[Deque Undoc for PDF Beta] Tip of the Day: Lions and Tigers and Type 3 Fonts, oh my!
- From: "Bill Diamond" <bill.diamond@xxxxxxxxx>
- To: <deque-ubeta@xxxxxxxxxxxxx>
- Date: Wed, 26 May 2004 17:06:08 -0400
Back in the earlier years of Postscript font design, the late 1980, Adobe
kept the design and architecture of the Type 1 family to itself. It
encouraged third party houses to develop fonts using Type 3 technology
instead. Type 3 was pretty popular as it was somewhat easier to design
fonts for and had some extras that Type 1 didn't, like the ability to embed
itself directly in the postscript output or to force characters to be drawn
as bitmaps or even to use bitmaps as elements in the fonts.
Type 3 fonts, though, had their drawbacks. Most notably, they couldn't be
managed with Adobe Type Manager. One concept that seemed like a great idea
at the time is a real liability when you're trying to make a document
accessible. That concept was the ability to download outline fonts as
bitmap fonts. From a pre-press perspective, Type 3 fonts sometimes post
major stumbling blocks when trying to rasterize a PDF for print - a process
usually referred to as "ripping".
One hint that you're facing a Type 3 font challenge is when you try to tag a
PDF and you get an odd error message that "too many unknown font encodings
encountered". This generally suggests there's a significant problem with
the use of Type 3 fonts and that Acrobat, which uses Type 1, OpenType, and
TrueType fonts, can't handle. So, what's the Undoc user to do?
You do have a variety of options. Certainly, if you can - try to get your
hands on the missing Type 3 fonts. Once those are installed, you may have
more success. Your second approach is a bit more taxing. It's likely that
your fonts have been subsetted within the PDF. This means that the
underlying postscript code that lurks inside your PDF has the postscript
instructions necessary to draw the characters that were used. If this is
true, you may be able to extract that font subset and recreate it. Products
like Macromedia's Fontographer can help you here. This is a process called
"font stripping".
To extract the troublesome fonts, save the PDF as an EPS file within
Acrobat. Use a text editor to open the EPS and looking for a line that
begins with
"%%BeginResource: font ...". Copy that line and everything below up to and
including the line that starts with "PDFVars /TermAll get exec end end".
Copy these lines into a new text file with the name of the font that you're
going to recreate. Remember to make sure you get the file name extension
right! This is platform specific, and on the Macintosh you'll need to get
the correct Type and Creator codes. Open the file with a font editor and
recreate the missing font. It probably won't be the full set of characters;
most people who create PDFs only subset the minimal set they need within the
PDF to reduce the file size and to respect the property rights of whoever
created the font.
Once you have all the fonts recreated and installed, you may then be able to
use the Acrobat Touch Up Text tool to try to change the font used.
In this case, "may be" is a cautionary statement. One feature of Type 3
fonts that I spoke about earlier was how these fonts could be drawn as
bitmaps - and this is also true when documents using such fonts are printed,
including to PDF, or to postscript that is then distilled into PDF. It is
quite possible to find that when the PDF was generated the fonts were
embedded as graphics. That can produce a better looking document that
behaves better as a PDF when it's scaled or resized, but what it produces is
a graphic element that's generated by the postscript programming language.
It's not a file with embedded text and associated styles and markups.
You can test your PDF with Adobe Acrobat by exporting the text in the file.
Use Acrobat's "File->Save As" menu and export to text. When you open the
text output you should see, well, text. Lots of text. If you don't see
text, but see lots of odd chacters like square boxes, tildes, and single
digit numbers, you probably have a postscript file with embedded Type 3 font
use to generate characters programmatically.
When you encounter this kind of a problem, there's still one thing you can
try. Go back into Acrobat and save your PDF as TIFF using the "File->Save
As" menu. Then, use an optical character recognition (OCR) tool to try to
recreate the file.
Bill
________________________
Bill Diamond
Director of Marketing and Alliances
Deque Systems (http://www.deque.com <http://www.deque.com/> )
11180 Sunrise Valley Drive, Suite 400
Reston , VA 20191
+001 703 225-0380 voice
+001 703 225-0387 fax
877-OK-TODAY USA Free Call
Other related posts:
