Digital BW, The Print

Roy, I was curious if you found the LAB Grayscale profile effective 
and if you updated your original profile?

I'm probably missing something, but it seems this would have to be a 
generic profile by its nature.  If used for viewing the tonal range 
on a monitor, how would it be able to differentiate between a high 
gamut glossy paper (like Kirkland Glossy which can hit an L* of about 
6) to lower gamut matte paper that can only display an L* or about 
20?  

Seems to me they would both display the same, but the output would be 
very different.

If you have found it effective, I'd love to know how you use it and 
how you handle different paper types.  I'd also like to get your 
latest profile if your results are positive.

Thanks, Lou

--- In DigitalBlackandWhiteThePrint@yahoogroups.com, "ldina" <lbdina@c...> wrote:
> 
> Roy, I was curious if you found the LAB Grayscale profile effective 
> and if you updated your original profile?
> 
> I'm probably missing something, but it seems this would have to be a 
> generic profile by its nature.  If used for viewing the tonal range 
> on a monitor, how would it be able to differentiate between a high 
> gamut glossy paper (like Kirkland Glossy which can hit an L* of about 
> 6) to lower gamut matte paper that can only display an L* or about 
> 20?  
> 
> Seems to me they would both display the same, but the output would be 
> very different.
> 
> If you have found it effective, I'd love to know how you use it and 
> how you handle different paper types.  I'd also like to get your 
> latest profile if your results are positive.
> 
> Thanks, Lou

Hi Lou,

I have been experimenting with this.  I find the Lab gray space better than
the gamma space.  At least I find the separations are better matched to the
print and to the eye.  But you are right about how different papers are
enough different that you can't switch back and forth without editing.

My current take is that the gray Lab space is best for editing, but that there
should be a way to do a mapping with perceptual intent rendering at
print time.   I think it's possible to have just a few profiles -- i.e. a
generic matte paper and a generic photo paper -- rather than making
a gazillion profiles for every combination.

Roy

Thanks, Roy. I look forward to hearing about your experiments and 
conclusions.  

Perhaps 3 different spaces would suffice, one for a high Dmax glossy 
paper, one for a low Dmax matte paper, and one in the middle.  Sure 
would make like easier and more predictable.  

Regards, Lou

--- In DigitalBlackandWhiteThePrint@yahoogroups.com, "Roy Harrington" 
<roy@h...> wrote:
> 
> --- In DigitalBlackandWhiteThePrint@yahoogroups.com, "ldina" 
<lbdina@c...> wrote:
> > 
> > Roy, I was curious if you found the LAB Grayscale profile 
effective 
> > and if you updated your original profile?
> > 
> > I'm probably missing something, but it seems this would have to 
be a 
> > generic profile by its nature.  If used for viewing the tonal 
range 
> > on a monitor, how would it be able to differentiate between a 
high 
> > gamut glossy paper (like Kirkland Glossy which can hit an L* of 
about 
> > 6) to lower gamut matte paper that can only display an L* or 
about 
> > 20?  
> > 
> > Seems to me they would both display the same, but the output 
would be 
> > very different.
> > 
> > If you have found it effective, I'd love to know how you use it 
and 
> > how you handle different paper types.  I'd also like to get your 
> > latest profile if your results are positive.
> > 
> > Thanks, Lou
> 
> Hi Lou,
> 
> I have been experimenting with this.  I find the Lab gray space 
better than
> the gamma space.  At least I find the separations are better 
matched to the
> print and to the eye.  But you are right about how different papers 
are
> enough different that you can't switch back and forth without 
editing.
> 
> My current take is that the gray Lab space is best for editing, but 
that there
> should be a way to do a mapping with perceptual intent rendering at
> print time.   I think it's possible to have just a few profiles -- 
i.e. a

> generic matte paper and a generic photo paper -- rather than making
> a gazillion profiles for every combination.
> 
> Roy

Roy Harrington wrote:
> 
> --- In DigitalBlackandWhiteThePrint@yahoogroups.com, "ldina" <lbdina@c...> wrote:
> 
>>Roy, I was curious if you found the LAB Grayscale profile effective 
>>and if you updated your original profile?
>>
>>I'm probably missing something, but it seems this would have to be a 
>>generic profile by its nature.  If used for viewing the tonal range 
>>on a monitor, how would it be able to differentiate between a high 
>>gamut glossy paper (like Kirkland Glossy which can hit an L* of about 
>>6) to lower gamut matte paper that can only display an L* or about 
>>20?  
>>
>>Seems to me they would both display the same, but the output would be 
>>very different.
>>
>>If you have found it effective, I'd love to know how you use it and 
>>how you handle different paper types.  I'd also like to get your 
>>latest profile if your results are positive.
>>
>>Thanks, Lou
> 
> 
> Hi Lou,
> 
> I have been experimenting with this.  I find the Lab gray space better than
> the gamma space.  At least I find the separations are better matched to the
> print and to the eye.  But you are right about how different papers are
> enough different that you can't switch back and forth without editing.
> 
> My current take is that the gray Lab space is best for editing, but that there
> should be a way to do a mapping with perceptual intent rendering at
> print time.   I think it's possible to have just a few profiles -- i.e. a
> generic matte paper and a generic photo paper -- rather than making
> a gazillion profiles for every combination.
> 
> Roy

Roy,

Is what you propose not already done in most RIPs (but in
colour mode) ?  The linearisation of the channels separate of
the profiling, the last to the Lab space/axis. In that way you
could use the LittleCMS or Argyll engine for profiling. The
fastest way now would be to use either Photoshop or Qimage and
a B&W adapted ICC profile. Qimage makes RGB print to file
output, as long as the greyscale isn't corrupted but only the
steps shifted along the curve it would be right. Doesn't
provide the right preview on color (tones) but the rest in an
acceptable way. The linearisation takes care of the paper +
printer + ink differences. The Lab curve goes on top of that
and you could probably limit them to the two choices you
mention, one covering up to 1.7 Dmax, the other to something
like 2.1.

Ernst

--- In DigitalBlackandWhiteThePrint@yahoogroups.com, Ernst Dinkla <E.Dinkla@c...> 
wrote:
> Roy Harrington wrote:
> > 
> > --- In DigitalBlackandWhiteThePrint@yahoogroups.com, "ldina" <lbdina@c...> 
wrote:
> > 
> >>Roy, I was curious if you found the LAB Grayscale profile effective 
> >>and if you updated your original profile?
> >>
> >>I'm probably missing something, but it seems this would have to be a 
> >>generic profile by its nature.  If used for viewing the tonal range 
> >>on a monitor, how would it be able to differentiate between a high 
> >>gamut glossy paper (like Kirkland Glossy which can hit an L* of about 
> >>6) to lower gamut matte paper that can only display an L* or about 
> >>20?  
> >>
> >>Seems to me they would both display the same, but the output would be 
> >>very different.
> >>
> >>If you have found it effective, I'd love to know how you use it and 
> >>how you handle different paper types.  I'd also like to get your 
> >>latest profile if your results are positive.
> >>
> >>Thanks, Lou
> > 
> > 
> > Hi Lou,
> > 
> > I have been experimenting with this.  I find the Lab gray space better than
> > the gamma space.  At least I find the separations are better matched to the
> > print and to the eye.  But you are right about how different papers are
> > enough different that you can't switch back and forth without editing.
> > 
> > My current take is that the gray Lab space is best for editing, but that there
> > should be a way to do a mapping with perceptual intent rendering at
> > print time.   I think it's possible to have just a few profiles -- i.e. a
> > generic matte paper and a generic photo paper -- rather than making
> > a gazillion profiles for every combination.
> > 
> > Roy
> 
> Roy,
> 
> Is what you propose not already done in most RIPs (but in
> colour mode) ?  The linearisation of the channels separate of
> the profiling, the last to the Lab space/axis. In that way you
> could use the LittleCMS or Argyll engine for profiling. The
> fastest way now would be to use either Photoshop or Qimage and
> a B&W adapted ICC profile. Qimage makes RGB print to file
> output, as long as the greyscale isn't corrupted but only the
> steps shifted along the curve it would be right. Doesn't
> provide the right preview on color (tones) but the rest in an
> acceptable way. The linearisation takes care of the paper +
> printer + ink differences. The Lab curve goes on top of that
> and you could probably limit them to the two choices you
> mention, one covering up to 1.7 Dmax, the other to something
> like 2.1.
> 
> Ernst

Hi Ernst,

Yes, that's basically the idea.  The grayscale makes things somewhat
simpler though --  I can do the icc profiles myself.
On the Mac its just like printing with a color icc profile.  
I'm not as sure what the best Windows workflow would be.

With Qimage can you apply the profile as you save the "Print to
File"?   I think with Photoshop you'd have to Convert to Profile
and save the tiff file.

The other approach would be to do the perceptual rendering in
QTR using a simple curve.

Roy

Roy Harrington wrote:

>>Is what you propose not already done in most RIPs (but in
>>colour mode) ?  The linearisation of the channels separate of
>>the profiling, the last to the Lab space/axis. In that way you
>>could use the LittleCMS or Argyll engine for profiling. The
>>fastest way now would be to use either Photoshop or Qimage and
>>a B&W adapted ICC profile. Qimage makes RGB print to file
>>output, as long as the greyscale isn't corrupted but only the
>>steps shifted along the curve it would be right. Doesn't
>>provide the right preview on color (tones) but the rest in an
>>acceptable way. The linearisation takes care of the paper +
>>printer + ink differences. The Lab curve goes on top of that
>>and you could probably limit them to the two choices you
>>mention, one covering up to 1.7 Dmax, the other to something
>>like 2.1.

> Yes, that's basically the idea.  The grayscale makes things somewhat
> simpler though --  I can do the icc profiles myself.
> On the Mac its just like printing with a color icc profile.  
> I'm not as sure what the best Windows workflow would be.
> 
> With Qimage can you apply the profile as you save the "Print to
> File"?   I think with Photoshop you'd have to Convert to Profile
> and save the tiff file.
> 
> The other approach would be to do the perceptual rendering in
> QTR using a simple curve.
> 
> Roy

The best Windows workflow would be a softproof display on 
Stephen's GUI which would show the color tone difference of 
the warm-cold tone slider and the steps shift that is applied 
with another slider for 1.6 <> 2.2 Dmax Lab curve addition. I 
guess one could do this with a small (gamut) colour engine to 
simulate all quad and CM(Y)K greyscale printers :-) The Lab 
curve addition should have an on-off switch to allow the 
workflow below.

If the above is too complex right now then Qimage's print to 
file + the softproof preview will show the effect of an ICC 
profile that uses only the Lab curve. With CM on, Qimage does 
that right now with any RGB printer profile on Greyscale and 
RGB files. On any file in the map you print from you can use 
softproof even before it is loaded in the print queue. 3 
profiles would be sufficient I guess. But it will not be 
correct to use the printer profile also for the (simulated) 
preview of the warm-cold tone setting as it will influence the 
RGB file that is exported to QTR. How much that influence will 
be if the file is converted to greyscale again has to be seen. 
  Of course the monitor profile could be manipulated for that 
but that's against all rules and an extra step + several 
monitor profiles are needed then.

This also is a good argument to add an RGB>Greyscale converter 
on Qimage's output or a filter like that on QTR-GUI's input 
instead of an additional greyscale engine in Qimage. Someone 
should mention that to Mike,
it is low on the list of to do.

With profile editors a "null" profile can be made easily by 
using the scanner target and the same image again as the 
result. With ICC profile editors one can map the colour in all 
renderings to the greyscale axis by setting the saturation 
sliders at total negative. With profile editors that work in 
PS one can add PS curve information to the profile. A long 
route so it is probably better to make an "artificial" 
profile. Don't ask me how.

Ernst