[reportlab-users] Issues related to setting parameters of graphs, lines, and rectangles, and they are not printing correctly

[J. R. Carroll]

OK, so I'm a day late replying!  Sorry - work and all =P

Alright - some updates - it turns out the problem of having stroke widths
that were really small was not a problem of reportlab, but rather the
grayscale conversion on our printers; it was trying to convert colors to
grayscale and it was causing some really nasty antialiasing issues which
made some grid lines disappear when printed - it created a plaid-like
pattern, and it was VERY frustrating to diagnose!

The solution was to print in a different color.  Since I was using the
colors dict from reportlab which has a ton of colors to select from without
having to input RGB/CMYK values, it was a pretty simple fix.  However, I
tried this at one of the earliest stages of debugging this nasty anomaly!
 The reason I dismissed colors so quickly was that I started off in pink -
saw the issue at the printer, thought "hmmm - maybe it's the colors" and
changed it to red.  Well red doesn't print very well on all of our printers
(Cannon, and HP) - and yes, looking back I see how quickly I dismissed it
especially with how close pink and red are in terms of printer ink -
but!!!!  there's more!!  =P  It's not the inks fault!  We have plenty of
ink, remember these were being printed in grayscale so color wasn't the
issue!  It was the color<=>grayscale conversion that was the issue.

The solution:  change the color

The next problem:  now we have to find a balance between "what looks good
on the screen" and "what looks good on print" (in both grayscale and
color).

The partial-solution:  I created a script that will retrieve all the colors
in reportlab, print a small sample grid + the color name + color RGB
values, neatly print them on several PDF pages (thanks to ReportLab!).  As
soon as I did this and sent to to multiple printers - we are seeing which
colors are working across the sample of available print medias we have, and
what looks good on the screen!  This script/PDF calibrator is really a
god-send.  I don't know if anyone else will ever have use for it, or if
there are better scripts out there that do the exact same thing, but this
is really helping us a lot over the next few days.

If anyone is interested in seeing the output and has access to several
printers (particular the business grayscale printers you are likely to find
in a corporation - i.e. Cannon and HP "monsters") download my script, and
open the resulting PDF (or I can email you the contents of the script in
case you don't feel comfortable downloading something from my personal
website).  Either way, I hope someone can make use of this solution - I
know my desk is happy that I'm not pounding my bloody forehead into it
anymore!

Script is available for download from:
http://www.jrcresearch.net/downloads/scripts.zip

If anyone has ideas on how to improve it - let me know.  I'm already
thinking of other neat printer calibration test prints I can add to this!

-J
----


J. R. Carroll
Independent Researcher through Hurtz Labs
Research Methods, Test Development, and Statistics
www.jrcresearch.net
Cell:  (650) 776-6613
Email: jrcarroll at jrcresearch.net
          jrcarroll at hurtzlab.com
          jrc.csus at gmail.com





On Wed, Nov 28, 2012 at 2:38 PM, Andy Robinson <andy at reportlab.com> wrote:


> On 28 November 2012 21:35, J. R. Carroll <jrcarroll at jrcresearch.net>wrote:

>

>> I'm on my way out, and would be happy to compile all this information

>> tomorrow morning, but I think I found the problem.  Apparently, neither

>> printer I was using (nor any printer I have access to) liked color.pink.  I

>> don't think it was ReportLab (and I'm sure you have more insight than I),

>> but the grayscale handling at the printer just mucked everything up (where

>> muck = some fancy tech lingo I can't think of right now).  As soon as I

>> changed the grid color from color.pink to color.black - the grid looks

>> perfect - really makes it look like I know what I am doing (jokes on them!).

>>

>> I think that this might be falling outside the ReportLab domain, but!

>> might you have any recommendations for handling color documents to a

>> black/white printer?  Do some colors print better than others - or is that

>> really a function of the printer?

>>

>>

> First question is whether the printer uses the CMYK colour model, or RGB.

>  High end ones for artwork tend to be CMYK, but for CAD I have no idea.

>  If it's RGB then some shade of red (or green or blue) is a good bet.  If

> CMYK, use magenta.  It should be easy to try out.

>

> Printers will tell you that they lay down overlapping, microscopic

> patterns of dots at various angles to each other (a bit like polarised

> filters) to achieve various colour effects.  This may not work well with

> 0.1 point lines.  If you need to mix two inks to get that line, my hunch is

> it's more likely to mess up.

>

> Please let us know tomorrow!

>

> - Andy

>

> _______________________________________________

> reportlab-users mailing list

> reportlab-users at lists2.reportlab.com

> http://two.pairlist.net/mailman/listinfo/reportlab-users

>

>

-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://two.pairlist.net/pipermail/reportlab-users/attachments/20121129/75bed573/attachment.htm>