {
1106 cmsSignalError(Tab ->InterpParams->ContextID, cmsERROR_RANGE, "cmsSmoothToneCurve: Degenerated, mostly poles.");
1107 return FALSE;
1108 }
1109
1110 // Seems ok
1111 for (i=0; i < nItems; i++) {
1112
1113 // Clamp to cmsUInt16Number
1114 Tab -> Table16[i] = _cmsQuickSaturateWord(z[i+1]);
1115 }
1116
1117 return TRUE;
1118 }
1119
1120 // Is a table linear? Do not use parametric since we cannot guarantee some weird parameters resulting
1121 // in a linear table. This way assures it is linear in 12 bits, which should be enought in most cases.
cmsIsToneCurveLinear(const cmsToneCurve * Curve)1122 cmsBool CMSEXPORT cmsIsToneCurveLinear(const cmsToneCurve* Curve)
1123 {
1124 cmsUInt32Number i;
1125 int diff;
1126
1127 _cmsAssert(Curve != NULL);
1128
1129 for (i=0; i < Curve ->nEntries; i++) {
1130
1131 diff = abs((int) Curve->Table16[i] - (int) _cmsQuantizeVal(i, Curve ->nEntries));
1132 if (diff > 0x0f)
1133 return FALSE;
1134 }
1135
1136 return TRUE;
1137 }
1138
1139 // Same, but for monotonicity
cmsIsToneCurveMonotonic(const cmsToneCurve * t)1140 cmsBool CMSEXPORT cmsIsToneCurveMonotonic(const cmsToneCurve* t)
1141 {
1142 int n;
1143 int i, last;
1144 cmsBool lDescending;
1145
1146 _cmsAssert(t != NULL);
1147
1148 // Degenerated curves are monotonic? Ok, let's pass them
1149 n = t ->nEntries;
1150 if (n < 2) return TRUE;
1151
1152 // Curve direction
1153 lDescending = cmsIsToneCurveDescending(t);
1154
1155 if (lDescending) {
1156
1157 last = t ->Table16[0];
1158
1159 for (i = 1; i < n; i++) {
1160
1161 if (t ->Table16[i] - last > 2) // We allow some ripple
1162 return FALSE;
1163 else
1164 last = t ->Table16[i];
1165
1166 }
1167 }
1168 else {
1169
1170 last = t ->Table16[n-1];
1171
1172 for (i = n-2; i >= 0; --i) {
1173
1174 if (t ->Table16[i] - last > 2)
1175 return FALSE;
1176 else
1177 last = t ->Table16[i];
1178
1179 }
1180 }
1181
1182 return TRUE;
1183 }
1184
1185 // Same, but for descending tables
cmsIsToneCurveDescending(const cmsToneCurve * t)1186 cmsBool CMSEXPORT cmsIsToneCurveDescending(const cmsToneCurve* t)
1187 {
1188 _cmsAssert(t != NULL);
1189
1190 return t ->Table16[0] > t ->Table16[t ->nEntries-1];
1191 }
1192
1193
1194 // Another info fn: is out gamma table multisegment?
cmsIsToneCurveMultisegment(const cmsToneCurve * t)1195 cmsBool CMSEXPORT cmsIsToneCurveMultisegment(const cmsToneCurve* t)
1196 {
1197 _cmsAssert(t != NULL);
1198
1199 return t -> nSegments > 1;
1200 }
1201
cmsGetToneCurveParametricType(const cmsToneCurve * t)1202 cmsInt32Number CMSEXPORT cmsGetToneCurveParametricType(const cmsToneCurve* t)
1203 {
1204 _cmsAssert(t != NULL);
1205
1206 if (t -> nSegments != 1) return 0;
1207 return t ->Segments[0].Type;
1208 }
1209
1210 // We need accuracy this time
cmsEvalToneCurveFloat(const cmsToneCurve * Curve,cmsFloat32Number v)1211 cmsFloat32Number CMSEXPORT cmsEvalToneCurveFloat(const cmsToneCurve* Curve, cmsFloat32Number v)
1212 {
1213 _cmsAssert(Curve != NULL);
1214
1215 // Check for 16 bits table. If so, this is a limited-precision tone curve
1216 if (Curve ->nSegments == 0) {
1217
1218 cmsUInt16Number In, Out;
1219
1220 In = (cmsUInt16Number) _cmsQuickSaturateWord(v * 65535.0);
1221 Out = cmsEvalToneCurve16(Curve, In);
1222
1223 return (cmsFloat32Number) (Out / 65535.0);
1224 }
1225
1226 return (cmsFloat32Number) EvalSegmentedFn(Curve, v);
1227 }
1228
1229 // We need xput over here
cmsEvalToneCurve16(const cmsToneCurve * Curve,cmsUInt16Number v)1230 cmsUInt16Number CMSEXPORT cmsEvalToneCurve16(const cmsToneCurve* Curve, cmsUInt16Number v)
1231 {
1232 cmsUInt16Number out;
1233
1234 _cmsAssert(Curve != NULL);
1235
1236 Curve ->InterpParams ->Interpolation.Lerp16(&v, &out, Curve ->InterpParams);
1237 return out;
1238 }
1239
1240
1241 // Least squares fitting.
1242 // A mathematical procedure for finding the best-fitting curve to a given set of points by
1243 // minimizing the sum of the squares of the offsets ("the residuals") of the points from the curve.
1244 // The sum of the squares of the offsets is used instead of the offset absolute values because
1245 // this allows the residuals to be treated as a continuous differentiable quantity.
1246 //
1247 // y = f(x) = x ^ g
1248 //
1249 // R = (yi - (xi^g))
1250 // R2 = (yi - (xi^g))2
1251 // SUM R2 = SUM (yi - (xi^g))2
1252 //
1253 // dR2/dg = -2 SUM x^g log(x)(y - x^g)
1254 // solving for dR2/dg = 0
1255 //
1256 // g = 1/n * SUM(log(y) / log(x))
1257
cmsEstimateGamma(const cmsToneCurve * t,cmsFloat64Number Precision)1258 cmsFloat64Number CMSEXPORT cmsEstimateGamma(const cmsToneCurve* t, cmsFloat64Number Precision)
1259 {
1260 cmsFloat64Number gamma, sum, sum2;
1261 cmsFloat64Number n, x, y, Std;
1262 cmsUInt32Number i;
1263
1264 _cmsAssert(t != NULL);
1265
1266 sum = sum2 = n = 0;
1267
1268 // Excluding endpoints
1269 for (i=1; i < (MAX_NODES_IN_CURVE-1); i++) {
1270
1271 x = (cmsFloat64Number) i / (MAX_NODES_IN_CURVE-1);
1272 y = (cmsFloat64Number) cmsEvalToneCurveFloat(t, (cmsFloat32Number) x);
1273
1274 // Avoid 7% on lower part to prevent
1275 // artifacts due to linear ramps
1276
1277 if (y > 0. && y < 1. && x > 0.07) {
1278
1279 gamma = log(y) / log(x);
1280 sum += gamma;
1281 sum2 += gamma * gamma;
1282 n++;
1283 }
1284 }
1285
1286 // Take a look on SD to see if gamma isn't exponential at all
1287 Std = sqrt((n * sum2 - sum * sum) / (n*(n-1)));
1288
1289 if (Std > Precision)
1290 return -1.0;
1291
1292 return (sum / n); // The mean
1293 }
1294