QTR-Quadtone RIP

Greetings, all.  I've been experimenting with QTR for my Epson 4000
with OEM (UC, K2) inks.  I haven't (yet) acquired a densitometer or
built custom profiles, so I'm using profiles that came with QTR.

I'm comparing an image run (nearly) directly into QTR with a version
run through Qimage into QTR.  The Qimage print appears darker in the
darkest 10% (based on 21-step and dark-shadow-detail image regions). 
This was discussed in this forum in January 2005 but apparently not
resolved.

Any new insights?  I like Qimage's resizing and print sharpening, and
I may be able to regain QTR's native ability by fiddling with ink
limit and gamma... but I'd rather be sure first that I'm not
experiencing pilot error.

Thanks in advance for your consideration.

Test environment:

   - Windows XP/SP2
   - Photoshop CS2
   - QTR v2.4.2, Quad400, 2880 dpi, matte ink, uni-d, ordered dither
   - QTR UC-EEnhMatte-cool (100%) UC-EEnhMatte-warm
   - QTR ink limit = gamma = 0
   - Qimage Studio Edition v2007.113
   - Qimage 360 ppi, Hybrid interpolation, Sharpen 5 (default)
   - Epson Enhanced Matte paper
   - Northlight Images B&W Test Image (8-bit JPG, Gray Gamma 2.2)

Comparison tests:

1.  In CS2, resize JPG to 7.07 x 10.00 inches (300 ppi).  Save as
8-bit TIF with embedded Gray Gamma 2.2.  Print with QTR.

2.  In Qimage, specify page of 7.5 x 10.00 inches (360 ppi), then
print JPG (same embedded Gray Gamma 2.2) to TIF with "QTR - RGB Matte
Paper (P, BPC)".  Verify embedded QTR profile with CS2.  Print with QTR.

I somehow suspect that a key question is whether QTR pays any
attention at all to the embedded profile in the incoming TIF.  Roy?

If it does not, then am I simply seeing small differences between the
original JPG/TIF (tagged with Gray Gamma 2.2) and Qimage's TIF that it
is color managing to (and tagging with) "QTR - RGB Matte Paper (P, BPC)"?

If so, then how should I adjust things in order to soft proof accurately?

Michael Miller
Atlanta, GA
(milogiacomorambaldi47 on Yahoo)

--- In QuadtoneRIP@yahoogroups.com, "milogiacomorambaldi47" 
<milogiacomorambaldi47@...> wrote:
>
> Greetings, all.  I've been experimenting with QTR for my Epson 4000
> with OEM (UC, K2) inks.  I haven't (yet) acquired a densitometer or
> built custom profiles, so I'm using profiles that came with QTR.
> 
> I'm comparing an image run (nearly) directly into QTR with a version
> run through Qimage into QTR.  The Qimage print appears darker in the
> darkest 10% (based on 21-step and dark-shadow-detail image regions). 
> This was discussed in this forum in January 2005 but apparently not
> resolved.
> 
> Any new insights?  I like Qimage's resizing and print sharpening, and
> I may be able to regain QTR's native ability by fiddling with ink
> limit and gamma... but I'd rather be sure first that I'm not
> experiencing pilot error.
> 
> Thanks in advance for your consideration.
> 
> Test environment:
> 
>    - Windows XP/SP2
>    - Photoshop CS2
>    - QTR v2.4.2, Quad400, 2880 dpi, matte ink, uni-d, ordered dither
>    - QTR UC-EEnhMatte-cool (100%) UC-EEnhMatte-warm
>    - QTR ink limit = gamma = 0
>    - Qimage Studio Edition v2007.113
>    - Qimage 360 ppi, Hybrid interpolation, Sharpen 5 (default)
>    - Epson Enhanced Matte paper
>    - Northlight Images B&W Test Image (8-bit JPG, Gray Gamma 2.2)
> 
> Comparison tests:
> 
> 1.  In CS2, resize JPG to 7.07 x 10.00 inches (300 ppi).  Save as
> 8-bit TIF with embedded Gray Gamma 2.2.  Print with QTR.
> 
> 2.  In Qimage, specify page of 7.5 x 10.00 inches (360 ppi), then
> print JPG (same embedded Gray Gamma 2.2) to TIF with "QTR - RGB Matte
> Paper (P, BPC)".  Verify embedded QTR profile with CS2.  Print with QTR.
> 
> I somehow suspect that a key question is whether QTR pays any
> attention at all to the embedded profile in the incoming TIF.  Roy?
> 
> If it does not, then am I simply seeing small differences between the
> original JPG/TIF (tagged with Gray Gamma 2.2) and Qimage's TIF that it
> is color managing to (and tagging with) "QTR - RGB Matte Paper (P, BPC)"?
> 
> If so, then how should I adjust things in order to soft proof accurately?
> 
> Michael Miller
> Atlanta, GA
> (milogiacomorambaldi47 on Yahoo)
>

Hi Michael,

The first comment I have is that most of the profiles were made with 1440 dpi not
2880dpi.   This can effect the overall density of the print.  You can use 2880 but
you'll probably have to adjust the Ink Limit Adj till you get it right.

QTR does not look at the embedded profile in the tiff file -- whether it comes from
PS or Qimage.  It just uses whatever the actual gray values are.  However when you
select QTR - RGB Matte Paper in Qimage this is a conversion from the source profile
GG 2.2 to the print profile.  This means all the values are re-calculated.   The main
area of difference is the shadows so that explains why your two prints are different.

I would think that the best screen match would be to use 1440dpi (or super) and
print using Qimage and QTR - RGB Matte Paper printing profile.   There's not much
difference with soft-proofing since it's just gray.   PS does allow a soft-proof with
Simulate Ink Black which will show the lower dMax if you use the Matte Paper profile.

Roy

Roy,

At last this makes sense.  To verify, I ran the original JPG with
embedded Gray Gamma 2.2 through Qimage with "printer ICC" OFF, which
theoretically should pass the GG2.2 pixels to QTR, but resized and
sharpened by Qimage (the whole point of using Qi).

Sure enough.  That final print matches the original QTR print of the
TIF made by PS from the JPG with embedded GG2.2 (Qimage out of the
loop entirely)!

One other point seems worth raising.  Your gray-readme suggests
editing in gray-lab, then printing from PS with gray-matte or
gray-photo.  Then you say that "in Qimage only RGB profiles can be
used".  Well, Qimage can certainly INPUT gray profiles (it was happy
to take my test image in GG2.2, for example).  It's just unwilling to
OUTPUT grayscale.

Hence, it would seem possible to edit and save from PS in gray-lab,
input that into Qimage (which will quietly convert to RGB), output
from Qimage to rgb-matte or rgb-photo, and then print with QTR.

I haven't tried that yet, but can you see any reason why that would
print differently than converting into rgb-matte/photo in PS before
running Qimage?  In other words, it shouldn't matter whether PS or
Qimage converts from gray-lab to rgb-matte/photo, right?

Finally, I'll try printing at 1440 and 1440/super to see if they
match rgb-matte/photo soft proofing in PS better than 2880.

I am so impressed with QTR, your responsiveness, and this forum 24
hours after learning about QTR.  I'm paying my shareware fee tonight!

Michael Miller
 
> Hi Michael,
> 
> The first comment I have is that most of the profiles were made with
1440 dpi not
> 2880dpi.   This can effect the overall density of the print.  You
can use 2880 but
> you'll probably have to adjust the Ink Limit Adj till you get it right.
> 
> QTR does not look at the embedded profile in the tiff file --
whether it comes from
> PS or Qimage.  It just uses whatever the actual gray values are. 
However when you
> select QTR - RGB Matte Paper in Qimage this is a conversion from the
source profile
> GG 2.2 to the print profile.  This means all the values are
re-calculated.   The main
> area of difference is the shadows so that explains why your two
prints are different.
> 
> I would think that the best screen match would be to use 1440dpi (or
super) and
> print using Qimage and QTR - RGB Matte Paper printing profile.  
There's not much
> difference with soft-proofing since it's just gray.   PS does allow
a soft-proof with
> Simulate Ink Black which will show the lower dMax if you use the
Matte Paper profile.

> 
> Roy
>

I've been wondering about these same issues. Right now I have my 
grayscale working space set to "QTR - Gray Lab" in Photoshop. For 
large prints where I want to maintain absolute control, I will resize 
and sharpen in PS, then save to TIFF with that profile embedded and 
then print from QTR. 

But sometimes I would like to be able to take advantage of the 
interpolation/layout/sharpening in QImage for convenience with 
smaller or less critical prints, by printing to file to create the 
TIFF for QTR. I guess I have the following options:

1) Set QImage to use qtr-rgb-lab.icc as the printer ICC.
2) Set QImage to use qtr-rgb-matte.icc (or qtr-rgb-photo.icc)
3) Set QImage to use the rgb version of a profile created from my 
custom curve and Eye One Photo.

Am I correct in thinking that option (1) will give identical results 
to printing the original file directly in QTR? Would one of the other 
two options give me better results? Also, what exactly is the 
difference between the matte/photo profiles? After all if the paper 
curves have been tweaked for good linearity I wouldn't think you need 
additional compensation in the profile for matte papers.
 
> QTR does not look at the embedded profile in the tiff file -- 
whether it comes from
> PS or Qimage.  It just uses whatever the actual gray values are.  
However when you
> select QTR - RGB Matte Paper in Qimage this is a conversion from 
the source profile
> GG 2.2 to the print profile.  This means all the values are re-
calculated.   The main
> area of difference is the shadows so that explains why your two 
prints are different.
> 
> I would think that the best screen match would be to use 1440dpi 
(or super) and
> print using Qimage and QTR - RGB Matte Paper printing profile.   
There's not much
> difference with soft-proofing since it's just gray.   PS does allow 
a soft-proof with
> Simulate Ink Black which will show the lower dMax if you use the 
Matte Paper profile.

> 
> Roy
>

--- In QuadtoneRIP@yahoogroups.com, "jkohn_home" <jkohn@...> wrote:

> But sometimes I would like to be able to take advantage of the 
> interpolation/layout/sharpening in QImage for convenience with 
> smaller or less critical prints, by printing to file to create the 
> TIFF for QTR. I guess I have the following options:
> 
> 1) Set QImage to use qtr-rgb-lab.icc as the printer ICC.
> 2) Set QImage to use qtr-rgb-matte.icc (or qtr-rgb-photo.icc)
> 3) Set QImage to use the rgb version of a profile created from my 
> custom curve and Eye One Photo.

I guess another option I should have mentioned is just disablig Printer 
ICC in QImage, which should work fine for images that I've already 
converted to qtr-rgb-lab.icc, but sometimes I'll be wanting to print 
RGB images without having to take the extra step of converting them to 
grayscale in Photoshop.

--- In QuadtoneRIP@yahoogroups.com, "milogiacomorambaldi47" 
<milogiacomorambaldi47@...> wrote:
>
> Roy,
> 
> At last this makes sense.  To verify, I ran the original JPG with
> embedded Gray Gamma 2.2 through Qimage with "printer ICC" OFF, which
> theoretically should pass the GG2.2 pixels to QTR, but resized and
> sharpened by Qimage (the whole point of using Qi).
> 
> Sure enough.  That final print matches the original QTR print of the
> TIF made by PS from the JPG with embedded GG2.2 (Qimage out of the
> loop entirely)!
> 
> One other point seems worth raising.  Your gray-readme suggests
> editing in gray-lab, then printing from PS with gray-matte or
> gray-photo.  Then you say that "in Qimage only RGB profiles can be
> used".  Well, Qimage can certainly INPUT gray profiles (it was happy
> to take my test image in GG2.2, for example).  It's just unwilling to
> OUTPUT grayscale.

If you use the printing profiles the editing space is not critical.  I still use the
gray-lab for my scanned negatives since there is no "inherent" space but 
GG2.2 is a perfectly good working space too.  If you shoot digitally and the 
camera produces AdobeRGB color files GG2.2 is a good match.

Qimage only handles RGB files internally so gray files are converted to RGB
on input.  RGB versions of my ICC profiles are really still grayscale profiles
with R=G=B so that Qimage can use them.   BTW, Adobe InDesign is another
major product that does not handle grayscale -- just RGB and CMYK.

> 
> Hence, it would seem possible to edit and save from PS in gray-lab,
> input that into Qimage (which will quietly convert to RGB), output
> from Qimage to rgb-matte or rgb-photo, and then print with QTR.

Exactly right.  Conversion can be done in PS or Qimage.

> 
> I haven't tried that yet, but can you see any reason why that would
> print differently than converting into rgb-matte/photo in PS before
> running Qimage?  In other words, it shouldn't matter whether PS or
> Qimage converts from gray-lab to rgb-matte/photo, right?
> 
> Finally, I'll try printing at 1440 and 1440/super to see if they
> match rgb-matte/photo soft proofing in PS better than 2880.
> 
> I am so impressed with QTR, your responsiveness, and this forum 24
> hours after learning about QTR.  I'm paying my shareware fee tonight!

Thanks,
Roy

> 
> Michael Miller
>  
> > Hi Michael,
> > 
> > The first comment I have is that most of the profiles were made with
> 1440 dpi not
> > 2880dpi.   This can effect the overall density of the print.  You
> can use 2880 but
> > you'll probably have to adjust the Ink Limit Adj till you get it right.
> > 
> > QTR does not look at the embedded profile in the tiff file --
> whether it comes from
> > PS or Qimage.  It just uses whatever the actual gray values are. 
> However when you
> > select QTR - RGB Matte Paper in Qimage this is a conversion from the
> source profile
> > GG 2.2 to the print profile.  This means all the values are
> re-calculated.   The main
> > area of difference is the shadows so that explains why your two
> prints are different.
> > 
> > I would think that the best screen match would be to use 1440dpi (or
> super) and
> > print using Qimage and QTR - RGB Matte Paper printing profile.  
> There's not much
> > difference with soft-proofing since it's just gray.   PS does allow
> a soft-proof with
> > Simulate Ink Black which will show the lower dMax if you use the
> Matte Paper profile.
> > 
> > Roy
> >
>

--- In QuadtoneRIP@yahoogroups.com, "Roy Harrington" <roy@...> wrote:
> 
> If you use the printing profiles the editing space is not critical.
 I still use the gray-lab for my scanned negatives since there is no
"inherent" space but GG2.2 is a perfectly good working space too.  If
you shoot digitally and the camera produces AdobeRGB color files GG2.2
is a good match.
> 
> Qimage only handles RGB files internally so gray files are converted
to RGB on input.  RGB versions of my ICC profiles are really still
grayscale profiles with R=G=B so that Qimage can use them.   BTW,
Adobe InDesign is another major product that does not handle grayscale
-- just RGB and CMYK.

To recap, running the highly tweaked GG2.2 test image (8-bit JPG)
straight to QTR works great.  Running the GG2.2 image through Qimage
(converting to QTR-RGB-Matte) to QTR blocks up the shadows.  Running
the GG2.2 image through Qimage (with no ICC conversion) to QTR works
great.

So... trying to move toward editing with a standard working space of
QTR-GrayLab and printing with QTR-RGB-Matte, I loaded the GG2.2 image
into PS, *converted* (not assigned) it to QTR-GrayLab, and saved it. 
Running that through Qimage (converting to QTR-RGB-Matte) to QTR
blocks up the shadows again!

What's happening?  Is it simply not possible accurately to convert
from GG2.2 to QTR-GrayLab on a highly tweaked 8-bit image?  The
histogram certainly changed shape somewhat.  Might this have worked if
the GG2.2 image were 16-bit?

If not, then I'm still confused.  Otherwise, then perhaps there's hope
for my previous workflow, which begins with Canon RAW files, then
Adobe Camera Raw, then 16-bit ProPhotoRGB, etc.  I'm hoping that
converting to QTR-GrayLab early in the process, editing entirely in
16-bit, converting to 8-bit only when all is ready, then converting to
QTR-RGB-Matte with Qimage, and finally printing with QTR... will do
the trick.  Sound plausible?

Michael

milogiacomorambaldi47 wrote:

> 
> If not, then I'm still confused.  Otherwise, then perhaps there's hope
> for my previous workflow, which begins with Canon RAW files, then
> Adobe Camera Raw, then 16-bit ProPhotoRGB, etc.  I'm hoping that
> converting to QTR-GrayLab early in the process, editing entirely in
> 16-bit, converting to 8-bit only when all is ready, then converting to
> QTR-RGB-Matte with Qimage, and finally printing with QTR... will do
> the trick.  Sound plausible?
> 
> Michael

If I recall it correctly Roy addressed the QTR profile 
compatibility of Qimage again when QTR's profiling application 
became part of QTR. However the process I found the most 
reliable (over a longer period) is in converting the Greyscale 
image to QTR-gray-matte profile (or another QTR based gray 
profile) in Photoshop, make it 8 bit, and load the Tiff in 
Qimage for further processing without color management on. It 
doesn't influence the extrapolation and sharpening functions 
of Qimage and keeps the file small till it drops in Qimage. In 
the past I have seen some shadow issues with QTR profiles in 
Qimage which probably had to do with the BPC interpretation in 
Qimage. The last has been changed later on in Qimage but 
Qimage is also a moving target with all its upgrades (a good 
thing) that I'm not willing to test every time its CM 
compatibility with QTR profiles. In theory it should work though.
One feature could tempt me to use Qimage's CM with quad prints 
again and that is the profiling per image on more images per 
print page. I've made some profiles for gloss varnished canvas 
and matte varnished canvas. Both can then be printed on the 
same canvas sheet but individually get the right profile for 
the varnish type used per print.

Ernst

-- 

                    --
           Ernst Dinkla


www.pigment-print.com
(         unvollendet         )

On Friday, August 11, 2006, at 04:40  AM, milogiacomorambaldi47 wrote:

> --- In QuadtoneRIP@yahoogroups.com, "Roy Harrington" <roy@...> wrote:
>>
>> If you use the printing profiles the editing space is not critical.
>  I still use the gray-lab for my scanned negatives since there is no
> "inherent" space but GG2.2 is a perfectly good working space too.  If
> you shoot digitally and the camera produces AdobeRGB color files GG2.2
> is a good match.
>>
>> Qimage only handles RGB files internally so gray files are converted
> to RGB on input.  RGB versions of my ICC profiles are really still
> grayscale profiles with R=G=B so that Qimage can use them.   BTW,
> Adobe InDesign is another major product that does not handle grayscale
> -- just RGB and CMYK.
>
> To recap, running the highly tweaked GG2.2 test image (8-bit JPG)
> straight to QTR works great.  Running the GG2.2 image through Qimage
> (converting to QTR-RGB-Matte) to QTR blocks up the shadows.  Running
> the GG2.2 image through Qimage (with no ICC conversion) to QTR works
> great.
>
> So... trying to move toward editing with a standard working space of
> QTR-GrayLab and printing with QTR-RGB-Matte, I loaded the GG2.2 image
> into PS, *converted* (not assigned) it to QTR-GrayLab, and saved it.
> Running that through Qimage (converting to QTR-RGB-Matte) to QTR
> blocks up the shadows again!

When you have multiple conversions, it's only the last one that makes 
any difference.
So:
GG2.2  ->  QTR Matte Paper
and
GG2.2  ->  GrayLab -> QTR Matte Paper
will yield the same print results.  L* is maintained during the 
conversions.
(the last conversion is a little different because of 
BlackPointCompensation)

It sounds like in your system, you can "highly tweak" based on your 
display and
get matching prints with just GG2.2.    Matching prints is the goal so 
I'd stop there.
If at some time you custom calibrated your whole system, maybe you'd 
get even
better matching.

>
> What's happening?  Is it simply not possible accurately to convert
> from GG2.2 to QTR-GrayLab on a highly tweaked 8-bit image?

The conversion should be fine -- the display didn't change did it?

> The histogram certainly changed shape somewhat.  Might this have 
> worked if
> the GG2.2 image were 16-bit?

The visible differences would be very small.  16-bit is good for 
editing since
you don't build up errors.  But when you are all done, 8 bit is good 
enough as
long as the conversion is done well -- (Photoshop does a good job of 
it).

>
> If not, then I'm still confused.  Otherwise, then perhaps there's hope
> for my previous workflow, which begins with Canon RAW files, then
> Adobe Camera Raw, then 16-bit ProPhotoRGB, etc.  I'm hoping that
> converting to QTR-GrayLab early in the process, editing entirely in
> 16-bit, converting to 8-bit only when all is ready, then converting to
> QTR-RGB-Matte with Qimage, and finally printing with QTR... will do
> the trick.  Sound plausible?

Since you report good prints with GG2.2, just do the final conversion to
GG2.2.   It doesn't matter what profile you use in the interim, 
especially
since it sounds like you'll be in 16-bit mode.   You could go from Raw
directly to GG2.2 and stay there from then on.   BTW,  I'm not very 
familiar
with the ProPhotoRGB but I did look it up and it happens to be a gamma 
1.8
space.

Roy

>
> Michael
>
>
-
Roy Harrington
roy@...
Black & White Photo Gallery
http://www.harrington.com