/usr/share/doc/png-definitive-guide/html/chapter11.html is in png-definitive-guide 20060430-1.
This file is owned by root:root, with mode 0o644.
The actual contents of the file can be viewed below.
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 1358 1359 1360 1361 1362 1363 1364 1365 1366 1367 1368 1369 1370 1371 1372 1373 1374 1375 1376 1377 1378 1379 1380 1381 1382 1383 1384 1385 1386 1387 1388 1389 1390 1391 1392 1393 1394 1395 1396 1397 1398 1399 1400 1401 1402 1403 1404 1405 1406 1407 1408 1409 1410 1411 1412 1413 1414 1415 1416 1417 1418 1419 1420 1421 1422 1423 1424 1425 1426 1427 1428 1429 1430 1431 1432 1433 1434 1435 1436 1437 1438 1439 1440 1441 1442 1443 1444 1445 1446 1447 1448 1449 1450 1451 1452 1453 1454 1455 1456 1457 1458 1459 1460 1461 1462 1463 1464 1465 1466 1467 1468 1469 1470 1471 1472 1473 1474 1475 1476 1477 1478 1479 1480 1481 1482 1483 1484 1485 1486 1487 1488 1489 1490 1491 1492 1493 1494 1495 1496 1497 1498 1499 1500 1501 1502 1503 1504 1505 1506 1507 1508 1509 1510 1511 1512 1513 1514 1515 1516 1517 1518 1519 1520 1521 1522 1523 1524 1525 1526 1527 1528 1529 1530 1531 1532 1533 1534 1535 1536 1537 1538 1539 1540 1541 1542 1543 1544 1545 1546 1547 1548 1549 1550 1551 1552 1553 1554 1555 1556 1557 1558 1559 1560 1561 1562 1563 1564 1565 1566 1567 1568 1569 1570 1571 1572 1573 1574 1575 1576 1577 1578 1579 1580 1581 1582 1583 1584 1585 1586 1587 1588 1589 1590 1591 1592 1593 1594 1595 1596 1597 1598 1599 1600 1601 1602 1603 1604 1605 1606 1607 1608 1609 1610 1611 1612 1613 1614 1615 1616 1617 1618 1619 1620 1621 1622 1623 1624 1625 1626 1627 1628 1629 1630 1631 1632 1633 1634 1635 1636 1637 1638 1639 1640 1641 1642 1643 1644 1645 1646 1647 1648 1649 1650 1651 1652 1653 1654 1655 1656 1657 1658 1659 1660 1661 1662 1663 1664 1665 1666 1667 1668 1669 1670 1671 1672 1673 1674 1675 1676 1677 1678 1679 1680 1681 1682 1683 1684 1685 1686 1687 1688 1689 1690 1691 1692 1693 1694 1695 1696 1697 1698 1699 1700 1701 1702 1703 1704 1705 1706 1707 1708 1709 1710 1711 1712 1713 1714 1715 1716 1717 1718 1719 1720 1721 1722 1723 1724 1725 1726 1727 1728 1729 1730 1731 1732 1733 1734 1735 1736 1737 1738 1739 1740 1741 1742 1743 1744 1745 1746 1747 1748 1749 1750 1751 1752 1753 1754 1755 1756 1757 1758 1759 1760 1761 1762 1763 1764 1765 1766 1767 1768 1769 1770 1771 1772 1773 1774 1775 1776 1777 1778 1779 1780 1781 1782 1783 1784 1785 1786 1787 1788 1789 1790 1791 1792 1793 1794 1795 1796 1797 1798 1799 1800 1801 1802 1803 1804 1805 1806 1807 1808 1809 1810 1811 1812 1813 1814 1815 1816 1817 1818 1819 1820 1821 1822 1823 1824 1825 1826 1827 1828 1829 1830 1831 1832 1833 1834 1835 1836 1837 1838 1839 1840 1841 1842 1843 1844 1845 1846 1847 1848 1849 1850 1851 1852 1853 1854 1855 1856 1857 1858 1859 1860 1861 1862 1863 1864 1865 1866 1867 1868 1869 1870 1871 1872 1873 1874 1875 1876 1877 1878 1879 1880 1881 1882 1883 1884 1885 1886 1887 1888 1889 1890 1891 1892 1893 1894 1895 1896 1897 1898 1899 1900 1901 1902 1903 1904 1905 1906 1907 1908 1909 1910 1911 1912 1913 1914 1915 1916 1917 1918 1919 1920 1921 1922 1923 1924 1925 1926 1927 1928 1929 1930 1931 1932 1933 1934 1935 1936 1937 1938 1939 1940 1941 1942 1943 1944 1945 1946 1947 1948 1949 1950 1951 1952 1953 1954 1955 1956 1957 1958 1959 1960 1961 1962 1963 1964 1965 1966 1967 1968 1969 1970 1971 1972 1973 1974 1975 1976 1977 1978 1979 1980 1981 1982 1983 1984 1985 1986 1987 1988 1989 1990 1991 1992 1993 1994 1995 1996 1997 1998 1999 2000 2001 2002 2003 2004 2005 2006 2007 2008 2009 2010 2011 2012 2013 2014 2015 2016 2017 | <!DOCTYPE HTML PUBLIC "-//W3C//DTD HTML 4.01 Transitional//EN"
"http://www.w3.org/TR/html4/loose.dtd">
<HTML>
<HEAD>
<TITLE>PNG Options and Extensions (PNG: The Definitive Guide)</TITLE>
<!-- META HTTP-EQUIV="Content-Type" CONTENT="text/html; charset=iso-8859-1" -->
<META HTTP-EQUIV="Content-Type" CONTENT="text/html; charset=utf-8">
<!-- http://www.w3.org/TR/REC-CSS2/box.html -->
<STYLE TYPE="text/css">
P { margin-bottom: 0em }
UL {
margin-bottom: 0em;
margin-top: 0em;
list-style: disc;
}
LI {
padding: 0px 0px 0px 0px;
margin: 0px 0px 0px 0px;
}
</STYLE>
<LINK REV="made" HREF="http://pobox.com/~newt/greg_contact.html">
<!-- Copyright (c) 1999 O'Reilly and Associates. -->
<!-- Copyright (c) 2002-2006 Greg Roelofs. -->
</HEAD>
<body bgcolor="#ffffff" text="#000000">
<hr> <!-- -=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=- -->
<a href="chapter10.html"><img width=24 height=13 border=0 align="left"
src="images/prev.png" alt="<-"></a>
<a href="chapter12.html"><img width=24 height=13 border=0 align="right"
src="images/next.png" alt="->"></a>
<div align="center">
<a href="chapter10.html"><font size="-1" color="#000000"
><b>PREVIOUS</b></font></a> <a
href="toc.html"><font size="-1" color="#000000"
><b>CONTENTS</b></font></a> <a
href="chapter12.html"><font size="-1" color="#000000"
><b>NEXT</b></font></a>
</div>
<hr> <!-- -=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=- -->
<h1 class="chapter">Chapter 11. PNG Options and Extensions</h1>
<div class="htmltoc"><h4 class="tochead">Contents:</h4><p>
<a href="#png.ch11.div.1">11.1. Background Color (bKGD)</a><br />
<a href="#png.ch11.div.2">11.2. Timestamp (tIME)</a><br />
<a href="#png.ch11.div.3">11.3. Latin-1 Text Annotations (tEXt, zTXt)</a><br />
<a href="#png.ch11.div.4">11.4. International Text Annotations (iTXt)</a><br />
<a href="#png.ch11.div.5">11.5. Histogram (hIST)</a><br />
<a href="#png.ch11.div.6">11.6. Suggested Palette (sPLT)</a><br />
<a href="#png.ch11.div.7">11.7. Significant Bits (sBIT)</a><br />
<a href="#png.ch11.div.8">11.8. Physical Pixel Dimensions (pHYs)</a><br />
<a href="#png.ch11.div.9">11.9. Physical Scale (sCAL)</a><br />
<a href="#png.ch11.div.10">11.10. Image Offset (oFFs)</a><br />
<a href="#png.ch11.div.11">11.11. Pixel Calibration (pCAL)</a><br />
<a href="#png.ch11.div.12">11.12. Fractal Parameters (fRAc)</a><br />
<a href="#png.ch11.div.13">11.13. GIF Conversion Info (gIFg, gIFx)</a><br />
<a href="#png.ch11.div.14">11.14. GIF Plain Text (gIFt)</a><br />
<a href="#png.ch11.div.15">11.15. Other Chunks</a><br />
</p></div>
<p>In addition to the core chunk types described thus far, the Portable Network
Graphics format supports a whole host of optional chunks for various purposes,
from text annotations to conversion information. These are described
in the following sections,
very roughly in order of importance to the average user.</p>
<div class="refentry">
<a name="png.ch11.div.1" />
<h2 class="sect1"><i>Background Color (bKGD)</i></h2>
<ul type="disc">
<li>Status: PNG Specification
<li>Location: After PLTE, before first IDAT
<li>Multiple: no
</ul>
<!--
<h4 class="refsect1">Description</h4>
-->
<p>
<a name="INDEX-821" />
<a name="INDEX-822" />
<a name="INDEX-823" />
In some applications, notably web browsers, there is a natural background
surrounding all images, against which images can be composited with transparency
information. But standalone image viewers typically have no preferred
background color or pattern and usually default to black, which may not be
appropriate for some images. PNG therefore supports the concept of a preferred
background color that can be used if nothing better is available.</p>
<p>The bKGD chunk is used for this purpose. Just as with the
transparency chunk, tRNS (see <a href="chapter08.html">Chapter 8, "PNG Basics"</a>), the format
of bKGD depends on the image type. For palette-based images it
contains a single byte, whose value is the palette index of the color
to be used for the background. For grayscale images, with or without
alpha, the chunk contains a 2-byte, unscaled gray value, just as
with tRNS--that is, the maximum allowed value is 2<sup class="superscript"><em class="emphasis">bit depth</em></sup>-1,
even though it is stored as a 16-bit integer. And for truecolor images, the
background chunk is exactly analogous to the grayscale version except that it
contains three 16-bit, unscaled values representing the red, green, and blue
components of the background color. There is no requirement in any of the
three cases that the background color be present in the actual image data.</p>
<p>Note that colored backgrounds are not supported in grayscale images; while
this is certainly a restriction, it appears not to be a particularly serious
one, to judge by the lack of public comment to date. Note also that the
background color should always be considered fully opaque, even if it happens
to match a color marked by the tRNS chunk as partly or fully transparent.</p>
</div>
<div class="refentry">
<a name="png.ch11.div.2" />
<h2 class="sect1"><i>Timestamp (tIME)</i></h2>
<ul type="disc">
<li>Status: PNG Specification
<li>Location: anywhere<a href="#FOOTNOTE-85">[85]</a>
<li>Multiple: no
</ul>
<!--
<h4 class="refsect1">Description</h4>
-->
<blockquote class="footnote">
<a name="FOOTNOTE-85" /><p>[85] Chunks with no explicit restrictions (``anywhere'') are nonetheless
implicitly constrained to come after the PNG signature and IHDR chunk,
before the IEND chunk, and not to fall between multiple IDAT chunks.</p>
</blockquote>
<p>
<a name="INDEX-824" /> <!-- .XX "timestamp" -->
<a name="INDEX-825" /> <!-- .XX "modification time" -->
<a name="INDEX-826" /> <!-- .XX "tIME chunk" -->
<!-- <a name="INDEX-827" /> GRR DUPLICATE INDEX (see below) -->
The timestamp chunk provides a way for the author (or image-editing software)
to record the time and date the image was last modified. The chunk contains
7 bytes of data, shown in
<a href="#png.ch11.tbl.1">Table 11-1</a>.
</p>
<a name="png.ch11.tbl.1" />
<div class="table" align="center">
<p>
<table width="502" border="0">
<tr>
<td>
<b class="emphasis-bold">Table 11-1.</b>
<i>tIME Chunk</i>
</td>
</tr>
</table>
</p>
<p>
<table width="502" border="1">
<tr>
<td><b class="emphasis-bold">Field</b></td>
<td><b class="emphasis-bold">Length and Valid Range
<?x-space 2p?><?x-space 2p?></b></td>
</tr>
<tr>
<td>Year</td>
<td>2 bytes (0-65,535)</td>
</tr>
<tr>
<td>Month</td>
<td>1 byte (1-12)</td>
</tr>
<tr>
<td>Day</td>
<td>1 byte (1-31)</td>
</tr>
<tr>
<td>Hour</td>
<td>1 byte (0-23)</td>
</tr>
<tr>
<td>Minute</td>
<td>1 byte (0-59)</td>
</tr>
<tr>
<td>Second</td>
<td>1 byte (0-60)</td>
</tr>
<?x-space 5p?>
</table>
</p>
</div>
<p><a name="INDEX-827" />
As this book is being written before the Third Millennium begins, the first
thing to notice is that PNG is not merely Y2K-compliant, but also Y2038,
Y10K, and pretty much everything else on up through
Y65K.<a href="#FOOTNOTE-86">[86]</a>
In addition, note that the seconds field is permitted to vary between
0 and 60; this allows for leap seconds, of which there have been
roughly two dozen since 1972. (There has never been more than one
leap second in any given minute, however.)
</p>
<blockquote class="footnote">
<a name="FOOTNOTE-86" /><p>[86] Presumably humanity will have come up with another image format or
two by then.</p>
</blockquote>
<p>On a less technical level, why does tIME store the modification time rather
than creation time? On the face of it, creation time would seem like a more
useful piece of information, and indeed, it is explicitly supported in PNG
via the text chunks described later. But whereas modification time is a
well-defined quantity--even a computer program can determine whether the
image data has been modified--creation time is ambiguous. If a scanned
photograph of the Mona Lisa is converted to PNG format, is its creation time
the time of image conversion, the time of the original scan, the time the
photograph was taken, or even the time the painting was created? The case
becomes even muddier if an artist creates a digital work partly based on the
scanned image. So creation time is supported via one or more text chunks,
which can also describe in precisely what sense the image was created.</p>
</div>
<div class="refentry">
<a name="png.ch11.div.3" />
<h2 class="sect1"><i>Latin-1 Text Annotations (tEXt, zTXt)</i></h2>
<ul type="disc">
<li>Status: PNG Specification
<li>Location: anywhere
<li>Multiple: yes
</ul>
<!--
<h4 class="refsect1">Description</h4>
-->
<p>
<a name="INDEX-828" />
<a name="INDEX-829" />
<a name="INDEX-830" />
<a name="INDEX-831" />
That brings us to PNG's original text chunks, which are perhaps its most
popular nonessential chunks. Regardless of how many words a picture is
worth, it is often useful or necessary to add a few more in order to record
pertinent information like title and author, store requisite legal
notices such as a copyright or disclaimer, or merely to transfer text
from one image to another.</p>
<p>PNG supports two types of Latin-1-based text chunks, uncompressed
(tEXt) and compressed (zTXt). There is also a new Unicode-based chunk
(iTXt) that I'll discuss next. For the first two, the format is
basically the same: an uncompressed keyword or key phrase, a null
(zero) byte, and the actual text. In zTXt the text is compressed; the
first byte after the null indicates the compression method, for which
only deflate is currently defined (method zero). The remainder is the
compressed stream, which for method zero must be in zlib 1.x format,
just as for image data. (The zlib 1.x format is described by revision
3.3 of the zlib specification, which is available from
<a href="http://www.zlib.org/zlib_docs.html">http://www.zlib.org/zlib_docs.html</a>/.)
<a name="INDEX-832" />
</p>
<p>Both keyword and raw text should be encoded with the Latin-1 (ISO/IEC 8859-1)
character set; neither may contain null bytes. Since the keyword is intended
to be recognizable by both humans
and computer programs, additional restrictions are placed on it: it may not
contain leading, trailing, or consecutive spaces, and it is restricted to
characters in the range 32-126 and 161-255 (which, in particular, rules
out both control characters and the nonbreaking space, decimal value 160).
The only other restriction on the main text of the chunk is that newlines
should be in Unix format, i.e., represented by a single line-feed character
(decimal value 10).</p>
<p>I mentioned in <a href="chapter07.html">Chapter 7, "History of the Portable Network Graphics Format"</a>, that the Unicode UTF-8 character set was one of the
items in the design of PNG that was voted down. In retrospect this was,
perhaps, a lamentable decision; it was finally addressed early in 1999
<a name="INDEX-833" />
<a name="INDEX-834" />
<a name="INDEX-835" />
<a name="INDEX-836" />with the iTXt chunk. But at the time, UTF-8 was very new and had not been
extensively tested in the field. In particular, it had little or no
operating-system support and no support in standard programming libraries,
either for encoding and decoding or for the translation and display of UTF-8
characters in the native character set(s) of existing systems. Since PNG's
design goals included both the use of well-tested technologies and the
avoidance of undue burdens on developers of PNG applications, support for
UTF-8 was dropped in favor of the more familiar Latin-1 character set.</p>
<p>The following list summarizes all of the keywords that are either
included in the specification itself or officially registered as extensions
to the spec:</p>
<dl>
<dt><b><em class="emphasis">Author</em></b></dt><dd><p><a name="INDEX-837" />The name of the author of the image. If the original image were a painting or
other nonelectronic medium, both the original artist and the person who
scanned the image might be listed.</p></dd>
<dt><b><em class="emphasis">Title</em></b></dt><dd><p><a name="INDEX-838" />A one-line title or caption. Longer captions should generally
use the Description keyword, but see the end of this section for an unofficial alternative.</p></dd>
<dt><b><em class="emphasis">Description</em></b></dt><dd><p><a name="INDEX-839" />A longer description of or caption for the image, perhaps including details
about the tools and settings used; the name, age, and/or location of the
subject matter; or the mood the artist was trying to convey. See also the
Software and Source keywords.</p></dd>
<dt><b><em class="emphasis">Creation Time</em></b></dt><dd><p><a name="INDEX-840" />The time the image was created, in whatever sense is most appropriate.
The recommended format is that prescribed by Internet RFC 822 (Section 5), as
amended by RFC 1123 (Section 5.2.14); specifically:
</p></dd>
</dl>
<blockquote><pre class="code">day month year hour:minute timezone</pre></blockquote>
<p>where <b class="emphasis-bold">day</b> is either one or two digits; <b class="emphasis-bold">month</b> is a three-letter
English abbreviation such as <b class="emphasis-bold">Jun</b>; <b class="emphasis-bold">year</b> is two or four digits
(though the latter is strongly recommended); <b class="emphasis-bold">hour</b> and <b class="emphasis-bold">minute</b>
are two digits each; and <b class="emphasis-bold">timezone</b> is either a three-letter abbreviation
(e.g., PST for Pacific Standard Time), or a one-letter U.S. military designation,
or a four-digit number with a leading positive or negative sign indicating the
hour:minute offset from Coordinated Universal Time (e.g., -0800 for Pacific
Standard Time, which is eight hours and zero minutes earlier than UTC). In
addition, the entire string may optionally be preceded by a <b class="emphasis-bold">weekday</b>
field, where <b class="emphasis-bold">weekday</b> is a three-letter English abbreviation (e.g.,
<b class="emphasis-bold">Fri</b>). A colon and two-digit <b class="emphasis-bold">seconds</b> field may also be
appended to the time (that is, <b class="emphasis-bold">hour:minute:second</b>). Note that this
is merely a recommendation; strings such as ``circa 1492'' are allowed, as is
explanatory text following an RFC-style date string.</p>
<dl>
<dt><b><em class="emphasis">Copyright</em></b></dt><dd><p><a name="INDEX-841" />The legal copyright notice for the image. For example, ``Copyright 1999
by Greg Roelofs. This image may be freely used and distributed provided that
it is not modified in any way and that this notice remains intact.''</p></dd>
<dt><b><em class="emphasis">Disclaimer</em></b></dt><dd><p><a name="INDEX-842" />
<a name="INDEX-843" />A legal disclaimer notice for the image. This might include a company's
standard boilerplate on all copyrighted works; in particular, it might be
lengthy enough to store in a compressed (zTXt) chunk, while the copyright
notice remains uncompressed.</p></dd>
<dt><b><em class="emphasis">Warning</em></b></dt><dd><p><a name="INDEX-844" />A warning about the content or effects of the image. For example, certain
types of popular material may not be suitable for minors, or a random-dot
stereogram (``Magic Eye'' 3D image) may induce headaches in some people.</p></dd>
<dt><b><em class="emphasis">Software</em></b></dt><dd><p><a name="INDEX-845" />The name and possibly the version of the software used to create the image. This
is most often generated automatically, but it need not be. More than one
software application may be listed.</p></dd>
<dt><b><em class="emphasis">Source</em></b></dt><dd><p><a name="INDEX-846" />Information about the device used to generate the image, such as a digital
camera or a scanner.</p></dd>
<dt><b><em class="emphasis">Comment</em></b></dt><dd><p><a name="INDEX-847" />A miscellaneous comment, often converted from a GIF comment (which lacks
keywords).</p></dd>
</dl>
<p>In addition to these official keywords, one of the technical
reviewers of this book and I have been known to make use of a few unofficial
<a name="INDEX-848" />keywords. The <B>Caption</B> keyword is used to provide a brief description
of an image that is more specifically tailored for use as a
publishable caption than the generic Description keyword; it is
also generally lengthier than is appropriate for the Title keyword.
<a name="INDEX-849" />The <B>E-mail</B> keyword stores the email address of the author in standard
Internet format (RFC 822, Section 6, as amended by RFC 1123, Sections
5.2.15 through 5.2.19); for example,
<em class="emphasis">roelofs@pobox.com</em> . And
<a name="INDEX-850" />the <B>URL</B> keyword is for a standard WWW Uniform Resource Locator (RFC
2068, Section 3.2); for example,
<a href="http://www.oreilly.com/">http://www.oreilly.com/</a> . If
the URL is reasonably self-explanatory, it is recommended that the
chunk consist of the single URL and nothing else, but this is not a
requirement. Multiple URLs should be separated by newline characters.
Note that spaces and other white space (tabs, newlines, and so forth)
are considered unsafe by the URL standard and therefore must be
escaped within a conforming URL. For example, a space character must
be encoded as <b class="emphasis-bold">%20</b>. This allows easy parsing
of optional explanatory text after a URL:
<?x-need 10?>the URL ends when the first white space
(space, tab, or newline) is encountered.
</p>
</div>
<div class="refentry">
<a name="png.ch11.div.4" />
<h2 class="sect1"><i>International Text Annotations (iTXt)</i></h2>
<ul type="disc">
<li>Status: PNG Extensions <a href="#FOOTNOTE-87">[87]</a>
<li>Location: anywhere
<li>Multiple: yes
</ul>
<!--
<h4 class="refsect1">Description</h4>
-->
<blockquote class="footnote">
<a name="FOOTNOTE-87" /><p>[87] As this book went to press, the iTXt chunk had
just been approved for
inclusion in the core PNG specification, but it was temporarily placed in
the PNG extensions document pending completion and approval of extensive
ISO-related changes to the core spec. (Note that these changes are almost
entirely of an organizational or editorial nature; the technical content of
the specification is expected to change only minimally from version
1.1.). Version 1.2 of the PNG specification is expected around mid-1999 or
later. In the meantime, iTXt can be found in version 1.1.1 (and possibly
later versions) of the extensions document, which is available electronically
from <a href="http://www.libpng.org/pub/png/pngdocs.html">http://www.libpng.org/pub/png/pngdocs.html</a>.
</p>
</blockquote>
<p>
<a name="INDEX-851" />
<a name="INDEX-852" />
<a name="INDEX-853" />
<a name="INDEX-854" />
<a name="INDEX-855" />
I previously noted that, as of early 1999, PNG was in the midst of
joint ISO/IEC standardization. One of the technical issues in the
first Committee Draft vote was the lack of support for non-Western
languages, specifically in the text chunks. In fact, the PNG
Development Group had already discussed a more general text chunk in
mid-1998, but its vote was deferred until there was external interest
in it. The ISO comments from Japan and the United States clearly fell
into the category of external interest, however, so the iTXt was
voted on and approved as part of the PNG specification in early
February 1999.</p>
<p>The layout of iTXt is a generalization of tEXt and zTXt, as shown in
<a href="#png.ch11.tbl.2">Table 11-2</a>.
</p>
<a name="png.ch11.tbl.2" />
<div class="table" align="center">
<p>
<table width="502" border="0">
<tr>
<td>
<b class="emphasis-bold">Table 11-2.</b>
<i>iTXt Chunk</i>
</td>
</tr>
</table>
</p>
<p>
<table width="502" border="1">
<tr>
<td><b class="emphasis-bold">Field</b></td>
<td><b class="emphasis-bold">Length and Valid Range
<?x-space 2p?><?x-space 2p?></b></td>
</tr>
<tr>
<td>Keyword</td>
<td>1-79 bytes (Latin-1 text)</td>
</tr>
<tr>
<td>Null separator</td>
<td>1 byte (0)</td>
</tr>
<tr>
<td>Compression flag</td>
<td>1 byte (0, 1)</td>
</tr>
<tr>
<td>Compression method</td>
<td>1 byte (0)</td>
</tr>
<tr>
<td>Language tag</td>
<td><em class="emphasis">k</em> bytes (ASCII text)</td>
</tr>
<tr>
<td>Null separator</td>
<td>1 byte (0)</td>
</tr>
<tr>
<td>Translated keyword</td>
<td><em class="emphasis">m</em> bytes (Unicode UTF-8 text)</td>
</tr>
<tr>
<td>Null separator</td>
<td>1 byte (0)</td>
</tr>
<tr>
<td>Text</td>
<td><em class="emphasis">n</em> bytes (Unicode UTF-8 text)</td>
</tr>
<?x-space 5p?>
</table>
</p>
</div>
<p>The first field is a keyword, with exactly the same restrictions and officially
registered values (Author, Description, and so on) as the tEXt and zTXt chunks.
Latin-1 (ISO/IEC 8859-1) was chosen so that existing PNG source code could
be used without modification to parse and optionally recognize the keyword.</p>
<p>The keyword is followed by a null separator byte and two compression-related
bytes. The first indicates whether the main text is compressed (if its value
is 1) or not (if it's 0). If the text is compressed, the next byte
indicates its compression method, which currently must be zero for the
zlib-encoded deflate algorithm. The two bytes could have been combined,
but for historical reasons relating to the method byte in IHDR, the split
approach was favored.</p>
<p>After the compression bytes is an optional <em class="emphasis">case-insensitive</em> field
indicating the (human) language used in the remaining two text fields. This
is necessary not only to render Unicode text properly but also to allow
decoders to distinguish between multiple iTXt chunks, which may consist of
the same text in different languages--but possibly identical keywords.
Unlike both the keyword and the main text, the language tag is plain ASCII
text (specifically, the ``invariant'' ASCII subset of ISO 646, which is itself
a subset of both Latin-1 and Unicode UTF-8) conforming to Internet Standard
RFC 1766. It consists of hyphen-separated ``words'' of between one and eight
characters each, where the first word is either a two-letter ISO language code
(ISO 639), the letter <em class="emphasis">i</em> for tags registered by the Internet Assigned
Numbers Authority (IANA)<a href="#FOOTNOTE-88">[88]</a>
or the letter <em class="emphasis">x</em> for private tags.
The second ``word'' is interpreted as an ISO 3166 country code if it is
exactly two characters long or as an IANA-registered code if it is between
three and eight characters. Subsequent ``words'' may be anything, as long
as they conform to the general rules. Examples of language tags include
<tt>cn</tt> (Chinese),
<tt>en-US</tt> (American English),
<tt>no-bok</tt> (Norwegian
<!-- <em class="emphasis">bokm<img src="images/U00E5.png" alt="[a-with-circle]" />l</em> or ``book language''), -->
<em class="emphasis">bokmål</em> or ``book language''),
<tt>i-navajo</tt> (Navajo), and
<tt>x-klingon</tt> (Klingon, from the fictional
Star Trek universe).
</p><blockquote class="footnote">
<a name="FOOTNOTE-88" /><p>[88] As this is written, indications are that IANA will eventually be replaced
by ICANN, the Internet Corporation for Assigned Names and Numbers. This
transition may not occur until 2000, however.</p>
</blockquote>
<p>A null separator byte terminates the language tag, which is followed by an
optional translation of the keyword into the given language. The translated
keyword is represented in the UTF-8 encoding of the Unicode character set,
which is described in the International Standard ISO/IEC 10646-1, in Internet
RFC 2279, and in the Unicode Consortium's reference, <em class="emphasis">The Unicode
Standard</em>. Like the primary keyword, it should not contain any newline
characters, and it is also followed by a null byte.
</p>
<p>The remaining chunk data is the main UTF-8 text, either zlib-compressed or
not, according to the compression flag. Since its length can be determined
from the chunk length, it is not null-terminated. As with the other
two text chunks, newlines should be represented by single line-feed characters
(decimal 10), and all other control characters (1-9, 11-31, and
127-159) are discouraged. Note, however, that UTF-8 <em class="emphasis">encodings</em> may
contain any of the bytes between 128 and 159; what is discouraged is the set
of Unicode characters whose four-byte <em class="emphasis">integer values</em> are 128-159.</p>
<p><a name="INDEX-856" />
<a name="INDEX-857" />That last point is confusing, so perhaps a quick primer on Unicode is in order.
The Unicode <em class="emphasis">character set</em> is a mapping between graphic characters (or
glyphs) and integers. The simplest representation is called UCS-4 and consists
of 4-byte integers, potentially allowing more than two billion characters
to be defined. On top of that are a number of possible transformations or
<em class="emphasis">encodings</em> of the character set; UTF-8 is one of the more popular ones,
encoding 4-byte UCS-4 characters into anywhere from 1 to 8 bytes.
All Unicode characters below 128 are encoded as single bytes in UTF-8, and
because Unicode characters 1-127 are identical to US-ASCII characters
1-127, the Unicode character set (and UTF-8 in particular) may be thought
of as a very large superset of 7-bit ASCII.</p>
<p><a name="INDEX-858" />Multibyte UTF-8 encodings, on the other hand, are composed entirely of byte
values between 128 and 253--which means that bytes 1-9, 11-31, and 127
will never be found in valid UTF-8-encoded text except when representing the
<em class="emphasis">characters</em> 1-9, 11-31, and 127. So about half of the control
characters that are discouraged in iTXt can be detected simply by checking
for those single bytes. The remaining half, characters 128-159, are all
encoded with 2-byte sequences that happen to begin with byte value 194:
194 128 through 194 159. The fact that character 128 is discouraged in
iTXt's UTF-8 text fields therefore means that the 2-byte encoding 194 128
is discouraged, but the 2-byte encoding 195 128 (À or ``Latin capital
letter A with grave accent'') is completely acceptable.
</p>
</div>
<div class="refentry">
<a name="png.ch11.div.5" />
<h2 class="sect1"><i>Histogram (hIST)</i></h2>
<ul type="disc">
<li>Status: PNG Specification
<li>Location: after PLTE, before first IDAT
<li>Multiple: no
</ul>
<!--
<h4 class="refsect1">Description</h4>
-->
<p>
<a name="INDEX-859" />
<a name="INDEX-860" />
A histogram is nothing more than a frequency-of-occurrence table, and the
PNG hIST chunk gives the approximate frequencies of occurrence for pixels
of various colors. This information is typically used to decide which
colors are the most important if the system is not capable of displaying
all of them. Rather than force the decoder to count pixels every time the
image is displayed, the histogram places the burden on the encoder, which
performs the task only once.</p>
<p>
<a name="INDEX-861" />
PNG's hIST implementation is tied to the PLTE chunk; if there is no palette,
hIST is not allowed. This and one or two other limitations were later
recognized and addressed by the sPLT chunk, which we'll discuss next; it is
generally favored over hIST, but the latter is smaller, and either may be
used.
The histogram must contain exactly as many entries as PLTE contains,
and each entry is a 16-bit unsigned integer. Since such integers can
only represent numbers in the range 0-65,535 and there may be
millions of pixels of a given color, the histogram entries often must
be scaled and are therefore inexact. The sole exception is the value
zero; it is guaranteed to mean that there are <em class="emphasis">no</em> pixels of the
corresponding color. A nonzero count that would otherwise be scaled
and rounded to zero must instead be rounded up to one.</p>
<p>
<a name="INDEX-862" />
<a name="INDEX-863" />
Truecolor images that include a PLTE chunk as a suggested quantization
are a special case. The histogram counts are dependent on the
algorithm used by the encoder for quantizing the pixels; if the
decoder happens to use a different algorithm, its counts would be
different, too. The upshot is that the histogram is particularly
approximate in this case. Because truecolor images
typically have far more colors than palette entries, the palette
entries that do appear should always represent at least one
pixel; thus there should be no zero counts in the histogram.
<?x-space -4p?></p>
</div>
<div class="refentry">
<a name="png.ch11.div.6" />
<h2 class="sect1"><i>Suggested Palette (sPLT)</i></h2>
<ul type="disc">
<li>Status: PNG Specification
<li>Location: before first IDAT
<li>Multiple: yes
</ul>
<!--
<h4 class="refsect1">Description</h4>
-->
<p>
<a name="INDEX-864" />
<a name="INDEX-865" />
The suggested-palette chunk, sPLT, grew out of an acknowledgment of
some limitations in PNG's PLTE, tRNS, and hIST chunks. I have already
noted that PLTE is allowed only in palette, RGB, and RGBA images and
that hIST is allowed only in images with PLTE; I also noted that tRNS
is disallowed in images with alpha channels, which rules out the use
of PLTE plus tRNS as a suggested gray/alpha or RGBA palette. sPLT
eliminates these restrictions by merging all three of the older chunks
into a general-purpose, suggested-RGBA-palette-plus-histogram
chunk. In addition, sPLT may contain any number of entries (as long as
it doesn't exceed the maximum chunk-size limit of two gigabytes); its
entries may have either 8-bit or 16-bit sample depths; and multiple
sPLT chunks encoding different suggested quantizations are allowed. A
palette-based image may even have an sPLT chunk, perhaps representing
a reduced palette for a particular web browser. The format of sPLT, given in
<a href="#png.ch11.tbl.3">Table 11-3</a>, is straightforward.
</p>
<a name="png.ch11.tbl.3" />
<div class="table" align="center">
<p>
<table width="502" border="0">
<tr>
<td>
<b class="emphasis-bold">Table 11-3.</b>
<i>sPLT Chunk</i>
</td>
</tr>
</table>
</p>
<p>
<table width="502" border="1">
<tr>
<td><b class="emphasis-bold">Field</b></td>
<td><b class="emphasis-bold">Length and Valid Range</b></td>
</tr>
<tr>
<td>Palette name</td>
<td>1-79 bytes (Latin-1 text)</td>
</tr>
<tr>
<td>Null separator</td>
<td>1 byte</td>
</tr>
<tr>
<td>Sample depth</td>
<td>1 byte (8 or 16)</td>
</tr>
<tr>
<td>Red value #1</td>
<td>1 byte (0-255) or 2 bytes (0-65,535)</td>
</tr>
<tr>
<td>Green value #1</td>
<td>1 byte (0-255) or 2 bytes (0-65,535)</td>
</tr>
<tr>
<td>Blue value #1</td>
<td>1 byte (0-255) or 2 bytes (0-65,535)</td>
</tr>
<tr>
<td>Alpha value #1</td>
<td>1 byte (0-255) or 2 bytes (0-65,535)</td>
</tr>
<tr>
<td>Relative frequency #1</td>
<td>2 bytes (0-65,535)</td>
</tr>
<tr>
<td> ...</td>
<td> ...</td>
</tr>
<?x-space 5p?>
</table>
</p>
</div>
<?x-space -2p?><p>
The number of sPLT entries is implicitly given by the size of the chunk and
the sample depth; in the more common case of 8-bit samples, it is obtained
by dividing the chunk size, less the length of the palette name and the two
subsequent bytes, by six. Entries are required to appear in decreasing
order of frequency, but there is no requirement that all of them be different
nor that all of them be used by the image. Furthermore, opaque images may
include nonopaque sPLT entries, grayscale images may include colored
entries, and the sample depth of sPLT is independent of that of the image.</p>
<p><a name="INDEX-866" />
Unlike the suggested practice for PLTE in RGBA images, the red, green, and
blue values in sPLT are neither premultiplied by the alpha values nor
precomposited against a background color. An encoder would still have to
inspect every pixel if it wanted to compute an optimal palette for display
of an RGBA image against a patterned background, but sPLT would enable a
statistical approach based on the background image's own histogram in
that case. And for solid backgrounds, sPLT provides the means to build
an optimal palette regardless of the choice of background color.</p>
<p><a name="INDEX-867" />
As with the hIST chunk, frequency values are scaled to the range 0-65,535
and therefore are likely to be approximate. Inflating ``important'' colors
based on the image's subject matter is allowed in sPLT, too. But whereas hIST
requires a 0 frequency to correspond exactly to 0 pixels, sPLT allows
the 0 value to represent infrequently used or unimportant colors. If all
of the frequency values are 0, however, the histogram is undefined.</p>
<p>Note that multiple sPLT chunks are required to have different palette names.</p>
</div>
<div class="refentry">
<a name="png.ch11.div.7" />
<h2 class="sect1"><i>Significant Bits (sBIT)</i></h2>
<ul type="disc">
<li>Status: PNG Specification
<li>Location: before PLTE and first IDAT
<li>Multiple: no
</ul>
<!--
<h4 class="refsect1">Description</h4>
-->
<p>
<a name="INDEX-868" />
<a name="INDEX-869" />
The significant-bits chunk is used to indicate the nature of the source
data in cases in which storing it in PNG form required a conversion. For
example, gray pixels in medical images are often 12 bits deep, but PNG
requires them to be scaled up to 16 bits for portability. Scaling the pixels
does not alter the fact that they contain only 12 bits of real information,
and the sBIT chunk stores this fact in a PNG file.</p>
<p>As with several other PNG chunks, the format of sBIT depends on the image
type. Grayscale images are the simplest; sBIT then contains a single byte
indicating the number of significant bits in the source data--in the
preceding example, 12. For grayscale images with an alpha channel, sBIT contains
2 bytes, one for the gray channel and one for alpha; RGB images require
3 bytes, and RGBA images require 4. Palette-based images are treated
like RGB except that the sBIT information refers to the palette entries, and
the palette's effective sample depth is always 8, regardless of how many
bits are used to index the palette. Note that the number of significant bits
for any given channel must be greater than zero and less than or equal to the
sample depth.</p>
<p>Ordinary PNG decoders need not worry about sBIT, but those that wish to
recover the original image data can do so by right-shifting each image
sample to leave only the number of bits indicated by sBIT. This implies
that the scaling procedure used by the PNG encoder must not change the original
bits; it can only append low-order bits to each sample.</p>
</div>
<div class="refentry">
<a name="png.ch11.div.8" />
<h2 class="sect1"><i>Physical Pixel Dimensions (pHYs)</i></h2>
<ul type="disc">
<li>Status: PNG Specification
<li>Location: before first IDAT
<li>Multiple: no
</ul>
<!--
<h4 class="refsect1">Description</h4>
-->
<p>
<a name="INDEX-870" />
<a name="INDEX-871" />
<a name="INDEX-872" />
The pHYs chunk encodes the absolute or relative dimensions of pixels.
For example, an image scanned at 600 dots per inch has pixels with known,
absolute sizes--namely, one six-hundredth of an inch in both
<em class="emphasis">x</em> and <em class="emphasis">y</em>
directions. Alternatively, an image created on a 1280 × 1024 display
will have nonsquare pixels, and the relative dimensions of each pixel,
<a name="INDEX-873" />also referred to as the <em class="emphasis">aspect ratio</em>, may be stored so the image
can be displayed as it was intended to be seen.
</p>
<p>The layout of the chunk is shown in
<a href="#png.ch11.tbl.4">Table 11-4</a>.
</p>
<a name="png.ch11.tbl.4" />
<div class="table" align="center">
<p>
<table width="502" border="0">
<tr>
<td>
<b class="emphasis-bold">Table 11-4.</b>
<i>pHYs Chunk</i>
</td>
</tr>
</table>
</p>
<p>
<table width="502" border="1">
<tr>
<td><b class="emphasis-bold">Field</b></td>
<td><b class="emphasis-bold">Length and Valid Range
<?x-space 2p?><?x-space 2p?></b></td>
</tr>
<tr>
<td>Pixels per unit, <em class="emphasis">x</em> axis</td>
<td>4 bytes (0-2,147,483,647)</td>
</tr>
<tr>
<td>Pixels per unit, <em class="emphasis">y</em> axis</td>
<td>4 bytes (0-2,147,483,647)</td>
</tr>
<tr>
<td>Unit specifier</td>
<td>1 byte (0, 1)</td>
</tr>
<?x-space 5p?>
</table>
</p>
</div>
<p>If the unit specifier byte is 1, the units are meters; if it is 0, the
units are unspecified, and only the relative dimensions are known. Currently,
no other values are valid. Note that the format of the chunk precludes
pixel sizes greater than one meter, which should not be a significant
hardship for most applications, but it allows pixels as small as 4.7
<!-- <img src="images/U00C5.png" alt="[A-with-circle]" />ngstroms, -->
Ångstroms,
which is roughly the size of a single atom.</p>
<p>For the previous scanning example, 600 dpi is equal to 23,622.05
pixels per meter, so both the <em class="emphasis">x</em> and <em class="emphasis">y</em> values would be
23,622, and the unit specifier would be 1. The second example is
slightly trickier. First, it is necessary to know that practically
all current computer displays have a physical aspect ratio of 4:3,<a href="#FOOTNOTE-89">[89]</a>
which means the viewable portion of the display (the glass) is three-quarters as high as it is wide. Thus, the horizontal pixels-per-unit in the
case of a 1280 × 1024 display is proportional to (1280/4) or 320, while the
vertical pixels-per-unit is proportional to (1024/3) or 341.333333. Because
we don't have an absolute scale, we are free to multiply these values by a
common factor; doing so will preserve some of the precision that would otherwise
be lost due to truncation of the decimal part of the second value (the .3333
part). One choice would be a power of 10, such as 1,000; then the stored
values would be 320,000 and 341,333, respectively. But in this case, we can
do better: we know that the fractional part is simply one-third, so multiplying
both values by 3 will preserve the aspect ratio exactly. Thus the chunk would
contain the values (3 × 1280/4) or 960, (3 × 1024/3) or 1,024, and 0 for
the unit specifier. Values of 15, 16, and 0 would work equally well.
</p><blockquote class="footnote">
<a name="FOOTNOTE-89" /><p>[89] This will change with the convergence of computers and high-definition
TV. Displays for the latter have a 16:9 aspect ratio, which apparently is
the geometric mean of standard television and computer displays (4:3) and
of modern, panoramic films (typically 2.35:1, but it varies).</p>
</blockquote>
<p><a name="INDEX-874" />A decoder that encounters a pHYs chunk with different values for the <em class="emphasis">x</em> and <em class="emphasis">y</em>
axes has several options. The simplest and least correct approach is to ignore
the chunk; most current viewers do this. A better approach is to interpolate
the pixels in one of the dimensions; this gives the correct overall appearance
but introduces noticeable artifacts--for the preceding example, it involves
either duplicating every 15th column stretching the image horizontally,
or deleting every 16th row shrinking the image vertically. The best
approach is to resample the image, a procedure that amounts to converting the
image to a continuous (or analog) representation and then overlaying the
desired pixel grid on that. This is, by far, the most expensive approach
in terms of CPU usage, but the results are excellent.</p>
</div>
<div class="refentry">
<a name="png.ch11.div.9" />
<h2 class="sect1"><i>Physical Scale (sCAL)</i></h2>
<ul type="disc">
<li>Status: officially registered (PNG Extensions)
<li>Location: before first IDAT
<li>Multiple: no
</ul>
<!--
<h4 class="refsect1">Description</h4>
-->
<p>
<a name="INDEX-875" />
<a name="INDEX-876" />
PNG's sCAL chunk is similar to pHYs, except that instead of measuring the
size of the image pixels relative to each other or to an original, physical
image, sCAL measures their size relative to the actual subject matter of the
image. For example, an astronomical image may span a certain number of radians
in each direction, or an aerial photograph of Earth may cover a given number of
kilometers.</p>
<p><a href="#png.ch11.tbl.5">Table 11-5</a> shows the format of sCAL; it is quite simple.
</p>
<a name="png.ch11.tbl.5" />
<div class="table" align="center">
<p>
<table width="502" border="0">
<tr>
<td>
<b class="emphasis-bold">Table 11-5.</b>
<i>sCAL Chunk</i>
</td>
</tr>
</table>
</p>
<p>
<table width="502" border="1">
<tr>
<td><b class="emphasis-bold">Field</b></td>
<td><b class="emphasis-bold">Length and Valid Range
<?x-space 2p?><?x-space 2p?></b></td>
</tr>
<tr>
<td>Unit specifier</td>
<td>1 byte (1, 2)</td>
</tr>
<tr>
<td>Units per pixel, <em class="emphasis">x</em> axis</td>
<td><em class="emphasis">m</em> bytes (positive floating-point)</td>
</tr>
<tr>
<td>Null separator</td>
<td>1 byte (0)</td>
</tr>
<tr>
<td>Units per pixel, <em class="emphasis">y</em> axis</td>
<td><em class="emphasis">n</em> bytes (positive floating-point)</td>
</tr>
<?x-space 5p?>
</table>
</p>
</div>
<p>Two units are defined: meters (unit specifier = 1) and radians (unit
specifier = 2). The size of a pixel in the given units, both
horizontally and vertically, is given by a pair of positive
floating-point numbers encoded as ASCII strings and separated by a
null byte. The most general form of a floating-point string includes
an optional leading sign (<b class="emphasis-bold">+</b> or <b class="emphasis-bold">-</b>), zero or more
decimal digits (0-9, the ``integer part''), an optional decimal
point followed by zero or more decimal digits (the ``fractional
part''), and an optional <b class="emphasis-bold">e</b> or <b class="emphasis-bold">E</b> followed by an
optional sign and one or more digits (the ``exponent part''). Either
the integer part or the fractional part must contain at least one
digit, but everything else may be omitted. Thus, <b class="emphasis-bold">1</b> and
<b class="emphasis-bold">.1</b> are valid floating-point numbers, as is
<b class="emphasis-bold">+123.4567e-089</b>. Note that the exponent is interpreted as a
power of 10 (10<sup class="superscript">-89</sup> in the third
example) to be multiplied by the integer and fractional parts; this is the
computer version of what is sometimes referred to as <em class="emphasis">scientific notation</em>.
</p>
</div>
<div class="refentry">
<a name="png.ch11.div.10" />
<h2 class="sect1"><i>Image Offset (oFFs)</i></h2>
<ul type="disc">
<li>Status: officially registered (PNG Extensions)
<li>Location: before first IDAT
<li>Multiple: no
</ul>
<!--
<h4 class="refsect1">Description</h4>
-->
<p>
<a name="INDEX-877" />
<a name="INDEX-878" />
For images that are available separately but envisioned as part of a greater
whole, the image-offset chunk, oFFs, can be used to specify the absolute
positioning of each. The most common example is positioning on a printed
page, especially in conjunction with the pHYs chunk.</p>
<p>The layout of the chunk is given in
<a href="#png.ch11.tbl.6">Table 11-6</a>.
</p>
<a name="png.ch11.tbl.6" />
<div class="table" align="center">
<p>
<table width="502" border="0">
<tr>
<td>
<b class="emphasis-bold">Table 11-6.</b>
<i>oFFs Chunk</i>
</td>
</tr>
</table>
</p>
<p>
<table width="502" border="1">
<tr>
<td><b class="emphasis-bold">Field</b></td>
<td><b class="emphasis-bold">Length and Valid Range
<?x-space 2p?><?x-space 2p?></b></td>
</tr>
<tr>
<td>Image position, <em class="emphasis">x</em> axis</td>
<td>4 bytes (-2,147,483,647 to +2,147,483,647)</td>
</tr>
<tr>
<td>Image position, <em class="emphasis">y</em> axis</td>
<td>4 bytes (-2,147,483,647 to +2,147,483,647)</td>
</tr>
<tr>
<td>Unit specifier</td>
<td>1 byte (0, 1)</td>
</tr>
<?x-space 5p?>
</table>
</p>
</div>
<p>Valid units are either pixels (unit specifier = 0) or microns<a href="#FOOTNOTE-90">[90]</a>
(unit specifier = 1). The image position is measured from the top and left
edges of the page (whether real or virtual); an image that is intended
to be partly cut off may have negative offsets.
</p>
<blockquote class="footnote">
<a name="FOOTNOTE-90" /><p>[90] Microns are more properly known as micrometers
<!-- (<img src="images/U03BC.png" alt="[mu]" />m); -->
(μm);
there are one million of them in a meter, or 25,400 in an inch.</p>
</blockquote>
</div>
<div class="refentry">
<a name="png.ch11.div.11" />
<h2 class="sect1"><i>Pixel Calibration (pCAL)</i></h2>
<ul type="disc">
<li>Status: officially registered (PNG Extensions)
<li>Location: after PLTE, before first IDAT
<li>Multiple: no
</ul>
<!--
<h4 class="refsect1">Description</h4>
-->
<p>
<a name="INDEX-879" />
<a name="INDEX-880" />
The pCAL chunk is currently the only registered scientific-visualization
extension to PNG, though it was moved into the regular PNG Extensions document
as part of the general PNG spec revision process in October 1998. It is also
the most mathematical of any approved chunk. Its purpose is to efficiently
encode the relevant conversions between the integer samples in a PNG file
and the physical quantity being represented by the image. Two conversions
are represented: a linear conversion between the PNG samples and the
original samples and a more general conversion from the original samples
to the physical values they represent. The first mapping is often the
identity mapping (i.e., the original samples are equal to the PNG samples),
but it need not be.</p>
<p>The layout of the pCAL chunk is presented in
<a href="#png.ch11.tbl.7">Table 11-7</a>.
</p>
<a name="png.ch11.tbl.7" />
<div class="table" align="center">
<p>
<table width="502" border="0">
<tr>
<td>
<b class="emphasis-bold">Table 11-7.</b>
<i>pCAL Chunk</i>
</td>
</tr>
</table>
</p>
<p>
<table width="502" border="1">
<tr>
<td><b class="emphasis-bold">Field</b></td>
<td><b class="emphasis-bold">Length and Valid Range</b></td>
</tr>
<tr>
<td>Calibration name</td>
<td>1-79 bytes (Latin-1 text)</td>
</tr>
<tr>
<td>Null separator</td>
<td>1 byte</td>
</tr>
<tr>
<td>Original zero, X<sub class="subscript">0</sub></td>
<td>4 bytes (signed integer)</td>
</tr>
<tr>
<td>Original maximum, X<sub class="subscript">1</sub></td>
<td>4 bytes (signed integer)</td>
</tr>
<tr>
<td>Equation type</td>
<td>1 byte</td>
</tr>
<tr>
<td>Number of parameters, N</td>
<td>1 byte</td>
</tr>
<tr>
<td>Unit name</td>
<td><em class="emphasis">n</em> bytes (Latin-1 text)</td>
</tr>
<tr>
<td>Null separator</td>
<td>1 byte</td>
</tr>
<tr>
<td>Parameter 0, P<sub class="subscript">0</sub></td>
<td><em class="emphasis">p</em><sub class="subscript">0</sub> bytes (ASCII floating-point text)</td>
</tr>
<tr>
<td>Null separator</td>
<td>1 byte</td>
</tr>
<tr>
<td> ...</td>
<td> ...</td>
</tr>
<tr>
<td>Parameter <em class="emphasis">L</em>, P<sub class="subscript">L</sub></td>
<td><em class="emphasis">p</em><sub class="subscript">L</sub> bytes (ASCII floating-point text)</td>
</tr>
<tr>
<td colspan="2">(Note: <em class="emphasis">L</em> = N-1)</td>
<td />
</tr>
<?x-space 5p?>
</table>
</p>
</div>
<p>The unit name is a label, such as <b class="emphasis-bold">kg/(m^3)</b> or <b class="emphasis-bold">Mpc</b>, that
applies to the
physical quantity represented by the image samples. Dimensionless data may
either include a descriptive string (e.g., ``fractal iteration count'') or
leave the unit field a null string. There are no restrictions on the length
of the unit name.</p>
<p>The X<sub class="subscript">0</sub> and X<sub class="subscript">1</sub> parameters encode the linear conversion.
For an 8-bit sample depth, the PNG samples range from 0 to 255; more generally,
they range from 0 to M, where M = 2<sup class="superscript"><em class="emphasis">bit depth</em></sup> - 1. Most often,
X<sub class="subscript">0</sub> will equal 0 and X<sub class="subscript">1</sub> will
equal M, indicating that the PNG samples are the same as the original samples.
But this need not be the case, and either of X<sub class="subscript">0</sub> or X<sub class="subscript">1</sub> may
be positive or negative; the only restriction is that they may not be equal to
each other. The conversion is done using integer arithmetic, according to the
following equation:</p>
<blockquote>
<pre class="code">
original_sample = (PNG_sample * (X<sub class="subscript">1</sub>-X<sub class="subscript">0</sub>) + M/2) / M + X<sub class="subscript">0</sub>
</pre>
</blockquote>
<p>The inverse mapping is:</p>
<blockquote>
<pre class="code">
PNG_sample = ((original_sample - X<sub class="subscript">0</sub>) * M + (X<sub class="subscript">1</sub>-X<sub class="subscript">0</sub>)/2) / (X<sub class="subscript">1</sub>-X<sub class="subscript">0</sub>)
</pre>
</blockquote>
<p>Note that integer arithmetic here means that fractional values are rounded
toward minus infinity, not toward zero; there's no difference for positive
values, but for negative values, there is. Also keep in mind that the PNG
samples are limited to the range [0,M] regardless of what the inverse mapping
might give.</p>
<p>The more general conversion, between original samples and actual physical
values, can be represented by one of four possible equation types: linear
(type 0), exponential (type 1), exponential with arbitrary base (type 2),
or hyperbolic sinusoidal (type 3). The number of parameters required by each
is 2, 3, 3, and 4, respectively, and the parameters are stored in the same
ASCII floating-point format as described for the sCAL chunk earlier. The
equations use floating-point arithmetic, not integer, and are given by:</p>
<blockquote>
<pre class="code">
[0] physical_value = P<sub class="subscript">0</sub> + P<sub class="subscript">1</sub> * original_sample/(X<sub class="subscript">1</sub>-X<sub class="subscript">0</sub>)
[1] physical_value = P<sub class="subscript">0</sub> + P<sub class="subscript">1</sub> * e<sup class="superscript">P<sub class="subscript">2</sub> * original_sample/(X<sub class="subscript">1</sub>-X<sub class="subscript">0</sub>)</sup>
[2] physical_value = P<sub class="subscript">0</sub> + P<sub class="subscript">1</sub> * P<sub class="subscript">2</sub><sup class="superscript">original_sample/(X<sub class="subscript">1</sub>-X<sub class="subscript">0</sub>)</sup>
[3] physical_value = P<sub class="subscript">0</sub> + P<sub class="subscript">1</sub> * sinh(P<sub class="subscript">2</sub>*(original_sample - P<sub class="subscript">3</sub>)/(X<sub class="subscript">1</sub>-X<sub class="subscript">0</sub>))
</pre>
</blockquote>
<p>Equation types 1 and 2 are equivalent in the sense that the same types of
functions can be represented by either one; both are defined for convenience.
For RGB or RGBA image types, the equations are applied to each of the color
sample values independently, while for palette images, the equations are
applied to the color sample values in the palette, not to the index values.
<?x-need 10?></p>
<p>Equation type 3 may seem odd, but it allows floating-point data to be
reduced to integer data in such a way that the resolution of the
integer data is roughly proportional to the magnitude of the original
floating-point data. That is, for 32-bit original data and 16-bit PNG
samples, the resolution near zero is around
10<sup class="superscript">-33</sup>, and near
<!--
<u>+</u>10<sup class="superscript">31</sup>
-->
±10<sup class="superscript">31</sup>
it is around 10<sup class="superscript">28</sup>.
To put it another way, the resolution everywhere is about 0.4% (or 1/256) of
the magnitude.</p>
</div>
<div class="refentry">
<a name="png.ch11.div.12" />
<h2 class="sect1"><i>Fractal Parameters (fRAc)</i></h2>
<ul type="disc">
<li>Status: officially registered (PNG Extensions)
<li>Location: anywhere
<li>Multiple: yes
</ul>
<!--
<h4 class="refsect1">Description</h4>
-->
<p><a name="INDEX-881" />The fRAc chunk is unique in that it was officially registered as a PNG
extension in 1995 but, as of early 1999, still had not actually been specified.
Intended to store parameters pertaining to the generation of fractal images,
the chunk is clearly useful only to a very specialized set of programs. As
a result, its design was left in the hands of experts--specifically, the
authors of Fractint, which is one of the most general fractal programs ever
written and probably the most popular. But for technical reasons relating
to Fractint's 16-bit origins, PNG support was not added as planned, so
design of the fRAc chunk was deferred pending a rewrite of the program as a
32-bit application.
</p>
</div>
<div class="refentry">
<a name="png.ch11.div.13" />
<h2 class="sect1"><i>GIF Conversion Info (gIFg, gIFx)</i></h2>
<ul type="disc">
<li>Status: officially registered (PNG Extensions)
<li>Location: anywhere
<li>Multiple: yes
</ul>
<!--
<h4 class="refsect1">Description</h4>
-->
<p>
<a name="INDEX-882" />
<a name="INDEX-883" />
Since PNG originated as an intended replacement for GIF, one requirement for
the new format was to be able to store all possible GIF information in one
form or another. Part of that requirement is addressed by chunks we've
already described. Within GIF's Logical Screen Descriptor (the global
header that immediately follows the GIF signature bytes), the Pixel Aspect
Ratio, Color Resolution, and Background Color Index fields map to pHYs, sBIT,
and bKGD, respectively. Note that Background Color Index only applies if
there is a Global Color Table, however. Within the Image Descriptor, the
Image Left Position and Image Top Position fields map to oFFs. And within
the Graphic Control Extension, the Transparent Color Index maps to tRNS.
This is summarized in
<a href="#png.ch11.tbl.8">Table 11-8</a>.
</p>
<a name="png.ch11.tbl.8" />
<div class="table" align="center">
<p>
<table width="502" border="0">
<tr>
<td>
<b class="emphasis-bold">Table 11-8.</b>
<i>Correspondence Between GIF Fields and Standard PNG Chunks</i>
</td>
</tr>
</table>
</p>
<p>
<table width="502" border="1">
<tr>
<td><b class="emphasis-bold">GIF Block</b></td>
<td><b class="emphasis-bold">GIF Variable Name</b></td>
<td><b class="emphasis-bold">PNG Chunk
<?x-space 2p?><?x-space 2p?></b></td>
</tr>
<tr>
<td rowspan=3>Logical Screen Descriptor</td>
<td>Pixel Aspect Ratio</td>
<td>pHYs</td>
</tr>
<tr>
<td>Color Resolution</td>
<td>sBIT</td>
</tr>
<tr>
<td>Background Color Index</td>
<td>bKGD</td>
</tr>
<?x-space 2p?><?x-space 2p?><tr>
<td rowspan=2>Image Descriptor</td>
<td>Image Left Position</td>
<td>oFFs</td>
</tr>
<tr>
<td>Image Top Position</td>
<td>oFFs</td>
</tr>
<?x-space 2p?><?x-space 2p?><tr>
<td>Graphic Control Extension</td>
<td>Transparent Color Index</td>
<td>tRNS</td>
</tr>
<?x-space 5p?>
</table>
</p>
</div>
<p>The remainder of the requirement that PNG be able to store all GIF information
is addressed by two of PNG's three GIF extension chunks. Both correspond
directly to GIF89a extensions: the Graphic Control Extension (gIFg) and the
Application Extension (gIFx). The third chunk, gIFt, turns out to be an
unintended special case; it is discussed separately later.</p>
<p><a name="INDEX-884" />GIF's Graphic Control Extension is most commonly used to indicate
transparency, for which it corresponds most closely to PNG's tRNS
chunk. But it is also used in multi-image GIFs to provide timing and
compositing information. Although this is more properly the realm of
MNG, PNG's multi-image cousin (which I'll discuss in <a href="chapter12.html">Chapter 12, "Multiple-Image Network Graphics"</a>, PNG
also supports the conversion of a multi-image GIF into several
single-image PNGs. The gIFg chunk is used to encode the
nontransparency information in the GIF extension block so that
lossless conversion back to an animated GIF is possible.
</p>
<p>The gIFg chunk, shown in
<a href="#png.ch11.tbl.9">Table 11-9</a>, contains only three fields.
</p>
<a name="png.ch11.tbl.9" />
<div class="table" align="center">
<p>
<table width="502" border="0">
<tr>
<td>
<b class="emphasis-bold">Table 11-9.</b>
<i>gIFg Chunk</i>
</td>
</tr>
</table>
</p>
<p>
<table width="502" border="1">
<tr>
<td><b class="emphasis-bold">Field</b></td>
<td><b class="emphasis-bold">Length and Valid Range
<?x-space 2p?><?x-space 2p?></b></td>
</tr>
<tr>
<td>Disposal method</td>
<td>1 byte (0-3)</td>
</tr>
<tr>
<td>User input</td>
<td>1 byte (0, 1)</td>
</tr>
<tr>
<td>Delay time</td>
<td>2 bytes (0-65,535)</td>
</tr>
<?x-space 5p?>
</table>
</p>
</div>
<p>The interpretation and value of each field are identical to those in part 23
of the GIF89a Specification, with the exception that the 2-byte delay time
<a name="INDEX-885" />
<a name="INDEX-886" />is stored in big-endian order (most significant byte first) in gIFg, whereas
GIF integers are stored in little-endian format. PNG decoders may treat the
delay time (measured in hundredths of a second) as the maximum amount of time
to display the image before going on to the next one, if any, but it is likely
that most decoders will ignore the chunk entirely.
</p>
<p><a name="INDEX-887" />GIF's Application Extension is simply a way for an application to include its
own information in the image; it corresponds exactly to a private chunk in a
PNG image. The format is given in
<a href="#png.ch11.tbl.10">Table 11-10</a>.
</p>
<a name="png.ch11.tbl.10" />
<div class="table" align="center">
<p>
<table width="502" border="0">
<tr>
<td>
<b class="emphasis-bold">Table 11-10.</b>
<i>gIFx Chunk</i>
</td>
</tr>
</table>
</p>
<p>
<table width="502" border="1">
<tr>
<td><b class="emphasis-bold">Field</b></td>
<td><b class="emphasis-bold">Length and Valid Range
<?x-space 2p?><?x-space 2p?></b></td>
</tr>
<tr>
<td>Application identifier</td>
<td>8 bytes (printable ASCII characters)</td>
</tr>
<tr>
<td>Authentication code</td>
<td>3 bytes</td>
</tr>
<tr>
<td>Application data</td>
<td><em class="emphasis">n</em> bytes</td>
</tr>
<?x-space 5p?>
</table>
</p>
</div>
<p>The contents of gIFx are a direct transcription of the GIF data, with the sole
exception that any GIF sub-blocks are deblocked into a flat stream.</p>
</div>
<div class="refentry">
<a name="png.ch11.div.14" />
<h2 class="sect1"><i>GIF Plain Text (gIFt)</i></h2>
<ul type="disc">
<li>Status: officially deprecated (PNG Extensions)
<li>Location: anywhere
<li>Multiple: yes
</ul>
<!--
<h4 class="refsect1">Description</h4>
-->
<p>
<a name="INDEX-888" />
<a name="INDEX-889" />
<a name="INDEX-890" />
GIF's Plain Text Extension is a way to define an image composed entirely of
text without actually storing the text as a bitmapped image. It defines a
rectangular grid of character cells into which text characters of the
specified foreground and background colors are placed, starting from the
upper left and proceeding left to right and top to bottom; the decoder
chooses the font that is the closest match to the specified size.</p>
<p>A casual reading of the GIF specification might suggest that the Plain Text
Extension defines a method for cheaply overlaying fixed-width text on top
of ordinary pixel data--and, indeed, that was probably the primary
motivation behind the extension. But a more careful inspection reveals that
the Plain Text Extension is treated as a separate subimage within the GIF
stream, on equal terms with any block of bitmap data. It may, in fact, be
the <em class="emphasis">only</em> graphic rendering block within the stream. And since PNG
images are required to include bitmap data (i.e., IDAT chunks), allowing
GIF Plain Text information to be included is perilously close to sanctioning
multi-image PNGs. Largely because of this, the gIFt chunk was officially
deprecated in October 1998. It is still allowed for backward compatibility
(the horses have already left the barn, so to speak), but the current
recommendation is that all decoders ignore the chunk and that encoders not
write it in the first place. In fact, it is quite possible that no encoder
or decoder ever did support gIFt; the Plain Text Extension was rarely used
even in GIF's heyday, and even gif2png
(see <a href="chapter05.html">Chapter 5, "Applications: Image Converters"</a>) never supported it.</p>
<p>In any case, the format of the gIFt chunk is as shown in
<a href="#png.ch11.tbl.11">Table 11-11</a>.
</p>
<a name="png.ch11.tbl.11" />
<div class="table" align="center">
<p>
<table width="502" border="0">
<tr>
<td>
<b class="emphasis-bold">Table 11-11.</b>
<i>gIFt Chunk</i>
</td>
</tr>
</table>
</p>
<p>
<table width="502" border="1">
<tr>
<td><b class="emphasis-bold">Field</b></td>
<td><b class="emphasis-bold">Length and Valid Range
<?x-space 2p?><?x-space 2p?></b></td>
</tr>
<tr>
<td>Text grid left position, pixels</td>
<td>4 bytes (0-2,147,483,647)</td>
</tr>
<tr>
<td>Text grid top position, pixels</td>
<td>4 bytes (0-2,147,483,647)</td>
</tr>
<tr>
<td>Text grid width, pixels</td>
<td>4 bytes (0-2,147,483,647)</td>
</tr>
<tr>
<td>Text grid height, pixels</td>
<td>4 bytes (0-2,147,483,647)</td>
</tr>
<tr>
<td>Character cell width, pixels</td>
<td>1 byte (0-255)</td>
</tr>
<tr>
<td>Character cell height, pixels</td>
<td>1 byte (0-255)</td>
</tr>
<tr>
<td>Text foreground color</td>
<td>3 bytes (R, G, B samples, 0-255 each)</td>
</tr>
<tr>
<td>Text background color</td>
<td>3 bytes (R, G, B samples, 0-255 each)</td>
</tr>
<tr>
<td>Plain text data</td>
<td><em class="emphasis">n</em> bytes</td>
</tr>
<?x-space 5p?>
</table>
</p>
</div>
<p>There are several differences from the GIF data structure. The actual text
in the GIF block is divided into sub-blocks of between 1 and 255 bytes;
the PNG plain text data is a single stream. In addition to the reversed order
for integer values (big-endian in PNG), gIFt's width and height fields for
the grid are 4 bytes each, twice as big as in GIF. The position fields are
also twice as wide, which makes little
sense from a preserve-the-GIF-data standpoint, but apparently was chosen for
consistency with PNG's image-offset chunk. Both the Plain Text Extension
and oFFs give positions relative to a logical page, not relative to the main
image; thus, in the presence of oFFs data, the gIFt positions should be adjusted
accordingly. Note that this may not be possible if the PNG image
uses microns in the oFFs chunk and has no pHYs chunk--in that case, there
is no conversion information between pixels (the only unit defined for gIFt)
and microns.</p>
<p><a name="INDEX-891" />Possibly the biggest difference, however, is that the Plain Text Extension
is affected by the Graphic Control Extension, which means it implicitly
includes transparency and timing effects. PNG's gIFt chunk does not include
any transparency information, so effectively there is no way to float the
gIFt text over the main image by giving it a transparent background color.
This limitation appears to have been an oversight in the design of the PNG
chunk and was another reason for its official deprecation. On the other
hand, if the gIFt chunk appears before the first IDAT chunk, a hypothetical
gIFt-aware PNG decoder might assume that the text amounts to a background
image and render the pixel data on top of it, applying any transparency
effects the main image possesses.</p>
</div>
<div class="refentry">
<a name="png.ch11.div.15" />
<h2 class="sect1"><i>Other Chunks</i></h2>
<!--
<h4 class="refsect1">Description</h4>
-->
<p>Several other chunks were proposed but never approved as official extensions,
mainly due to the perceived lack of need for them.
<a name="INDEX-892" />The alignment chunk (aLIG, had it been approved) would have provided centering
and baseline information about an image so that it could be aligned more cleanly
with surrounding text; this would have been most useful for images
<a name="INDEX-893" />with transparent edges. The fingerprint chunk (fING) would have provided a
16-byte MD5 fingerprint of the raw image data, a type of cryptographic
signature that could be used to test whether two images were identical.
Neither aLIG nor fING was ever put up for a vote, and both proposals have long
since expired.</p>
<p><a name="INDEX-894" />There were also three proposed scientific-visualization chunks, all of which
were rejected in formal voting. The false-color chunk (fALS) would have
provided false-color information for grayscale images, such as might be used
to highlight a tumor in a medical scan or a shock front in a hydrodynamic
<a name="INDEX-895" />
<a name="INDEX-896" />
<a name="INDEX-897" />simulation. The calibration chunks (xSCL and ySCL, but also known as xCAL
and yCAL in later proposals) were similar to sCAL in providing information
about the physical characteristics of an image subject but would have allowed
offsets and different units along the two axes; they thus would have provided
full calibration data, not just scaling information.</p>
<p>Note that any of these chunks may be resurrected in the future, as PNG becomes
more widely used and as the needs of various PNG-using communities evolve.
</p>
</div>
<PRE>
</PRE>
<hr> <!-- -=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=- -->
<a href="chapter10.html"><img width=24 height=13 border=0 align="left"
src="images/prev.png" alt="<-"></a>
<a href="chapter12.html"><img width=24 height=13 border=0 align="right"
src="images/next.png" alt="->"></a>
<div align="center">
<a href="chapter10.html"><font size="-1" color="#000000"
><b>PREVIOUS</b></font></a> <a
href="toc.html"><font size="-1" color="#000000"
><b>CONTENTS</b></font></a> <a
href="chapter12.html"><font size="-1" color="#000000"
><b>NEXT</b></font></a>
</div>
<hr> <!-- -=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=- -->
</body></html>
|