QTR-Quadtone RIP

Roy, we are almost on the same page.

 

I edit in GG2.2 or Adobe 1998 using all 100 L*(or K=100 to K=100 as you wrote) with no consideration of output profile. I am in complete agreement with you; there is no other way to edit. If the darkest pixel is not already at L*0 and the lightest at L*100 after Raw conversion, I make a black point and white point adjustment that places them there. When I’m finished editing the image, it is beautiful in the full dynamic range of the workspace and the screen. 

 

The target adjustments are only applied just prior to printing. 

 

Only after completing the editing stage do I flatten the file and save it as a Master. Back in the 8 bit days I did not flatten the Master preferring instead to apply all adjustments simultaneously, but in a 16 bit file I am not afraid of data loss, since I have 32,768 levels of grey and humans can only perceive 100.  

 

Then I begin the targeting stage. I make groups for every separate output (glossy paper, matte paper or web site). Each group bears the name of its output profile. Inside each group I make a series of target adjustments. In B&W these are essentially, black point target adjustment and reproduction curve target adjustment. The black point target adjustment can be applied blindly; once a paper / printer combination has been tested it will never change. To continue your example of a matte paper with Dmax of L*16 (which as you say is 83% in a workspace with a gamma of 2.2), the black point adjustment (usually levels) uses the output slider to make an L*0 = L*16 adjustment. This compresses the curve. The second target adjustment (which is purely aesthetic) is done using soft proofing on an image to image basis.  I create a temporary duplicate of the file and compare the workspace view to the soft proof view (configured with relative rendering & no BPC) with the black point target adjustment active, (and therefore the simulate black ink off).  Should I think certain forms or contours of the image were particularly disadvantaged by the uniform compression of the curve after the BP adjustment I rectify the curve accordingly. 

 

Then I make a workprint and… for me (and I know this is controversial), I never look back at the screen again, after the first print the only viable reference is the print.  

 

Then, should I want to print the master file on glossy paper I just deactivate the matt paper group and begin a new set of target adjustments for the glossy paper. If the glossy paper is Museo Silver Rag, on the P800 this black point target adjustment is L*0 = L*2, and as you can imagine the target reproduction curve adjustment is minimal if needed at all.

 

Before I discovered QTR I was operating in a “same as source” work flow in B&W. I had to develop a working relationship between ink densities and file data for each paper/ printer combination. My only other option was to print B&W with colour inks and I’m sure everyone on this list understands why I didn’t want to do that. 

 

Eugene

 

 

De : QuadtoneRIP@yahoogroups.com [mailto:QuadtoneRIP@yahoogroups.com] 
Envoyé : August-01-16 9:51 PM
À : QuadtoneRIP@yahoogroups.com
Objet : Re: [QuadtoneRIP] Re: QIDF versus ICC

 

  

Eugene,

 

I think I get what you are saying -- see if this is what you mean.

 

You have a printer/ink/paper combo that goes from L*16 to L*100.

For editing you are using GG2.2 so L*16 is about K=83.

So while editing you are using and seeing K=0 through K=83 thus

seeing specifically the L* range of the paper.  Then when you print

you expect L=100 to map to white paper (no inks) and K=83 to

map to the L=16 dMax -- and probably all the L 's in the middle match.

Sounds kind of nice in general -- but I guess I've never heard anyone

doing it this way.  I think people in general use generic workspaces

like GG2.2 -- edit as they like on the screen.  Then that's a "master"

image file that you can print to matte paper, photo paper or web sites.

The mappings are always K=100 to K=100 even if K=100 on the

paper is L=16 or L=4.  ICC CM is the tool that does this mapping in

a way that hopefully does a "best" job given that compression.

 

In your way you give up 17% of the range in GG2.2 -- 8-bit that would

be an issue but with 16-bit its probably not serious.  (there still are 8-bit

places in workflow though).   More of an issue though is that now you

have to have separate masters for different targets -- i.e. matte master

and photo master, and maybe for different individual papers.

Essentially you are editing in a "print" space rather that a generic

space that you can convert as needed for any print space.  I see the

whole ICC profiling purpose is to get away from that.

 

This is my impression about how most people do things.  I'm certainly

more in the B&W world than color, and I don't do ad's that have to

match specific colors, nor try to make exact L-value targets or something.

Of course, I don't profess any absolute of how you should do a workflow.

 

Roy

 

On Sun, Jul 31, 2016 at 7:58 AM, Le Mois de la Photo à Uqbar  <mailto:info@...> info@... [QuadtoneRIP] < <mailto:QuadtoneRIP@yahoogroups.com> QuadtoneRIP@yahoogroups.com> wrote:

Roy,



“You need to try it.  I don't know what you are really trying to accomplish”.



This surprises me , it makes me feel like I’m from another planet. But maybe I am in some respects  I’m getting the feeling that my perception of the  ICC workflow is actually a completely different school of thought to the  QTR workflow .

In the ICC workflow that I am familiar with the media type settings ( ink level curves) do not remap file data to maximum black, colour management does.



So in that light this is what I mean to accomplish is



Obviously The L*0 to ( roughly  L*3 -8 for glossy papers or L*17- L*22 for Matte papers)  values of the image file have to be moved to the maximum ink level of the paper /printer combination. In the dark room we made test strips to determine precisely where Dmax yielded to the first visible tone. In the ICC work flow I only know of 3 ways to accomplish this,  either;  a conversion using perceptual rendering to the output profile, a conversion using relative rendering in conjunction with Adobe Black Point Compensation or a conversion using relative with no BPC where the black point has been manually targeted beforehand. Perceptual rendering and BPC are like automatic modes. I have never fully understood the behaviour of BPC especially where colour data is concerned, and since I’m a control freak, I prefer to do it myself.



When I print a chart of the 100 L* values in an ICC workflow with relative intent and no BPC I can verify ( under lighting conditions the final print will be displayed in)  exactly where DMAX is lost and the length of the toe,( the distance between the first appearance of visible line and a usable threshold density). This allows me to see which paper/ printer combinations take as much as 5 L* to move from the first visible line to usable threshold density, and which paper/ printer combinations can make clean and clear tonal distinction rapidly.  The elasticity of this dmax to usefull threshold zone allows me to place whatever deep shadow detail I do not want to be printed with anything less than useful threshold density, precisely there.  Then once I’ve compressed the curve by  targeting the black point and threshold I can use a target curve with soft proofing to realign the  L* axis should I feel the uniform compression fused contours, or simply to readjust local contrast,  especially if  vital vibrations from tonal juxtapositions were diminished by the uniform compression.



As a said previously I have the feeling the QTR workflow attempts to do this with ink level curves.



I can’t believe that this approach is alien to you I have to believe I’m not expressing myself properly. I hope this detailed description helps.



Eugene

 

 

-- 

Roy Harrington
roy@...
www.harrington.com





[Non-text portions of this message have been removed]