[pdftex] Strange font name in PDF

Reinhard Kotucha reinhard.kotucha at web.de
Mon May 9 23:46:09 CEST 2011

On 2011-05-10 at 06:27:52 +1000, Ross Moore wrote:

 > On 10/05/2011, at 1:11 AM, narke wrote:
 > > Hi,
 > > 
 > > In the pdflatex generated file, I found the font names look strange,
 > > it seem has some kind of random string as prefix, such as
 > > LUSKLS-CMR10.
 > This is a perfectly standard name for a font subset;
 > in this case a subset of CMR10.
 > The prefix implies that it is not the complete font.
 > Different subsets of the same font must have a different
 > prefix. When a PDF application finds that a document will
 > contain different subsets of the same font, then these 
 > can be combined.
 > This naming scheme dates from PDF v1.1.
 > It is described in 
 >     ¤7.7.4   PDF Reference Manual v1.3  (1999)  Adobe Systems Inc.
 >     ¤5.5.3   PDF Reference 5th ed. PDF v1.6  (2004)  Adobe Systems Inc.
 >     ¤9.6.4   Document management Ñ Portable document format Ñ Part 1: PDF 1.7  (2008)
 > and other versions of the PDF specification documents.
 > > 
 > > I tried to read pdftex manul, but it seem not easy to me.  I lack some
 > > tex knowledge, just know how to use pdflatex to produce pdf from latex
 > > soruce.
 > > 
 > > Now these strange fonts name are causing problems to me. When I tried
 > > to convert a tikz pictures to emf picture using pstoedit, the fonts
 > > mapping looked not correct.
 > Sounds like  pstoedit  isn't programmed to combine font subsets
 > automatically. There is no instance of 'subset' in the docs:
 >    http://www.pstoedit.net/pstoedit/pstoedit.htm
 > > 
 > > Is there a way letting me have standard fonts names in produced PDF files?
 > These *are* standard PDF font subset names.
 >  pstoedit  has a  -fontmap  option which may do what you want;
 > that is, allow you to map the prefixed font-subset names to the
 > full font name, and thereby (hopefully) include the whole font
 > (provided you have it on your system). 
 > Others, with experience using  pstoedit , may be able to confirm
 > whether this is possible or not, or give advice on what else
 > to try.

The most obvious solution is to create PostScript instead of PDF:
Run latex and dvips instead of pdflatex. 

pstoedit prefers PostScript files anyway.  AFAIK, dvips creates only
one subset per font and thus doesn't need those random strings.


Reinhard Kotucha                                      Phone: +49-511-3373112
Marschnerstr. 25
D-30167 Hannover                              mailto:reinhard.kotucha at web.de
Microsoft isn't the answer. Microsoft is the question, and the answer is NO.

More information about the pdftex mailing list