Dan Margulis Applied Color Theory
Clarification of Basic Concepts
Date: Mon, 10 Apr 2006 00:41:28 -0400
From: "Russell Proulx"
Subject: Clarification of basic concepts
Sorry if these issues have been addressed (perhaps ad
nauseam) in the past. I'm trying to correct potential misunderstandings
I've been living with (and spewing misinformation to my students..).
What is the minimum (highlight) value of K that will
print (sheetfed/web 'on decent paper' /newsprint)?
Do 300ppi scans for 133lpi films offer any advantage
over 200ppi scans for the same output?
Does evaluating unsharp masking at greater than100%
monitor resolution make any sense? Is there any 'real world' advantage to
adding Gaussian Blur to a JPG to smooth out artifacts that are only
apparent @300-800% ?
Thanks for your understanding and sharing your
knowledge.
Russell
___________________________________________________________________________
Date: Mon, 10 Apr 2006 07:06:36 -0400
(EDT)
From: Mark Segal
Subject: Re: Clarification of basic concepts
Russell, the one part of your query that I can respond
to is that about unsharp masking. It definitely does NOT make sense to
evaluate results at greater than 100% magnification because pixelation
begins to obscure the effect of the sharpening. In fact it is not even
recommended to observe sharpening effects at 100%. 50% or 75%
magnifications are adequate, and they are appropriate because at these
percentages you avoid the anti-aliasing that Photoshop applies to
intermediate percentages. While it is generally advised that monitors tend
to exaggerate sharpness effects compared with what is visible in the
corresponding print, my experience (using Photokit Sharpener Pro with a CRT
Trinitron monitor, and outputting to an Epson 4800) is the reverse.
Mark Segal
___________________________________________________________________________
Date: Mon, 10 Apr 2006 08:52:12 -0400
From: Terry Wyse
Subject: Re: Clarification of basic concepts
On Apr 10, 2006, at 12:41 AM, Russell Proulx wrote:
What is the minimum (highlight) value of K that will
print
(sheetfed/web 'on decent paper' /newsprint)?
Any properly calibrated CtP system for offset printing
should be able to hold right down to a 1-1.5% dot on the press sheet. In
terms of SETTING a highlight dot in an image for printing the first
discernible highlight detail area (diffuse highlight), I would stay in the
range of 3-5% just like always.
Do 300ppi scans for 133lpi films offer any advantage
over
200ppi scans for the same output?
POSSIBLY but probably not, depending on the subject
matter. I can remember running some tests years ago where there APPEARED to
be some improvement in detail/sharpness up to about 2.3 x linescreen
(that's about 306ppi @ 133pi) but USM plays a huge role in this. Sloppy
sharpening of a high resolution image vs. careful sharpening of a lower
resolution image (even down to 1.4 x linescreen) can make all the
difference. Stick to around 1.5-2.0 x linescreen and you and your students
will be fine.
Does evaluating unsharp masking at greater than100%
monitor
resolution make any sense?
Nope. In fact, if your scans/images are at the normal 2
x LPI, it makes more sense to view at 50%. That'll give you about the right
look compared to the final printing. Viewing at 100% or larger, you won't
like what you see in terms of USM but you have to keep in mind that half of
what you see at 00% is going to get "thrown away" in the final
processing/imaging anyway. Again, I'm assuming this is all relating to
halftone printing. If the final output is inkjet or photo printing,
different rules apply.
Is there any 'real world'
advantage to adding Gaussian Blur to a JPG to smooth
out
artifacts that are only apparent @300-800% ?
No opinion since I never use JPEGs.
Regards,
Terry Wyse
___________________________________________________________________________
Date: Mon, 10 Apr 2006 10:23:57 -0400
From: todie
Subject: Re: Clarification of basic concepts
My rule is (as much as possible) "no photographic
image under 5MB" (no interpolation), even if the resolution goes
higher than 300ppi for small B&Ws, and even if the RIP may be a Scitex
which could screw up resolutions over 300.
Laurentiu Todie
___________________________________________________________________________
Date: Mon, 10 Apr 2006 08:59:50 -0500
From: "RJay Hansen"
Subject: Re: Clarification of basic concepts
On 4/9/06, Russell Proulx wrote:
What is the minimum (highlight) value of K that will
print
(sheetfed/web 'on decent paper' /newsprint)?
We still use imagesetters, and we hold dots in the 1 -
2% range fairly consistantly, but I agree w/Terry, the 3 - 5% range will be
held consistantly, and that's what I set highlights I want to hold to.
Do 300ppi scans for 133lpi films offer any advantage
over
200ppi scans for the same output?
The old rule of thumb of course, was 2 X line screen
which would be 266ppi for a 133 line screen. But that is definitely not a
hard and fast rule. Results really vary with the subject matter, and
generally, you can get by with less than that 2X rule. As Terry said,
sticking to 1.5 - 2X is probably "safe", but I've seen many
images we get come in at less than that and print significantly better (and
acceptably for many purposes) than I would have predicted from the
resolution.
Does evaluating unsharp masking at greater than100%
monitor
resolution make any sense?
I've always been told you should sharpen at 100%, while
viewing the result at 50% will give you a fairly good idea of how it will
look when printed.
Is there any 'real world'
advantage to adding Gaussian Blur to a JPG to smooth
out
artifacts that are only apparent @300-800% ?
I don't have a definitive answer to that (as I've never
done it), but it seems highly unlikely to me.
RJay
___________________________________________________________________________
Date: Mon, 10 Apr 2006 09:11:13 -0700
From: Tom Thompson
Subject: Re: Clarification of basic concepts
Is there any 'real world'
advantage to adding Gaussian Blur to a JPG to smooth
out
artifacts that are only apparent @300-800% ?
I would try the "Reduce Noise" filter in
Photoshop. It has a "Remove JPEG Artifact" button that may be
useful.
The other questions have been adequately answered
already.
Tom Thompson
System Administrator
Effective Graphics Inc.
___________________________________________________________________________
Date: Mon, 10 Apr 2006 12:27:55 -0400
From: Lee Clawson
Subject: Re: Clarification of basic concepts
on 4/10/06 12:41 AM, Russell Proulx wrote:
What is the minimum (highlight) value of K that will
print
(sheetfed/web 'on decent paper' /newsprint)?
I'll allow 1 to 2% with 5% as the first printing
highlight tone that has any detail. With newsprint (depending on image)
I'll push a bit higher on the 5%.
Do 300ppi scans for 133lpi films offer any advantage
over
200ppi scans for the same output?
I'll stay close to the 2X line rule if the images are
large, have fine detail and/or are printed on good quality paper. For
smaller images and those going to lower quality magazine stock or newsprint
that are "normal" in tone and detail I'll use as low as 225dpi.
The bulk of the images we work with are 250 to 300 dpi
for offset output. Lower for screen printing and large display graphics and
(years ago) much, much, higher for output to film transparency.
All of the above gets a qualification: the images are
placed and used at 100% in page layout app's. For example, images for
posters may start out at lower dimensions but higher dpi (way above 300)
knowing the the effective enlarged res is more normal. And if we know that
we'll use the image in many output dimensions we may request a scan above
300.
Does evaluating unsharp masking at greater than100%
monitor
resolution make any sense? Is there any 'real world'
advantage to adding Gaussian Blur to a JPG to smooth
out
artifacts that are only apparent @300-800% ?
I've heard of blurring and re-sharpening JEPG's but
never do it. But that may be more that we don't see to many lower res JPEG
camera files
Generally I'll view amounts of USM added at 100% or a
bit higher and then make final judgements at lower viewing mags. Output for
screen printing is one exception (they like low USM) and there may be
others.
Lee Clawson
2/\V/\7 Studio
___________________________________________________________________________
Date: Tue, 11 Apr 2006 09:58:38 -0000
From:Stephen Marsh
Subject: Re: Clarification of basic concepts
Mark Segal wrote:
Russell, the one part of your query that I can respond
to is that
about unsharp masking. It definitely does NOT make
sense to evaluate
results at greater than 100% magnification because
pixelation begins
to obscure the effect of the sharpening.
Agreed Mark, and a side note on reduced views...
I found that 25% seemed to offer a more accurate
interpolation of the 100% view than 50% does (for repeating patterns at
least) - although 50% seems to show a good view of what to expect for
a final offset print run on good stock/screen (I did not try 75%).
Stephen Marsh.
___________________________________________________________________________
Date: Wed, 12 Apr 2006 09:03:49 +1000
From: Ross Mitchell
Subject: Re: Clarification of basic concepts
Does evaluating unsharp masking at greater than100%
monitor
resolution make any sense? Is there any 'real world'
advantage to adding Gaussian Blur to a JPG to smooth
out
artifacts that are only apparent @300-800% ?
Same answer for both
sharpening and JPG artifacts.
In both cases the
suitable screen magnification is to mimic the viewing proportions you are
outputting to. Most images will be viewed as full images at arms
length of a little less, in these cases 100% or 50% give you the same
viewing angles on screen. However if I were outputting to a mural or
a wall hanging which would be viewed very close rather than as a whole
image, then sharpening above 100% would be correct.
Removing JPG
artifacts is something I would very much want to avoid having to do, this
certainly means never saving to JPG and subsequently re-editing. But
if the artifacts are there in the capture, then I would deal with them
under a viewing geometry which matched the image usage.
My criteria is that
the only reality is what I you can see in the image, if I can see it then
I have to work with it and live with the result.
Thanks for your understanding and sharing your
knowledge.
I have been on the
list for over a year. This is the first time I have had anything to
contribute.
Regards
Ross Mitchell