[GRASS-user] Odd scaling issue with ps.map output and ghostscript ps2pdf

GRASS GRASS user
1

On several occasions I've created maps with ps.map at 1:24000 scale, converted
to PDF form with ghostscript's "ps2pdf" script, and sent them off to a printer
(i.e. paid a shop to print them) only to find that they've come back in
1:25000 scale instead. The first time this happened was after I'd imported
the PDF into Inkscape to clean up labels, so I attributed it to Inkscape.

This time, however, I cleverly avoided using Inkscape to monkey with the map,
and found it came back to me in 1:25000 again anyway. And when I investigated,
found that the PDF file was indeed shrunk from the PostScript --- when I
viewed on screen at 1:1 scale, I found that my 1:24000 UTM interpolator worked
great with the on-screen postscript file, but I needed to create a 1:25000
interpolator for the PDF version of the same file.

I'm really puzzled how this can happen, and am guessing that somehow ghostscript
is using a different DPI to create the PDF file than it should.

Has anyone else ever seen this odd behavior? I'd really like to understand
it and avoid it again. For both delivering to a print shop and for importing
to Inkscape, PDF format is better than PostScript, so I'm really keen to fix
this problem.

For those who want a look, the map in question is temporarily available in
both PDF and PostScript versions at <ftp://ftp.swcp.com/pub/tmp/russo/&gt;\.
The files at that ftp site are automatically deleted after 4 days, though.

Can anyone suggest a way around this scaling of the map by ghostscript? Or
point me to a reason it should be happening?

--
Tom Russo KM5VY SAR502 DM64ux http://www.swcp.com/~russo/
Tijeras, NM QRPL#1592 K2#398 SOC#236 http://kevan.org/brain.cgi?DDTNM
"It is better to live on your feet than to die with your knees."
  -- Mil Millington on running, in Instructions for Living Someone Else's Life

2

Tom Russo wrote:

On several occasions I've created
maps with ps.map at 1:24000 scale, converted
to PDF form with ghostscript's "ps2pdf" script, and sent
them off to a printer
(i.e. paid a shop to print them) only to find that they've
come back in
1:25000 scale instead. The first time this happened
was after I'd imported
the PDF into Inkscape to clean up labels, so I attributed
it to Inkscape.

This time, however, I cleverly avoided using Inkscape to
monkey with the map,
and found it came back to me in 1:25000 again anyway.
And when I investigated,
found that the PDF file was indeed shrunk from the
PostScript --- when I
viewed on screen at 1:1 scale, I found that my 1:24000 UTM
interpolator worked
great with the on-screen postscript file, but I needed to
create a 1:25000
interpolator for the PDF version of the same file.

I'm really puzzled how this can happen, and am guessing
that somehow ghostscript
is using a different DPI to create the PDF file than it
should.

Has anyone else ever seen this odd behavior? I'd
really like to understand
it and avoid it again. For both delivering to a print
shop and for importing
to Inkscape, PDF format is better than PostScript, so I'm
really keen to fix
this problem.

For those who want a look, the map in question is
temporarily available in
both PDF and PostScript versions at <ftp://ftp.swcp.com/pub/tmp/russo/&gt;\.
The files at that ftp site are automatically deleted after
4 days, though.

Can anyone suggest a way around this scaling of the map by
ghostscript? Or
point me to a reason it should be happening?

(sorry about yahoo's broken line wrap)

no idea, but from now on I'll keep an eye out for it. :slight_smile:
perhaps an A4 vs. Letter papersize adjustment?

how are you converting to PDF? "ps2pdf13" or another GS wrapper?

what does acrobat's "About the Document" say about the PDF paper size
versus what is in the header lines of the .ps file?

does it happen for self-printed PDFs, or only ones from the print shop?

how are you/them printing? via Acrobat's File->Print? is the "adjust to
fit on page" tickbox ticked?

Hamish

3

On Tue, 27 Apr 2010, Hamish wrote:

no idea, but from now on I'll keep an eye out for it. :slight_smile: perhaps an A4 vs.
Letter papersize adjustment?

Hamish/Tom:

   For my most recent project I did not care at what scale the maps were
printed as long as the area of interest fit on the page. I noticed that
while I could specify a scale, ps.map adjusted that scale when the
PostScript was generated. Like Tom, I used ps2pdf and the scale did not
change in this step.

   While this does not lead to insight into the source of Tom's frustration
it's an observation that suggests ps.map changes the scale to fit the page
size and, perhaps, tweaking the margins or slightly shrinking the displayed
area would keep the scale at the desired 24K.

Rich

4

On Tue, Apr 27, 2010 at 04:32:09AM -0700, we recorded a bogon-computron collision of the <hamish_b@yahoo.com> flavor, containing:

Tom Russo wrote:
> On several occasions I've created
> maps with ps.map at 1:24000 scale, converted
> to PDF form with ghostscript's "ps2pdf" script, and sent
> them off to a printer
> (i.e. paid a shop to print them) only to find that they've
> come back in
> 1:25000 scale instead. The first time this happened
> was after I'd imported
> the PDF into Inkscape to clean up labels, so I attributed
> it to Inkscape.
>
> This time, however, I cleverly avoided using Inkscape to
> monkey with the map,
> and found it came back to me in 1:25000 again anyway.
> And when I investigated,
> found that the PDF file was indeed shrunk from the
> PostScript --- when I
> viewed on screen at 1:1 scale, I found that my 1:24000 UTM
> interpolator worked
> great with the on-screen postscript file, but I needed to
> create a 1:25000
> interpolator for the PDF version of the same file.

[...]

(sorry about yahoo's broken line wrap)

no idea, but from now on I'll keep an eye out for it. :slight_smile:
perhaps an A4 vs. Letter papersize adjustment?

how are you converting to PDF? "ps2pdf13" or another GS wrapper?

what does acrobat's "About the Document" say about the PDF paper size
versus what is in the header lines of the .ps file?

does it happen for self-printed PDFs, or only ones from the print shop?

how are you/them printing? via Acrobat's File->Print? is the "adjust to
fit on page" tickbox ticked?

I'm now convinced that there's nothing wrong with the PDF, and that this
is indeed a "shrink to fit" issue. Contrary to what I said at first, there's
no scaling issue visible on screen when I view either postscript or pdf using
the same program (ghostview), and it's spot-on at 1:24000.

My guess is that my map comes too close to the margins, and the print shop is
using "shrink to fit" to avoid cutting off content. I'm going to give them
a call today to find out for sure.

Sorry for the noise I created here.

--
Tom Russo KM5VY SAR502 DM64ux http://www.swcp.com/~russo/
Tijeras, NM QRPL#1592 K2#398 SOC#236 http://kevan.org/brain.cgi?DDTNM
"It is better to live on your feet than to die with your knees."
  -- Mil Millington on running, in Instructions for Living Someone Else's Life

5

On Tue, Apr 27, 2010 at 06:19:38AM -0700, we recorded a bogon-computron collision of the <rshepard@appl-ecosys.com> flavor, containing:

On Tue, 27 Apr 2010, Hamish wrote:

> no idea, but from now on I'll keep an eye out for it. :slight_smile: perhaps an A4 vs.
> Letter papersize adjustment?

Hamish/Tom:

   For my most recent project I did not care at what scale the maps were
printed as long as the area of interest fit on the page. I noticed that
while I could specify a scale, ps.map adjusted that scale when the
PostScript was generated.

I knew of this issue, and was very careful not to try to include more map
than could fit on the 11x17 landscape page at 1:24000. ps.map does print out
what scale it actually uses, and I always watch to see that it is using what
I asked for.

In this instance I am now completely convinced it's an issue at the print shop,
probably because I'm trying to use margins that are too small and they're
"helpfully" adjusting the size of the image to fix my mistake so the edges
aren't cut off.

--
Tom Russo KM5VY SAR502 DM64ux http://www.swcp.com/~russo/
Tijeras, NM QRPL#1592 K2#398 SOC#236 http://kevan.org/brain.cgi?DDTNM
"It is better to live on your feet than to die with your knees."
  -- Mil Millington on running, in Instructions for Living Someone Else's Life

6

On Tue, 27 Apr 2010, Tom Russo wrote:

In this instance I am now completely convinced it's an issue at the print
shop, probably because I'm trying to use margins that are too small and
they're "helpfully" adjusting the size of the image to fix my mistake so
the edges aren't cut off.

Tom,

   It may also be that the printer they're using has non-printing margins
greater than what you specify. Try printing on larger paper (e.g., C size;
they must have a plotter there) and trimming to size after.

Rich

7

On Tue, Apr 27, 2010 at 09:34:07AM -0700, we recorded a bogon-computron collision of the <rshepard@appl-ecosys.com> flavor, containing:

On Tue, 27 Apr 2010, Tom Russo wrote:

> In this instance I am now completely convinced it's an issue at the print
> shop, probably because I'm trying to use margins that are too small and
> they're "helpfully" adjusting the size of the image to fix my mistake so
> the edges aren't cut off.

Tom,

   It may also be that the printer they're using has non-printing margins
greater than what you specify. Try printing on larger paper (e.g., C size;
they must have a plotter there) and trimming to size after.

I spoke with the fellow at the print shop this morning. They had indeed
"shrunk to fit" precisely because the very lowest line of text on my map
collar was too close to the edge of the page. Since I had a declination
diagram on the map that had a caption saying "Diagram not to scale" they had
mistakenly concluded that the entire document was not to scale and therefore
scaled it rather than clip the edges. (The guy admitted that he hadn't noticed
that elsewhere on the map it said "Scale 1:24000" in larger and more
prominent letters.)

I went back and looked at the PDF from the last time this happened, and it, too
had some text that went perilously close to the edge of the paper.

So this was a cross between my fault for letting collar text get too close
to the edge, and theirs for not telling me "we scaled your document so the
content wouldn't be clipped." Now that I know what I did to cause the
problem, and that it was indeed a setting that they needed not to select, I'll
remember not to let my collar text get so close to the edges, and will also
remember to state "I want this printed without being shrunk to fit, even if a
little text is truncated" when I send them a job.

Again, sorry for having made noise on this list with something that has nothing
whatsoever to do with ps.map or ps2pdf.

(And in answer to Hamish's question about whether this only happens at the
print shop and not with self-printed maps, yes, it does --- because I only
send them stuff that's too large a format for me to print myself --- but
I don't do PDF conversions for printing myself, because I have postscript
printers that don't need that step. So there were multiple issues
leading me to the incorrect conclusion about where the problem lay.)

--
Tom Russo KM5VY SAR502 DM64ux http://www.swcp.com/~russo/
Tijeras, NM QRPL#1592 K2#398 SOC#236 http://kevan.org/brain.cgi?DDTNM
"It is better to live on your feet than to die with your knees."
  -- Mil Millington on running, in Instructions for Living Someone Else's Life