[pdftex] incompatibility caused by Adobe EPS 6.0 patterns?
crowell03 at lightandmatter.com
Sat Sep 13 11:00:46 CEST 2003
I posted recently with a question about what preflight checks I could
do on a book being printed from pdftex output. Thanks, everyone,
for the suggestions!
The only actual problem that did show up was unexpected. The printer's
Agfa imagesetter (am I getting the terminology right? -- produces plates (?)
directly from PDF?) gave errors on certain pages of the book. I'm going
to try to explain the situation in as much detail as possible, but some
of the information I got back was kind of vague. I produce my books through
a guy who subcontracts out the printing and binding, so I'm one step removed
from everything. The guy I deal with is wonderful, but is not all that
comfortable with computers in general. The software on the Agfa machine
is proprietary, and apprently gives lousy, uninformative error messages.
Of course these guys are used to Distiller output, and have never heard of
pdftex, and I'm sure the Agfa software was only ever tested on Distiller
I /think/ the situation is as follows. On two pages of my book, I have
line art that was done with Adobe Illustrator and includes what Illustrator
calls a ``pattern'' of bricks. It has a variety of patterns you can use,
fish scales, etc. I normally save these files as EPS 1.0 and convert them
to PDF using epstopdf, but I know from experience that when you use certain
features, including gradients and patterns, you have to save them as EPS 6.0
(which is much more bloated). The two pages that use the pattern feature seem to
be the ones that are causing the problem, although apparently the imagesetter
only tells you which 4-page plate it occurs on, and at first we didn't even
realize that these pages had this common denominator.
Well, the final result seems ok anyway -- they say the output actually looks
fine despite the error. But of course it's a little nerve-wracking, since
the process is direct digital->plate, so I can only see proofs made on a
different device, not proofs made on the Agfa. In this particular book,
I went ahead and replaced the brick pattern with hand-drawn bricks, which
may eliminate the error in the future. I have 5 more books, however, that I
plan to convert from PageMaker to TeX, and would like to know if I'm
understanding the problem correctly.
Does anyone know if my analysis makes any sense? Does a particular feature
like patterns in Adobe EPS 6.0 map onto a particular feature in PDF, and
if so, could there be a real problem with the PDF output from pdftex?
One possible solution that occurred to me was to print the pdftex output
to a PostScript file, and then run Distiller on that. However, that
doesn't seem like a great solution to me since, (a) it means there's
one more obstacle to getting rid of my computers that run a proprietary
OS, (b) other information might get lost or garbled in that process,
and (c) it's no substitute for understanding what's actually going on.
The book in question (now with hand-drawn bricks on pp. 120 and 130) is
at http://www.lightandmatter.com/area1book2.html (both source
More information about the pdftex