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
|
/* makesRGB.c -- build sRGB-to-linear and linear-to-sRGB conversion tables
*
* Last changed in libpng 1.6.0 [February 14, 2013]
*
* COPYRIGHT: Written by John Cunningham Bowler, 2013.
* To the extent possible under law, the author has waived all copyright and
* related or neighboring rights to this work. This work is published from:
* United States.
*
* Make a table to convert 8-bit sRGB encoding values into the closest 16-bit
* linear value.
*
* Make two tables to take a linear value scaled to 255*65535 and return an
* approximation to the 8-bit sRGB encoded value. Calculate the error in these
* tables and display it.
*/
#define _C99_SOURCE 1
#include <stdio.h>
#include <math.h>
#include <stdlib.h>
/* pngpriv.h includes the definition of 'PNG_sRGB_FROM_LINEAR' which is required
* to verify the actual code.
*/
#include "../../pngpriv.h"
#include "sRGB.h"
/* The tables are declared 'const' in pngpriv.h, so this redefines the tables to
* be used.
*/
#define png_sRGB_table sRGB_table
#define png_sRGB_base sRGB_base
#define png_sRGB_delta sRGB_delta
static png_uint_16 png_sRGB_table[256];
static png_uint_16 png_sRGB_base[512];
static png_byte png_sRGB_delta[512];
static const unsigned int max_input = 255*65535;
double
fsRGB(double l)
{
return sRGB_from_linear(l/max_input);
}
double
sRGB(unsigned int i)
{
return fsRGB(i);
}
double
finvsRGB(unsigned int i)
{
return 65535 * linear_from_sRGB(i/255.);
}
png_uint_16
invsRGB(unsigned int i)
{
unsigned int x = nearbyint(finvsRGB(i));
if (x > 65535)
{
fprintf(stderr, "invsRGB(%u) overflows to %u\n", i, x);
exit(1);
}
return (png_uint_16)x;
}
int
main(int argc, char **argv)
{
unsigned int i, i16, ibase;
double min_error = 0;
double max_error = 0;
double min_error16 = 0;
double max_error16 = 0;
double adjust;
double adjust_lo = 0.4, adjust_hi = 0.6, adjust_mid = 0.5;
unsigned int ec_lo = 0, ec_hi = 0, ec_mid = 0;
unsigned int error_count = 0;
unsigned int error_count16 = 0;
int test_only = 0;
if (argc > 1)
test_only = strcmp("--test", argv[1]) == 0;
/* Initialize the encoding table first. */
for (i=0; i<256; ++i)
{
png_sRGB_table[i] = invsRGB(i);
}
/* Now work out the decoding tables (this is where the error comes in because
* there are 512 set points and 512 straight lines between them.)
*/
for (;;)
{
if (ec_lo == 0)
adjust = adjust_lo;
else if (ec_hi == 0)
adjust = adjust_hi;
else if (ec_mid == 0)
adjust = adjust_mid;
else if (ec_mid < ec_hi)
adjust = (adjust_mid + adjust_hi)/2;
else if (ec_mid < ec_lo)
adjust = (adjust_mid + adjust_lo)/2;
else
{
fprintf(stderr, "not reached: %u .. %u .. %u\n", ec_lo, ec_mid, ec_hi);
exit(1);
}
/* Calculate the table using the current 'adjust' */
for (i=0; i<=511; ++i)
{
double lo = 255 * sRGB(i << 15);
double hi = 255 * sRGB((i+1) << 15);
unsigned int calc;
calc = nearbyint((lo+adjust) * 256);
if (calc > 65535)
{
fprintf(stderr, "table[%d][0]: overflow %08x (%d)\n", i, calc,
calc);
exit(1);
}
png_sRGB_base[i] = calc;
calc = nearbyint((hi-lo) * 32);
if (calc > 255)
{
fprintf(stderr, "table[%d][1]: overflow %08x (%d)\n", i, calc,
calc);
exit(1);
}
png_sRGB_delta[i] = calc;
}
/* Check the 16-bit linear values alone: */
error_count16 = 0;
for (i16=0; i16 <= 65535; ++i16)
{
unsigned int i = 255*i16;
unsigned int iexact = nearbyint(255*sRGB(i));
unsigned int icalc = PNG_sRGB_FROM_LINEAR(i);
if (icalc != iexact)
++error_count16;
}
/* Now try changing the adjustment. */
if (ec_lo == 0)
ec_lo = error_count16;
else if (ec_hi == 0)
ec_hi = error_count16;
else if (ec_mid == 0)
{
ec_mid = error_count16;
printf("/* initial error counts: %u .. %u .. %u */\n", ec_lo, ec_mid,
ec_hi);
}
else if (error_count16 < ec_mid)
{
printf("/* adjust (mid ): %f: %u -> %u */\n", adjust, ec_mid,
error_count16);
ec_mid = error_count16;
adjust_mid = adjust;
}
else if (adjust < adjust_mid && error_count16 < ec_lo)
{
printf("/* adjust (low ): %f: %u -> %u */\n", adjust, ec_lo,
error_count16);
ec_lo = error_count16;
adjust_lo = adjust;
}
else if (adjust > adjust_mid && error_count16 < ec_hi)
{
printf("/* adjust (high): %f: %u -> %u */\n", adjust, ec_hi,
error_count16);
ec_hi = error_count16;
adjust_hi = adjust;
}
else
{
adjust = adjust_mid;
printf("/* adjust: %f: %u */\n", adjust, ec_mid);
break;
}
}
/* For each entry in the table try to adjust it to minimize the error count
* in that entry. Each entry corresponds to 128 input values.
*/
for (ibase=0; ibase<65536; ibase+=128)
{
png_uint_16 base = png_sRGB_base[ibase >> 7], trybase = base, ob=base;
png_byte delta = png_sRGB_delta[ibase >> 7], trydelta = delta, od=delta;
unsigned int ecbase = 0, eco;
for (;;)
{
png_sRGB_base[ibase >> 7] = trybase;
png_sRGB_delta[ibase >> 7] = trydelta;
/* Check the 16-bit linear values alone: */
error_count16 = 0;
for (i16=ibase; i16 < ibase+128; ++i16)
{
unsigned int i = 255*i16;
unsigned int iexact = nearbyint(255*sRGB(i));
unsigned int icalc = PNG_sRGB_FROM_LINEAR(i);
if (icalc != iexact)
++error_count16;
}
if (error_count16 == 0)
break;
if (ecbase == 0)
{
eco = ecbase = error_count16;
++trybase; /* First test */
}
else if (error_count16 < ecbase)
{
if (trybase > base)
{
base = trybase;
++trybase;
}
else if (trybase < base)
{
base = trybase;
--trybase;
}
else if (trydelta > delta)
{
delta = trydelta;
++trydelta;
}
else if (trydelta < delta)
{
delta = trydelta;
--trydelta;
}
else
{
fprintf(stderr, "makesRGB: impossible\n");
exit(1);
}
ecbase = error_count16;
}
else
{
if (trybase > base)
trybase = base-1;
else if (trybase < base)
{
trybase = base;
++trydelta;
}
else if (trydelta > delta)
trydelta = delta-1;
else if (trydelta < delta)
break; /* end of tests */
}
}
png_sRGB_base[ibase >> 7] = base;
png_sRGB_delta[ibase >> 7] = delta;
if (base != ob || delta != od)
{
printf("/* table[%u]={%u,%u} -> {%u,%u} %u -> %u errors */\n",
ibase>>7, ob, od, base, delta, eco, ecbase);
}
else if (0)
printf("/* table[%u]={%u,%u} %u errors */\n", ibase>>7, ob, od,
ecbase);
}
/* Only do the full (slow) test at the end: */
min_error = -.4999;
max_error = .4999;
error_count = 0;
for (i=0; i <= max_input; ++i)
{
unsigned int iexact = nearbyint(255*sRGB(i));
unsigned int icalc = PNG_sRGB_FROM_LINEAR(i);
if (icalc != iexact)
{
double err = 255*sRGB(i) - icalc;
if (err > (max_error+.001) || err < (min_error-.001))
{
printf(
"/* 0x%08x: exact: %3d, got: %3d [tables: %08x, %08x] (%f) */\n",
i, iexact, icalc, png_sRGB_base[i>>15],
png_sRGB_delta[i>>15], err);
}
++error_count;
if (err > max_error)
max_error = err;
else if (err < min_error)
min_error = err;
}
}
/* Re-check the 16-bit cases too, including the warning if there is an error
* bigger than 1.
*/
error_count16 = 0;
max_error16 = 0;
min_error16 = 0;
for (i16=0; i16 <= 65535; ++i16)
{
unsigned int i = 255*i16;
unsigned int iexact = nearbyint(255*sRGB(i));
unsigned int icalc = PNG_sRGB_FROM_LINEAR(i);
if (icalc != iexact)
{
double err = 255*sRGB(i) - icalc;
++error_count16;
if (err > max_error16)
max_error16 = err;
else if (err < min_error16)
min_error16 = err;
if (abs(icalc - iexact) > 1)
printf(
"/* 0x%04x: exact: %3d, got: %3d [tables: %08x, %08x] (%f) */\n",
i16, iexact, icalc, png_sRGB_base[i>>15],
png_sRGB_delta[i>>15], err);
}
}
/* Check the round trip for each 8-bit sRGB value. */
for (i16=0; i16 <= 255; ++i16)
{
unsigned int i = 255 * png_sRGB_table[i16];
unsigned int iexact = nearbyint(255*sRGB(i));
unsigned int icalc = PNG_sRGB_FROM_LINEAR(i);
if (i16 != iexact)
{
fprintf(stderr, "8-bit rounding error: %d -> %d\n", i16, iexact);
exit(1);
}
if (icalc != i16)
{
double finv = finvsRGB(i16);
printf("/* 8-bit roundtrip error: %d -> %f -> %d(%f) */\n",
i16, finv, icalc, fsRGB(255*finv));
}
}
printf("/* error: %g - %g, %u (%g%%) of readings inexact */\n",
min_error, max_error, error_count, (100.*error_count)/max_input);
printf("/* 16-bit error: %g - %g, %u (%g%%) of readings inexact */\n",
min_error16, max_error16, error_count16, (100.*error_count16)/65535);
if (!test_only)
{
printf("PNG_CONST png_uint_16 png_sRGB_table[256] =\n{\n ");
for (i=0; i<255; )
{
do
{
printf("%d,", png_sRGB_table[i++]);
}
while ((i & 0x7) != 0 && i<255);
if (i<255) printf("\n ");
}
printf("%d\n};\n\n", png_sRGB_table[i]);
printf("PNG_CONST png_uint_16 png_sRGB_base[512] =\n{\n ");
for (i=0; i<511; )
{
do
{
printf("%d,", png_sRGB_base[i++]);
}
while ((i & 0x7) != 0 && i<511);
if (i<511) printf("\n ");
}
printf("%d\n};\n\n", png_sRGB_base[i]);
printf("PNG_CONST png_byte png_sRGB_delta[512] =\n{\n ");
for (i=0; i<511; )
{
do
{
printf("%d,", png_sRGB_delta[i++]);
}
while ((i & 0xf) != 0 && i<511);
if (i<511) printf("\n ");
}
printf("%d\n};\n\n", png_sRGB_delta[i]);
}
return 0;
}
|