Dan Margulis Applied Color Theory
Minimum CMYK Highlight Value
Date: Wed, 18 Jun 2003 09:25:41 -0400
From: Paul Graham
Subject: CMYK minimum highlight values
hi all,
I'm completing the scans for a new photo book that
includes a lot of extremely high key images - many pictures operate in
shades of near whites (often there is nothing below 210 in any RGB
channel...). The printer says that when it is in CMYK, he must have 6,4,4,0
as a minimum to work with, or perhaps I could push it to 5,3,3,0. But, when
I adjust my files to this, by dragging down the highlight point, I get dull
flat looking images. I don't have this issue on my Epson proofing printer,
but it seems that I have to flatten the contrast of everything for the CMYK
press.
so:
how true is this mimimal ink requirement, and what
happens if I supply a 3,1,1,0 super light sky, for example. ( I know in
theory its about leaving control for the press operator to adjust the grey
balance etc, but what happens in practise with an accurate file?) any
advice for operating in this upper segment of the tonal scale?
Incidentally, he's printing the images at 500 dpi,
which is pretty amazing... and yes, it is a really top quality printer,
with state of the art equipment.
thanks,
Paul
________________________________________________________________________
Date: Wed, 18 Jun 2003 14:23:59 -0000
From: Ron Scratch
Subject: Re: CMYK minimum highlight values
These guidelines are straight off the SWOP web page:
http://www.swop.org/
Minimum Printable Dot (Minimum Tone Value)
"For film, the exposure process will determine the
smallest film dot that can be effectively reproduced on plate. This is
typically 2% for plates requiring negative film and 4% for plates requiring
positive film. It is important to prepare input material, including proofs,
with these limits in mind. With computer-to-plate it is possible to
accurately produce 1% dots on plate. In preparing digital files this should
be kept in mind. For critical work it is important to use a proofing
systemthat reflects this minimum tone reproduction characteristic.Although
developments in digital plating and engraving technologies have improved
tone reproduction control in the extreme highlights (less than a 5% dot),
designers should still be cautious in placing image components in this
tonal range. This is because all-digital production is not yet universal
and process control cannot always guarantee precise reproduction below a 5%
dot, depending on the process involved. This caution may be relaxed in
situations where appropriate agreement has been made among all parties in
the production process."
END of SWOP guidlines.
Generally when a printer gives you specific guidelines
for minimum tones, they should be followed because they are familiar with
there press conditions. However the 6,4,4,0 number sounds conservative to
me.
On the issue that- (often there is nothing below 210 in
any RGB channel...). I'm confused on why you would need to do any
adjustment at all with an RGB value of 210. I have files that run between
235-240 and I still can maintain a 5% minimum in Cyan. Perhaps you
can expand on your process to let the group know if you can make
adjustments at the scanner or perhaps use a different profile to achive
these required minimums.
Ron Scratch
________________________________________________________________________
Date: Wed, 18 Jun 2003 09:15:48 -0600
From: Ron Kelly
Subject: Re: CMYK minimum highlight values
Paul:
No doubt others with more experience than me will wade
in on this question as well.
I have been making my own seps for a few years, and I
have learned to fight for constrast as a general rule.
I have been using a highlight of 2,1,1,0 with no
problems. My printer regularly gets this for sheetfed coated at 175 lpi and
no complaints from them, or me, for the most part. Works great on snow
images, for example.
I must admit, though, that high key images are a
challenge. Keeping them neutral, with detail, and not flat looking is
possibly the toughest situation I've been in.
I have struggled mightily to get a good result in a
particular shot of Athabasca Falls which is mostly white water. Over the
years I've printed this blue, green, yellow, and blown out. It's tough.
One thing for sure: you can't rely on your monitor for
this type of picture, no matter how well calibrated.
I humbly suggest correcting by the numbers, followed by
an accurate proof as your best guidelines.
Be wary of anyone who tells you just to get some piece
of equipment and make sure that it is adjusted properly.
Ron Kelly
________________________________________________________________________
Date: Wed, 18 Jun 2003 16:40:01 -0000
From: Paul Graham
Subject: Re: CMYK minimum highlight values
On the issue that- (often there is nothing below 210 in
any RGB
channel...). I'm confused on why you would need to do
any adjustment at
all with an RGB value of 210. I have files that run
between 235-240 and
I still can maintain a 5% minimum in Cyan.
Perhaps you can expand on
your process to let the group know if you can make
adjustments at the
scanner or perhaps use a different profile to achive
these required minimums.
Ron Scratch
maybe that was confusing - all I simply meant was that
even the *darkest* detail in the images is above 210. obviously most of it
is way up there in the 240 to 254 range... which is where I get problem
with fulfillng the 6,3,3,0 requirement. so maybe I should follow the later
advice and just blast through to something like 2,1,1,0 or so?
I really don't like the highlight flatness the lower
press values force on the images
paul
________________________________________________________________________
Date: Wed, 18 Jun 2003 13:10:31 -0400
From: Dan Remaley
Subject: RE: Re: CMYK minimum highlight values
Hi all, it would be difficult to hold or control a 1%
dot in a detail area, the tendency would be to "drop out" to pure
white, like the paper. Consider that the plate readers are accurate to +/-
1%. Film setters same (+/- 1% issue). The 'safe' response is 6-3-3 in a
'detail' area.
Dan
Dan Remaley
Process Control Mgr.
Graphic Arts Technical Foundation
412.741.6860x450
________________________________________________________________________
Date: Wed, 18 Jun 2003 17:34:58 -0000
From: "jrscratch
Subject: Re: CMYK minimum highlight values
Apologies for the misunderstanding. Your last post you
had said: But, when I adjust my files to this, by dragging down the
highlight point, I get dull flat looking images. It seems that I have to
flatten the contrast of everything for the CMYK press.
I went into Photoshop and made a gradient in a range of
RGB from 245 to 210. In curves I took the Highlight point and dragged it
down to bring the minimums to 6,4,4,0 and as expected I essentially made a
lot of the tonal values the same, essentially making the transition of tone
flatter. So I tried a slightly different approach, I pulled my highlight
down as before except with the same curve I pulled the Shadow end straight
over like a pivot, which essentially exaggerated the differences between
tonal values. This kept the flatness issue less apparent. I do not know if
this can help you, but give that a try and let me know how much the
integrity of the image was compromised.
Thanks
Ron Scratch
________________________________________________________________________
Date: Wed, 18 Jun 2003 14:07:03 EDT
From:Dan Margulis
Subject: Re: CMYK minimum highlight values
Paul Graham writes,
I'm completing the scans for a new photo book that
includes a lot of
extremely high key images - many pictures operate in
shades of near
whites (often there is nothing below 210 in any RGB
channel...).
That's the whole problem right there. In principle, you
should have *something* in each channel that's way, way, way darker than
210--more like 010. That's a rule that applies to about 99% of all images,
and without it the images will be unacceptably flat.
Now examples do exist where there really isn't anything
that you can possibly portray as being dark, in which case such a shadow
value wouldn't be appropriate. It sounds like you may be dealing with that
kind of image. However, in that case you take a deep breath and make the
darkest point of the image as dark as you can stand. As I said, in 99% of
all images the darkest value should be in the neighborhood of 10 in each
channel. I can probably count on the fingers of one hand all the images
I've ever seen where the darkest point should be as light as 128. A darkest
value of 210, I've never heard of. That's what's killing the contrast of
the picture, not whatever the printer is doing.
To fix this, go to Image: Adjustments>Auto Contrast.
This will make the image ridiculously contrasty, but now go to Edit: Fade
and reduce the percentage of the move. Choose the highest percentage that
looks acceptable.
The printer says that when it is in CMYK, he must have
6,4,4,0 as a
minimum to work with, or perhaps I could push it to
5,3,3,0. But, when I
adjust my files to this, by dragging down the highlight
point, I get
dull flat looking images.
This is because in your original files you tend to have
nothing that's darker than about 20%. So, if you shave a point or two off
the highlight you've lost a very substantial portion of the real estate you
need to portray tonality.
Incidentally, he's printing the images at 500 dpi,
which is pretty
amazing...and yes, it is a really top quality printer,
with state of the art
equipment.
The phrase "500 dpi" is ambiguous. I can't
imagine he's using a 500-line screen, which indeed would affect how light
of a highlight he could hold. If you're saying you're providing a
500-pixel-per inch image file, that won't affect the contrast issue.
If he indeed has state-of-the-art equipment he probably
does hold a 1% dot most of the time but you can't bet your life that he
will the day he runs your job. If highlight detail is absolutely critical
you should probably stick with 5c3m3y, but if you can afford to have some
of the highlight go lower than this you should do it.
Again, however, the problem is not the printer, but the
lack of range in the original file.
Dan Margulis
________________________________________________________________________
Date: Wed, 18 Jun 2003 19:01:18 -0000
From: Paul Graham
Subject: Re: CMYK minimum highlight values
thanks, all -
these are very special tweaked images and I know way
out of the regular loop - they are adjusted to mimick that feeling when you
walk out of the movies and into the sunlight, and get optically stunned.
That is why there is nothing at all below 200. It's all about brightness,
blinding whiteness. But that doesn't mean I don't want to do the best I can
in that small range, hence my original question of how far I could push it
in minimum press values.
I wish Photoshop had a way that you could enlarge a
section of the curves box and just accurately manipulate that area. it
would be a great feature.
The phrase "500 dpi" is ambiguous. I can't
imagine he's using a 500-line
screen, which indeed would affect how light of a
highlight he could hold.
It's a 250 line/inch (100 line/cm) screen and the scans
are supplied
at 500 dpi.
you're saying you're providing a 500-pixel-per inch
image file, that won't affect
the contrast issue.
of course, it was just an interesting aside as to the
quality.
it may be wasted on the high key images, but not on the
'normal' ones
which are exceptionally sharp medium format scans.
If he indeed has state-of-the-art equipment he probably
does hold a 1% dot
most of the time but you can't bet your life that he
will the day he runs your
job. If highlight detail is absolutely critical you
should probably stick with
5c3m3y, but if you can afford to have some of the
highlight go lower than this
you should do it.
I can't stand looking at them with (eg) a near white
sky set at 6,4,4, and even 5,3,3, is touchy. so I guess I'll push it and
see how it goes.
we are 'wet proofing' on the machine first anyway, so
we'll see...
incidentally the varnish layer really helps hold the
image onto the paper and prevent the burn out to paper white, keeping that
distinction between image and paper (nothing is full bleed)
bests,
paul
________________________________________________________________________
Date: Wed, 18 Jun 2003 15:25:33 -0400
From: Lee Clawson
Subject: Re: CMYK minimum highlight values
Paul,
many pictures operate in shades of near whites.........
leaving control for the press operator to adjust the
grey balance etc,........
If I understand this you have RGB photographs that have
nothing darker than 210 (all highlight tone and color). And when converted
to CMYK will range from 0 to 3% at the lightest and the darkest area just
about 15 to 20%. Is that the darkest or just the range you care about most
??? Also, do you need to hold subtle pastels or is tonal variations most
important???
On work like this I'd make sure my printer saw (vs a
phone call) very early on what I wanted. My advice is too make a section
proof (small cropped sections from a variety of the pictures) early on, if
possible do this on the same paper used for the book.
The printer says that when it is in CMYK, he must have
6,4,4,0 as a
minimum to work with, or perhaps I could push it to
5,3,3,0. But, when I
adjust my files to this, by dragging down the highlight
point, I get
dull flat looking images. I don't have this issue on my
Epson proofing
printer, but it seems that I have to flatten the
contrast of everything
for the CMYK press.
I'm confused. What contrast was there to be lost??
Anyway try adding a bit of tone overall. Keep in mind that specular (paper)
whites are ok.
Lee
________________________________________________________________________
Date: Wed, 18 Jun 2003 18:24:37 -0700
From: Jono Moore
Subject: Re: Re: CMYK minimum highlight values
on 18.6.03 12:01 PM, peegeenyc wrote:
these are very special tweaked images and I know way
out of the regular loop -
they are adjusted to mimick that feeling when you walk
out of the movies and
into the sunlight, and get optically stunned.
Sounds like a fun project. :-)
It's a 250 line/inch (100 line/cm) screen and the scans
are supplied at 500
dpi.
Wow, overkill on the resolution. Unless the
printer really wants them that way, 350-400dpi would probably be plenty.
As has been mentioned in recent posts, as the line
screen goes up you get diminishing returns on the formula of dpi=2*lpi. The
only thing that really goes up is storage space.
We recently produced an art book for our local museum
of around 250 black on black duotones. Our problem was the opposite of
yours, 95%+ of our image detail was in the shadows. To make for more fun,
some of them also had highlight patches that we had to hold (white shell
inlays on black shale/slate type stone).
It was printed at 420lpi stochastic and the printer
told us we didn't need any more than 300dpi. There is a lot of nice detail
in the printed piece.
One thing it looks like we could have used more of was
sharpening. Someone please correct me if I'm wrong but does stochastic
screening want heavier sharpening? It was our first experience with
stochastic.
I'm sure a lot of folks on the list would have cried at
the shape some of the negatives were in...but that's another story.
I can't stand looking at them with (eg) a near white
sky set at 6,4,4, and
even 5,3,3, is touchy. so I guess I'll push it and see
how it goes. we are
'wet proofing' on the machine first anyway, so we'll
see...
What you're trying to do is really going to depend on
your printer's equipment and what kind of paper you are printing on. I'd
say their settings are on the conservative side as well, but you never
know.
I used to run the prepress dept for an
"average" small print shop (~20 employees) that had a 4-colour
Heidleberg. We had pretty good equipment for running film (new Xitron Adobe
Extreme RIP and Agfa SelectSet 7000 'setter) and I think we could reliably
hold a 3% dot, maybe even 2% on a good day. The weak point was burning
plates.
Being as you are looking for a sort-of blasted out
look, setting your highlights lower might be fine, but if the detail is
critical then you gotta know whether or not the press/paper combo can hold
it.
--
Jono Moore
udoprinting.com
________________________________________________________________________
Date: Thu, 19 Jun 2003 08:12:28 +0300
From: "Mihai Necsulescu"
Subject: Re: Re: CMYK minimum highlight values
Hi Jono,
We are a pre-press shop with CTcP (Computer To
conventional Platesetter). We found that most of the time 1.5 x Lpi
resolution is enough and sharpening is also more efficient then on higher
resolutions. We done some prints with 420 lpi regular screening and also
with stochastic screening. We found that stochastic may look a little bit
grainy compared with 300 lpi or more and the feeling for the customer has
confirmed what we observed. If you go at 300 lpi or further then you are
close to contiguous tone. Regarding the detail it depend much of the source
image and if you don't re-scale images then 300 dpi is enough for 300 lpi
or more. Sharpening can play a major role here and you should experiment
the amount and looking at the plate or film prior to print. Anyway, 300 dpi
was the physical resolution of BW printers and further to that all lines
appear straight to naked eye so I presume that is not necessary to have
more information in that picture. Remember that the eye is integrating the
image and that is the trick beside any offset printing! If you want to
print an image that can be viewed under a magnifier glass then you must go
further but otherwise is not necessary. The output resolution is another
issue because the performance of that can severely influence the output
capability to reproduce a high lpi or FM. And if you do not have CTP
capability then the copy frame and plate processor can also have major
impact to your final output. On out CTcP plates we expose a digital control
stripe to monitor all these processes and we have plates that hold a 17
microns dot on press but several printing shops from here don't monitor
even the linearity of imagesetter films. Our customers usually provide us
digital files and it was amazing that they have observed the difference in
printed matters because we hold reliably 1 and 99% on each plate. A bad
calibrated imagesetter, a dusty copy frame not calibrated can easy throw
out 5% on a 175 lpi film. Because we provide CTcP plates to many printing
shops we observed that many old machines can output a high quality print if
they run perfect plates compared with new machines that use bad calibrated
imagesetter, copy frame and plate processors.
Best Regards,
Mihai Necsulescu
manager
Zoom-Soft srl
Sabinelor 106 Str., bl. 115, sc. 1, ap. 1, sect. 5
Bucharest - Romania
Tel: +40-21-410.06.75, Fax: +40-21-411.35.39
mobile: +40-722.360.399
________________________________________________________________________
Date: Thu, 19 Jun 2003 09:38:15 +0100
From: Martin Bailey
Subject: Re: Re: CMYK minimum highlight values
At 02:24 19/06/2003, Jono Moore wrote:
One thing it looks like we could have used more of was
sharpening. Someone
please correct me if I'm wrong but does stochastic
screening want heavier
sharpening? It was our first experience with
stochastic.
I'm not sure if it needs heavier sharpening, but most
stochastic certainly needs sharpening applied over a smaller radius.
Regards
Martin Bailey
-------------------------------------------------------------
Senior Technical Consultant
+44 1223 873800
Global Graphics Software
http://www.globalgraphics.com
Developers of Harlequin, Jaws, Jaws PDF &
MaxWorkFlow
-------------------------------------------------------------
________________________________________________________________________
Date: Thu, 19 Jun 2003 09:18:33 -0400
From: "John"
Subject: Re: CMYK minimum highlight values
Dan Remaley wrote:
Hi all, it would be difficult to hold or control a 1%
dot in a detail area,
the tendency would be to "drop out" to pure
white, like the paper. Consider
that the plate readers are accurate to +/- 1%. Film
setters same (+/- 1%
issue). The 'safe' response is 6-3-3 in a 'detail'
area.
I'm not sure what you mean by "control" a 1%
dot. If you platesetter will in fact image a 1% dot, and your press will
print it, you have a 1% dot of ink on paper that in all honesty would be
very hard to screw up. A gain factor of 100% would make it a 2% dot... that
I doubt anyone could tell the difference of. The dot values of 40% to
60% are the hardest to control on press and have major bearing on the final
outcome. Highlight values are the easiest to control on the press.
There was a day (and not that long ago) where some
films and plates were in fact a bit shaky under a 4% dot value. To the
point that sometimes they in fact would not even reproduce on the plate, or
if they did would "fall off" the plate quickly with very few wear
impressions on them. To make matters worse, when they broke, they sometimes
left and abrubt fall of line, causing choppy and blotchy highlight areas.
The catch all fix for all this was the 'safe' response of 6-3-3 in a
'detail' area (knowing that there should be no black there anyway) to cover
up the fact that anything less, might be lost. Times change. Depending on
the subject matter, if you can hold it small, print it. A lot has to do
with subject matter... you need dots to make and show detail, but if the
subject lends itself to bright & contrasty... I'll take that
everyday over a flat image with dirty whites.
John Rawlins
________________________________________________________________________
Date: Thu, 19 Jun 2003 13:26:57 -0000
From: Paul Graham
Subject: Re: CMYK minimum highlight values
If I understand this you have RGB photographs that have
nothing darker than
210 (all highlight tone and color). And when converted
to CMYK will range
from 0 to 3% at the lightest and the darkest area just
about 15 to 20%.
Is that the darkest or just the range you care about
most ???
Also, do you need to hold subtle pastels or is tonal
variations most
important???
that is the darkest. everything is the entire image is
above 200. not just what I am concerned about - everything.
On work like this I'd make sure my printer saw (vs a
phone call) very early
on what I wanted. My advice is too make a section proof
(small cropped
sections from a variety of the pictures) early on, if
possible do this on
the same paper used for the book.
we've already done some machine proofs on different
papers, and have ended up with a super-bright paper, which obviously gives
us the most headroom...
I'm confused. What contrast was there to be lost??
Anyway try adding a bit
of tone overall. Keep in mind that specular (paper)
whites are ok.
well, ok, there isn't 'contrast' in the normal sense,
but when you only have 200 to 250 range, losing even 5 points is 10% of
your tonal scale, so that's why I don't want to suppress the brightest
near-white skies to 6,3,3,0. there is a big difference, on screen and in
proofs when you don't have to cap the images to an artifically low ceiling.
I'm using high-pass sharpening. it seems to work best.
or are there any other recommendations?
________________________________________________________________________
Date: Thu, 19 Jun 2003 10:46:28 -0400
From: Dan Remaley
Subject: RE: Re: CMYK minimum highlight values
From: John
I'm not sure what you mean by "control" a 1%
dot. If you platesetter will in
fact image a 1% dot, and your press will print it, you
have a 1% dot of ink
on paper that in all honesty would be very hard to
screw up. A gain factor
of 100% would make it a 2% dot... that I doubt anyone
could tell the
difference of. The dot values of 40% to 60% are
the hardest to control on
press and have major bearing on the final outcome.
Highlight values are the
easiest to control on the press.
What I'm saying is that a 1% value is the minimum dot -
if there are any calibration issues, processing issues, pressure issues at
press - that value could change from 0% or 3%! A detailed
highlight looking like the paper ' white' is a problem. I agree that the
midtone (50%) is where the largest change takes place
At the Sheetfed Conference this year, in Chicago, I
spoke with a web pressman from Quad Graphics. They use a color control
system they developed with system Brunner. It measures the gray balance on
the color bar for run control, he said the 2,000 th.sheet matches the
50,000 th. sheet.
Dan Remaley
Process Control Mgr.
Graphic Arts Technical Foundation
412.741.6860x450
Print: The Original Information Technology at
<www.gain.net>
________________________________________________________________________
Date: Thu, 19 Jun 2003 11:36:37 -0700
From: Alison Walker
On Wednesday, June 18, 2003, at 12:01 PM, Paul Graham
wrote:
I wish Photoshop had a way that you could enlarge a
section of the
curves box and just accurately manipulate that area. it
would be a
great feature.
This is an interesting comment. How are you using the
Curves box? It does have two sizes, click on the button in the lower right
corner if you are working in Photoshop 7. It is my understanding that
at the smaller size you have 100 increments of adjustment control for
working in CMYK. And at the larger size you have 256 increments of control
for working in RGB. I have found that working in the large sized Curves box
gives me more control in either space. Try the larger size and see if you
find you have a bit more control when dialing in your lighter tones.
Good luck,
Alison Walker
________________________________________________________________________
Date: Fri, 20 Jun 2003 04:11:42 -0000
From: "Roy Harrington"
Subject: Re: CMYK minimum highlight values
Paul Graham wrote:
I wish Photoshop had a way that you could enlarge a
section of the
curves box and just accurately manipulate that area.
Here's a way you can simulate what you want using
Adjustment Layers.
On top of your image, create an adj layer that expands
your range of 255 to 204 to the full range 255 to 0 -- Levels with
the black arrow Input Levels at 204 will do it. (I pick 204 because
its exactly 80% of 255.)
Now you can have a Curves Adj Layer that works on the
full range. This is where you do all your work.
Finally, another Level Adj Layer that re-compresses the
255 to 0 range back to 255 to 204. Levels black arrow at the Output
Levels set to 204 gets you back.
Roy
________________________________________________________________________
Date: Fri, 20 Jun 2003 18:45:18 -0000
From: Paul Graham
Subject: Re: CMYK minimum highlight values
great suggestion, thanks Roy, and all who advised,
paul graham
Adobe Photoshop training classes are taught in the US by Sterling Ledet & Associates, Inc.