From e0c6872cf40896c7be36b11dcc744620f10adf1d Mon Sep 17 00:00:00 2001 From: Norbert Preining Date: Mon, 2 Sep 2019 13:46:59 +0900 Subject: Initial commit --- systems/doc/dvipdfmx/dvipdfmx.tex | 2240 +++++++++++++++++++++++++++++++++++++ 1 file changed, 2240 insertions(+) create mode 100644 systems/doc/dvipdfmx/dvipdfmx.tex (limited to 'systems/doc/dvipdfmx/dvipdfmx.tex') diff --git a/systems/doc/dvipdfmx/dvipdfmx.tex b/systems/doc/dvipdfmx/dvipdfmx.tex new file mode 100644 index 0000000000..b346ef1315 --- /dev/null +++ b/systems/doc/dvipdfmx/dvipdfmx.tex @@ -0,0 +1,2240 @@ +\documentclass[a4paper,xetex,oneside]{book} +\usepackage{xltxtra} +\usepackage{fontspec} +\usepackage{microtype} +\usepackage{xunicode} +\usepackage{unicode-math} +%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% +%% PLATFORM DEPENDENT INSTRUCTION HERE +%% Font setup for body text. +\setmainfont{Constantia} +\setsansfont{Cambria} +\setmonofont[Scale=0.92]{Consolas}% Adjust xheight difference +\setmathfont{Cambria Math} +%% The following section is for showing some fancy examples. +%% Some fonts used here may not be availabel on your system. +%% Please modify this. Just replacing with empty macros is OK. +\newcommand{\jpzerofourexamples}{{% +\setmainfont[Scale=5,RawFeature=+jp04]{SourceHanSerifJP-Light.otf}葛祇逢}} +\newcommand{\jpninezeroexamples}{{% +\setmainfont[Scale=5,RawFeature=+jp90]{SourceHanSerifJP-Light.otf}葛祇逢}} +%% We use PostScript raw code here to test dvipdfmx's capability. +%% TFM files `uprml' and `uprmlv' are distributed with upTeX. +%% TrueType font `yumindb.ttf' is bundled with Windows 10. +\newcommand{\jphoritext}{% +\makebox[112bp][l]{% +\raisebox{88bp}[112bp][0bp]{% +\special{pdf:mapline uprml UniJIS-UTF8-H yumindb.ttf} +\special{ps: uprml findfont 16 scalefont setfont + currentpoint moveto (「こんにちは」) show}% +}}} +\newcommand{\jpverttext}{% +\makebox[16bp][l]{% +\raisebox{112bp}[112bp][0bp]{% +\special{pdf:mapline uprmlv UniJIS-UTF8-V yumindb.ttf} +\special{ps: uprmlv findfont 16 scalefont setfont + currentpoint moveto (「こんにちは」) show}% +}}} +%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% +\usepackage{xcolor,hyperref,hyperxmp} +\hypersetup{% + bookmarksnumbered=true,% + colorlinks=true,% + urlcolor=[rgb]{0.2,0.2,0.4}, + linkcolor=[rgb]{0.2,0.2,0.4}, + citecolor=[rgb]{0.2,0.2,0.4}, + pdftitle={The Dvipdfmx User's Manual},% + pdfauthor={The dvipdfmx project team},% + pdfsubject={A User's Manual for Dvipdfmx and Xdvipdfmx.},% + pdfkeywords={dvipdfmx, XeTeX, TeX, LaTeX}, + pdflang=en, + pdfcopyright={Copyright © The dvipdfmx project team. + Permission is granted to copy, distribute and/or modify this document + under the terms of the GNU Free Documentation License, Version 1.3 + or any later version published by the Free Software Foundation; + with no Invariant Sections, no Front-Cover Texts, and no Back-Cover Texts. + A copy of the license is included in the section entitled "GNU + Free Documentation License".}, + pdflicenseurl={http://www.gnu.org/licenses/fdl.html} +} +\usepackage{listings} +\lstset{ + keepspaces=true, + basicstyle={\ttfamily}, + frame={tb}, + breaklines=true, + columns=[l]{fullflexible}, + numbers=none, + xrightmargin=2em, + xleftmargin=2em, + aboveskip=2em, + belowskip=2em +} +\usepackage{mflogo} +\usepackage{lipsum} +\usepackage{array} +\usepackage{marginnote} +\renewcommand*{\marginfont}{\footnotesize\itshape} +\usepackage[noorphans,font=itshape]{quoting} +\newcommand{\package}[1]{\textit{#1}} +\newcommand{\code}[1]{\mbox{\texttt{#1}}} +\newcommand{\keyword}[1]{\textit{#1}} +\newcommand{\option}[1]{\mbox{`\texttt{#1}'}} +\newcommand{\dvipdfm}{\texttt{dvipdfm}} +\newcommand{\dvipdfmx}{\texttt{dvipdfmx}} +\newcommand{\xdvipdfmx}{\texttt{xdvipdfmx}} +\newcommand{\deprecated}[1]{\marginnote{\addfontfeatures{Color=CC3333}#1}} +\newcommand{\newfeature}[1]{\marginnote{\addfontfeatures{Color=3366CC}#1}} +\newcommand{\lnum}[1]{{\addfontfeatures{RawFeature=+lnum}#1}} +% For placeing drawings via \special +\newcommand{\specialbox}[3]{% +\makebox[#1][l]{\raisebox{-#2}[0bp][#2]{\special{#3}}}} +\usepackage{fancyhdr} +\pagestyle{fancy} +%%DVIPDFMX SETTINGS +\AtBeginDocument{% + \special{dvipdfmx:config O 1}% + \special{dvipdfmx:config V 7} +} +\begin{document} + +\begin{titlepage} + \begin{raggedleft} + {\Huge\bfseries The Dvipdfmx User's Manual}\\[\baselineskip] + \Large The Dvipdfmx Project Team\\ + December 23, 2018\par + \end{raggedleft} +\end{titlepage} + +\tableofcontents + +\chapter{Getting Started} + +\section{Introduction} + +The \dvipdfmx\ (formerly \dvipdfm-cjk) project provides an extended version of +the \dvipdfm, a DVI to PDF translator developed by Mark~A.~Wicks. + +The primary goal of this project is to support multi-byte character encodings +and large character sets such as for East Asian languages. +This project started as a combined work of the \dvipdfm-jpn project by +Shunsaku Hirata and its modified one, \dvipdfm-kor, by Jin-Hwan Cho. + +Extensions to \dvipdfm\ include, +\begin{itemize} + \item Support for OpenType and TrueType font, including partial support + for the OpenType Layout for finding glyph variants and vertical writing. + \item Support for CJK-\LaTeX\ and H\LaTeX\ with Subfont Definition Files. + \item Support for various legacy multi-byte encodings via PostScript CMap + Resources. + \item Unicode related features: Unicode as an input encoding and + auto-creation of ToUnicode CMaps. + \item Support for p\TeX\ (a Japanese localized variant of \TeX) including + vertical writing extension. + \item Some extended DVI specials. + \item Reduction of output files size with on-the-fly Type1 to CFF (Type1C) + conversion and PDF object stream. + \item Advanced raster image support including alpha channels, embedded + ICC profiles, 16-bit bit-depth colors, and so on. + \item Basic PDF password security support for PDF output. +\end{itemize} +Some important features are still missing: +\begin{itemize} + \item Linearization. + \item Color Management. + \item Resampling of images. + \item Selection of compression filters. + \item Variable font and OpenType 1.8. + \item and many more... +\end{itemize} + +\dvipdfmx\ is now maintained as part of \TeX\ Live. Latest source code can +be found at the \TeX\ Live SVN repository. For an instruction on accessing the +development sources for \TeX\ Live, see,\medskip + +\url{http://www.tug.org/texlive/svn/} +\medskip + +This document, ``The \dvipdfmx\ User's Manual'', was originally prepared for +\TeX\ Live 2017. Current maintainer of this document is Shunsaku Hirata. +Latest version and contact information can be found at:\medskip + +\url{http://github.com/shirat74/dvipdfm-x-doc} +\medskip + +\noindent{}Please send questions or suggestions. + +\subsection{\xdvipdfmx} + +\xdvipdfmx\ is an extended version of \dvipdfmx, and is now incorporated into +\dvipdfmx. + +The \xdvipdfmx\ extensions provides support for the Extended DVI (.xdv) format +generated by \XeTeX\ which includes support for platform-native fonts and the +\XeTeX\ graphics primitives, as well as Unicode text and OpenType font. + +\XeTeX\ originally used a Mac-specific program called \code{xdv2pdf} as a +backend program instead of \xdvipdfmx. The \code{xdv2pdf} program supported +a couple of special effects that are not yet available through \xdvipdfmx: +The Quartz graphics-based shadow support, AAT ``variation'' fonts such as Skia, +transparency as an attribute of font, and so on. +It would be nice if they continue to be supported. Suggestions and help are +welcomed. + + +\subsection{Legal Notice} + +Copyright © The dvipdfmx project team. +Permission is granted to copy, distribute and/or modify this document +under the terms of the GNU Free Documentation License, Version 1.3 +or any later version published by the Free Software Foundation; +with no Invariant Sections, no Front-Cover Texts, and no Back-Cover Texts. +A copy of the license is included in the section entitled +``\hyperref[SEC:FDL]{GNU Free Documentation License}''. + +\section{Installation and Usage} + +Typical usage and installation steps are not different from the original +\dvipdfm. Please refer documents from \dvipdfm\ distribution for detailed +instruction on how to install and how to use \dvipdfm. The \dvipdfm\ manual is +available from its CTAN site:\medskip + +\url{http://www.ctan.org/tex-archive/dviware/dvipdfm} +\medskip + +The minimal requirements for building \dvipdfmx\ is the \keyword{kpathsea} library. +the \keyword{zlib} library for compression and the \keyword{libpng} library for PNG +inclusion are highly recommended. +Optionally, the \keyword{libpaper} library might be used to handle paper size. + +This document mainly focuses on the additions and modifications to \dvipdfm. +Please refer the +``\href{http://mirrors.ctan.org/dviware/dvipdfm/dvipdfm.pdf}{Dvipdfm User's Manual}'' +available from the CTAN site mentioned above for basic usage. + +Some additional command line options recognized by \dvipdfmx\ are listed in +Table~\ref{TABLE:options}. In addition to this, the \code{-V} option for specifying +the output PDF version now accepts the version specification of a form \code{2.0}. Try +\begin{lstlisting} +dvipdfmx --help +\end{lstlisting} +for the list of command line options and their explanations. + +\begin{table} + \centering + \begin{tabular}{lp{8cm}}\hline + Option & Description \\ \hline\hline + \code{-C} \textit{number} & Specify miscellaneous option flags. See, + section of ``\hyperref[SEC:compatibility]{Incompatible Changes}'' for + details. \\ + \code{-S} & Enable PDF encryption. \\ + \code{-K} \textit{number} & Set encryption key length. The default value + is 40.\\ + \code{-P} \textit{number} & Set permission flags for PDF encryption. + The \textit{number} is a 32-bit unsigned integer representing permission + flags. + See, section of ``\hyperref[SEC:encryption]{Encryption Support}''. + The default value is \code{0x003C}.\\ + \code{-I} \textit{number} & Life of image cache in hours. By specifying + value \code{0} \dvipdfmx\ erases cached images, and value \code{-1} + erases all cached images and does not leave newly generated one. The + default value is \code{-2}. (ignore image cache) \\ + \code{-M} & Process \MP\ generated PostScript file.\\ + \code{-E} & Always try to embed fonts \emph{regardless of + licensing}.\\ + \code{-O} \textit{number} & Set maximum depth of open bookmark item.\\ + \hline + \end{tabular} + \caption{Additional command line options recognized by \dvipdfmx.}% + \label{TABLE:options} +\end{table} + + +\section{Quick Guide} + +\dvipdfmx\ is supposed to be used by users of \LaTeX\ packages for typesetting +CJK languages like H\LaTeX\ and CJK-\LaTeX, and \TeX\ variants such as \XeTeX, +p\TeX, and up\TeX. +This section is intended to be a quick guide for each users. + +\subsection{\texorpdfstring{\XeTeX}{XeTeX}} + +\XeTeX\ users +normally do not invoke the \dvipdfmx\ command directly. To control the +behavior of \dvipdfmx, please consider using the \code{dvipdfmx:config} +special explained in the section of ``\hyperref[SEC:specials]{Specials}''. +Some part of this document is irrelevant for \XeTeX\ users. + +\subsection{p\TeX} +p\TeX\ users are at least required to install several auxiliary files +mentioned in the section of ``\hyperref[SEC:auxfiles]{Auxiliary Files}'' +and to setup font-mappings. Just install the \package{adobemappings} and +\package{glyphlist} for auxiliary files. (As \TeX\ Live basic installation +requires them, they are probably already installed for \TeX\ Live users.) + +For \TeX\ Live users, setting up fontmaps can be easily done with the help +of the \package{ptex-fontmaps} package and the \keyword{updmap} program. +To use with the IPAex fonts (contained in the \package{ipaex} package), for +example, run, +\begin{lstlisting} +kanji-config-updmap --sys ipaex +\end{lstlisting} +where the \option{--sys} option indicates the system-wide configuration. +After successful invocation of the above command, the IPAex fonts will +be used by default. The current setting can be checked via, +\begin{lstlisting} +kanji-config-updmap --sys status +\end{lstlisting} +For more information on the updmap program, try, +\begin{lstlisting} +kanji-config-updmap --help +\end{lstlisting} +or refer the documentation of the updmap program. + +Alternatively, just for a quick test of installation, try the following: +\begin{lstlisting} +\documentclass{article} +\begin{document} +\special{pdf:mapline rml H KozMinProVI-Regular} +...Some Japanese text goes here... +\end{document} +\end{lstlisting} +In this example, PDF viewer which can handle substitute font is required since +\dvipdfmx\ does not embed fonts. + +For using Japanese text in PDF document information and annotations, put +the following \code{special} command, +\begin{lstlisting} +\AtBeginDocument{\special{pdf:tounicode 90ms-RKSJ-UCS2}} +\end{lstlisting} +in the preamble. +The above \code{special} command instructs \dvipdfmx\ to convert text encoded +in Shift-JIS to Unicode. For EUC-JP, replace 90ms-RKJK-UCS2 with EUC-UCS2. + +\subsection{up\TeX} +up\TeX\ users are basically the same as p\TeX\ users but there are two choices +for setting fontmaps. Setup fontmaps as mentioned above for p\TeX, or use +keyword \code{unicode} in the encoding field of the fontmap file. + +The former might be easier as the auto-creation of fontmap files can be done +with the updmap program and the \package{ptex-fontmaps} package. But in +this method there are some difficulties when using fonts which employ character +collections (glyph repertoire) other than Adobe-Japan\lnum{1} in the case of PostScript +flavored OpenType fonts. +In the later case, the \package{adobemappings} package is not required +and newer PostScript flavored OpenType fonts which do not employ Adobe-Japan\lnum{1} +can be easily used too. + +Using \code{unicode} is more simpler and intuitive thus it is recommended to +use this method.\footnote{For \TeX\ Live 2017. Earlier versions have buggy +support.} +A typical example fontmap entries for using Adobe's SouceHan series +might look like: +\begin{lstlisting} +urml unicode SourceHanSerifJP-Light.otf +urmlv unicode SourceHanSerifJP-Light.otf -w 1 +ugbm unicode SourceHanSansJP-Medium.otf +ugbmv unicode SourceHanSansJP-Medium.otf -w 1 +\end{lstlisting} + +As in p\TeX, the following \code{special} instruction might be necessary for PDF +document information and annotations to be shown correctly: +\begin{lstlisting} +\AtBeginDocument{\special{pdf:tounicode UTF8-UCS2}} +\end{lstlisting} +Here, input encoding is assumed to be UTF-8. + +\subsection{CJK-\LaTeX} + +CJK-\LaTeX\ users are required to have \keyword{Subfont Definition Files} +to be installed. They are available as part of the \package{ttfutils} package. + +To use TrueType Arphic fonts provided by the \package{arphic-ttf} package: +\begin{lstlisting} +\documentclass{article} +\usepackage{CJKutf8} +...Other packages loaded here... +\AtBeginDocument{% + \special{pdf:tounicode UTF8-UCS2}% + \special{pdf:mapline bsmiu@Unicode@ unicode bsmi00lp.ttf}% + } +\begin{document} +\begin{CJK}{UTF8}{bsmi} +...some Chinese text goes here... +\end{CJK} +\end{document} +\end{lstlisting} +Here, \code{pdf:mapline} special is used to setup a font-mapping. + +\section{Auxiliary Files}\label{SEC:auxfiles} + +This section is mostly for supporting legacy encodings and legacy font format +such as PostScript Type1 font. \XeTeX\ users may skip this section. + +\dvipdfmx\ has a capacity to handle various input encodings from 7-bit +encodings to variable-width multi-byte encodings. It also has some sort of +Unicode support. Several auxiliary files which are not common to \TeX\ users +are needed to enable those features. This section shortly describes about them. + +\subsection{PostScript CMap Resources} + +\keyword{PostScript CMap Resources}\footnote{See, +``\href{http://www.adobe.com/content/dam/Adobe/en/devnet/font/pdfs/5014.CIDFont\_Spec.pdf}% +{Adobe CMap and CIDFont Files Specification}''} +are required for supporting legacy encodings such as Shift-JIS, EUC-JP, \lnum{Big5}, +and other East Asian encodings. \dvipdfmx\ internally identifies glyphs with +identifiers (CIDs\footnote{PostScript terminology ``Character IDentifier''.}) +represented as an integer ranging from 0 to 65535 in the CID-based glyph access. +PostScript CMap Resources describes the mapping between sequences of input +character codes and CIDs. \dvipdfmx\ has an extensible support for multi-byte +encodings via PostScript CMap Resources. + +CMap files for standard East Asian encodings, for use with Adobe's character +collections, are included in the \package{adobemapping} package. +The latest version of those CMap files maintained by Adobe can be found at +Adobe's GitHub Project page:\medskip + +\url{http://github.com/adobe-type-tools/cmap-resources} +\medskip + +Those files are mandatory for supporting p\TeX. up\TeX\ users may also +want to install them but they are not required. + +\subsection{Subfont Definition Files} + +CJK fonts usually contain several thousands of glyphs. For using such fonts +with (original) \TeX, which can only handle 8-bit encodings, it is necessary to +split a font into several \keyword{subfonts}. The Subfont Definition File (SFD) +specify the way how those fonts are split into subfonts. \dvipdfmx\ uses SFD +files to convert a set of subfonts back to a single font. + +SFD files are not required for use with \TeX\ variants which can handle +multi-byte character encodings and large character sets such as p\TeX, +up\TeX,\XeTeX, and Omega. +H\LaTeX\ and CJK-\LaTeX\ users are required to have those files to be +installed. SFD files are available as a part of the \package{ttfutils} package +for \TeX\ Live users. + +\subsection{The Adobe Glyph List and ToUnicode Mappings} + +The Adobe Glyph List\footnote{See, +``\href{http://github.com/adobe-type-tools/agl-specification}{Adobe Glyph List Specification}''} +(AGL) describes correspondence between PostScript glyph names (e.g., \code{AE}, +\code{Aacute},...) and Unicode character sequences representing them. +Some features described in the section ``Unicode Support'' requires AGL file. + +\dvipdfmx\ looks for the file \code{glyphlist.txt} when conversion from +PostScript glyph names to Unicode sequences is necessary. +This conversion is done in various situations; when creating ToUnicode CMaps +for 8-bit encoding fonts, finding glyph descriptions from TrueType and OpenType +fonts when the font itself does not provide a mapping from PostScript glyph +names to glyph indices (version 2.0 ``post'' table), and when the encoding +\code{unicode} is specified for Type1 font. + +The AGL file is included in the \package{glyphlist} package. The latest version +can be found at Adobe's GitHub site:\medskip + +\url{http://github.com/adobe-type-tools/agl-aglfn} +\medskip + +ToUnicode Mappings are similar to AGL but they describe correspondence +between CID numbers (instead of glyph names) and Unicode values. +The content of those files are the same as CMap Resources. +They are required when using TrueType fonts emulated as a CID-keyed font. +They should be found in the same directory as ordinary CMap files. + +ToUnicode Mapping files are included in the \package{adobemapping} package. +Those files are not required for \XeTeX\ users. + +\section{Overview of Extensions} + +This section gives a quick overview of \dvipdfmx's extended capabilities. + +\subsection{CJK Support} + +There are many extensions made for supporting CJK languages. Features described +here is mainly for CJK languages but their use is actually not limited to it. +Those features are implemented in a generic way so that it can be beneficial to +users who are not involved in CJK languages. + +\subsubsection{Legacy Multi-byte Encodings} + +\dvipdfmx\ has an extensible support for encodings by means of +PostScript CMap Resources. Just like \code{enc} files are written for 8-bit +encodings, one can write their own CMap files to support custom encodings. +See, Adobe's technical notes\cite{ADOBE} for details on PostScript CMap Resources. + +\subsubsection{Vertical Writing} + +\dvipdfmx\ supports the vertical writing extension used by p\TeX\ and up\TeX. +A DVI instruction to set the writing mode is supported. The OpenType Layout +GSUB Feature is supported for selecting vertical version of glyphs. + +\begin{figure} +\centering +\jphoritext\hspace{24pt}\jpverttext% +\caption{An example of horizontal and vertical text; +left and right corner brackets are replaced with their vertical counterparts.}% +\label{FIG:verttext} +\end{figure} + +\subsection{Unicode Support} + +Unicode support here consists of two parts: Supporting Unicode as an input +encoding and making output PDF files ``Unicode aware''. + +\subsubsection{Unicode as Input Encoding} + +\dvipdfmx\ recognizes an additional keyword \code{unicode} in fontmap files +to declare that Unicode values are used in the input DVI file. Unicode support is +basically limited to the Basic Multilingual Plane (BMP) since there are no +support for code ranges that requires more than three bytes in TFM and extended +TFM formats. + +\subsubsection{ToUnicode CMap Support} + +In PDF, it is often the case that text is not encoded in Unicode. +However, modern applications usually want them represented in Unicode to make +it usable as text information. +The ToUnicode CMap is a bridge between PDF text string encodings and Unicode +encodings, and makes it possible to extract text in PDF files as +Unicode encoded strings. It is important to make resulting PDF search-able and +copy-and-past-able. Dvipdfmx supports auto-creation of ToUnicode CMaps. + +It will not work properly for multiply encoded glyphs due to fundamental +limitations of Unicode conversion mechanism with ToUnicode CMaps. + +\subsection{Other Extensions} + +\dvipdfmx\ can generate encrypted PDF documents to protect its contents from +unauthorized access. It is limited to password-based authentication, and +public-key based authentication is not supported. The 256-bit AES encryption is +also supported for PDF version 1.7 and 2.0 setting although it may not be supported +by PDF viewers. + +There are various other improvements over \dvipdfm. The most notable one is +more improved PDF input and output support: The cross-reference stream and +object stream introduced in \lnum{PDF-1.5} are also supported. + + +\chapter{Graphics}\label{SEC:graphics} + +\section{Image Inclusion} + +The basics of incorporating images into output PDF is the same as in \dvipdfm. +To do this, \LaTeX\ users can simply use the \package{graphicx} package. +(possibly with the driver option \code{dvipdfmx}) +This section is \emph{not} for providing a how-to guide to include images but +just for supported graphics and image formats along with supported features. + +Graphics support was mostly rewritten in \dvipdfmx. +Support for BMP and JPEG\lnum{2000} was added. An effort to preserve more +information originally found in included images, e.g., embedded ICC Profiles +and XMP Metadata, was made. + +However, \dvipdfmx\ does not support various features common to graphics +manipulation programs such as resampling, color conversion, and selection of +compression filters. Thus, it is recommended to use other programs specialized +in image manipulation for preparation of images. + +\subsection{Supported Graphics File Formats} + +Supported formats are, PNG, JPEG, \lnum{JPEG2000}, BMP, PDF, and \MP\ generated +EPS. All other format images, such as SVG and PostScript, must be converted to +PDF before inclusion. The \option{-D} option, as in \dvipdfm, can be used for +filtering images. + +\subsubsection{Notes on PNG Support} + +PNG is supported as in \dvipdfm\ with many improvements. + +PNG support includes most of important features of PNG format such as color +palette, transparency, 16-bit bit-depth color, embedded ICC Profiles, +calibrated color, and embedded XMP Metadata. + +In including PNG images, \dvipdfmx\ first decompresses image data and then +compresses (if requested) it again. +For better compression ratio, a preprocessing filter (Predictor filter) might +be applied before compression. +\dvipdfmx\ supports the TIFF Predictor 2 and the PNG optimum filter. +However, there is yet no way to specify which predictor function is to be used +and currently PNG optimum filter is always employed. + +Predictor filter is a preprocessing filter to image data for improving compression. +Using a predictor filter usually gives better compression +but in many cases compression speed becomes significantly slower. +Try \option{-C 0x20} command line option to disable predictor filters and to +check if slowness is due to filtering. + +For the PNG optimum filter, a heuristic approach, ``minimum sum of absolute +differences'', is employed for finding the most optimal filter to be used. +See, discussion in the PNG Specification ''Filter selection'':\medskip + +\url{http://www.w3.org/TR/2003/REC-PNG-20031110/\#12Filter-selection} +\medskip + +As built-in support for the sRGB color space is absent in PDF, +the sRGB color can only be represented precisely by means of the sRGB ICC Profile. +However, for sRGB color PNG images, \dvipdfmx\ uses an approximate calibrated +RGB color space instead. +For approximating the sRGB color, the gamma and CIE \lnum{1931} chromaticity values +mentioned in the PNG Specification is used. +See, the following page for more information:\medskip + +\url{http://www.w3.org/TR/2003/REC-PNG-20031110/\#11sRGB} +\medskip + +\dvipdfmx\ also supports calibrated color with the \code{gAMA} and the \code{cHRM} chunk. +These chunks carry information on more accurate color representation. +Some software programs, however, write only \code{cHRM} but do not record the gamma value +although the PNG specification recommends to do so. Gamma value 2.2 is assumed if only +\code{cHRM} is present but \code{gAMA} is not. + +Some PNG features are unavailable depending on output PDF version setting. Please refer +Table~\ref{TABLE:PNGfeat} for more details. + +\begin{table} + \centering + \begin{tabular}{lp{8cm}}\hline + Feature & PDF Version Required \\ \hline\hline + 16-bit Color Depth & Version 1.5 \\ + Transparency & Full support for alpha channel requires PDF version 1.4. + Color key masking (a specific color is treated as fully transparent) + requires 1.3.\\ + XMP Metadata & Version 1.4 \\ + ICC Profile & Version 1.3 \\ + \hline + \end{tabular} + \caption{PNG features and corresponding PDF versions required.}% + \label{TABLE:PNGfeat} +\end{table} + +\subsubsection{JPEG and \lnum{JPEG2000}} + +JPEG format is supported as in \dvipdfm. In addition to this, \lnum{JPEG2000} +is also supported. + +JPEG and \lnum{JPEG2000} image inclusion is basically done as ''pass-through'', +that is, image data is not decompressed before inclusion. So, although these +formats are basically lossy, there should be no quality loss when including +images. + +JPEG is relatively well supported. \dvipdfmx\ supports embedded ICC Profiles +and CMYK color. Embedded XMP metadata is also preserved in the output PDF. +JFIF or Exif data might be used to determine image's physical size. + +As the PDF specification does not require information irrelevant to +displaying images to be embedded, \dvipdfmx\ does not embed whole data. +Especially, not all application specific data is retained. Application +specific data such as JFIF, Exif, and \code{APP14} Adobe marker are +preserved. +Please note that XMP and Exif data which may contain location information +where the photo was taken is always preserved in the output PDF file. + +\lnum{JPEG2000} is also supported. It is restricted to JP2 and JPX baseline +subset as required by the PDF specification. It is not well supported and still in +an experimental stage. J2C format and transparency are not supported. + +\subsubsection{PDF Support} + +PDF inclusion is supported as in \dvipdfm, with various important enhancement +over \dvipdfm\ for more robust inclusion. \dvipdfmx\ can handle cross-reference +streams and object streams introduced in \lnum{PDF-1.5}. +\dvipdfmx\ also supports inclusion of PDF pages other than the first page. +However, tagged PDF may cause problems and annotations are not kept. + +As there are no clear way to determine the natural extent of the graphics content +to be clipped, \dvipdfmx\ preferably try to find the \emph{crop box} to decide +image size. If there are no crop box \emph{explicitly} specified,% +\footnote{There are some accusations by Japanese \TeX\ users as +''violating the PDF spec.'' regarding this point. However, what we are +talking about is how to guess the \emph{natural} or \emph{intended} size of +images but not the default value of the PDF crop box itself.} then it +tries to refer other boundary boxes such as the \emph{art box}. If there are no +possible boundaries of the graphics content explicitly specified, the \emph{media box}, +which is the boundaries of the physical medium on which the page is to be printed, +is used as the last resort. + +The \code{pdf:image} special supports additional keys, ``\code{page}'' and +``\code{pagebox}''. The \code{page} key takes an integer value representing +the page number of PDF page to be included, and the \code{pagebox} takes one +of the keywords \code{mediabox}, \code{cropbox}, \code{artbox}, \code{bleedbox}, +or \code{trimbox} for selecting page's boundary box to be used. For example, +\begin{lstlisting} +\special{pdf:image pagebox artbox page 3 (foo.pdf)} +\end{lstlisting} +includes 3rd page of `foo.pdf' with the boundary box set to the art box. + +\subsubsection{Other Image Formats} + +For \MP\ generated Encapsulated PostScript (EPS) files, multi-byte encoding +support is added. +\dvipdfmx\ also supports ``\MP\ mode''. When \dvipdfmx\ is invoked with +\option{-M} option, it enters in \MP\ mode and processes a \MP\ +generated EPS file as an input.% +\footnote{\code{prologue} should be set to \code{2}.} + +BMP support is also added. It is limited to uncompressed or RLE-compressed +raster images. Extensions are not (won't be) supported. + +For image formats not natively supported, the \code{-D} option can be +used to convert images to PDF format before inclusion, as in \dvipdfm. +In \dvipdfmx, the letter \code{v} in the \code{-D} option argument is expanded +to the output PDF version. + +\subsection{Image Cache} + +Caching of images generated via filtering command specified with \option{-D} +option is supported. It solves the problems that image inclusion becomes very +slow when external filtering program such as GhostScript is invoked each time +images are included. + +Use \option{-I} option to enable this feature: +\begin{lstlisting} +-I 24 +\end{lstlisting} +where the integer represents the life of cache files, 24 hours in this example. +Zero and negative values have a special meaning. +Value 0 for ``erase old cached images while leaving newly created one'', +-1 for ``erase all cached images'', and -2 for ``ignore image cache''. +Default value is set to -2. + +\section{Graphics Drawing} + +\dvipdfmx\ does not offer a high level interface to draw graphics objects. +A possible way to draw graphics is to write raw PDF graphics drawing codes and +then to insert them into the output via \code{special} commands. + +To show an example, the following code: +\begin{lstlisting} +\special{pdf:content + 1 0 0 1 0 0 cm + 0 100 m + 80 100 120 80 120 0 c + S +} +\end{lstlisting} +draws a Bézier curve, +\begin{center} +\specialbox{120bp}{100bp}{pdf:content 1 0 0 1 0 0 cm 0 100 m 80 100 120 80 120 0 c S} +\end{center} +The \code{pdf:content} special is used here which is useful for inserting an +isolated graphics object. + +The above example illustrates a typical example of PDF graphics drawing. +It consists of three parts; setting graphics state, constructing a path, and painting +a path. A Graphic object are specified as a sequence of operators and their operands +using \emph{postfix notation}. \keyword{Graphics state operators} comes first, +\code{cm} in this example sets the current transformation matrix (CTM). Then, +\keyword{path construction} operators follow; move to position $(0, 100)$, +append a Bézier curve from current point to $(120, 0)$ with control points +$(80, 100)$ and $(120, 80)$. Finally, a \keyword{path painting} operator +comes to draw the constructed path. +In this example the stroking operator \code{S} +is used. + +\subsection{The \code{pdf:content} Special} + +The \code{pdf:content} special can be used for drawing an \emph{isolated} +graphics object. It sets the origin of graphics drawing operators supplied to +this command to the position where it is inserted. +The whole content is enclosed by a pair of graphics state save-restore +operators. So for example, a color change made within a \code{pdf:content} +command takes an effect only within the content of this special. + +\subsection{Guide to PDF Graphics Operators} + +PDF employs essentially the same imaging model as PostScript. +So, it is easy to learn about PDF graphics drawing for people who are +well accustomed to PostScript. This section is intended to be a short guide +for PDF graphics operators. + +\subsubsection{Graphics State Operators} + +The \code{cm} operator modifies CTM by concatenating the specified matrix. +Operands given to this operators are six numbers each representing +transformation matrix elements: +translation represented as $[1, 0, 0, 1, t_x, t_y]$, +scaling $[s_x, 0, 0, s_y, 0, 0]$, +rotation $[\cos\theta, \sin\theta, -\sin\theta, \cos\theta, 0, 0]$. + +To uniformly scale the object, just use +\begin{lstlisting} +2.0 0 0 2.0 0 0 cm +\end{lstlisting} + +The \code{w} operator sets the line width, e.g., `\code{2 w}' sets the line +width to 2. Here is an example of drawing a rotated rectangle: +\begin{lstlisting} +0.866 0.5 -0.5 0.866 30 2 cm 5 w 0 0 100 50 re S +\end{lstlisting} +\begin{center} +\specialbox{120bp}{96bp}{pdf:content 0.866 0.5 -0.5 0.866 30 2 cm 5 w 0 0 100 50 re S}% +\end{center} + +Transformations can be sequentially applied; for the above example, +\begin{lstlisting} +1 0 0 1 30 2 cm 0.866 0.5 -0.5 0.866 0 0 cm +5 w 0 0 100 50 re S +\end{lstlisting} +gives the same result. + +To specify colors, use \code{RG}, \code{rg}, \code{K}, and \code{k} operators, +for RGB or CMYK color for stroking (upper-case) and nonstroking (lower-case). +\begin{lstlisting} +0.866 0.5 -0.5 0.866 30 2 cm 5 w +1 0.4 0 0 K 1 0 0 0 k +0 0 100 50 re B +\end{lstlisting} +where the \code{B} operator fill and then stroke the path. +\begin{center} +\specialbox{120bp}{96bp}{pdf:content 0.866 0.5 -0.5 0.866 30 2 cm 5 w +1 0.4 0 0 K 1 0 0 0 k 0 0 100 50 re B} +\end{center} + +A dash pattern can be specified with the \code{d} operator. +Operands for this operator are the \keyword{dash array} and +the \keyword{dash phase}. +For example, to specify a dash pattern with 6-on 4-off starting with phase 0: +\begin{lstlisting} +[6 4] 0 d 2 w 0 0 m 320 0 l S +\end{lstlisting} +draws the following dashed line: +\begin{center} +\specialbox{320bp}{30bp}{pdf:content +1 0 0 1 0 20 cm [6 4] 0 d 2 w 0 0 m 320 0 l S} +\end{center} + +Other important operators are \code{q} and \code{Q}, which saves and restores +the graphics state. +\begin{lstlisting} +1 0 0 1 30 2 cm +q +0.866 0.5 -0.5 0.866 0 0 cm +[6 4] 0 d 2 w 0 0 100 50 re S +Q +-30 0 m 90 0 l S +0 -2 m 0 96 l S +\end{lstlisting} +In the above example, \code{d}, \code{w}, and rotation only take effect within +the \code{q}-\code{Q} block. The portion drawing two straight lines is unaffected +by these graphics state change. +\begin{center} +\specialbox{120bp}{96bp}{pdf:content +1 0 0 1 30 2 cm +q +0.866 0.5 -0.5 0.866 0 0 cm +[6 4] 0 d 2 w 0 0 100 50 re S +Q +-30 0 m 90 0 l S +0 -2 m 0 96 l S +} +\end{center} + +For a (incomplete) list of graphics state operators, see +Talbe~\ref{TAB:operatorsGS}. + +\begin{table} + \centering + \begin{tabular}{llp{7.6cm}}\hline + Operands & Operator & Description\\ \hline\hline + --- & \code{q} & Save the current graphics state.\\ + --- & \code{Q} & Restore the previously saved graphics state.\\ + $a$ $b$ $c$ $d$ $e$ $f$ & \code{cm} & Modify the current transformation matrix + by concatenating the specified matrix.\\ + \textit{width} & \code{w} & Set the line width.\\ + \textit{array} \textit{phase} & \code{d} & Set the line dash pattern.\\ + $r$ $g$ $b$ & \code{RG} & Set the stroking color space to RGB and + set the stroking color as specified.\\ + $r$ $g$ $b$ & \code{rg} & Set the nonstroking color space to RGB and + set the nonstroking color as specified.\\ + $c$ $m$ $y$ $k$ & \code{K} & Set the stroking color space to CMYK and + set the stroking color as specified.\\ + $c$ $m$ $y$ $k$ & \code{k} & Set the nonstroking color space to CMYK and + set the nonstroking color as specified.\\ + \hline + \end{tabular} + \caption{A few examples of graphics state operators and color operators.}% + \label{TAB:operatorsGS} +\end{table} + +\subsubsection{Path Construction Operators} + +A path construction normally starts with a \code{m} operator which moves the +current point to the specified position and then sequences of other path +construction operators follow. The path currently under construction is +called the \keyword{current path}. A sequence of path construction operators +appends segments of path to the current path and then move the +\keyword{current point} to the end point of appended path. +A typical sequence of path construction looks like, +\begin{lstlisting} +100 50 m +100 78 78 100 50 100 c +22 100 0 78 0 50 c +0 22 22 0 50 0 c +78 0 100 22 100 50 c +S +\end{lstlisting} +\begin{center} +\specialbox{120bp}{80bp}{pdf:content +100 50 m +100 78 78 100 50 100 c +22 100 0 78 0 50 c +0 22 22 0 50 0 c +78 0 100 22 100 50 c +S} +\end{center} +This example is an approximated circle drawn by four Bézier curves. + +Table~\ref{TAB:operators} shows a list of path construction operators. +Those who are accustomed to the PostScript language should note that in PDF +the current path is not a part of the graphics state, +and hence is \emph{not} saved and restored along with the other graphics state +parameters. + +\begin{table} + \centering + \begin{tabular}{llp{6.9cm}}\hline + Operands & Operator & Description \\ \hline\hline + $x$ $y$ & \code{m} & Begin a new path by moving the current point specified by given operands.\\ + $x$ $y$ & \code{l} & Append a line segment from the current point to the point specified.\\ + $x_1$ $y_1$ $x_2$ $y_2$ $x_3$ $y_3$ & \code{c} & Append a Bézier curve to the current path. + Two Control points and the end point given as operands.\\ + $x_2$ $y_2$ $x_3$ $y_3$ & \code{v} & Append a Bézier curve to the current path. Using the current + point and first two operand as the Bézier control points.\\ + $x_1$ $y_1$ $x_3$ $y_3$ & \code{y} & Append a Bézier curve to the current path. The second + control point coincides with the end point.\\ + --- & \code{h} & Close the current path by appending a straight line segment from the current point + to the starting point of the path.\\ + $x$ $y$ \textit{width} \textit{height} & \code{re} & Append a rectangle. First two operands for the + position of lower-left corner, third and forth operand representing width and + height.\\ + \hline + \end{tabular} + \caption{List of path construction operators. All operators move the current point to the end + point of appended path.}\label{TAB:operators} +\end{table} + +\subsubsection{Path Painting Operators} + +There are basically four kind of path painting operators: \code{S}, \code{f}, +\code{B}, and \code{n}. The first three for ``stroke'', ``fill'', and +``fill then stroke'' operators respectively, and the last one \code{n} +paints nothing but end the path object. For filling operators, there are +two kind of operators depending on how ``insideness'' of points are determined: +the \keyword{non-zero winding number rule} and the \keyword{even-odd rule}. +The even-odd rule operators are \code{f*} and \code{B*}. + +The following example illustrates the difference: +\begin{lstlisting} +0 0 100 100 re 20 20 60 60 re f +1 0 0 1 120 0 cm +0 0 100 100 re 20 20 60 60 re f* +\end{lstlisting} +\begin{center} +\specialbox{220bp}{100bp}{pdf:content +0 0 100 100 re 20 20 60 60 re f +1 0 0 1 120 0 cm +0 0 100 100 re 20 20 60 60 re f*} +\end{center} +The ``interior'' of the ``inner'' square has a non-zero even winding number. +(In this example, counter-clockwise direction is assumed for both of +two \code{re} operators.) + +\chapter{Specials} + +\section{PDF Specials}\label{SEC:specials} + +\dvipdfmx\ recognizes various special commands originally introduced in +\dvipdfm. Please refer to the ``Dvipdfm User's Manual''\cite{DVIPDFM} for detailed +information on PDF specials. + +\subsection{Additions to PDF Specials} + +Several \code{special} commands are added for more flexible PDF generation: +creation of arbitrary stream objects, controlling \dvipdfmx\ behavior, and some +specials which might be useful for graphics drawing. + +\subsubsection{PDF Object Manipulation} + +PDF object manipulation is a key feature of PDF specials. +The \code{pdf:fstream} special is added in \dvipdfmx\ which enables creation of +PDF stream object from an existing file. The syntax of this special is, +\begin{lstlisting} +pdf:fstream @identifier (filename) <> +\end{lstlisting} +where identifier and filename (specified as a PDF string object) are +mandatory and a dictionary object, following the filename, which is to be added +to the stream dictionary is optional. + +For example, to incorporate XMP Metadata from a file \code{test.xmp}, +\begin{lstlisting} +\special{pdf:fstream @xmp (test.xmp) << + /Type /Metadata + /Subtype /XML +>>} +\special{pdf:put @catalog << /Metadata @xmp >>} +\end{lstlisting} + +Similarly, \code{pdf:stream} special can be used to create a PDF stream +object from a PDF string instead of a file. +\begin{lstlisting} +pdf:stream @identifier (stream contents) <> +\end{lstlisting} + +This special might be useful for creating a simple image inline. +\begin{lstlisting} +\special{pdf:stream @myim01 + <5500AAC05500AAC05500AAC05500AAC05500> + << + /Type /XObject + /Subtype /Image + /BitsPerComponent 1 + /ColorSpace /DeviceGray + /Width 9 + /Height 9 + >> +} +\special{pdf:put @resources << + /XObject << /MyIM01 @myim01 >> +>>} +\special{pdf:content 81 0 0 81 0 0 cm /MyIM01 Do} +\end{lstlisting} +The above example draws an image like Figure~\ref{FIG:stream}. +\begin{figure} + \centering + \makebox[81bp][l]{\raisebox{-81bp}[0bp][81bp]{% + \special{pdf:stream @myim01 <5500AAC05500AAC05500AAC05500AAC05500> + << + /Type /XObject + /Subtype /Image + /BitsPerComponent 1 + /ColorSpace /DeviceGray + /Width 9 + /Height 9 + >>} + \special{pdf:put @resources << + /XObject << /MyIM01 @myim01 >> + >>} + \special{pdf:content 81 0 0 81 0 0 cm /MyIM01 Do} +}} +\caption{An image created by \code{pdf:stream} special.}% +\label{FIG:stream} +\end{figure} + +\subsubsection{Controlling Font Mappings} + +\code{pdf:mapline} and \code{pdf:mapfile} specials can be used to append a +fontmap entry or to load a fontmap file: +\begin{lstlisting} +pdf:mapline foo unicode bar +pdf:mapfile foo.map +\end{lstlisting} + +\subsubsection{Specifying Output PDF Version} + +\code{pdf:majorversion} and \code{pdf:minorversion} specials can be used to +specify major and minor version of output PDF. +\begin{lstlisting} +pdf:minorversion 3 +\end{lstlisting} + +Please note that this command must appear on the first page, +otherwise it will be ignored. + +\subsubsection{Custom File Identifiers} + +\newfeature{Addition in \TeX\ Live 2019} + +A custom file identifier (the \code{ID} entry in the trailer dictionary) +can be specified via \code{pdf:trailerid} special as +\begin{lstlisting} +pdf:trailerid [(0123456789abcdef) (0123456789abcdef)] +\end{lstlisting} +An array of two 16-byte PDF string objects must be supplied as +a file identifier. This special command must appear on the first page. + +\subsubsection{Encryption} + +To protect output PDF with encryption, use \code{pdf:encrypt} special +\begin{lstlisting} +pdf:encrypt userpw (foo) ownerpw (bar) length 128 perm 20 +\end{lstlisting} +where user-password (\code{userpw}) and owner-password (\code{ownerpw}) must be +specified as PDF string objects. (which can be empty) +Numbers specifying key-length and permission flags here are decimal numbers. +See, section ``\hyperref[SEC:encryption]{Encryption Support}'' for a brief +description of permission flags. + +\subsubsection{PDF Document Creation} + +\newfeature{Addition in \TeX\ Live 2019} + +As a convenience, the \code{pageresources} special is added, which puts given page +resources into subsequent page's \keyword{Resource Dictionary}. For example, +\begin{lstlisting} +\special{pdf:pageresources << + /ExtGState << /MyGS01 << /ca 0.5 /CA 0.5 >> >> +>>} +\end{lstlisting} +puts an ExtGState resource named \code{MyGS01} into the current page's and all +subsequent pages' resource dictionary. + +Other notable extensions are \code{code}, \code{bcontent}, and +\code{econtent}. The \code{code} special can be used to insert raw PDF graphics +instructions into the output. It is different from \dvipdfm's +\code{content} special in that it does not enclose contents with a \code{q} +and \code{Q} (save-restore of graphics state) pair. +A typical usage of this special is: +\begin{lstlisting} +\special{pdf:code q 1 Tr} +...some text goes here... +\special{pdf:code Q} +\end{lstlisting} +which changes text rendering mode to 1, as shown in Figure~\ref{FIG:trmode}. + +\begin{figure} +\centering +\mbox{\special{pdf:code q 1 w 1 Tr}% +{\fontsize{200pt}{200pt}\selectfont\textchi}% +\special{pdf:code Q}}% +\caption{A character drawn in the PDF text rendering mode 1.}% +\label{FIG:trmode} +\end{figure} + +Be careful on using this special as it is very easy to generate +broken PDF files. The \code{bcontent} and \code{econtent} pair is somewhat +fragile and might be incompatible to other groups of special commands. +It may not always be guaranteed to work as `expected'. + +\subsection{ToUnicode Special} + +PDF allows users to attach various additional information such as document information, +annotation, and navigation information (like bookmarks) to their document. +All human-readable text, \keyword{text string}, contained in such information must +be encoded either in \keyword{PDFDocEncoding} or UTF-16BE with Unicode byte order +marker. However, as many users can't prepare their document with text strings +properly encoded, there is a special kind of special command, \code{pdf:tounicode}, +which can be used to convert text strings into the appropriate Unicode form. +Note that this feature is provided just as a remedy for users incapable of encoding +text strings properly. + +For example, +\begin{lstlisting} +\special{pdf:tounicode 90ms-RKSJ-UCS2} +\end{lstlisting} +declares that \emph{some} of text strings should be re-encoded according to specified +conversion CMap \code{90ms-RKSJ-UCS2}. + +Conversion is done only for arguments to several PDF special commands such as +\code{docinfo}, \code{ann}, and \code{out} but not for that of general object creation +specials. Note that not all PDF string objects are subject to this conversion. +By default, only dictionary entries listed below are converted. +\begin{lstlisting} +Title Author Subject Keywords Creator Producer Contents Subj +TU T TM +\end{lstlisting} + +\newfeature{Addition in \TeX\ Live 2019} + +The list of dictionary entries subject to conversion can be extended by supplying +additional dictionary keys as an array object: +\begin{lstlisting} + \special{pdf:tounicode 90ms-RKSJ-UCS2 [/RC /DS]} +\end{lstlisting} + +If the name of conversion CMap contains one of the keywords RKSJ, \lnum{B5}, GBK, and KSC, +PDF string objects are treated specially when they are parsed. A two byte sequence +starting with the first byte's high bit set is treated ``as is'' so that +the \code{0x5c} byte appears in the second byte is not treated as an escape sequence. +This behavior is not compliant to the PDF specification. + +\subsection{PDF Special Examples} + +This section shows several examples of special command usage. +It is intended to be a hint for advanced users, so uninterested users can safely skip +this section. In many cases, using \dvipdfmx\ PDF specials requires knowledge on PDF. +Please refer to Adobe's ``PDF Reference''\cite{ADOBE}. + +\subsubsection{Annotations} + +In this section, some useful special commands for creating \keyword{annotations} +are explained. Note that viewer support is required for annotations to be +displayed correctly. + +First start with a very simple \keyword{Text Annotation} for attaching a comment. +This feature is supported by many PDF viewer applications. +\marginnote{\special{pdf:ann width 20bp height 20bp + << + /Type /Annot + /Subtype /Text + /Name /Comment + /T (Text Annotation Example) + /Subj (An Example of Text Annotations) + /Contents (A Quick Brown Fox Jumped Over The Lazy Dog.) + >> +}} + +\begin{lstlisting} +\special{pdf:ann width 20bp height 20bp + << + /Type /Annot + /Subtype /Text + /Name /Comment + /T (Text Annotation Example) + /Subj (An Example of Text Annotations) + /Contents (A Quick Brown Fox Jumped Over The Lazy Dog.) + >> +} +\end{lstlisting} +\code{pdf:ann} special is used to create an annotation. A small icon shall +be shown on the side margin. Here, dictionary entry \code{T} is for the tilte, +\code{Subj} for the subject of this annotation, and \code{Contents} for +the text string to be shown in the body of this annotation. + +Likewise, \keyword{Rubber Stamp Annotation}, which places a rubber stamp +like figure, +\begin{center} +\specialbox{150bp}{50bp}{pdf:ann width 150bp height 50bp + << + /Type /Annot + /Subtype /Stamp + /Name /Approved + >> +} +\end{center} + +\begin{lstlisting} +\special{pdf:ann width 150bp height 50bp + << + /Type /Annot + /Subtype /Stamp + /Name /Approved + >> +} +\end{lstlisting} +Other keywords such as \code{Expired}, \code{Final}, \code{Draft}, and so on, +can be used in place of \code{Approved}. + +One can create stamps of their own style. For this purpose, other specials +\code{pdf:bxobj} and \code{pdf:exobj} can be used for designing stamps. +Those specials ``capture'' all typeset material enclosed by them into a PDF +\keyword{Form XObject}, which is a reusable graphics object like included images. + +For a simple example, +\begin{lstlisting} +\special{pdf:bxobj @MyStamp + width 280pt height 0pt depth 40pt} +\addfontfeature{Scale=4,Color=FF9933}My Own Stamp +\special{pdf:exobj} +\end{lstlisting} +It captures typeset material ``My Own Stamp'' (this example uses \code{fontspec} +package's command for changing font size and text color) into the object labeled as +\code{MyStamp} for later reuse. +Then, \code{AP} (\keyword{appearance dictionary}) entry for controlling the appearance +of annotations is used as, +\begin{lstlisting} +\special{pdf:ann width 280pt height 40pt + << + /Type /Annot + /Subtype /Stamp + /AP << /N @MyStamp >> + >> +} +\end{lstlisting} +The image captured into the object \code{MyStamp} is used as ``Normal'' +(\code{AP} dictionary entry \code{N}) appearance. +(\code{R} for ``Rollover'' and \code{D} for ``Down'' can be used.) + +\parbox[t][0pt][c]{280pt}{% +\special{pdf:bxobj @MyStamp width 280pt height 0pt depth 40pt}% +\addfontfeature{Scale=4,Color=FF9933}My Own Stamp% +\special{pdf:exobj}% +}% + +The result is: +\begin{center} +\specialbox{280bp}{40bp}{pdf:ann width 280pt height 40pt + << + /Type /Annot + /Subtype /Stamp + /AP << /N @MyStamp >> + >> +} +\end{center} + +With the following code, \dvipdfmx\ reads the source file and creates a stream +object named \code{SourceFile}, and then creates file attachment annotation. +\marginnote{% +\special{pdf:fstream @SourceFile (\jobname.tex)}% +\special{pdf:ann width 10bp height 20bp + << + /Type /Annot + /Subtype /FileAttachment + /FS << + /Type /Filespec + /F (\jobname.tex) + /EF << /F @SourceFile >> + >> + /Name /PushPin + /C [0.8 0.2 0.2] + /T (The Dvipdfmx User's Manual) + /Subj (The Dvipdmfx User's Manual) + /Contents (XeLaTeX source file of this manual.) + >> +}}% +\begin{lstlisting} +\special{pdf:fstream @SourceFile (\jobname.tex)}% +\special{pdf:ann width 10bp height 20bp + << + /Type /Annot + /Subtype /FileAttachment + /FS << + /Type /Filespec + /F (\jobname.tex) + /EF << /F @SourceFile >> + >> + /Name /PushPin + /C [0.8 0.2 0.2] + /T (The Dvipdfmx User's Manual) + /Subj (The Dvipdfmx User's Manual) + /Contents (XeLaTeX source file of the manual.) + >> +} +\end{lstlisting} +A push-pin image must be shown on the margin if viewer supports this kind of +annotation. +PDF viewer applications are required to provide predefined icon appearances +at least for the following standard icons: \code{Graph}, \code{PushPin}, +\code{PaperClip}, and \code{Tag}. + +\subsubsection{Special Color Space} + +This section shows various examples of using \keyword{Special color spaces}. +Examples in this section have a common structure. They consist of essentially +three parts. The first part is for defining color space itself. +PDF object creation commands like \code{pdf:obj} and \code{pdf:stream} are used +for this purpose. Next is for registering color space resources in the page's +\keyword{Resource Dictionary}. It can be done via \code{pdf:put} command as, +\begin{lstlisting} +\special{pdf:put @resource << + /Category << ...key-value pairs... >> +>>} +\end{lstlisting} +where \code{@resource} is a special keyword representing current page's +Resource Dictionary and \code{Category} (to be replaced by actual category +name) is a category name such as \code{ColorSpace}. +Finally, graphics objects are placed, with or with a combination of +text and, PDF drawing operators inserted by the \code{pdf:code} or the +\code{pdf:contents} special. + +%%EXAMPLE: Separation Color Space +The first example is the \keyword{Separation} color space: +\special{pdf:stream @TintTransform1 + ({0 exch dup 0.62 mul exch 0}) + << /FunctionType 4 + /Domain [0.0 1.0] + /Range [0.0 1.0 0.0 1.0 0.0 1.0 0.0 1.0] + >> +}% +\special{pdf:stream @TintTransform2 + ({dup 0.78 mul exch dup 0.05 mul exch 0.71 mul 0}) + << /FunctionType 4 + /Domain [0.0 1.0] + /Range [0.0 1.0 0.0 1.0 0.0 1.0 0.0 1.0] + >> +}% +\special{pdf:obj @Orange [ + /Separation /Orange /DeviceCMYK @TintTransform1 + ] +}% +\special{pdf:obj @Green [ + /Separation /Green /DeviceCMYK @TintTransform2 + ] +}% +\begin{center} +\mbox{% +\special{pdf:put @resources << + /ColorSpace << /CS01 @Orange /CS02 @Green >> + >> +}% +\fontsize{40pt}{40pt}\selectfont +\special{pdf:code q /CS01 cs 1.0 scn}Orange\special{pdf:code Q} and +\special{pdf:code q /CS02 cs 1.0 scn}Green\special{pdf:code Q} +} +\end{center} +\begin{lstlisting} +\special{pdf:stream @TintTransform1 + ({0 exch dup 0.62 mul exch 0}) + << /FunctionType 4 + /Domain [ 0.0 1.0 ] + /Range [ 0.0 1.0 0.0 1.0 0.0 1.0 0.0 1.0 ] + >> +} +\special{pdf:stream @TintTransform2 + ({dup 0.78 mul exch dup 0.05 mul exch 0.71 mul 0}) + << /FunctionType 4 + /Domain [ 0.0 1.0 ] + /Range [ 0.0 1.0 0.0 1.0 0.0 1.0 0.0 1.0 ] + >> +} +\special{pdf:obj @Orange [ + /Separation /Orange /DeviceCMYK @TintTransform1 + ] +} +\special{pdf:obj @Green [ + /Separation /Green /DeviceCMYK @TintTransform2 + ] +} +\mbox{% + \special{pdf:put @resources << + /ColorSpace << /CS01 @Orange /CS02 @Green >> + >> + }% + \special{pdf:code q /CS01 cs 1.0 scn} + Orange + \special{pdf:code Q} + and + \special{pdf:code q /CS02 cs 1.0 scn} + Green + \special{pdf:code Q} +} +\end{lstlisting} + +\code{TintTransform}'s defined here are functions for transforming +\keyword{tint} values into approximate colors in the +\keyword{alternate color space} +(\code{DeviceCMYK} in this example). PostScript calculator functions are used +for converting a single component value representing ``Orange'' or ``Green'' +into four component CMYK values to approximate those colors. The ``Orange'' +color $v$ is approximated as $(0, 0.62v, v, 0)$ in CMYK color space for +alternate display here. + +The \code{cs} operator for selecting color space and the \code{scn} operator +for color values are used in the \code{pdf:code} special. +Be sure that the \code{pdf:put} command, which puts color space resources +into the current page's Resource Dictionary, goes into the same page as +subsequent drawing commands. + +\dvipdfmx\ currently does not have an easy interface for using various color +space families such as CIE-Based color spaces (e.g., calibrated colors and +color space with an ICC profile) and Special color spaces (e.g., indexed, +separation, shading and patterns). + +%%EXAMPLE: Shading +Another example is a \keyword{shading pattern}: +\begin{lstlisting} +\special{pdf:put @resources << + /Shading << + /SH01 << + /ShadingType 2 + /ColorSpace @Orange + /Coords [0 0 320 20] + /Extend [true true] + /Function << /FunctionType 2 /Domain [0 1] /N 1.0 >> + >> + >> +>>} +\special{pdf:content 0 0 320 20 re W n /SH01 sh} +\end{lstlisting} +where the ``Orange'' separation color space defined before is used again. +This example shows an axial shading (\code{ShadingType} 2) pattern. +\begin{center} +\makebox[320pt][l]{ + \special{pdf:put @resources << + /Shading << + /SH01 << + /ShadingType 2 + /ColorSpace @Orange + /Coords [0 0 320 20] + /Extend [true true] + /Function << /FunctionType 2 /Domain [0 1] /N 1.0 >> + >> + >> + >>} + \raisebox{-20pt}[0pt][20pt]{\special{pdf:content 0 0 320 20 re W n /SH01 sh}} +} +\end{center} + +The shading pattern requires coordinate values to be mapped into color values. +Type 2 (Exponential Interpolation) \keyword{Function} is used for defining how +this mapping should occur here. +The above example, with the exponent $N=1$, is just a simple linear-gradient. + +The final examples is a \keyword{tiling pattern}. +\begin{lstlisting} +\special{pdf:stream @MyPattern + (0.16 0 0 0.16 0 0 cm 4 w + 50 0 m 50 28 28 50 0 50 c S 100 50 + m 72 50 50 28 50 0 c S + 50 100 m 50 72 72 50 100 50 c S + 0 50 m 28 50 50 72 50 100 c S + 100 50 m 100 78 78 100 50 100 c 22 100 0 78 0 50 c + 0 22 22 0 50 0 c 78 0 100 22 100 50 c S + 0 0 m 20 10 25 5 25 0 c f 0 0 m 10 20 5 25 0 25 c f + 100 0 m 80 10 75 5 75 0 c f + 100 0 m 90 20 95 25 100 25 c f + 100 100 m 80 90 75 95 75 100 c f + 100 100 m 90 80 95 75 100 75 c f + 0 100 m 20 90 25 95 25 100 c f + 0 100 m 10 80 5 75 0 75 c f + 50 50 m 70 60 75 55 75 50 c 75 45 70 40 50 50 c f + 50 50 m 60 70 55 75 50 75 c 45 75 40 70 50 50 c f + 50 50 m 30 60 25 55 25 50 c + 25 45 30 40 50 50 c f + 50 50 m 60 30 55 25 50 25 c 45 25 40 30 50 50 c f) + << + /BBox [0 0 16 16] + /PaintType 2 + /PatternType 1 + /Resources << + /ProcSet [/PDF] + >> + /TilingType 3 + /Type /Pattern + /XStep 16 + /YStep 16 + >> +} +\end{lstlisting} + +The above example defines a tiling pattern. The content stream containing +painting operators, \code{m} for ``move-to'', \code{c} for ``curve-to'', +\code{f} for ``fill'', and \code{S} for ``stroke'', defines the appearance of +the \keyword{pattern cell} for this tiling pattern. With the following code, +\begin{lstlisting} +\special{pdf:put @resources + << + /ColorSpace << /CS01 [/Pattern /DeviceRGB] >> + /Pattern << /PT01 @MyPattern >> + >> +} +\special{pdf:content + q 0.8 0.3 0.3 RG /CS01 cs 0.8 0.3 0.3 /PT01 scn + 0 0 320 100 re f +} +\end{lstlisting} +a box filled with the tiling pattern defined above is drawn. +\special{pdf:stream @MyPattern + (0.16 0 0 0.16 0 0 cm 4 w + 50 0 m 50 28 28 50 0 50 c S 100 50 + m 72 50 50 28 50 0 c S + 50 100 m 50 72 72 50 100 50 c S + 0 50 m 28 50 50 72 50 100 c S + 100 50 m 100 78 78 100 50 100 c 22 100 0 78 0 50 c + 0 22 22 0 50 0 c 78 0 100 22 100 50 c S + 0 0 m 20 10 25 5 25 0 c f 0 0 m 10 20 5 25 0 25 c f + 100 0 m 80 10 75 5 75 0 c f + 100 0 m 90 20 95 25 100 25 c f + 100 100 m 80 90 75 95 75 100 c f + 100 100 m 90 80 95 75 100 75 c f + 0 100 m 20 90 25 95 25 100 c f + 0 100 m 10 80 5 75 0 75 c f + 50 50 m 70 60 75 55 75 50 c 75 45 70 40 50 50 c f + 50 50 m 60 70 55 75 50 75 c 45 75 40 70 50 50 c f + 50 50 m 30 60 25 55 25 50 c + 25 45 30 40 50 50 c f + 50 50 m 60 30 55 25 50 25 c 45 25 40 30 50 50 c f) + << + /BBox [0 0 16 16] + /PaintType 2 + /PatternType 1 + /Resources << + /ProcSet [/PDF] + >> + /TilingType 3 + /Type /Pattern + /XStep 16 + /YStep 16 + >> +}% +\begin{center} +\raisebox{-100bp}[0bp][100bp]{\makebox[320bp][l]{\special{pdf:put @resources + << + /ColorSpace << /CS01 [/Pattern /DeviceRGB] >> + /Pattern << /PT01 @MyPattern >> + >> +}% +\special{pdf:content + 0.8 0.3 0.3 RG /CS01 cs 0.8 0.3 0.3 /PT01 scn + 0 0 320 100 re f +}}}% +\end{center} + +\subsubsection{Transparency} + +%%EXAMPLE: Transparency +\XeTeX's transparency feature is currently lost in \xdvipdfmx, but the same +effect can be achieved by setting graphics state parameters with +\code{ExtGState} resources and \code{gs} operator. Here is a simple +transparency example: +\special{pdf:obj @gs01 << + /Type /ExtGState /CA 0.5 /ca 0.5 /AIS false +>>}% +\begin{figure}[ht] +\centering +\mbox{% + %\raisebox{120pt}{\parbox[t]{0.8\textwidth}{% + % \addfontfeature{Color=444444}\lipsum[1]}% + %}% + %\hspace{-0.8\textwidth}% + \special{pdf:put @resources << + /ExtGState << /GS01 @gs01 >> + >>}% + \fontsize{220pt}{220pt}\selectfont + \special{pdf:code q /GS01 gs 1.0 0.8 0.2 rg}% + α% + \special{pdf:code 0.4 0.8 0.4 rg}% + \hspace{-0.3em}% + β% + \hspace{-0.3em}\raisebox{0.5ex}{% + \special{pdf:code 0.4 0.4 0.8 rg}% + π% + }% + \special{pdf:code 0.6 0.2 0.8 rg}% + \hspace{-0.2em}% + γ% + \special{pdf:code Q}% +} +\end{figure} +\begin{lstlisting} +\special{pdf:obj @GS01 << + /Type /ExtGState /CA 0.5 /ca 0.5 /AIS false +>>}% +\mbox{% + \special{pdf:put @resources << + /ExtGState << /GS01 @GS01 >> + >>}% + \special{pdf:code q /GS01 gs 1.0 0.8 0.2 rg}% + α% + \special{pdf:code 0.4 0.8 0.4 rg}% + \hspace{-0.3em}% + β% + \hspace{-0.3em}\raisebox{0.5ex}{% + \special{pdf:code 0.4 0.4 0.8 rg}% + π% + }% + \special{pdf:code 1.0 0.2 0.4 rg}% + \hspace{-0.2em}% + γ% + \special{pdf:code Q}% +} +\end{lstlisting} +where values for \code{CA} and \code{ca} represent opacity of stroke and +fill color respectively. Again, \code{pdf:put} command must go into the same +page as subsequent graphics and text drawing operators. + +\section{Dvipdfmx Extensions} + +\newfeature{Addition in \TeX\ Live 2016} + +\noindent{}A new special \code{dvipdfmx:config} was introduced in +\TeX Live 2016 which makes it possible to invoke a command line option. +Several single letter command line options except \option{D} are supported. +For example, +\begin{lstlisting} +dvipdfmx:config C 0x10 +\end{lstlisting} +sets the \dvipdfmx's compatibility flags. See, the section +``\hyperref[SEC:compatibility]{Incompatible Changes}'' for an explanation of +compatibility flags. + +\section{PS Specials} + +PS (PostScript) specials can be used to insert a raw PostScript code for drawing +graphics objects and transforming subsequent text and graphics. +Please note that support for PostScript operators in \dvipdfmx\ is very +limited. It is just enough for interpreting PostScript figures output by \MP. +Only a basic set of operators for arithmetic and math, stack operation and manipulation, +graphics state, path construction and painting, glyph and font, are supported. +See, Table~\ref{TABLE:PS} for the list of +recognized PostScript operators. + +\begin{table} + \centering + \begin{tabular}{p{3cm}>{\raggedright\arraybackslash}p{8cm}}\hline + Classification & Operators \\ \hline\hline + Arithmetic \& Math & \code{add} \code{sub} \code{mul} \code{div} \code{neg} \code{truncate}\\ + Stack Operation & \code{clear} \code{pop} \code{exch}\\ + Graphis State & \code{gsave} \code{grestore} + \code{setlinewidth} \code{setdash} \code{setlinecap} \code{setlinejoin} \code{setmiterlimit} \code{setgray} \code{setrgbcolor} \code{setcmykcolor} \\ + Coordinate System & \code{concat} \code{scale} \code{translate} \code{rotate} \code{idtransform} \code{dtransform}\\ + Path Construction & \code{currentpoint} \code{newpath} \code{closepath} \code{moveto} \code{rmoveto} \code{lineto} \code{rlineto} \code{curveto} \code{rcurveto} \code{arc} \code{arcn} \code{clip} \code{eoclip} \\ + Painting & \code{stroke} \code{fill} \\ + Glyph \& Font & \code{show} \code{findfont} \code{scalefont} \code{setfont} \code{currentfont} \code{stringwidth}\\ + \hline + \end{tabular} + \caption{List of PostScript operators recognized by \dvipdfmx.}% + \label{TABLE:PS} +\end{table} + +It might be enough for the purpose of basic graphics drawings but as there are +no support for conditionals and controls it is not enough for complicated tasks, +especially, the PSTricks package is not supported. + +In \dvipdfmx, text handling is extended to support CJK text. +The following code draws Japanese text like shown in Figure~\ref{FIG:verttext}: +\begin{lstlisting} +\special{pdf:mapline uprml UniJIS-UTF8-H yumindb.ttf} +\special{ps: uprml findfont 16 scalefont setfont + currentpoint moveto + (...some Japanese text goes here...) show +} +\end{lstlisting} + +\chapter{Fonts and Encodings} + +\section{Fonts and Encodings Support} + +In \dvipdfmx, all font formats supported by \dvipdfm\ are also supported with +many improvements: The CFF conversion for PostScript Type1 fonts\footnote{PostScript +Type1 font support is restricted to the binary format as in \dvipdfm.} is +implemented which greatly reduces the output file size. Embedded TrueType fonts are +now subsetted. The OpenType font format is also supported.\footnote{Its +implementation is based on the OpenType specification version 1.4. +Newly added features such as color fonts and variable fonts are not supported yet.} + +There are various enhancements made for supporting Unicode and legacy multi-byte +character encodings for East Asian languages. + + +\section{Font Mappings} + +The Syntax of font-mapping (fontmap) files is basically the same as in \dvipdfm. +There are few extensions available in \dvipdfmx. In addition to the 8-bit \code{enc} +file and keywords \code{builtin} and \code{none}, \dvipdfmx\ accepts a PostScript CMap +Resource name and the keyword \code{unicode} in the encoding field. + +When the keyword \code{unicode} is specified in the encoding field of fontmap +files, it is assumed that Unicode values are used in the input DVI file. + +\begin{lstlisting} + urml unicode SourceHanSerifJP-Light.otf +\end{lstlisting} + +Although the DVI format allows 3-byte and 4-byte character codes to be used, +\dvipdfmx\ only supports up to 2-byte range since there is no TFM format supporting +3-byte or 4-byte codes. + +For PostScript Type1 fonts which do not support Unicode natively, an auxiliary file, +the Adobe Glyph List, is required to make it possible to use fonts with Unicode access. + +As a general framework for supporting legacy multi-byte encodings, \dvipdfmx\ employs +PostScript CMap Resources for handling input strings encoded in various +character encodings. A CMap name can be specified in the encoding field just like +the encoding name for 8-bit encodings. For example, to specify the CMap ``UniJIS-UCS2-H'', + +\begin{lstlisting} + urml UniJIS-UCS2-H HiraMinPro-W3.otf +\end{lstlisting} + +For information on the Adobe Glyph List and PostScript CMap Resources, see, +the section \ref{SEC:auxfiles}, ``Auxiliary Files''. + +\subsection{Extended Syntax and Options} + +Few options are available in \dvipdfmx\ in addition to the original dvipdfm's +one. Please note that all features which makes \dvipdfmx\ to use +non-embedded fonts are deprecated, as by doing so it makes \dvipdfmx\ to create +PDF files which can be non-compliant to the ISO standards. + +\subsubsection{SFD Specification} + +For bundling up a font split into multiple subfonts via SFD back into +a single font, dvipdfmx supports extended syntax of the form +\begin{lstlisting} +tfm_name@SFD@ encoding filename options +\end{lstlisting} +A typical example looks like: +\begin{lstlisting} +gbsn@EUC@ GB-EUC-H gbsn00lp +\end{lstlisting} +where TFMs \code{gbsn00}, \code{gbsn01}, \code{gbsn02}... are mapped into a +single font named \code{gbsn00lp} via the rule described in the SFD file +\code{EUC}. + +\subsubsection{TrueType Collection Index} + +TrueType Collection index number can be specified with \code{:n:} +in front of the TrueType font name: +\begin{lstlisting} +min10 H :1:mincho +\end{lstlisting} +In this example, the option \code{:1:} tells \dvipdfmx\ to select first +TrueType font from the TTC font \code{mincho.ttc}. Alternatively, the +\option{-i} option can be used in the option field to specify TTC index: +\begin{lstlisting} +min10 H mincho -i 1 +\end{lstlisting} + +\subsubsection{Non-embedding Switch} + +\deprecated{Use of this option is deprecated.} +\noindent{}The character \option{!} in front of the font name can be used to +indicate that the font shall not be embedded. This feature greatly reduces the +size of the final PDF output, but the PDF file may not be viewed exactly the same +in other systems on which appropriate fonts are not installed. +\bigskip + +\noindent{}NOTE: \dvipdfmx\ always converts input encodings to CIDs and then +uses Identity CMaps\footnote{Predefined CMaps \code{Identity-H} and +\code{Identity-V} for the identity mapping.} +in the output PDF. However, \lnum{ISO~32000-1:2008} describes as +\begin{quoting} +The Identity-H and Identity-V CMaps shall not be used with a non-embedded font. +Only standardized character sets may be used. +\end{quoting} +which had never appeared in Adobe's PDF References. This makes all PDF files +generated by \dvipdfmx\ with non-embedded CID-keyed fonts non-compliant to +the ISO standards. + +\subsubsection{`Standard' CJK Fonts} + +\deprecated{This feature is deprecated.} +\noindent{}Use of this feature shall be avoided for new documents. It is +described here since it might still be useful for some situations. + +\dvipdfmx\ recognizes several `Standard' CJK fonts although there are no such +notion in PDF. In older days where there were not so many freely available CJK +fonts, it was sometimes useful to create PDF files without embedding fonts and +let PDF viewers or printers to use substitute fonts (tend to be higher quality) +installed in their systems. \dvipdfmx\ `knows' several fonts which might be +available in PostScript printers and PDF applications such as Acrobat Reader, +and uses them without actually having it. +See, Table~\ref{TABLE:StdCJKFont}, for the list of available `Standard' CJK +fonts. + +\begin{table} + \centering + \begin{tabular}{lll}\hline + Character Collection & Font Family & Description \\ \hline\hline + Adobe-Japan1 & Ryumin-Light & PS printers \\ + & GothicBBB-Medium & \\ + Adobe-CNS1 & MHei-Medium-Acro & Acrobat Reader 4 \\ + & MSung-Light-Acro & \\ + Adobe-GB1 & STSong-Light-Acro & \\ + & STHeiti-Regular-Acro & \\ + Adobe-Japan1 & HeiseiMin-W3-Acro & \\ + & HeiseiKakuGO-W5-Acro & \\ + Adobe-Korea1 & HYGoThic-Medium-Acro & \\ + & HYSMyeongJo-Medium-Acro & \\ + Adobe-CNS1 & MSungStd-Light-Acro & Acrobat Reader 5 \\ + Adobe-GB1 & STSongStd-Light-Acro & \\ + Adobe-Korea1 & HYSMyeongJoStd-Medium-Acro \\ + Adobe-CNS1 & AdobeMingStd-Light-Acro & Adobe Reader 6 \\ + Adobe-GB1 & AdobeSongStd-Light-Acro & \\ + Adobe-Japan1 & KozMinPro-Regular-Acro & \\ + & KozGoPro-Medium-Acro & \\ + Adobe-Korea1 & AdobeMyungjoStd-Medium-Acro & \\ + Adobe-CNS1 & AdobeHeitiStd-Regular & Adobe Reader 7 \\ + Adobe-Japan1 & KozMinProVI-Regular & Adobe Reader 8\\ + \hline + \end{tabular} + \caption{List of available `Standard' CJK font. Most of them are + available as a part of Adobe Asian Font Packs for each versions of + Adobe or Acrobat Reader.}\label{TABLE:StdCJKFont} +\end{table} + +Only fixed-pitch glyphs (i.e., quarter, third, half, and full widths) are +supported for those fonts. + +\subsubsection{Stylistic Variants} + +\deprecated{Use of this option is deprecated.} +\noindent{}Keywords \code{,Bold}, \code{,Italic}, and \code{,BoldItalic} can be +used to create synthetic bold, italic, and bolditalic style variants from other +font using PDF viewer's (or OS's) function. +\begin{lstlisting} +jbtmo@UKS@ UniKSCms-UCS2-H :0:!batang,Italic +\end{lstlisting} + +Availability of this feature highly depends on the implementation of PDF +viewers. This feature is usually not supported for embedded fonts. +Notice that this option automatically disables font embedding thus +use of it is deprecated. + +\subsection{Specifying Unicode Plane} + +As there are no existing TFM formats supporting 3-byte or 4-byte character +encodings, the only way to use Unicode characters other than the BMP is to +map the code range 0-65535 to different planes via (e.g., to plane 1) +the \option{-p 1} fontmap option. This option is +available only when \code{unicode} is specified in the encoding field. + +\subsection{OpenType Layout Feature} + +The OpenType Layout Feature fontmap options mentioned below are only meaningful +when \code{unicode} is specified in the encoding field. + +With the \option{-w} option, writing mode can be specified. +\option{-w 1} denotes the font is for vertical writing. It automatically +enables an OpenType Layout Feature related to vertical writing, namely, +\code{vert} or \code{vrt2}, to choose proper glyphs for vertical text. + +\newfeature{Addition in \TeX\ Live 2017.} + +The \option{-l} (lower case el) option can be used to enable various +OpenType Layout GSUB Features. For example, \option{-l jp04} enables +\code{jp04} feature to select \lnum{JIS2004} forms for Kanjis. +Features can be specified as a ``:'' separated list of OpenType Layout +Feature tags like \option{-l vkna:jp04}. Script and language may be +additionally specified as +\option{-l kana.JAN.ruby}. + +An example can be +\begin{lstlisting} +uprml-v unicode SourceHanSerifJP-Light.otf -w 1 -l jp90 +\end{lstlisting} +which declares that font should be treated as for vertical writing and +use \lnum{JIS1990} form for Kanjis. + +\begin{figure} +\centering +\jpzerofourexamples\hspace{30pt}\jpninezeroexamples% +\caption{\lnum{JIS2004} vs. \lnum{JIS1990} form.}\label{FIG:jp90} +\end{figure} + +This feature is limited to the single substitution, there are no way to select +a glyph from multiple candidates, such as in the \code{aalt} feature, and specifying +general many-to-many glyph substitutions does not take effect. + +\section{Other Improvements} + +This section briefly describes other improvements made for \dvipdfmx. +There is an extension to glyph name handling in the \code{enc} file for +seamless support of both PostScript Type1 and TrueType fonts. +PostScript Type1 font support is enhanced although this format might be +considered obsolete. + +\subsection{Extended Glyph Name Syntax} + +\dvipdfmx\ accepts the following syntax for glyph names in the \code{enc} file: +\code{uni0130}, \code{zero.onum} and \code{T\_h.liga}. +Each represents a glyph accessed with Unicode value \code{U+0130}, +oldstyle number for zero and ``Th'' ligature accessed via the OpenType +Layout GSUB Feature \code{onum} and \code{liga}, respectively. +Note that \dvipdfmx\ does not understand glyph names which directly +use a glyph index such as \code{index0102} or \code{gid2104}. + +When \dvipdfmx\ encounters a glyph name, e.g., \code{T\_h.liga}, it first looks +for OpenType \code{post} table if such glyph name exists; if it exists, then +\dvipdfmx\ simply uses \code{post} table and maps the glyph name to the glyph index; +if not, \dvipdfmx\ tries to convert \code{T\_h} to a Unicode sequence (U+0054 +U+0068 in this example) via the AGL mapping; then, OpenType \code{cmap} table is used +to further convert the resulting Unicode sequence to the sequence of glyph indices; +finally, the OpenType Layout Feature \code{liga} is applied to get the desired glyph. + +A glyph name of a form \code{a.swsh2} can be specified to denote the 2nd swash +variant form of the letter `a'. + +\subsection{CFF Conversion} + +\dvipdfmx\ supports on-the-fly PostScript Type1 to CFF (Type1C) conversion +which greatly reduces size of the resulting PDF file when using Type1 fonts. +Conversion is essentially `lossless' and there should not be any quality loss. +However, due to differences in the ability of rasterizers, there might be +noticeable differences on rendering result. + +When an older Type1 font is used, \dvipdfmx\ may give the following warning +message: +\begin{lstlisting} +Obsolete four arguments of "endchar" will be used for Type1 +"seac" operator. +\end{lstlisting} +It happens when there is an accented character made as a composite glyph using +the ``seac'' operator. +This warning is issued as conversion can't be done without relying on the +\emph{deprecated} usage of the \code{endchar} operator. However, as mentioned in +``Appendix C Compatibility and Deprecated Operators'' of Adobe Technical +Note \#5177, +``\href{http://wwwimages.adobe.com/content/dam/Adobe/en/devnet/font/pdfs/5177.Type2.pdf}{Type 2 Charstring Format}'', +PDF applications should support this operator and hence this warning +message can be ignored. + +Use of Type1 font should be avoided as much as possible. +Please consider using OpenType version instead. + +\section{Font Licensing} + +In OpenType font format, information regarding how a font should be treated +when creating a document can be recorded.% +\footnote{See, +``\href{http://www.microsoft.com/typography/otspec/os2.htm}{OpenType Specification: +OS/2 -- OS/2 and Windows Metrics Table}''}. +\dvipdfmx\ uses this information to decide whether font embedding is permitted. + +This font licensing information is indicated by the flag called \code{fsType} +recorded in OpenType font files; each bits representing different restrictions on +font embedding. If multiple flag bits are set in \code{fsType}, the least restrictive +license granted takes precedence in \dvipdfmx. +The \code{fsType} flag bit recognized by \dvipdfmx\ is as follows: +\begin{itemize} + \item Installable embedding + \item Editable embedding + \item Embedding for Preview \& Print only +\end{itemize} +\dvipdfmx\ issues the following warning message for fonts with `Preview \& Print only' +setting: +\begin{verbatim} + This document contains 'Preview & Print' only licensed font +\end{verbatim} + +For a font with this type of licensing, font embedding is allowed solely for the +purpose of (on-screen) viewing and/or printing; further editing of the document +or extracting embedded font data for other purposes are not allowed. +One way to ensure this condition is to protect your document with a non-empty +password. + +All other flags are treated as more restrictive license than any of +the above flags and treated as ``No embedding allowed''; e.g., if both +of the editable-embedding flag and unrecognized license flag is set, +the font is treated as editable-embedding allowed, however, if only +unrecognized flags are set, the font is not embedded. + +Font Embedding flags are preserved in the embedded font if they are embedded +as a TrueType font or a CIDFontType2 CID-keyed font. +For all fonts embedded as a PostScript font (Type1C and CIDFontType0 +CID-keyed font), they are not preserved. +Only \code{Copyright} and \code{Notice} in the \keyword{FontInfo} dictionary +are preserved in this case. + +Some font vendors put different embedding restrictions for different +condition; e.g., font embedding might not be permitted for the commercial +use unless you acquire the ``commercial license'' separately. +Please read EULA carefully before making decision on the font usage. + +See, for example, +\href{http://www.adobe.com/products/type/font-licensing/font-embedding-permissions.html}{Adobe's site on font embedding permissions} +for the font in the Adobe Type Library. +Microsoft also has a +\href{http://www.microsoft.com/typography/RedistributionFAQ.mspx}{FAQ page on Font Redistribution}. + +For Japanese font in general, embedding permission tend to be somewhat +restrictive. Japanese users should read the statement regarding font +embedding from Japan Typography Association (in Japanese):\medskip + +\url{http://www.typography.or.jp/act/morals/moral4.html} +\medskip + +\dvipdfmx\ does not support full embedding. Only subset embedding is supported. + +\chapter{Encryption} + +\section{Encryption Support}\label{SEC:encryption} + +\dvipdfmx\ offers basic PDF password security support including the 256-bit AES encryption. +Only the ``Standard'' security handler is supported and the Public-key security handler is not +supported. + +Encryption is enabled by \option{-S} command line option. +\begin{lstlisting} + dvipdfmx -S -K 128 -P 0x14 +\end{lstlisting} +where \code{-K} and \code{-P} options are used to specify encryption key length and +permission flags respectively, and are briefly explained in Table~\ref{TABLE:enc-options}. + +\begin{table} + \centering + \begin{tabular}{lp{8cm}}\hline + Option & Description \\ \hline\hline + \code{-S} & Enable PDF encryption. \\ + \code{-K} \textit{number} & Set encryption key length. The default value + is 40.\\ + \code{-P} \textit{number} & Set permission flags for PDF encryption. + The \textit{number} is a 32-bit unsigned integer representing permission + flags. + See, Table~\ref{TABLE:flags}. The default value is \code{0x003C}.\\ + \hline + \end{tabular} + \caption{Command line options for encryption.}% + \label{TABLE:enc-options} +\end{table} + +When \dvipdfmx\ is invoked with encryption via the \code{-S} option, +passwords will be asked. +However, in some circumstances, it might be desirable to avoid being prompted for +passwords. In that case, use the \code{pdf:encrypt} special to supply passwords in +the \TeX\ file, as, +\begin{lstlisting} + \special{pdf:encrypt userpw (foo) ownerpw (bar) length 128 perm 20} +\end{lstlisting} +Here, user and owner passwords are supplied as PDF string objects (\code{foo} and \code{bar} +in the example above) which can be empty. + +Up to two passwords can be specified for a document -- an owner password and a user password. +If a user attempts to open an encrypted document with user password being set, PDF application +should prompt for a password. Users are allowed to access the contents of the document only when +either password is correctly supplied. +Depending on which password (user or owner) was supplied, additional operations +allowed for the opened document is determined; full access for users who opened +with the correct owner password or additional operations +controlled by permission flags for users who opened with the correct user password. + +Although PDF specification allows various character encodings other than \code{US-ASCII} +for entering password, \dvipdfmx\ is unable to handle it properly. +Thus it must be assumed that \code{US-ASCII} is used for password strings. + +Access permission flags can be specified via \option{-P} command-line option. +Each bits of (32-bit unsigned) integer number given to this option represents user +access permissions; e.g., bit position 3 for allowing ``print'', 4 for +``modify'', 5 for ``copy or extract'', and so on. See, Table~\ref{TABLE:flags}. +For example, \code{-P 0x34} allows printing, copying and extraction of text, and adding and +modifying text annotation and filling in interactive form fields (but disallows +modification of the contents of the document). +\begin{table} + \centering + \begin{tabular}{lp{8.3cm}}\hline + Bit Position & Meaning \\ \hline\hline + 3 & (Revision 2) Print the document. \\ + & (Revision 3 or greater) Print the document. Print quality depending on bit 12.\\ + 4 & Modify the contents of the document by operations other than those controlled + by bits 6, 9, and 11. \\ + 5 & Copy or extract text and graphics from the document. \\ + 6 & Add or modify text annotations, fill in interactive form fields. + Creation and modification of interactive form field is also + allowed if bit 4 is set.\\ + 9 & (Revision 3 or greater) Fill in existing interactive form fields + (including signature fields), even if bit 6 is clear.\\ + 10 & \em{Deprecated in PDF 2.0} \\ + & (Revision 3 or greater) Extract text and graphics (in support of accessibility to + users with disabilities or for other purposes).\\ + 11 & (Revision 3 or greater) Assemble the document + (insert, rotate, or delete pages and create document outline items or thumbnail + images), even if bit 4 is clear.\\ + 12 & (Revision 3 or greater) High-quality printing. + When this bit is clear (and bit 3 is set), printing shall be limited to a low-level, + possibly of degraded quality.\\ + \hline + \end{tabular} + \caption{Flag bits and their short explanation. + Revision 2 is used when encryption key length is 40 bits or when PDF output + version is less than 1.5. Otherwise, Revision 3 or greater is used.}\label{TABLE:flags} +\end{table} + +The \option{-K} option can be used to specify the encryption key length. +The key length must be multiple of 8 in the range 40 to 128, or 256 (for PDF +version 1.7 plus Adobe Extension or PDF version 2.0). Please note +that when key length 256 is specified for PDF version 1.7 output, it requires +Adobe's Extension to \lnum{PDF-1.7} and hence PDF applications may not support it. +PDF version 1.4 is required for key length more than 40 bits. AES encryption algorithm requires +PDF version 1.6. + +To show some examples:\\ +128-bit AES encryption with print-only (high-quality) setting, +\begin{lstlisting} + dvipdfmx -V 5 -S -K 128 -P 0x804 input.dvi +\end{lstlisting} +256-bit AES encryption with print (low-quality), adding and modifying text annotations +allowed, +\begin{lstlisting} + dvipdfmx -V 2.0 -S -K 256 -P 0x24 input.dvi +\end{lstlisting} + +The default value for \option{-K} is 40 and for \option{-P} is \code{0x003C0} +(all bits from bit-position 3 to 6 set). + +\chapter{Compatibility} + +\section{Incompatible Changes}\label{SEC:compatibility} + +There are various minor incompatible changes to \dvipdfm. + +The \option{-C} command line option may be used for compatibility to +\dvipdfm\ or older versions of \dvipdfmx. The \option{-C} option takes flags +meaning +\begin{itemize} + \item bit position 2: Use semi-transparent filling for tpic shading + command, instead of opaque gray color. (requires PDF 1.4) + \item bit position 3: Treat all CID-keyed font as fixed-pitch font. This is + only for compatibility. + \item bit position 4: Do not replace duplicate fontmap entries. + \dvipdfm\ behavior. + \item bit position 5: Do not optimize PDF destinations. Use this if you + want to refer from other files to destinations in the current file. + \item bit position 6: Do not use predictor filter for Flate compression. + \item bit position 7: Do not use object stream. +\end{itemize} + +The remap option \option{-r} in fontmaps is no longer supported and is +silently ignored. The command line option \option{-e} to disable partial +(subset) font embedding is not supported. + +\section{Important Changes} + +Here is a list of important changes since the \TeX\ Live 2016 release: +\begin{itemize} +\item Changes to make PDF/A creation easier: Always write CIDSet and CharSet +for embedded fonts. Do not compress XMP metadata. +\item Merge from libdpx for p\TeX-ng by Clerk Ma. +\item Addition of \code{STHeiti-Regular-Acro} for CJK `Standard' fonts. +\item Command line option \option{-p} takes precedence over \code{papersize} +and \code{pagesize} specials. +\item Fixed serious bugs in supporting `\code{unicode}' encoding: +OpenType Layout Feature \code{vert} and \code{vrt2} was not enabled. +Support for format 2 CFF charsets was broken. +\item Added simplified version of OpenType Layout support: The `\option{-l}' +option in fontmaps. +\end{itemize} +The full \code{ChangeLog} entries can be viewed via the web interface of the +\TeX\ Live SVN repository: +\medskip + +\url{http://www.tug.org/svn/texlive/trunk/Build/source/texk/dvipdfm-x} +\medskip + +There was an undocumented feature for supporting OpenType Layout but it was +dropped. Simplified support for the OpenType Layout was introduced instead. + + +%\renewcommand{\refname}{Further Reading} +\renewcommand{\bibname}{Further Reading} +\begin{thebibliography}{99} +\bibitem{DVIPDFM} ``\href{http://mirrors.ctan.org/dviware/dvipdfm/dvipdfm.pdf}% +{Dvipdfm User's Manual}'' written by Mark~A.~Wicks. +\bibitem{ADOBE} Adobe's PDF References and a free copy of +\lnum{ISO 32000-1:2008} standard are available from +``\href{http://www.adobe.com/devnet/pdf.html}{PDF Technology Center}'' +on \href{http://www.adobe.com/devnet.html}{Adobe Developer Connection}. +\bibitem{MICROSOFT} The OpenType Specification is available from Microsoft's +site: +``\href{http://www.microsoft.com/en-us/Typography/OpenTypeSpecification.aspx}% +{OpenType Specification}''. +\bibitem{PNGSPEC} ``\href{https://www.w3.org/TR/2003/REC-PNG-20031110/}% +{Portable Network Graphics (PNG) Specification (Second Edition)}''. +\bibitem{CHOF} An article regarding DVI specials: Jin-Hwan Cho, +``\href{http://www.tug.org/TUGboat/tb30-1/tb94cho.pdf}{DVI specials for PDF generation}'', +TUGboat, 30(1):6-11, 2009. +\end{thebibliography} + +\appendix + +\chapter{GNU Free Documentation License}\label{SEC:FDL} + +This document is distributed under the term of the GNU Free Documentation +License. See, the attached file for copying conditions.% +\marginnote{% +\special{pdf:fstream @fileobj (fdl-1.3.txt)}% +\special{pdf:ann width 10bp height 20bp + << /Type /Annot + /Subtype /FileAttachment + /FS << + /Type /Filespec + /F (fdl-1.3.txt) + /EF << /F @fileobj >> + >> + /Name /PushPin + /C [0.8 0.2 0.2] + /T (GNU Free Documentation License) + /Subj (GNU FDL) + /Contents (Plain text version of the GNU Free Documentation License.) + >> +}}% + +Or, in case that PDF viewers can not extract attached files, please visit the +following site: +\medskip + +\url{http://www.gnu.org/licenses/fdl.html} + + +\end{document} -- cgit v1.2.3