Digital BW, The Print

Just interested to know what value you guys take to be absolute mid 
grey. The good old 21 step wedge in greyscale mode gives a midpoint 
value of 50% black. No surprises there but convert it to RGB values and 
it is 148,148,148 and not 128 (the midpoint between 0 and 255).
OK I know I'm simple so it should take a simple answer.

Pete.

50% K varies in colour according to workspace.  That's gamma.

> From: petexp2 <kafoozalem@...>

> Just interested to know what value you guys take to be absolute mid
> grey. The good old 21 step wedge in greyscale mode gives a midpoint
> value of 50% black. No surprises there but convert it to RGB values and
> it is 148,148,148 and not 128 (the midpoint between 0 and 255).
> OK I know I'm simple so it should take a simple answer.
> 
> Pete.
>

> ... interested to know what value you guys take to be absolute mid grey.


The "absolute" midpoint, as Steve has said, is not absolute, but only
relative to the model used by the workspace setting in your system -- of
which there are several.

I think of the 2 main variables in the approaches as (1) the midpoint
luminance and (2) the shadow compression.

Many of the models nominally put the 50% point in the "middle" of the scale.

Lab Luminance (*L), as I understand it, starts at 0% light being reflected
-- absolute black.  While Kodak used L=50 for its "gray card" midpoint, that
turns out to be rather far from the effective or perceptual midpoint in our
actual prints.  Our matte prints have a black point that is considerably
above the Lab L=0.  So, the apparent visual midpoint will be lighter than
Lab L=50.
  

As a practical matter, many of the B&W approaches, like mine, have used the
actual dmax of a good matter paper and placed the midpoint half way to the
paper white.  This approach results in one of the lighter midpoints due to
the light dmax.  I get close to this point with the 2400 Advanced B&W mode.
This approach is not "color managed" as such.

Since most of the industry and equipment is focused on getting good colors,
there may be significant value in tapping into the popular and
well-supported color spaces.  For color spaces a number of factors seem to
have influenced how they defined their "centers" and contoured the shadows
(& ramps).  It may be easiest to just go with the main markets and
standards.

I use Adobe Photoshop Elements 2 as a proxy for the color world spaces.  PS
Elements is often given away with printers, the program has the most
important image editing powers, and it is the most consistent with the
industry-dominating (and expensive) Photoshop CS2.  

Elements 2 has three color space options.  When one clicks on Elements'
Edit>Color Settings, the following 3 options appear:  
(1) "No color management," 
(2) "Limited color management -- optimized for Web graphics," and 
(3) "Full color management -- optimized for Print."  

In the past I mostly used the "No color management" route.  The the 2400 ABW
k3 driver mode seems to also take this route.  

However there may be advantages to going with one of the popular "Color
Management" ("CM") routes.   

The first, "Limited color management" option in Elements results in images
that display fairly appropriately when they are placed on the web.  The
shadows are relatively compressed.  In Elements' "Print Preview," the "Print
Space Profile" is shown as "Working Gray -- Gray Gamma 2.2."

The second Elements' Edit>Color Settings CM option is what Adobe here calls,
"Full color management."  This setting is said to be, "optimized for Print."
It has relatively open shadows and a long grayscale and is close to what I
used, in part because I was coming from the PiezoBW starting point.  This
option results in a Print Space Profile called, "Working Gray -- Dot Gain
20%."  

You can switch between the options and see the monitor views change.  

For the older, non-CM approach, we often used "Custom Dot Gain" curves to
make the monitor view match the print view better.  That rather complex
manual procedure still result in the most accurate profiling of the monitor
for B&W, but the procedure is only in full Photoshop, and it's complex.
Realistically, most will want the convenience of an automatic monitor view
match that the Color Management approaches allow.  



> ...  so it should take a simple answer.

Well ... switching between Dot Gain 20% and Gray Gamma 2.2 with a 21-step
test file on the monitor is a good way to easily see the differences.  

I find it useful to accept industry standards and those I'm accustomed to. 

Adobe Photoshop Elements' options seem to be relatively clear and simple. 

Paul
www.PaulRoark.com

Thank you Paul for such an informative answer. I really did think it 
would be simpler but you have helped me answer it to my own 
satifaction. In fact playing around with a 21 step wedge in Photoshop 
I note that the when you convert it from gamma 2.2 grayscale to RGB 
it has a midtone value of 128. Alternatively when you convert it from 
dot gain 20% grayscale to RGB the mid point value is 148 RGB. I will 
certainly favour dot gain 20% as a starting point for making prints 
and bear in mind what you say about the compressed tonality of a 
matte paper.

Pete.

--- In DigitalBlackandWhiteThePrint@yahoogroups.com, "Paul Roark" 
<paul.roark@v...> wrote:
>
>  
> > ... interested to know what value you guys take to be absolute 
mid grey.
> 
> 
> The "absolute" midpoint, as Steve has said, is not absolute, but 
only
> relative to the model used by the workspace setting in your system -
- of
> which there are several.
> 
> I think of the 2 main variables in the approaches as (1) the 
midpoint
> luminance and (2) the shadow compression.
> 
> Many of the models nominally put the 50% point in the "middle" of 
the scale.
> 
> Lab Luminance (*L), as I understand it, starts at 0% light being 
reflected
> -- absolute black.  While Kodak used L=50 for its "gray card" 
midpoint, that
> turns out to be rather far from the effective or perceptual 
midpoint in our
> actual prints.  Our matte prints have a black point that is 
considerably
> above the Lab L=0.  So, the apparent visual midpoint will be 
lighter than
> Lab L=50.
>   
> 
> As a practical matter, many of the B&W approaches, like mine, have 
used the
> actual dmax of a good matter paper and placed the midpoint half way 
to the
> paper white.  This approach results in one of the lighter midpoints 
due to
> the light dmax.  I get close to this point with the 2400 Advanced 
B&W mode.
> This approach is not "color managed" as such.
> 
> Since most of the industry and equipment is focused on getting good 
colors,
> there may be significant value in tapping into the popular and
> well-supported color spaces.  For color spaces a number of factors 
seem to
> have influenced how they defined their "centers" and contoured the 
shadows
> (& ramps).  It may be easiest to just go with the main markets and
> standards.
> 
> I use Adobe Photoshop Elements 2 as a proxy for the color world 
spaces.  PS
> Elements is often given away with printers, the program has the most
> important image editing powers, and it is the most consistent with 
the
> industry-dominating (and expensive) Photoshop CS2.  
> 
> Elements 2 has three color space options.  When one clicks on 
Elements'
> Edit>Color Settings, the following 3 options appear:  
> (1) "No color management," 
> (2) "Limited color management -- optimized for Web graphics," and 
> (3) "Full color management -- optimized for Print."  
> 
> In the past I mostly used the "No color management" route.  The the 
2400 ABW
> k3 driver mode seems to also take this route.  
> 
> However there may be advantages to going with one of the 
popular "Color
> Management" ("CM") routes.   
> 
> The first, "Limited color management" option in Elements results in 
images
> that display fairly appropriately when they are placed on the web.  
The
> shadows are relatively compressed.  In Elements' "Print Preview," 
the "Print
> Space Profile" is shown as "Working Gray -- Gray Gamma 2.2."
> 
> The second Elements' Edit>Color Settings CM option is what Adobe 
here calls,
> "Full color management."  This setting is said to be, "optimized 
for Print."
> It has relatively open shadows and a long grayscale and is close to 
what I
> used, in part because I was coming from the PiezoBW starting 
point.  This
> option results in a Print Space Profile called, "Working Gray -- 
Dot Gain
> 20%."  
> 
> You can switch between the options and see the monitor views 
change.  
> 
> For the older, non-CM approach, we often used "Custom Dot Gain" 
curves to
> make the monitor view match the print view better.  That rather 
complex
> manual procedure still result in the most accurate profiling of the 
monitor
> for B&W, but the procedure is only in full Photoshop, and it's 
complex.
> Realistically, most will want the convenience of an automatic 
monitor view
> match that the Color Management approaches allow.  
> 
> 
> 
> > ...  so it should take a simple answer.
> 
> Well ... switching between Dot Gain 20% and Gray Gamma 2.2 with a 
21-step
> test file on the monitor is a good way to easily see the 
differences.  
> 
> I find it useful to accept industry standards and those I'm 
accustomed to. 
> 
> Adobe Photoshop Elements' options seem to be relatively clear and 
simple. 

> 
> Paul
> www.PaulRoark.com
>

The notion of "middle grey" really becomes, in many ways, a bit of a moot
point with today's ability to edit an image.  In the old days, one had to
(ideally) make a decision with regard to middle grey (or at least ONE point
on the exposure scale in which case it's handy to make it near the middle)
at exposure because there was very limited ability to adjust that colour at
a later date and very limited ability to manage tonal transformations from
exposed film to print - most image values "fell where they fell".  This
discipline of fixing one point with confidence through to print is obviously
the core of any system for systematically correct exposures in print (and,
notably, the Zone System).

But today we can "shift exposure" (speaking in broad terms) very easily.  So
in the end the focus can, if need be or if recommended, fall back towards
the image editing (or RAW processing) stage to rendering an image to a
pleasing state overall.  With digital capture, exposing for middle grey may
actually lead to a disadvantageous situation due to the linear nature of the
sensor - you may well be much better off intentionally over-exposing the
part of the image you envision ending up as whatever you decide "middle
grey" to be and pulling it back in the RAW converter.

As a result, the issue comes down to the broader one of how you map your
pleasing image file (as portrayed on screen) to a pleasing image on paper -
how each and EVERY grey gets maps to the new dynamic range.  The tools
available to a pure analogue photographer to manage the tonal range
transformation from processed film to print were, in a relative sense, very
limited (choice of paper grade and dodging and burning etc).  Today you can
do wonders with things such as PS curves and colour management.

So when you start asking questions with regard to the numbers attached to
the same colour in different workspaces (or start with just one colour:
whatever you decide "middle grey" looks like) you are really beginning to
dip your toe into the relationship between file pixel values, how those file
pixel values are interpreted into colours depending on the encoding (colour
space) we attach to them, and what the printer does when it receives those
file pixel values.  (Remember your printer doesn't know what a colour is.
It just fires certain jets in a certain order depending on the 8 bit number
it gets hit with.)  That relationship is part of the topic we call "colour
management".  The colour you decide to call "middle grey" has a number
attached to it which differs according to colour space.   Send any of those
numbers to the printer and you'll likely generate yet a different shade of
grey.  The question should be: when you send off towards your printer the
number that in your workspace looks "middle grey" to you, how can you ensure
that the printer receives not necessarily that same number but a number
which will cause it to print the "right" grey?  That same question needs to
be answered for all possible shades of grey you see in your image - not just
"middle grey".

> From: petexp2 <kafoozalem@...>
> Reply-To: <DigitalBlackandWhiteThePrint@yahoogroups.com>
> Date: Mon, 31 Oct 2005 19:47:38 -0000
> To: <DigitalBlackandWhiteThePrint@yahoogroups.com>
> Subject: [Digital BW] Re: Mid grey : 50%K or 128RGB?
> 
> Thank you Paul for such an informative answer. I really did think it
> would be simpler but you have helped me answer it to my own
> satifaction. In fact playing around with a 21 step wedge in Photoshop
> I note that the when you convert it from gamma 2.2 grayscale to RGB
> it has a midtone value of 128. Alternatively when you convert it from
> dot gain 20% grayscale to RGB the mid point value is 148 RGB. I will
> certainly favour dot gain 20% as a starting point for making prints
> and bear in mind what you say about the compressed tonality of a
> matte paper.
> 
> Pete.
>

Steve wrote, in part:

> The notion of "middle grey" really becomes, in many ways, a bit of a moot
> point with today's ability to edit an image.
> 
> ... today we can "shift exposure" (speaking in broad terms) very easily.

I think a big change is the wider availability of high bit depth.  If you
move things around too much with a 256 step file, you're liable to lose too
much information and may well end up with an uneven print ramp (too few
steps may look like an increase is grain or as lines in a smooth gradient).
If the file is 16 bit, the 65K steps are so vast compared to what we need
that we have much more wiggle room.

Paul
www.PaulRoark.com

You make some interesting observations here Steve. Yes we are far
less tied to the notion of correct exposure than photographers of
yesteryear and I for one am finding that I can rework negatives I
once printed in the darkroom to a far higher standard in Photoshop.
Those darkroom users who use multigrade papers which enable split
grade printing (ie. changing filtration during the burning of various
areas of a print)have a bit more flexibility than their forebears but
there are still limits.

 Mentally I still use the zone system as a reference point in all I 
do. I am not slavishly tied to it but armed with the accurate spot 
meter on my Contax RX (film dinosaur)I can quickly check the exposure 
value of a shadow area or a white cloud to make sure it will be 
possible to extract a suitable level of detail.

I have found it useful to carry these principles forward into
photoshop where I have subjectively assigned RGB values to key zones.
In fact my method was rather quaint. I carefully studied a finished
print and worked out which zone key areas fell on and back in
photoshop cut squares out of each area and reassembled them into a
row of 10 textured squares. It was then natural to measure the RGB
values of each square in photoshop and assign upper and lower limits
to each. For example the all important Zone 2(very dark shadow with a
hint of detail) came out at RGB 40 to 60 and the similarly zone 8
(textured white) gives way to zone 9 (white with no texture but just
distinguishable from paper white) at an RGB value of 245. 

Whilst this was cobbled together it is very useful to ME because it 
matches MY PERCEPTION of zones. There may be more mathematically 
correct ways to follow this approach but it works for me. My question 
about mid greys really arose from the fact my perceived mid grey was 
rather dark and obviously need ammending. I am currently in the 
process of printing tests for my new R2400 and I have found it useful 
to construct a test chart which contains my zone strip as well as a 
21 step wedge and a colour chart and one of my colour prints. However 
in the end the right result is the one which looks right to you. In 
the past I have often got the tonality right for a test strip taken 
from part of an image only to find when I made the whole image it was 
obviously wrong. Our perception of tonal values is very subjective -- 
you just have to surround a square of mid grey with black and compare 
it to one surrounded by white to prove that. One thing I find 
liberating with Photoshop as opposed to the traditional darkroom is 
that we are not slavishly tied to the film curve characteristics of 
the film we are using (or digital sensor). Maybe at the extremes we 
can't make XP2 look like Technical Pan but we can certainly emulate 
Tri X.

Pete (Dinosaur)


--- In DigitalBlackandWhiteThePrint@yahoogroups.com, "Paul Roark" 
<paul.roark@v...> wrote:
>
> Steve wrote, in part:
> 
> > The notion of "middle grey" really becomes, in many ways, a bit 
of a moot
> > point with today's ability to edit an image.
> > 
> > ... today we can "shift exposure" (speaking in broad terms) very 
easily.
> 
> I think a big change is the wider availability of high bit depth.  
If you
> move things around too much with a 256 step file, you're liable to 
lose too
> much information and may well end up with an uneven print ramp (too 
few
> steps may look like an increase is grain or as lines in a smooth 
gradient).
> If the file is 16 bit, the 65K steps are so vast compared to what 
we need

> that we have much more wiggle room.
> 
> Paul
> www.PaulRoark.com
>