httpd-docs mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From Joshua Slive <>
Subject Re: Antwort: Re: latex status
Date Tue, 17 Jun 2003 19:23:45 GMT

On Tue, 17 Jun 2003 wrote:
> Which exact purpose is this PDF file to serve in the first place?
> I am far from understanding PDF, but for one of my projects we
> generated some PDF version of a changes documentation in different
> resolutions (one very detailled to provide high-quality printing,
> another one still sufficient for online reading) and the size dif-
> fered by a factor of nearly 10 ... for the same content, that is.
> So maybe fine-tuning the "output quality" might help here?

There are two basic types of ps/pdf fonts: outline and bitmap.  The
outline fonts are described by their shape, while the bitmap fonts are
described by -- you guessed it -- a bitmapped picture.  (In postscript
terminology, outline fonts are "Type 1"  and bitmap fonts are "Type 3".)

Bitmap fonts look good only at a single resolution, so if you want to
read them on screen, you need to include a low-resolution bitmap, and
if you want a 600dpi printout, you need a high-resolution bitmap.  But
outline fonts can be scaled to any resolution.  We are using outline fonts
in our pdf file, so there will be very little impact from changing the

Of course, there are some other font issues.  For example, I'm not sure if
pdflatex is doing proper font-subsetting.  For the fonts that it includes
(because it can't find the right character in the standard fonts), it
should only be including the outlines of the characters it uses.  But it
may instead be including all the characters.

At some point, I can try an experiment.  Rather than using pdflatex to
generate pdf directly, I can do
latex -> dvips -Pwww -> distiller
If this comes out with something substantially smaller, it means that
pdflatex is doing something wrong.

> But then, I am not used to getting a great compression factor for
> PDFs anyway, as the language used therein looks like it contains
> much less redundancy than, say, HTML.
> I just "gzip -9"ed a couple of PDF files here and didn't get any-
> where near to a factor 2 of savings. (And things get even worse
> when the PDF file contains graphics.)

Actually, our mystery is why is the post-compression so GOOD?  PDF is
supposed to be a compressed format, so it is surprising that we can
recompress using zip/gzip and get a substantially smaller file.  It
implies that pdflatex is not doing a great job.


To unsubscribe, e-mail:
For additional commands, e-mail:

View raw message