Powerpoint Postscript errors using two-color gradient backgrounds

[Michael C. Newell]

We are having a problem with PowerPoint on OS-X (V10.2.6). One of our users
was trying to use a two-color gradient background on his slides and they
were not printing. When we enabled Postscript error printing on the
printers (various color HP models) the gradient printed and embedded in it
was the error:

ERROR: undefinedresult
OFFENDING COMMAND: div

STACK:

0.0
0.0
-dictionary-

Clearly PPT is generating Postscript that wants to do a divide by zero.
This only happens with the two color gradient background; if we change the
slides to use one color gradient (or a solid color) the error does not
occur. This happens independent of the PPD (we tried Generic Postscript and
just for fun the Lexmark PPD), and it also occurs if we just create a new
presentation with a two color gradient background.

Can anyone out there shed some light on this?

Thanks!

Mike Newell
NASA HQ ISEM contract chief engineer
(e-mail address removed)

 

[Steve Rindsberg]

Hm. Just tried a twocolor grad and it prints ok to a B/W PS printer (HP
Laserjet 4V/M) from PPT-X

I wonder if we can isolate it further -- what colors did you use, what
gradient direction, was there an outline or no, was transparency applied to
either of the colors ... etc.

As a workaround, what happens if you save to PDF then print the PDF?

 

[Michael C. Newell]

It's always printed fine on

What version of PPT are you using? We're using 10.1.4. I'm not sure what
updates they've done, but this particular user we're having problems with is
an early adopter and may be somewhat back-rev. He is able to print his
charts if he drops to classic mode PPT 2001 without a hitch. They also
print fine from any version of Windows PPT that we've tried...

As far as colors are concerned we've tried lots of different combinations.
All fail. We've also tried reducing the image size (in case the gradient is
running off the printable area of the printer) and it still fails.
Weierd...

Printing to PDF then printing the PDF seems to work.

Thanks!

Mike

This just in as I was typing this - our Mac person tried printing it on an
HP 8550 and it printed fine. Fails on all our 4500 and 4550 printers
though... This user, of course, has a 45xx printer. Sigh...

 

[Steve Rindsberg]

What version of PPT are you using? We're using 10.1.4. I'm not sure
what

updates they've done, but this particular user we're having problems with is
an early adopter and may be somewhat back-rev.

About PowerPoint says PowerPoint X SR1
I haven't grabbed any updates to that, however.

As far as colors are concerned we've tried lots of different combinations.
All fail. We've also tried reducing the image size (in case the gradient is
running off the printable area of the printer) and it still fails.
Weierd...

Between that and your other info, it makes me wonder if it's not some
peculiarity of the PS RIP in the printer.
PS is actually a kind of programming language; the PS an app sends out can
query the printer and ask it questions like "What resolution are you set
to?" and then adjust the way they do things like gradients accordingly. On
the PC, PPT's PS doesn't do this - a pity, as it would be a major
improvement, but it may be limited to what the operating system provides.
I've never torn apart the PS from Mac/PPT, but it might be doing something
of the sort. The resolution can effect the way you optimize gradients ....
if it's querying the printer for that and getting back a zero or null
instead of a valid value ... well .... you've seen the smoke, right? ;-)

 

[Michael C. Newell]

Yeah - I actually got enormously interested in PostScript many years ago
when the printers first came out. In fact there's a book called "Thinking
Postscript" that turns out to have been written by a former college
classmate of mine; I bought it as a reference when it first came out and
didn't realize until years later who the author was. Years ago (yeah -
I'm old!) when upgrading Laserwriters there were some anomolies on the
macros and I used to have to go in a patch the PPDs to produce the right
code...

The problem is clearly a divide by zero in the Postscript generated by
Powerpoint. Why it only occurs on HP 4xxx printers is interesting. I'm
going to play with routing the job through Ghostscript and generating PCL to
see if the problem goes away for this user. Lots of effort so someone can
print a two-color gradient background...

Thanks!

Mike