Digital BW, The Print

Hi Roy


> From: Roy Harrington <roy@...>


 
> The new space does several different things.   For printing through QTR
> there is no effect.  Remember in the Print with Preview we select Same
> as Source for the print.   We still have a grayscale image so this means
> the exact pixel values are sent to QTR.

Yes and there is no colour management in the RIP.

> For display though -- what you see on the screen -- there's still a bunch
> of color management going on.  The pixels go from the gray space (now
> just L) to Lab to the RGB of the monitor.  Before they went from gamma 2.2
> to Lab to RGB.   

Yes

>So the new way we see the difference between 95 and 100
> on the screen as well as on the print.
> 
> The bottom line is that you should Assign to the new space NOT Convert to it.
> 

I am still not convinced of this but let me think about it.  I agree that
the file that gets sent to QTR (via Same as Source) needs to be LAB or
LAB-grey.  But if I took a photograph such that a portion of the subject is
Kodak middle grey (L=50) and this successfully makes it through to my B&W
image on screen (ie is still L=50) before going to LAB-grey, if I ASSIGN
LAB-grey the L value of this portion of the image will change.

(As an example, pull up a grey gamma 2.2 step wedge.  The 50% step is L=54.
Assign LAB-Grey and the 50% step changes colour and becomes L=50.  What was
L=50 also moves to something less than that.  My image has changed and I
have lost my reference point that I noted at the time of exposure.)


>> 
>> (I am very intrigued as to how you did this by the way.)
> 
> With mirrors of course.  Actually I tried a few things to find out what
> was in the icc file for a gray space and then matched it to the known
> formulas.  
> Now I have to go back and have it all calculated.
> 

You are a genius.

>> 
>> 2.  Now that we have our image in LAB-lite we still have the issue of this
>> space not being the same as the printer space.  For starters, an image using
>> the full tonal range (such as an ordinary step wedge) has deeper blacks than
>> we can print and hence we don't have WYSIWYG.)  So no "built-in" proofing
>> :-(  See my last post to Keith in the other thread which I will reproduce
>> here:
>> 
>> "Even if the printer RIP automatically spaces LAB values from
>> dMin to dMax, ie 0 gets mapped to 16, 5 to 20 etc, you still end up with the
>> same result: the mid-point shifts.
>> 
>> (By the way, take a look at this sequence of co-ordinates, and their density
>> equivalents, and look how all print values are shifted.  Plot the step vs
>> density figures of the two papers and overlay LAB.  Even if the RIP doesn't
>> make this linear, as I believe was suggested, but curved in equal increments
>> of LAB, each paper has a different gamma and non have the same value or
>> curvature as LAB. Only as paper white moves closer to perfect white and ink
>> dMax approaches perfect black do the curves begin to converge and have the
>> same gamma.)"
> 
> You are right that absolute values will vary based on dMin, dMax, the kind
> of display you use, the lighting etc...  I lean toward the everything relative
> point of view though.   With color management there are the different
> renderings and Perceptual is usually the one of choice.   This corresponds to
> the compressing the gamut mode.  The Absolute Colormetric is more like
> your view but seems rarely used.
> 

Not quite.  In my methodology there has been no "compression of gamut" per
se at the point of "calibrating" the printer - merely a recognition of what
the range is.  This is important.  I make no judgement call at this point as
to how out-of-gamut values should be treated.  I recognise that the printer
can print a good portion of the LAB-grey space and I simply require that it
do so for the portion it can and I note down the portion that it can't.

With the recognition that the printer can only render say that portion of
the LAB-grey space defined by 16<=L=>96, I then look at my image.  How the
image is then compressed with a simple curve to 16<=L=>96 has an infinite
number of possibilities - all of which I control simply with one neat PS
curve.  It is actually better than colorsync in that you have an infinite
number of possible curves rather than 4 options.  For example, if I chose to
clip certain values I am in effect using a mix of Absolute Colormetric and
Perceptual.

OK you could say that I make the call at the time of calibration that IF I
don't do anything else, out-of-gamut values will be clipped to dMax or dMin
respectively!  But you gotta take the package in the same way I must take
the current package of a transformation function embedded in the RIP and the
transformation function I must manually do during soft proofing to get the
image to "look right" again.  If I do nothing with the current setup I don't
get a satisfactory image - I have to go back and rejigger it with a
softproof in hand.  Worse yet, if my image is fully in gamut (eg my
photography class example) at the moment I still have to go back and
rejigger it to get the image right.

Cheers

Steve