Date   

Re: Curves vs Sel Color

Terry Wyse <terry@...>
 

Hi Dave,

Wouldn't just using the Saturation tool to de-saturate do essentially the
same thing? This would remove color without altering the overall
luminosity/contrast. Or maybe I'm not understanding the question completely.
Yea, that's it! :-)

Terry

on 3/30/01 7:25 AM, Dave Badger wrote:

When I want to pull color out overall from low contrast images (for example,
glassware), I have a tenancy to use Selective Color Neutrals-Whites-Black
instead of Curves.

I know when you pull a curve down, your decreasing the steepness of the
curve and reducing the change in tones from one step to the next, i.e.
reducing contrast even further. But is Selective Color Neutrals (a misnomer
as it affects all colors) doing same thing or is it likely to hurt the
contrast less?

Dave Badger
_____________________________
Terence L. Wyse
PrePress Systems Specialist
All Systems Integration, Inc.
114 Cummings Park
Woburn, MA 01801
Main office:
781.935.3322 voice
781.935.6622 fax
http://www.allsystems.com
terry@allsystems.com
_____________________________


Curves vs Sel Color

Dave Badger <dbadge@...>
 

When I want to pull color out overall from low contrast images (for example,
glassware), I have a tenancy to use Selective Color Neutrals-Whites-Black
instead of Curves.

I know when you pull a curve down, your decreasing the steepness of the
curve and reducing the change in tones from one step to the next, i.e.
reducing contrast even further. But is Selective Color Neutrals (a misnomer
as it affects all colors) doing same thing or is it likely to hurt the
contrast less?

Dave Badger


knock out on existing film

jonathan clymer <jeclymer@...>
 

A while back I designed an ad which was run in the trade
magazine Broadcast Engineering. I supplied film for the ad.
This month, the ad is running again and I noticed that a
booth number for the NAB trade show has been added at the
bottom of the page. The text is white against the original
dark blue textured background. I see no evidence of film
stripping.

How is this accomplished?

Jonathaan Clymer


Re: Blue/Green Screens

samarsh@...
 

--- In colortheory@y..., pbeck@a... wrote:
OK...I'll bite. Educate me. Crop Vs. Clip

Paul Beck
12 year photog, 7 year photoshop (high level user)
I thought at this lists level my statement would speak for itself. Guess not...

* CROP - a rectangular destructive edit which removes pixels.

* CLIP - a PostScript masking feature which is non desctructive, does not change canvas size, is editale and can be regular or irregular (any shape).

* Crop is an old photographic term.

* Clip is a PostScript Page Description programming language term.

I don't know if this comes down to local graphics 'dialect'...

Some users say they want to crop an image - when they mean make the background transparent.

To any literal and logical user, the two are totally different things.

But it does not seem to be the case. Another case of blue=cyan?

Stephen Marsh.

P.S. Sorry if yahoo screws up this post, egroups worked fine.


blue/green screens

Dan Margulis <76270.1033@...>
 

Andrew Engelhardt writes,

just wondering if anyone has any experience shooting or dealing with
images that have been shot against a blue screen? We have to try to
facilitate a high volume of single product images with clipping paths in a
fairly fast turn-around time and we're looking at how to speed up the
process of getting paths on everything. We're hoping that the blue or green
screen background might help either Photoshop or another program like Mask
Pro with creating a path quickly and accurately. These are files that will
be repurposed for a bunch of different uses from all kinds of printing
conditions to web use, thus the paths on just about everything. The only
problem we can think of so far is when the colour of the product is close
to
that of the background.>>

Having the background in a contrasting color is definitely helpful, however
AFAIK there isn't a way of completely eliminating the human factor. If it's
a simple shape probably nothing is going to be faster than the pen tool. If
complex, and if there are a *lot* of such images, AND if the blue color is
fairly constant and fairly pronounced, then I'd consider a script that did
the following:

1) Make copy of image

2) Convert copy to LAB.

3) Apply drastic curves to A and B channels, making them much steeper and
moving them to the right (i.e. magenta-yellow direction). This will make
the background a brilliant blue and make give everything else a red cast.

4) Virtually wipe out the L channel by reducing the black point to about
10%.

6) Return copy to RGB.

7) Move the shadow point of the RGB master curve so far as to restore the
darkness of the original image that was wiped out by Step 4. This will
insure enormous edge contrast in color.

8) Allow operator to click a representative portion of background. Select
by Color Range, Fuzziness 200.

9) Save Selection as a separate document.

10) Discard the altered copy of the original.

The saved selection is going to be rather close to a mask and it will be
easy for the operator to edit out extraneous parts.

If I were distributing these images, I'd consider doing so with masks, not
clipping paths. Making efficient clipping paths is a lot of work; if the
user needs a clipping path and already has a mask, the path is easily
generated.

Dan Margulis


Blue/Green Screens

pbeck@...
 

OK...I'll bite. Educate me. Crop Vs. Clip

Paul Beck
12 year photog, 7 year photoshop (high level user)


Re: blue/green screens

Shangara Singh <eye.eye@...>
 

on 28/3/01 7:52 pm, Andrew Engelhardt at aengelhardt@londondrugs.com wrote:

Hi all, just wondering if anyone has any experience shooting or dealing with
images that have been shot against a blue screen?
Any input or advice would be greatly appreciated.
Light the screen evenly from top to bottom, side to side, use soft lights if
you can afford them.

Depending on the subject size you need to place it well in front of the
screen - minimum I'd say of 6'.

If going for a blue screen sometimes putting a slightly warm gel on the
backlight can help to kill the contamination; a light magenta if going for a
green screen though I've never lit a green screen only blues.

Get a proper blue screen and make sure when you mount it there are no
wrinkels in it. If painting a cyc get blue screen paint and apply several
coats for an even look.

Underexpose the screen about a stop and a half to get some saturation.


--
Respectfully,


Shangara Singh.

A ___adobe
C ___certified
E ___expert/Photoshop 5

Graphics Photography Multimedia Websites
&#92;_______&#92;__________&#92;________&#92;_______
</>Websites: http://www.oxyopia.co.uk


Re: Blue/Green Screens

samarsh@...
 

Andrew Engelhardt wrote:

Hi all, just wondering if anyone has any experience shooting or dealing with images that have been shot against a blue screen? >
===

Hello Andrew, sorry - no experience. Try this link:

http://macworld.zdnet.com/1998/05/create/4317.html

===

We have to try to facilitate a high volume of single product images with clipping paths in a fairly fast turn-around time and we're looking at how to speed up the process of getting paths on everything. We're hoping that the blue or green screen background might help either Photoshop or another program like Mask Pro with creating a path quickly and accurately. These are files that will be re-purposed for a bunch of different uses from all kinds of printing conditions to web use, thus the paths on just about everything. The only problem we can think of so far is when the colour of the product is close to that of the background. Any input or advice would be greatly appreciated. >
===

There seems to be a common misconception about cross application
transparency - and transparency in general.

There are three common areas for transparency:

i) Photoshop layer compositing

ii) Clipping Paths for PostScript 'Print' Output (Adobe blur the lines
with PDF however, this is both monitor and print...but it works on non
PS devices - I don't understand how PDF 'masks' to QuickDraw or PCL,
Quartz seems obvious for PDF however)

iii) Web and Multi Media (monitor) based transparency output

Most users knew the three 'general rules' as above.

Now APS6 changes things by using vector based layer shapes for layer
transparency within Photoshop - while saving a clipping path in a EPS or
TIFF for transparency outside of Photoshop is another thing (I presume -
I currently use version 5.5, so only have past experience to base things
on).

Does Photoshop 6 use the term CLIPPING MASK or CLIP PATH for these
vector shape transparency effects? I hope they make some sort of naming
difference between them. It's like users who say crop when they mean
clip. It leads to confusion.

Traditionally, for PostScript output and transparency - there were only
two methods available:

i) Clip Paths on a contone image (Greyscale/RGB/CMYK)

ii) Bitmap mode TIFF or EPS (1 bit file)

Both these methods only gave 1 bit transparency. you could not have
varying tones of opacity like in Photoshop - it was all or nothing.

At this point, the only professional *print* tool that I know of that
can use a raster (pixel) alpha mask for transparency is Adobe
Illustrator 9 (or Acrobat 5 perhaps?). This is from a layer mask in a
native layered PSD file from Photoshop.

The good news is that the transparency can now vary from 0-255 levels,
instead of only 0 or 255.

I understand FreeHand and Canvas had transparency before Illustrator -
but it was not useful for this 'deep etching' purpose, AFAIK.

Now for the 'bad news'. This raster transparency for print has not been
proven to me. The manual shows it - but I have not had the time to test
this, let alone trust it to a live job going to film - or even worse,
digital output at a publication which we have no quality control over.

Who knows how this 'magic' variable raster mask based transparency will
behave if saved as EPS and imported into layout on *another* background
for separations? Will all RIPs behave the same? I don't even know if it
works within Illustrator 9.0.1 (Don't let me get started about AI9 - and
I use it on the PC, which seems to have had a better ride than the Mac).
I will shortly try the 9.0.2 update...

As for third party masking like Knockout and MaskPro - as far as I know,
these only perform raster based alpha masking...which is perhaps the
opposite to a clip path.

Most of these developers do not claim to create PostScript clip paths -
or cross application transparency for print.

They simply promise to deliver an alpha channel. It is up to the user to
understand all the issues of transparency and output.

As far as I know, this transparency is all for layer compositing (where
a merged flat image is finally delivered) or web/multi media monitor
output, once agian a final flat image with an alpha channel.

Now if I have misunderstood, and you know about the limitations of
transparency for different applications, output and media - then you
must have been referring to using a selection loaded off one of these
alpha channel masks to turn into a path, for use as a PostScript Clip
Path.

This is something that I never suggest doing (if quality is a concern).
Autotracing in this method does have it's creative uses - just not for
etching an image from it's background.

I do not know of any quick way to etch an image with a precise vector
path. Even human operators can make a bad clip and expose some
background or clip too much of the image.

As for transparency for web or multi media - that is another subject,
possibly off topic or off list.

Waiting to see what the regulars have to say (go easy).

So, to conclude:

At this point in time, it seems like you still need a Photoshop jockey
to manually etch out the tedious clip path, if quality and accuracy are
a concern (as I sincerely hope they are).

There are other thoughts on actually using clip paths that I can offer,
but these are best left for later.

At the moment you have to figure out what type of transparency you
really need.

Test, verify - then retest. Don't believe the hype.

Sincerely,

Stephen Marsh.


Re: scanning in RGB vs. CMYK: a major issue here!

Chris Murphy <lists@...>
 

The problem is as follow: several printers and publishers are saying
that they prefer if my company provide them the files as CMYK because
the files are easier to print. Also they do not care to give me the
output printer type or paper kind, so I do not know which profile
embed into my files. They keep say that CMYK is much better, because
with the K channel I get images with more vivid colors, more details,
better overall.
Translation of what printer is really saying: "I don't know what I'm
doing when it comes to RGB images and therefore I don't want to be
responsible for making disgusting awful separations. I want you to be
responsible for making disgusting awful separations. I know so little
about desktop color separation I can't even give you a hint as to what
settings to use in Photoshop to make a halfway decent separation for my
printing conditions."

What I would do if you are using Photoshop 6 is use the "U.S. Web Coated
(SWOP) v2" profile if you have no idea how things are going to print.
Otherwise you might try one of the coated or uncoated profiles in that
same list if you know in advance you will be printing to coated or
uncoated stock.

As for embedding profiles, I would recommend staying away from it when
dealing with outside printers who don't know color management let alone
how to make decent separations at the desktop level. An embedded profile
will not help them, so it won't help you. If you like them, use them for
your own purposes and save a no profile version for the printer/service
bureau.



Chris Murphy
Color Remedies (tm)
Boulder, CO
303-415-9932


blue/green screens

Andrew Engelhardt
 

Hi all, just wondering if anyone has any experience shooting or dealing with images that have been shot against a blue screen? We have to try to facilitate a high volume of single product images with clipping paths in a fairly fast turn-around time and we're looking at how to speed up the process of getting paths on everything. We're hoping that the blue or green screen background might help either Photoshop or another program like Mask Pro with creating a path quickly and accurately. These are files that will be repurposed for a bunch of different uses from all kinds of printing conditions to web use, thus the paths on just about everything. The only problem we can think of so far is when the colour of the product is close to that of the background. Any input or advice would be greatly appreciated.

Thanks,
Andrew Engelhardt
Digital Prepress & Imaging
Marketing Dept.
London Drugs Ltd.
(604) 272-7602
 



I have unsubscribed from this list through Listbot--

Julie Hyry <julie_hyry@...>
 

I unsubscribed from this list several weeks ago using
listbot, and haven't received any mail until today.
Please unsubscribe me!

__________________________________________________
Do You Yahoo!?
Get email at your own domain with Yahoo! Mail.
http://personal.mail.yahoo.com/?.refer=text


Sorry about the rant (was: scanning in RGB vs. CMYK: a major issue here!)

Terry Wyse <terry@...>
 

on 3/28/01 8:44 AM, Terry Wyse wrote:


Short rant:
It's high time printers get a clue and start "empowering" their customers to
give them correctly separated images. Whether it's ICC color management or
simply a proper "generic" PS CMYK setup doesn't matter. I get a bit weary of
hearing printers/prepress shops bitch and moan about the quality of the
images they're getting from their customers when they're providing them ZERO
help in improving the situation. The fact is, with "consumer" scanners
getting better/cheaper all the time and with the proliferation of digital
camera images, the number of supplied RGB images is likely to increase. So
HELP your customers get them right and they'll love you for it or you can
continue to tell them how lousy their images are and watch them take their
work to somebody who will spend some time working with them.
End of rant.

I just got scolded by the scanner supervisor of a large printer in the
Boston area where myself and my company helped them implement some color
management policies towards, we think, a very positive result. I was scolded
for lumping all prepress/printers in one "clueless on color management"
basket. My mistake. Naturally, there are those out there doing a good job
and being proactive with their clients in helping them supply good color. I
can tell you this, I believe those that either "get a clue" or hire outside
color consultants and "buy a clue" will see not only an increase in quality
but even an increase in customer loyalty (hard to come by these days). I've
seen this first hand.

I think I was a bit cranky because the last few days I've been working with
a photographic client that is running Corel PhotoPaint 9 on a Win98
platform. Not fun. I'd estimate that PhotoPaint 9 is about a
version-and-a-half behind Photoshop 6 in implementing color management. It's
been a struggle trying to get the results we expect.

So my apologies to those that I may have offended that have a clue and are
helping out their customers. To the REST of you, for shame..... :-)

Regards,
Terry "I'd like to buy a clue?" Wyse

_____________________________
Terence L. Wyse
PrePress Systems Specialist
All Systems Integration, Inc.
114 Cummings Park
Woburn, MA 01801
Main office:
781.935.3322 voice
781.935.6622 fax
http://www.allsystems.com
terry@allsystems.com
_____________________________


Re: scanning in RGB vs. CMYK: a major issue here!

Andrew Rodney <andrew@...>
 

on 3/28/01 6:54 AM, Dan Margulis at 76270.1033@... wrote:

Putting a profile in your RGB file does not solve the problem of how to get
the file into CMYK.

It helps in we need a description of the RGB to get to CMYK. That was true before ICC profiles (using a monitor preference in versions previous to Photoshop 5) and it’s true today. It’s of no use of she is going to do the conversions (the user will know the recipe of RGB to get to CMYK. Still embedding the profile in the original RGB file does no harm and actually does some good should a conversion need to be made in the future).

If the RGB file is given to another person, having a profile is quite useful. Without it, what assumption should be made about the recipe of RGB for the CMYK conversion? The fact that a good number of printers will get hostile seeing an embedded profile or not know what to do with a file with an embedded profile doesn’t change the facts that a description of RGB is always required to get to CMYK.

It must be understood that different printers will get different results
from the same CMYK file.

That’s an understatement! Yesterday I was showing a student just how varied a certain manufacturer of contract proofs could be from shop to shop. Since I have made dozens of profiles from the same proofing system (but at different printers all over the country), I can use ProfileMaker Pro’s measure tool to look at all 600 spectral readings from patches used to build the profiles and get some very useful stats. In best case, two completely different shops running the same proofing system (Matchprint) were showing an average deltaE of 4 which is quite good. But more often than not, the average deltaE was well over 10 and in some cases double that. Now consider sending the same CMYK file to shops using vastly different proofing systems and the numbers go sky high. A deltaE variance of 6 or less is said to be “acceptable” matching for most clients. Some would reject that much difference in colors. But a deltaE of over 10 means big surprise when that proof does come back!

Andrew Rodney


scanning in RGB vs. CMYK: a major issue here!

Dan Margulis <76270.1033@...>
 

Andrea writes,

the CreoScitex local site trainer is saying:
The scanner acquire the files in RGB
The RGB color space has more color info than CMYK
It is fine to acquire the files in CMYK but the CreoScitex trainer
say that it is not necessary>>

All this is approximately correct.

The problem is as follow: several printers and publishers are saying
that they prefer if my company provide them the files as CMYK because
the files are easier to print.>>

All files *must* be CMYK when they get to press, so if you do not provide
CMYK files, the printer must do the extra work of converting from RGB into
CMYK.

Also they do not care to give me the output printer type or paper kind,
so I do not know which profile embed into my files.>>

You don't need to embed any profile. But you do need to know how to make
the conversion between RGB and CMYK. For this it would be helpful if the
printer would give you more information, as otherwise you have to guess at
certain settings.

They keep say that CMYK is much better, because with the K channel I get
images with more vivid colors, more details, better overall.>>

I think that they are saying that if you are an expert color corrector,
CMYK offers certain opportunities to improve the image that aren't present
in RGB.

I am lost! Shall I indeed now go to CMYK or I can use a professional
system like GretagMacbeth color profile and spectrolino to apply the proper
profile to my RGB files?>>

Putting a profile in your RGB file does not solve the problem of how to get
the file into CMYK. It certainly doesn't sound like you should trust your
printer to make this conversion. Rather, you should learn how to configure
Photoshop to do it yourself.

The first question you must ask yourself is whether you will assume a
"generalized" CMYK or whether you are going to attempt to make different
separations for different printing conditions, which is much harder. Most
people assume a generalized CMYK, that is, they find a separation setting
that will create a Matchprint or Cromalin that is fairly accurate. Then
they separate all the files that way, send them to the printer and hope for
the best. This is in fact what you must do if you don't know who the
printer is going to be or what type of press or paper will be used.

To create your own setting, enter CMYK Setup (Photoshop 5.x) or Color
Settings: CMYK: Custom CMYK (Photoshop 6). Choose as your options
Eurostandard (coated), 14% dot gain, Black type: GCR, Black Generation:
Light; Maximum Black=85%; Total Ink=300%; UCA=0%.

With these settings loaded, convert a few files from RGB into CMYK and have
Cromalins made. Now you must judge how they look in comparison to what you
expected. If all look too dark, your dot gain setting was too low and must
be adjusted upward. If all look too light, the dot gain setting must be
reduced. There is also the possibility of adjusting the individual dot gain
curves in the event of color variation. If necessary you now pull a second
set of proofs to confirm your revised settings.

It must be understood that different printers will get different results
from the same CMYK file. If the file is printed on a sheetfed press there
will be less dot gain (a lighter result) than if printed on a web press. If
the file is processed in Italy, where it is customary to image with
positive film, the printing generally will be lighter than it would be in
the United States, where negative film is customary. If you know these
factors in advance, you can adjust for them during the separation. It may
be that you need several different separation settings, if you have
continuing relations with several different printers.

Dan Margulis

Dan Margulis


Re: scanning in RGB vs. CMYK: a major issue here!

Terry Wyse <terry@...>
 

on 3/28/01 4:34 AM, Andrea de Polo wrote:

Hello everyone,

I have a CreoScitex EverSmart Supreme scanner; I have already scanned
70.000 color and b/w images at 304 dpi, 2000x3000 pixels, RGB-TIFF.
Now we got the following scenario: the CreoScitex local site trainer
is saying:

The scanner acquire the files in RGB
The RGB color space has more color info than CMYK
It is fine to acquire the files in CMYK but the CreoScitex trainer
say that it is not necessary


The problem is as follow: several printers and publishers are saying
that they prefer if my company provide them the files as CMYK because
the files are easier to print. Also they do not care to give me the
output printer type or paper kind, so I do not know which profile
embed into my files. They keep say that CMYK is much better, because
with the K channel I get images with more vivid colors, more details,
better overall.
They're saying "more vivid colors, more details, better overall" ?? Well...
I suppose the argument could be made that a CMYK image will have better
shadow detail and possibly sharpness do to the addition of the K channel
("the K is the key!") but to say more vivid colors is just plain wrong in
most all cases.

I think your confusion Andrea is that it sounds to you, based on what you're
hearing, that you have an "option" to go either RGB or CMYK when providing
these files to your printer. The fact is that your images WILL end up in
CMYK at some point regardless. In the printing/prepress world, you simply
can't print RGB without a conversion to CMYK somewhere in the mix. The
question then becomes "WHEN do I convert from RGB to CMYK?".

All things being equal, I think you should consider capturing your image in
RGB from the scanner and performing as much image edits/color corrections as
you can in RGB while at the same time use Photoshop 6's very cool soft
proofing feature to preview the results in CMYK so you know what to expect
of the final results. When you think you've got it, either YOU perform the
RGB-to-CMYK transform (using an appropriate CMYK setup) or have the printer
do the conversion for you and provide you with a proof of the result (note:
give them the RGB image with an embedded profile). Either way, you will have
to know what is the "proper" CMYK setup or profile to use for this printer.
Without it, either you will convert it incorrectly or, even if they perform
the conversion, you may be looking at an incorrect soft proof during your
image editing of the RGB file.

So how do you get the correct CMYK setup?
In my opinion, the printer should either 1) provide you with an ICC profile
generated from their proof/press conditions or 2) at the very least give you
an idea of what they're expecting for UCR/GCR, total ink, dot gain and black
limit. If they can't at least give you that basic info then find somebody
else to work with.

I am lost! Shall I indeed now go to CMYK or I can use a professional
system like GretagMacbeth color profile and spectrolino to apply the
proper profile to my RGB files?
If you want to create your own profile(s), that's fine but I don't feel
anybody should have to, at their expense, create profiles for a specific
printer. If you do end up creating your own profiles (from what? a proof or
press sheets), I would NOT recommend giving these profiles to them without
compensating you for doing THEIR job. If they want to receive proper CMYK
seps from you, then they should give you the means to achieve that and if
that means they provide you with custom ICC profiles of their press
conditions, then so be it. If they can't/won't, give them the RGB and let
them take responsibility for the CMYK conversion.

Short rant:
It's high time printers get a clue and start "empowering" their customers to
give them correctly separated images. Whether it's ICC color management or
simply a proper "generic" PS CMYK setup doesn't matter. I get a bit weary of
hearing printers/prepress shops bitch and moan about the quality of the
images they're getting from their customers when they're providing them ZERO
help in improving the situation. The fact is, with "consumer" scanners
getting better/cheaper all the time and with the proliferation of digital
camera images, the number of supplied RGB images is likely to increase. So
HELP your customers get them right and they'll love you for it or you can
continue to tell them how lousy their images are and watch them take their
work to somebody who will spend some time working with them.
End of rant.

My $.03 worth (since I went a little long) :-)

Terry

_____________________________
Terence L. Wyse
PrePress Systems Specialist
All Systems Integration, Inc.
114 Cummings Park
Woburn, MA 01801
Main office:
781.935.3322 voice
781.935.6622 fax
http://www.allsystems.com
terry@allsystems.com
_____________________________


scanning in RGB vs. CMYK: a major issue here!

Andrea de Polo <andrea@...>
 

Hello everyone,

I have a CreoScitex EverSmart Supreme scanner; I have already scanned 70.000 color and b/w images at 304 dpi, 2000x3000 pixels, RGB-TIFF. Now we got the following scenario: the CreoScitex local site trainer is saying:

  • The scanner acquire the files in RGB
  • The RGB color space has more color info than CMYK
  • It is fine to acquire the files in CMYK but the CreoScitex trainer say that it is not necessary

The problem is as follow: several printers and publishers are saying that they prefer if my company provide them the files as CMYK because the files are easier to print. Also they do not care to give me the output printer type or paper kind, so I do not know which profile embed into my files. They keep say that CMYK is much better, because with the K channel I get images with more vivid colors, more details, better overall.

I am lost! Shall I indeed now go to CMYK or I can use a professional system like GretagMacbeth color profile and spectrolino to apply the proper profile to my RGB files?

Currently my files, without a proper profile look for b/W images bluish on a cromalin test and the color not proper for color images; TIA; Andrea
--
------------------------------------------------
Fratelli Alinari Photo Archives and Museum
http://www.alinari.com
The world's oldest picture library
tel: +39-055-2395201
fax: +39-055-2382857
A DCT (Digital Copyright Technologies) VAR company
http://www.dct-group.com
"A leading watermarking and security firm"
------------------------------------------------


Mixed Artificial Lighting

Chris Brown <cb@...>
 

I deal with lots of images from museums and art galleries. Often these
have very strong mixed casts which I find impossibly difficult to fix with
any sort of success rate. A typical recent example is a dark bronze statue
lit in a mixture of daylight, mercury halide and possibly something else.
Its cast is bright yellow in the top half, a strong magenta in the lower
half and some very blue highlights for good measure. The nature and
distribution of the various casts make even a selection, erm,
challenging.
The way I have drastically reduced or eliminated the effects of different
colored lights and various color temperatures in a scene is to:

1) Filter any lights, as best possible, to make the main subject in the
scene film-neutral (i.e. Rosco CC gels)
2) Filter at the lens to assist in acheiving step #1 (i.e. Wratten filters)
3) When scanning, eliminate any remaining color contamination (using the
scanning software) on the scene's main subject.

At this point there will be surrounding areas outside my main subject area
where the color and temperature of the light is not pleasing. I then use
Alpha channels and paths in PS to isolate these areas into groups of similar
contamination. Then, using Curves, Levels, and Selective Color, I eliminate
the unwanted colors.

Tedious but effective.

Chris
@
Chris Brown Photography
http://www.chrisbrownphoto.com
Vox: (217) 356-0540 * Fax: (217) 356-1394


Re: Mixed artificial lighting

David.Clark@...
 

Fuji NPS film is specifically designed with an extra dye layer to handle
mixed lighting conditions.

I'm looking forward to Dan's explanation on repair, but if you can get the
photographer to use NPS your life will be much simpler. I have shot
antique shows under mixed lighting that required almost no adjustment.


Mixed Artificial Lighting

Dan Margulis <76270.1033@...>
 

I deal with lots of images from museums and art galleries. Often these
have very strong mixed casts which I find impossibly difficult to fix with
any sort of success rate. A typical recent example is a dark bronze statue
lit in a mixture of daylight, mercury halide and possibly something else.
Its cast is bright yellow in the top half, a strong magenta in the lower
half and some very blue highlights for good measure. The nature and
distribution of the various casts make even a selection, erm,
challenging.>>

You are on the right track by realizing that a selection is necessary. This
type of work is not like altering the general appearance or lighting, which
can be done with curves. But here you are taking something with two
different light sources and making them appear to merge into one. This is a
move away from the art, just as much as if you had decided to change the
statue from bronze to silver. The only thing is, it's easy to select the
statue, but very difficult to select the cast area unless you know the
secret.

Assuming that there are only two casts, the way to tackle this is to
correct for one of them by means of curves. Then load a mask for the area
affected by the second cast, and desaturate the area with Adjust
Hue/Saturation.

That mask can be made easily from the A or, more commonly, B channel of
LAB. For example, suppose you are attempting to eliminate a blue cast in a
certain area.

1) Make a copy of the image, and convert the copy to LAB. Make a copy of
the B channel separately and throw the LAB away.

2) This B channel will look like a gray blur, but the cast area will be
marginally darker than its surroundings. Apply a very steep straight line
curve to the channel, trying to darken the cast area and blow everything
else out. You will probably need to apply curves twice to do this, but the
result should be a white background and a dark gray gradation where there's
a cast.

3) In the event that there's something in the image that's *supposed* to be
blue, erase it from the channel.

4) Apply a Gaussian blur to further soften the edges.

5) Return to the original image, and with Select: Load Selection, load this
other channel. In this particular case, where the area you want to select
is black in the mask rather than white, you'll need to click the "Invert"
box in the Load Selection dialog.

6) With the mask loaded, you can desaturate or apply curves or selective
color, depending on the character of the iamge.

Dan Margulis


Mixed Artificial Lighting

Dan Margulis <76270.1033@...>
 

Folks,

Because this is a complicated color problem that comes up from time to
time, especially for images shot on location, I forward this question,
which did not come from a list member. My response will be in a separate
post.

Dan Margulis

--------------- Forwarded Message ---------------

To: Dan Margulis, 76270,1033
Date: Mon, Mar 19, 2001, 6:37 AM

RE: Mixed Artificial Lighting

Hi,

I have "Professional Photoshop 6" which I have found incredibly useful and
informative and for which I thank and congratulate you.

Something I would really like a few pointers on though. I deal with lots of
images from museums and art galleries. Often these have very strong mixed
casts which I find impossibly difficult to fix with any sort of success
rate. A typical recent example is a dark bronze statue lit in a mixture of
daylight, mercury halide and possibly something else. Its cast is bright
yellow in the top half, a strong magenta in the lower half and some very
blue highlights for good measure. The nature and distribution of the
various casts make even a selection, er, challenging. Mercury Halide in
particular seems impossible as it doesn't seem to me to be linear.

I realise you must be very busy and are probably plagued by such requests,
but any advice or links (or even which sections of your book I should read
again) on the subject of strong mixed casts would be greatly appreciated.

Many thanks,