Digital BW, The Print

Hi all, 

First a quick update, then a question:

I've got Roy's QTR software up and cooking under redhat 9.
Thanks to Roger Sopher and Roy for help and preliminary
installation notes. Installation was almost painless--we 
discovered one idiosyncrasy with configuring the CUPS printing
system, which involved changing some processing rules; I've
created a file of additional CUPS MIME rules that resolve 
the issue for me, which Roy and Roger now have.

I'm driving an Epson 2200 loaded with standard UC inks.

To get started, I'm printing tif's from Linux using the
qtcups utility, and the output is lovely. So far, I've sunk
about a buck (3 blank CD's used to burn the RH9 ISO images)
and 30 hours of sweat equity into this project. Not a dongle
in sight, either. :-) I'm using a worthless old PII 333 MHz 
PC to run RH9. 128 MB memory, 6GB HD.

Now my question: 

When I print the 21Step image using the curve EEM_2200-cool I 
see two very distinct hitches in the continuous ramp, one above
the 60% patch and one above the 90% patch. After running my first
calibration series last night, I got the same thing in my own
profile, shifted to slightly different positions.

Except for the discontinuity at those two points, the ramp
looks great, cool and smooth with delicate highlights and a 
velvety black at 100%. 

Am I doing something wrong? Do others see this? The settings I 
used were:

Dither: Adaptive Hybrid
Image Type: Photograph
Ink Type: Seven Color Photo
Resolution 1440x720 Highest Quality. 
Blend Curves 1 and 2: 100-0
Quadtone Rip Curve 1: EEM_2200-cool

-David

Hi David,

Glad to see things are getting better.  I've got a hunch about 
the hitches in the gray ramp:  the 2200 printer has two dotsizes
for 1440x720 resolution.  The gimp-print software has to transition
from one dotsize to the next smoothly and it does this based on a weighting
of the dotsizes.  I have a feeling that gimp-print hasn't had a very accurate
way to measure this weighting, because they've changed it several times
for the 2200.  Below is the code that has this weighting -- the 0.661 is
the important value.  The value should probably be lower, in GP I've seen
it go to 0.56, then 0.50.  I've worked with Carl a bit trying to get a way to
measure it accurately and on his machine 0.46 seemed best.  You can
just edit the source file, recompile and install rastertoprinter-quad.
This will require a re-measurement for the profile.

Roy


---------file: RasterSrc/main/print-escp2-data.c ------------

/***************************************************************\
*                                                               *
*            ULTRACHROME (2100/2200, 7600, 9600)                *
*                                                               *
\***************************************************************/

static const stp_simple_dither_range_t standard_4pl_pigment_dither_ranges[] =
{
  { 0.661, 0x1, 0, 2 },
  { 1.00,  0x2, 0, 3 }
};

DECLARE_INK(standard_4pl_pigment, 1.0);

--------------------



--- In DigitalBlackandWhiteThePrint@yahoogroups.com, "David Wroblewski" 
<dawroblewski@y...> wrote:

> Hi all, 
> 
> First a quick update, then a question:
> 
> I've got Roy's QTR software up and cooking under redhat 9.
> Thanks to Roger Sopher and Roy for help and preliminary
> installation notes. Installation was almost painless--we 
> discovered one idiosyncrasy with configuring the CUPS printing
> system, which involved changing some processing rules; I've
> created a file of additional CUPS MIME rules that resolve 
> the issue for me, which Roy and Roger now have.
> 
> I'm driving an Epson 2200 loaded with standard UC inks.
> 
> To get started, I'm printing tif's from Linux using the
> qtcups utility, and the output is lovely. So far, I've sunk
> about a buck (3 blank CD's used to burn the RH9 ISO images)
> and 30 hours of sweat equity into this project. Not a dongle
> in sight, either. :-) I'm using a worthless old PII 333 MHz 
> PC to run RH9. 128 MB memory, 6GB HD.
> 
> Now my question: 
> 
> When I print the 21Step image using the curve EEM_2200-cool I 
> see two very distinct hitches in the continuous ramp, one above
> the 60% patch and one above the 90% patch. After running my first
> calibration series last night, I got the same thing in my own
> profile, shifted to slightly different positions.
> 
> Except for the discontinuity at those two points, the ramp
> looks great, cool and smooth with delicate highlights and a 
> velvety black at 100%. 
> 
> Am I doing something wrong? Do others see this? The settings I 
> used were:
> 
> Dither: Adaptive Hybrid
> Image Type: Photograph
> Ink Type: Seven Color Photo
> Resolution 1440x720 Highest Quality. 
> Blend Curves 1 and 2: 100-0
> Quadtone Rip Curve 1: EEM_2200-cool
> 
> -David

David,

I've sent you an email with an attachment containing the 2200 profiles  
I'm using now, after the dot weighting adjustment that Roy mentioned in  
his post.  You might want to re-linearize the profiles after you make  
the GP dot weighting adjustment.

Carl

On Wednesday, October 15, 2003, at 10:16  AM, David Wroblewski wrote:

> Hi all,
>
> First a quick update, then a question:
>
> I've got Roy's QTR software up and cooking under redhat 9.
> Thanks to Roger Sopher and Roy for help and preliminary
> installation notes. Installation was almost painless--we
> discovered one idiosyncrasy with configuring the CUPS printing
> system, which involved changing some processing rules; I've
> created a file of additional CUPS MIME rules that resolve
> the issue for me, which Roy and Roger now have.
>
> I'm driving an Epson 2200 loaded with standard UC inks.
>
> To get started, I'm printing tif's from Linux using the
> qtcups utility, and the output is lovely. So far, I've sunk
> about a buck (3 blank CD's used to burn the RH9 ISO images)
> and 30 hours of sweat equity into this project. Not a dongle
> in sight, either. :-) I'm using a worthless old PII 333 MHz
> PC to run RH9. 128 MB memory, 6GB HD.
>
> Now my question:
>
> When I print the 21Step image using the curve EEM_2200-cool I
> see two very distinct hitches in the continuous ramp, one above
> the 60% patch and one above the 90% patch. After running my first
> calibration series last night, I got the same thing in my own
> profile, shifted to slightly different positions.
>
> Except for the discontinuity at those two points, the ramp
> looks great, cool and smooth with delicate highlights and a
> velvety black at 100%.
>
> Am I doing something wrong? Do others see this? The settings I
> used were:
>
> Dither: Adaptive Hybrid
> Image Type: Photograph
> Ink Type: Seven Color Photo
> Resolution 1440x720 Highest Quality.
> Blend Curves 1 and 2: 100-0
> Quadtone Rip Curve 1: EEM_2200-cool
>
> -David
>
>
>
>
>
>
>
>
>
> ------------------------ Yahoo! Groups Sponsor  
> ---------------------~-->
> Buy Ink Cartridges or Refill Kits for your HP, Epson, Canon or Lexmark
> Printer at MyInks.com. Free s/h on orders $50 or more to the US &  
> Canada.
> http://www.c1tracking.com/l.asp?cid=5511
> http://us.click.yahoo.com/mOAaAA/3exGAA/qnsNAA/ucIolB/TM
> --------------------------------------------------------------------- 
> ~->
>
> Please visit the Group Homepage to check the Files, Bookmarks, Polls  
> and other resources as they are often being updated. The page is at:
>
> http://groups.yahoo.com/group/DigitalBlackandWhiteThePrint
>
> If you wish to receive no emails or just a daily digest, or you wish  
> to unsubscribe, please edit your Membership preferences by visiting  
> this same page.
>
> Please follow these basic guidelines:
> - Include your full name with your message.
> - Include the address of your website, if you have one.
> - As threads develop, trim off excess portions of earlier messages to  
> keep them short.
> - As the topic of a thread changes remember to change the subject  
> header.
> - Good manners are required at all time. No personal attacks or flames
> - Complete your Yahoo profile.
> - Before posting a question, search the message archives and the  
> various resources on the homepage.
>
>
>
>
> Your use of Yahoo! Groups is subject to  
> http://docs.yahoo.com/info/terms/
>
>

Making the change now... will let you know what happens. 

Is the dot size issue specific to that one resolution?
Or are there similar transition points for other resolutions?

-d



--- In DigitalBlackandWhiteThePrint@yahoogroups.com, "Roy 
Harrington" <roy@h...> wrote:
> 
> Hi David,
> 
> Glad to see things are getting better.  I've got a hunch about 
> the hitches in the gray ramp:  the 2200 printer has two dotsizes
> for 1440x720 resolution.  The gimp-print software has to transition
> from one dotsize to the next smoothly and it does this based on a 
weighting
> of the dotsizes.  I have a feeling that gimp-print hasn't had a 
very accurate
> way to measure this weighting, because they've changed it several 
times
> for the 2200.  Below is the code that has this weighting -- the 
0.661 is
> the important value.  The value should probably be lower, in GP 
I've seen
> it go to 0.56, then 0.50.  I've worked with Carl a bit trying to 
get a way to
> measure it accurately and on his machine 0.46 seemed best.  You can
> just edit the source file, recompile and install rastertoprinter-
quad.
> This will require a re-measurement for the profile.
> 
> Roy
> 
> 
> ---------file: RasterSrc/main/print-escp2-data.c ------------
> 
> /***************************************************************\
> *                                                               *
> *            ULTRACHROME (2100/2200, 7600, 9600)                *
> *                                                               *
> \***************************************************************/
> 
> static const stp_simple_dither_range_t 
standard_4pl_pigment_dither_ranges[] =
> {
>   { 0.661, 0x1, 0, 2 },
>   { 1.00,  0x2, 0, 3 }
> };
> 
> DECLARE_INK(standard_4pl_pigment, 1.0);
> 
> --------------------
> 
> 
> 
> --- In DigitalBlackandWhiteThePrint@yahoogroups.com, "David 
Wroblewski" 

> <dawroblewski@y...> wrote:
> > Hi all, 
> > 
> > First a quick update, then a question:
> > 
> > I've got Roy's QTR software up and cooking under redhat 9.
> > Thanks to Roger Sopher and Roy for help and preliminary
> > installation notes. Installation was almost painless--we 
> > discovered one idiosyncrasy with configuring the CUPS printing
> > system, which involved changing some processing rules; I've
> > created a file of additional CUPS MIME rules that resolve 
> > the issue for me, which Roy and Roger now have.
> > 
> > I'm driving an Epson 2200 loaded with standard UC inks.
> > 
> > To get started, I'm printing tif's from Linux using the
> > qtcups utility, and the output is lovely. So far, I've sunk
> > about a buck (3 blank CD's used to burn the RH9 ISO images)
> > and 30 hours of sweat equity into this project. Not a dongle
> > in sight, either. :-) I'm using a worthless old PII 333 MHz 
> > PC to run RH9. 128 MB memory, 6GB HD.
> > 
> > Now my question: 
> > 
> > When I print the 21Step image using the curve EEM_2200-cool I 
> > see two very distinct hitches in the continuous ramp, one above
> > the 60% patch and one above the 90% patch. After running my first
> > calibration series last night, I got the same thing in my own
> > profile, shifted to slightly different positions.
> > 
> > Except for the discontinuity at those two points, the ramp
> > looks great, cool and smooth with delicate highlights and a 
> > velvety black at 100%. 
> > 
> > Am I doing something wrong? Do others see this? The settings I 
> > used were:
> > 
> > Dither: Adaptive Hybrid
> > Image Type: Photograph
> > Ink Type: Seven Color Photo
> > Resolution 1440x720 Highest Quality. 
> > Blend Curves 1 and 2: 100-0
> > Quadtone Rip Curve 1: EEM_2200-cool
> > 
> > -David

Roy and Carl--thanks!

The 0.661 -> 0.46 change totally eliminates the hitches I 
was seeing in the gray ramp.

(1) Even without changing profiles, modifying 0.661 to 0.46
removed the hitches. I immediately get a smooth tonal range
when printing 21Step. 

(2) With Carl's new EEM_2200-cool profile, using his linearization, 
I get equally smooth results. The only difference between theis and
(1) is where the "cool" toning is overtaken by pure black, which 
varies by a few percent. Both look great to me.

When I can, I'll linearize this for my printer and report those 
results. But I haven't even blended any curves yet! I have to play
with that first.

-david

ps. Advice: don't install new profiles while a print job is 
running. :-) The current print job gets confused, reverts to lots 
of funny ascii chars, and you have to cycle power on the 
printer to stop it. Maybe I'm missing some trick, but after 
you cycle power on the printer, Linux can't talk to it 
until you reboot.


--- In DigitalBlackandWhiteThePrint@yahoogroups.com, Carl Schofield 
<scho@m...> wrote:
> David,
> 
> I've sent you an email with an attachment containing the 2200 
profiles  
> I'm using now, after the dot weighting adjustment that Roy 
mentioned in  
> his post.  You might want to re-linearize the profiles after you 
make  

> the GP dot weighting adjustment.
> 
> Carl

--- In DigitalBlackandWhiteThePrint@yahoogroups.com, "David Wroblewski" 
<dawroblewski@y...> wrote:
> Making the change now... will let you know what happens. 
> 
> Is the dot size issue specific to that one resolution?
> Or are there similar transition points for other resolutions?
> 
> -d
> 

Unfortunately, what dots are appropriate is different for each
different resolution so that changes ink limits, dot weightings and
dot gain.   Kind of a mess, my take is to do everything at 1440x720
unless some pressing need comes along.

Roy

--- In DigitalBlackandWhiteThePrint@yahoogroups.com, "Roy 
Harrington" <roy@h...> wrote:

Okay, I'll stick with 1440x720 for the time being. I can't see
any dots in the highlights, so I'm not sure why I'd
want to drive it any higher anyway.

I've been experimenting with blending Carl's warm/cool curves.
I love it--works like a charm! 

david


> --- In DigitalBlackandWhiteThePrint@yahoogroups.com, "David 
Wroblewski" 

> <dawroblewski@y...> wrote:
> > Making the change now... will let you know what happens. 
> > 
> > Is the dot size issue specific to that one resolution?
> > Or are there similar transition points for other resolutions?
> > 
> > -d
> > 
> 
> Unfortunately, what dots are appropriate is different for each
> different resolution so that changes ink limits, dot weightings and
> dot gain.   Kind of a mess, my take is to do everything at 1440x720
> unless some pressing need comes along.
> 
> Roy

Hi Costa, that's interesting. Coupla questions.

First, how do you set up an image to print a single-ink
tone ramp as you describe? I know this must be basic
but I've never seen it explained in my half dozen photoshop
books. QTR even has a file, inkpattern.psd, that prints 7
different single-ink step wedges, but it uses some CMYK magic
that I just cannot get my head around. Sorry to be the
class dunce.

Second, doesn't your method assume that all the inks have the
same transition point(s)? Is that a safe assumption?

-david

--- In DigitalBlackandWhiteThePrint@yahoogroups.com,
    "ccolbertbw" <ccolbert@u...> wrote:
> I went through the same sort of adjustments for the SC 980.  I found that
> printing a smooth (1%) gradient with a single ink made it much easier to
judge
> whether the dot size values were correct. This is because it is not
transitions
> between inks that are the issue here, but instead the transition from one
dot
> size to the next (within each ink). In other words a gradient with only
one ink
> will show the "hitches" just the same.  After changing the dot  values,
the
> single ink gradient looked smooth and then the full ink set fell right in
line.

>
> Costa Colbert
>

--- In DigitalBlackandWhiteThePrint@yahoogroups.com, "David Wroblewski" 
<dawroblewski@y...> wrote:
> Hi Costa, that's interesting. Coupla questions.
> 
> First, how do you set up an image to print a single-ink
> tone ramp as you describe? I know this must be basic
> but I've never seen it explained in my half dozen photoshop
> books. QTR even has a file, inkpattern.psd, that prints 7
> different single-ink step wedges, but it uses some CMYK magic
> that I just cannot get my head around. Sorry to be the
> class dunce.

David - 
  Remember that this is using QTR and gimp-print. The final files produced by Roy's
calibration procedure just tell the driver how much ink to put out for each pixel value. 
These are just text files, so, it is easy enough to just replace all values with zero for 
all inks except the one of interest. Easier yet, you should also be able to just do the 
calibration pretending you have just one ink.  

I started doing it this way when I could not get a smooth multi-ink ramp. Then  I 
figured out that the bad transitions were at points where only the lightest ink was 
being put out. To test this idea I turned off the other inks. The result was still bumpy. 
Roy helped me figure out which dot value in the gimp-print driver needed 
adjustment. After I got the right value the ramp became very smooth. Of course this 
is just the part of the ramp where the lightest ink is being put out, but smooth as can 
be. Then I added the other inks, linearized, and got a great curve. I think the value of 
using one ink is that you don't get distracted by ink transitions. 

> Second, doesn't your method assume that all the inks have the
> same transition point(s)? Is that a safe assumption?

  The gimp-print software only lets you specify the dot value once. It is not per ink. 
That's why a single value of 0.45 or 0.46 is being mentioned in this thread. I agree it 
seems like the optimal value could be different for each ink.  It is possible that there 
is a different best value for the different inks and that the overall best single value is 
a compromise. However, I think all that is really needed is a decent value where there 
are no obvious discontinuities. The linearization process can then take over and mix 
the inks to give you a smooth ramp. 

hope that helps.

Costa
> 
> -david

Robert's reply on the messages is as follows:

I went through the same sort of adjustments for the SC 980.  I
   found that printing a smooth (1%) gradient with a single ink
made
   it much easier to judge whether the dot size values were
   correct. This is because it is not transitions between inks
that
   are the issue here, but instead the transition from one dot
size to
   the next (within each ink). In other words a gradient with
only one
   ink will show the "hitches" just the same.  After changing the
dot
   values, the single ink gradient looked smooth and then the
full ink
   set fell right in line.

Agreed.  That's how I calibrate printers.


    I did the experiment of printing out a 0.45 version
   of the 21Step.tif stepwedge included in the QTR distribution.

   There's an improvement, but it is extremely subtle, even held
up to
   a very bright light. The 0.45 ramp is just a _little_ smoother
in
   the 50% and 80%-90% range. Thinking that might be natural
variation
   from print to print, I printed it again and got the same
   result. All three stepwedges (0.46, 0.45 #1 and 0.45 #2) were
   printed with the same sheet of paper, with the same carts.

   So 0.45 wins for my setup. But wow it's subtle.


OK, we'll stay with 0.45.

   This is probably within the variability of the printers, the
   measurements and the dithering code.  I re-looked at the data
and
   graphs where I came up with 0.46 and 0.45 looks just as good
(the
   two dotsize curves intersect in slightly different locations).
It
   would be nice if the gimp-print guys had a standard for
measuring
   this because other printers are way off.

I'm working on a technique for doing this.  I know other printers
are
off; if people want to tune them, I'll be happy to accept the
revised
parameters.

-- 
Robert Krawitz
<rlk@...>

Transferred by Ernst

--- "ccolbertbw" <ccolbert@u...> wrote:
>   Remember that this is using QTR and gimp-print. The 
>   final files produced by Roy's calibration procedure 
>   just tell the driver how much ink to put out for each 
>   pixel value. These are just text files, so, it is easy 
>   enough to just replace all values with zero for all 
>   inks except the one of interest. 

Aha. Fiendishly clever. So one could produce profiles called,
"JustC", "JustY", "JustM", "JustK", and so on, which direct 
QTR to use the one ink, then print a grayscale ramp.

I also see your point that running a calibration pretending 
there is only 1 ink is a good quick method.

I need to spend some time doing some basic exercises, such as
setting up a curve file that prints each ink, another that prints 
them all just like the Epson driver, etc, before I understand 
this completely. I know that's somewhat academic--I could just 
start printing with the curves I have now, having corrected the
constant in the code. But I would like to understand how this all
works a little better.

Thanks, that was helpful.

-david

While maybe a bit off topic, I need some feedback.

I just bought a copy of Picture Window Pro 3.5. The drill is you 
download it and as demo and when you buy it they send you a "key" that 
turns the demo copy into a real copy. I also bought the CD. I never 
recieved the key.

When the CD came I thought I would just install it and be up and going, 
but it also requires the "key."  I have contacted them by Email and left 
a message last week. Today I call and the phone rings once and goes 
busy. Has something happened to the company, are they still in business? 
Anyone know how I can get the off dead center?

I've used PWPro 3.1 and liked it. It is more intutive and easier to use 
than Photoshop. But I sure don't like their lack of support. Anyone have 
any suggestions.

Truman

--- In DigitalBlackandWhiteThePrint@yahoogroups.com, "David Wroblewski" 
<dawroblewski@y...> wrote:
> --- "ccolbertbw" <ccolbert@u...> wrote:
> >   Remember that this is using QTR and gimp-print. The 
> >   final files produced by Roy's calibration procedure 
> >   just tell the driver how much ink to put out for each 
> >   pixel value. These are just text files, so, it is easy 
> >   enough to just replace all values with zero for all 
> >   inks except the one of interest. 
> 
> Aha. Fiendishly clever. So one could produce profiles called,
> "JustC", "JustY", "JustM", "JustK", and so on, which direct 
> QTR to use the one ink, then print a grayscale ramp.
> 
> I also see your point that running a calibration pretending 
> there is only 1 ink is a good quick method.
> 
> I need to spend some time doing some basic exercises, such as
> setting up a curve file that prints each ink, another that prints 
> them all just like the Epson driver, etc, before I understand 
> this completely. I know that's somewhat academic--I could just 
> start printing with the curves I have now, having corrected the
> constant in the code. But I would like to understand how this all
> works a little better.
> 
> Thanks, that was helpful.
> 
> -david

David,

The calibration page inkpattern.psd just makes it easy to see the inks
separately.  So it basically gives you the "JustC", "JustY", "JustM", "JustK"
on one page.  The idea of calibration is to express each lighter ink as a
percentage of the darker ink.   Well the dots are exactly analogous,
you want express a smaller dot as a percentage of the larger dot. 
Underlying its really just a volume measurement of the droplet of
ink -- the in this case the small dot is 45% of the volume of the large
dot.  Since we can't actually see or measure the individual dots, the
method I've been using is matching the densities.  I do a patch of
100% coverage using just the small dot, then I see what percent
coverage using just the large dot is necessary to get the same density 
on the paper.  I've carried it a little further by doing a whole step
wedge for each dotsize and trying to fit the whole curve together.

Unfortunately, doing a calibration page for the dots is a little more 
complicated, but all in all its the same principle.  If you interested in
the gory details of this I'd be glad to share it offline.

Roy