Replacing Type 3 fonts with Type 1 fonts in PDF files

[Rogério Brito]

Dear people,

I have some documents that were produced with (La)TeX quite some time
ago. Unfortunately, some of them use the Computer Modern fonts in low
resolution (say, embedded with 300dpi) and, to make matters worse, some
of the documents I have have already been converted from PS to PDF with
the use of ps2pdf (which calls ghostscript as an auxiliary program).

I do know that Heiko has a program called pkfix (which is a Perl
script) to operate on PS files for substitution of the bitmapped fonts
with scalable fonts, but it unfortunately doesn't work for documents
created with some early versions of dvips.

I have already used Heiko's program, but I still have some PDF files
that include bitmapped fonts. Is there any way to batch-replace the
fonts of a bunch of documents with fonts from, say, the cm-super
package (which I have installed on my computer)?

I am currently using Debian's teTeX 2.0, but I can upgrade to teTeX 3.0
if that would make anything easier. I would prefer to have a solution
involving only Free Software, but depending on the solution, this
requirement may be dropped.

Any comments are more than welcome regarding this issue.

Thank you very much, Rogério Brito.

--
Rogério Brito : rbr...@ime.usp.br : http://www.ime.usp.br/~rbrito
Homepage of the algorithms package : http://algorithms.berlios.de
Homepage on freshmeat: http://freshmeat.net/projects/algorithms/

[Bob Tennent]

On 2 Oct 2005 04:52:10 -0700, Rogério Brito wrote:

> I have some documents that were produced with (La)TeX quite some time
> ago. Unfortunately, some of them use the Computer Modern fonts in low
> resolution (say, embedded with 300dpi) and, to make matters worse, some
> of the documents I have have already been converted from PS to PDF with
> the use of ps2pdf (which calls ghostscript as an auxiliary program).
>
> I do know that Heiko has a program called pkfix (which is a Perl
> script) to operate on PS files for substitution of the bitmapped fonts
> with scalable fonts, but it unfortunately doesn't work for documents
> created with some early versions of dvips.

Use (current) pdf2ps and then apply pkfix.

Bob T.

[Rogério Brito]

Hi, Bob. Thank you very much for your comment.

Unfortunately, it doesn't seem to work here:

- - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - -
rbrito@dumont:/tmp$ gs --version
8.15
rbrito@dumont:/tmp$ pdffonts comb.pdf | grep -c "Type 3"
54
rbrito@dumont:/tmp$ pdf2ps comb.pdf
rbrito@dumont:/tmp$ pkfix comb.ps output.ps
PKFIX 1.3, 2005/02/25 - Copyright (c) 2001, 2005 by Heiko Oberdiek.
==> no fonts converted
rbrito@dumont:/tmp$
- - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - -

Would there be anything else that I could try? I'm using gs from
Debian's testing distribution (that's gs-gpl).

I also have the option of using gs-esp, but I suspect that the results
would be the same as it appears that the pkfix script relies heavily on
the output generated by dvips.

I am quite open to any other idea.

Thanks in advance for any help, Rogério.

[Robin Fairbairns]

"=?iso-8859-1?q?Rog=E9rio_Brito?=" <rbr...@gmail.com> writes:
>Hi, Bob. Thank you very much for your comment.
>
>Unfortunately, it doesn't seem to work here:
>
>- - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - -
>rbrito@dumont:/tmp$ gs --version

>8=2E15

>rbrito@dumont:/tmp$ pdffonts comb.pdf | grep -c "Type 3"
>54
>rbrito@dumont:/tmp$ pdf2ps comb.pdf
>rbrito@dumont:/tmp$ pkfix comb.ps output.ps
>PKFIX 1.3, 2005/02/25 - Copyright (c) 2001, 2005 by Heiko Oberdiek.

>=3D=3D> no fonts converted

>rbrito@dumont:/tmp$
>- - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - -
>
>Would there be anything else that I could try? I'm using gs from
>Debian's testing distribution (that's gs-gpl).

this isn't surprising: look at the pkfix docs, and it explains it only
works with dvips output, and that from a "sufficiently recent" version
of dvips.

istr you can do this job (distilling from an existing bitmapped ps
file to pdf with corresponding outlines only) with adobe acrobat;
however, i don't actually know how to do it (acrobat is a maze of
twisty little options, all [slightly] different).
--
Robin (http://www.tex.ac.uk/faq) Fairbairns, Cambridge

[Piet van Oostrum]

>>>>> r...@cl.cam.ac.uk (Robin Fairbairns) (RF) wrote:

>RF> istr you can do this job (distilling from an existing bitmapped ps
>RF> file to pdf with corresponding outlines only) with adobe acrobat;
>RF> however, i don't actually know how to do it (acrobat is a maze of
>RF> twisty little options, all [slightly] different).

I wouldn't bet on that (acrobat can do it), for the same reason the pdf2ps
route doesn't work. Very old dvips-generated Postscript files don't have
any information about what fonts where used. Moreover, the encoding of the
fonts used is more or less random (I think it uses codes starting from 1 or
so for only the characters used, kind of subsetting). These documents also
are not searchable, and you can't copy and paste text from it for the same
reason. So the only way to find out what character is used is to look at
the bitmap and do a kind of OCR. There used to be a program that did this
long ago, but it needed copies of the bitmap fonts that were equal to the
fonts used, i.e. the same mode etc. I don't think it was very successful.
--
Piet van Oostrum <pi...@cs.uu.nl>
URL: http://www.cs.uu.nl/~piet [PGP 8DAE142BE17999C4]
Private email: pi...@vanoostrum.org