[pdftex] Font Embedding and Font Substitutions
Hartmut Henkel
hartmut_henkel at gmx.de
Mon Nov 27 22:24:32 CET 2006
On Mon, 27 Nov 2006, Paarvai Naai wrote:
> I have a question regarding pdflatex's font embedding and font
> substitution. I have a latex file that includes two graphics.
> pdffonts returns the following on each:
>
> name type emb sub uni object ID
> ------------------------------------ ------------ --- --- --- ---------
> GXJYGS+CMMI10 Type 1C yes yes no 11 0
> QAKHSM+CMMI7 Type 1C yes yes no 13 0
>
> name type emb sub uni object ID
> ------------------------------------ ------------ --- --- --- ---------
> CPKLQV+CMMI7 Type 1C yes yes no 19 0
> EKIDYN+CMMI10 Type 1C yes yes no 11 0
>
> Then I run pdflatex on my latex file and get an ouput that shows the
> following with pdffonts:
>
> name type emb sub uni object ID
> ------------------------------------ ------------ --- --- --- ---------
> EIOZYT+CMMI10 Type 1 yes yes no 31 0
> PMUOFJ+CMMI7 Type 1 yes yes no 282 0
> EIOZYT+CMMI10 Type 1 yes yes no 982 0
> PMUOFJ+CMMI7 Type 1 yes yes no 985 0
> PMUOFJ+CMMI7 Type 1 yes yes no 1022 0
> EIOZYT+CMMI10 Type 1 yes yes no 1031 0
>
> So it's great that the CMMI7 and CMMI10 fonts have been consolidated
> (note how they all have the same prefix). However, they still show up
> twice in pdffonts. So what does this mean? It seems like everything
> should have been consolidated to only show up once.
it could be like this, but optimization currently just doesn't go that
far. E. g. when embedding PDF files, each font call still produces one
separate font dictionary and one font descriptor object. Currently only
the fontfile stream objects are optimized in that one embedded subsetted
fontfile can collect all glyphs from several fonts. In this case several
font descriptor objects point to the same fontfile object.
> I am using gs 8.54 and epstopdf to create the two images and pdflatex
> 1.40beta to generate the final document.
Regards, Hartmut
More information about the pdftex
mailing list