Dan Margulis Applied Color Theory

Scanning Color Negatives

Scanning color negatives
Posted by: "Chris Brown"
Fri Dec 15, 2006 9:32 pm (PST)

Hi all,

It's been a while and I hope no one minds my inquiry.

Is it better to tag a scan of a color negative with the ICC profile of the scanner or not?

I've fallen into a conundrum about this because at some point the file must be tagged as it moves through the color managed tool of Photoshop and on to getting separated, printed by an inkjet printer or displayed on someone's monitor.

I know that neg films benefit from a scanner's profile, but if the image is not tagged with the scanner profile, what is it tagged with?

Thanks,
Chris Brown
___________________________________________________________________________

Re: Scanning color negatives
Posted by: "J Walton"
Sat Dec 16, 2006 5:26 am (PST)

On 12/15/06, Chris Brown wrote:

Is it better to tag a scan of a color negative with the ICC profile of the scanner or not?

I have found that ICC profiles are of little use when dealing with color negatives. Perhaps some others on the list have another experience.

Scanning color negs is, to me, a great exercise in color correction technique because you really have to think about how the image is supposed to look. I think Dan's approach to color is by far the best when scanning color negs.

I've fallen into a conundrum about this because at some point the file must be tagged as it
moves through the color managed tool of Photoshop and on to getting separated, printed by an inkjet printer or displayed on someone's monitor.

I'd say you tag it with your working space after making it look good on screen.

-----
J Walton
___________________________________________________________________________

Re: Scanning color negatives
Posted by: "Michael King"
Sat Dec 16, 2006 6:24 am (PST)

Color negatives are also a real challange because they use a very narrow part of the full color/tone range of the scanner because of the base film color. The ideal approach is to build a scanner profile for each particular negative film.

DPL from Azek allows you to do that (custom profile not ICC). It also reprograms the Howtek/ Aztek scanner dacs to use the full 12 bits over just the tone range of the negative.

Its an example of a great solution to a difficult problem. But its only for Howtek/Aztek scanners.

Mike King

PS I have relationship with Aztek other than being a customer- its just a
great solution to this problem.
___________________________________________________________________________

Re: Scanning color negatives
Posted by: "MARK SEGAL"  
Sat Dec 16, 2006 8:31 am (PST)

Scanners cannot be profiled for colour negatives. Buy Silverfast Ai Studio 6.5 version for your scanner, use Negafix specified and adjusted best possible for your film stock and specify an RGB(98) colour working space at the scan stage. Open the image in Photoshop and convert to whatever colour working space you prefer using.

I wrote a couple of articles on this whole process which you will find on www.Luminous-Landscape.com. You can do a search there using my name - I don't have the specific links on hand.

Mark Segal
___________________________________________________________________________

Re: Scanning color negatives
Posted by: "Chris Brown"
Sat Dec 16, 2006 8:32 am (PST)

Michael King wrote:

Color negatives are also a real challange because they use a very narrow
part of the full color/tone range of the scanner because of the base film
color. The ideal approach is to build a scanner profile for each particular
negative film.

By scanner profile you mean software settings. Correct?

As far as scanning the narrow range of a neg film, I agree that some cheaper scanners are not as sensitive as, say, drum scanners. I'm fully aware of the ability of a drum scanner's higher sensitivity to a film's subtle changes in density levels, and thus produce better scans.

My question is in regards to tagging a scan of neg film with an ICC profile. Ever since Photoshop v5, we have had to work within the boundaries of a color managed workflow (and I do remember Margulis' rants about this - quite educational and fun to read). This means files coming through Photoshop require an ICC profile for best results when viewed on a monitor or printed on a modern inkjet printer (they all support ICC profile use, too).

If I am to use this color managed workflow, I need a scan tagged with an ICC profile. The use of negative films, which don't lend themselves to being tagged, can pose a problem in this workflow.

~ Chris Brown
___________________________________________________________________________

Re: Scanning color negatives
Posted by: "Chris Brown"
Sat Dec 16, 2006 8:35 am (PST)

J Walton,

Thanks for your response. I have a few more comments/questions.

Scanning color negs is, to me, a great exercise in color correction
technique because you really have to think about how the image is
supposed to look. I think Dan's approach to color is by far the best
when scanning color negs.

Hmm. My book is at my studio. What does he recommend? Tag it and run? ;-D

I'd say you tag it with your working space after making it look good on screen.

But when I tag a scan of a neg with an arbitrary colorspace, there's no way of knowing what I'm actually doing to the image/colors. At least when I tag it with my scanner profile, I have a known starting point for reference.

I've asked others and most everyone tags their neg scans with an arbritrary working space (I'm in the minority by using my scanner profile). Some use ProPhoto, some use Adobe '98, etc. This has no relation to the film gamut or the input device.

One of my questions is, what happens to the reference white point? In my scanner profile I have a calculated white point based on the physical properties of my scanner. With a color space such as ProPhoto, I have a white point that's different and levels of saturation that exceed anything the film can reproduce.

Even though my scanner profile is based on transparency film, isn't that profile better than an arbitrary one?

~ Chris Brown
___________________________________________________________________________

Re: Scanning color negatives
Posted by: "Les De Moss"
Sat Dec 16, 2006 8:36 am (PST)

Hummmm. We typically convert to our working space upon opening a new (unprofiled) drum scan in Photoshop. Why would one choose to tag after the color correction as opposed to before? Is there an underlying difference according to the sequence of color correction and conversion to working space?

Les De Moss
DigiGraphics LLC._,___
___________________________________________________________________________

Scanning color negatives
Posted by: "Andrew Rodney"
Sat Dec 16, 2006 8:36 am (PST)

On 12/15/06 10:14 PM, "Chris Brown" wrote:

Is it better to tag a scan of a color negative with the ICC profile of the
scanner or not?

If it©ˆs a good profile producing preferable color appearance, yes. Now technically there are no true ICC profiles (input or scanner profiles) for negs although every now an then you see one in the software package (example, the Flextight Input profile for Imacon scanners). This is a setting used to make a scan and tells the software how to convert from one color space (probably a native scanner space and TRC) to a working space. Fuji had a system of scanning a supplied target to build a neg with their higher end scanners years ago, worked pretty well too. Depends on the scanner driver. One of the best is SilverFast. Anyway, use the system as intended, if you get color appearance you like, use the profile as a tag or just convert into a well behaved RGB working space.

Andrew Rodney
___________________________________________________________________________

Re: Scanning color negatives
Posted by: "john castronovo"
Sat Dec 16, 2006 8:36 am (PST)

From: "J Walton"

I have found that ICC profiles are of little use when dealing with
color negatives. Perhaps some others on the list have another
experience.

Yes, the scanner profile is useless with color negs, so you should scan raw and deal with it in Photoshop by inverting, neutralizing the base, setting white and shadow points, neutralizing the color, adding an 's' curve for good contrast, etc. Doo all this through the use of adjustment layers so you can play with it until it's good. You may also need to address the issue of grain reduction.

Scanning color negs is, to me, a great exercise in color correction
technique because you really have to think about how the image is
supposed to look. I think Dan's approach to color is by far the best
when scanning color negs.

This is what professional photo labs always had to do with color negatives - interpret them to make them look good. There's no absolutely correct way to interpret a neg, so it's all very subjective.

I'd say you tag it with your working space after making it look good
on screen.

Yes, you MUST tag it with your working space, assuming it's still RGB, or else convert it for hard proofing and printing.

john
___________________________________________________________________________

Re: Scanning color negatives
Posted by: "Les De Moss"
Sat Dec 16, 2006 8:56 am (PST)
-
From: J Walton

I have found that ICC profiles are of little use when dealing with
color negatives. Perhaps some others on the list have another
experience.

Yes. First, there are so many different film types including legacy films dating back to who-knows-when, second, processing variation, and third, the age of the film. All conspire against a comprehensive library of accurate profiles. Add to that the decreasing use of neg film and subsequent need for drum scans - there is little reason to devote time and energy to this.

Scanning color negs is, to me, a great exercise in color correction
technique because you really have to think about how the image is
supposed to look. I think Dan's approach to color is by far the best
when scanning color negs.

In experimenting over the years, we have found we achieve the best results by scanning color negatives as positives, setting the 'white' point on an unexposed area of the base, and later inverting in Photoshop for color correction. I don't know if this is technically the best way, or what hidden costs may be involved, but avoiding the scanner software in making the conversion from negative to positive image seems to work best for us.

Les De Moss
DigiGraphics
___________________________________________________________________________

Re: Scanning color negatives
Posted by: "MARK SEGAL"  
Sat Dec 16, 2006 12:15 pm (PST)

As usual, different things work better for different folks. I find the process you describe is very difficult to work with. These inversions, whether done in the scanner or in Photoshop (using LAB curves) are never completely successful and they require alot of after-the-fact fiddling with curves to get the colors "right". When you use a first-class piece of software such as Silverfast it does all the work for you very well once you climb down its learning curve.

Mark Segal
___________________________________________________________________________

Re: Scanning color negatives
Posted by: "Andrew Rodney"
Sat Dec 16, 2006 12:15 pm (PST)

On 12/16/06 8:07 AM, "Chris Brown" wrote:

But when I tag a scan of a neg with an arbitrary colorspace, there's no way of
knowing what I'm actually doing to the image/colors. At least when I tag it with my
scanner profile, I have a known starting point for reference.

You know by looking at the image based on this description of the numbers (all the profile does is provide a scale, a definition of the current fixed numbers). Do you like the color appearance? If so, tag and move on. I©ˆm using this workflow in the case where the scanner is removing the orange mask and inverting the image. You could of course scan the neg as a neg and try the inversion in Photoshop. It©ˆs a lot of work that is often not necessary when the driver does the conversions. Again, SilverFast and a few other products do a very good job of this. But you have to look at the initial rendering and decide if you like it or not. There©ˆs no way to produce a colorimetrically correct version of a color neg.

I've asked others and most everyone tags their neg scans with an arbritrary
working space (I'm in the minority by using my scanner profile). Some use ProPhoto, some use Adobe '98, etc. This has no relation to the film gamut or the input device.

You can©ˆt (shouldn©ˆt) tag a file as ProPhoto RGB, sRGB or any other color space unless that©ˆs either the correct description for the numbers or that tagging for whatever reason produces a desirable color appearance on a calibrated display.

One of my questions is, what happens to the reference white point? In my scanner profile I have a calculated white point based on the physical properties of my scanner.
With a color space such as ProPhoto, I have a white point that's different and levels of
saturation that exceed anything the film can reproduce.

You have to convert from input color space (of which the profile is supposed to define) to the working space. Then it©ˆs all fine. There isn©ˆt a scanner or camera on this planet that can produce ProPhoto RGB (or for that matter, sRGB or any other of the well know synthetic RGB working spaces. These are all mathematically constructed color spaces and done for the reason we don©ˆt want editing spaces that are based on any real world, specific device).

Even though my scanner profile is based on transparency film, isn't that
profile better than an arbitrary one?

No. The transparency profile is based on the behavior of the scanner when using this type of film stock. Profiles only define device behavior anyway. Whenever you deviate from the conditions of the scanner, that profile is invalid. In theory you could use a transparency scanner profile on a neg but the results would be a neg (orange mask, inverted) which doesn©ˆt buy you anything. Think of a color neg as more similar to a Raw file where you control the color rendering. The profile for this so called raw file doesn©ˆt affect anything but a default rendering which you may or may not like.

On 12/16/06 8:18 AM, "john castronovo" wrote:

Yes, the scanner profile is useless with color negs, so you should scan
raw and deal with it in Photoshop by inverting, neutralizing the base,
setting white and shadow points, neutralizing the color, adding an 's'
curve for good contrast, etc. Doo all this through the use of adjustment
layers so you can play with it until it's good. You may also need to
address the issue of grain reduction.

Way too much work. Now if you don©ˆt have adequate scanning software and do a few negs a year, OK but a far better solution is to handle this with good software that can invert the image, remove the orange mask and produce a desired color rendering. There are plenty of solutions for doing this.

This is what professional photo labs always had to do with color
negatives - interpret them to make them look good. There's no absolutely
correct way to interpret a neg, so it's all very subjective.

Exactly what we encounter with Raw files...

Andrew Rodney
___________________________________________________________________________

Re: Scanning color negatives
Posted by: "john castronovo"
Sun Dec 17, 2006 4:20 am (PST)

Not too much work at all for a high quality scan. We use an action that inverts the image and sets up the layers for us. After a few clicks which are unique to each negative, we're done except for the fine tuning which would be needed no matter how you do it. We drum scan a lot of negs this way. Coming from the background of a professional photo lab, I know all too well that I don't want to depend on canned negative interpretation. Our inexpensive CCD scans are all done using high speed software however, but that doesn't give us the accuracy or control of what I'm talking about here.

Due to the wide variability of exposure conditions, film types and processing, color negatives require a great deal more flexibility in the printing or scanning than when scanning transparencies or prints. Aside from the basics of inverting and neutralizing the base, there isn't much that one can depend upon, and if you don't know the original subject colors, you're pretty much lost and looking for clues.

There's no way to tell, for example, if you're looking at a white box on a gray background, or a tan box on a brown one. Possibly it was shot with mixed tungsten light on daylight film and the usual numbers for that film type are all wrong too. Ultimately, you're left making subjective judgments on the display, hopefully in adjustment layers so you can go back for corrections if you're wrong about your assumptions. If you commit to those assumptions in the scanner software, you'll have a harder time of it later on.

john castronovo
___________________________________________________________________________

Scanning Color Negatives
Posted by: "Raymond E. McKinley"
Sat Dec 16, 2006 12:11 pm (PST)

Chris

It seems that you are confused about the benefits of a scanner profile. A good profile simply substantially reduces the inherent color cast of the scanner. This allows you to spend time tweaking your image rather than waste time trying to bring the appearance close to that of the lightbox.

The profile doesn't increase the gamut of your scanner. As the other posters have remarked because of the wide range of colors of each negative base, profiling negatives doesn't work, since you would have to generate a profile for each emulsion and I have only read of one case where someone claimed that this method worked using EZ Color. However when I e-mailed the author asking for more info or evidence that this method worked I received no reply.

Using Nikon Scan to scan my negatives, I've gotten good results using what I've learn't from Dan and others regarding color correction. I scan the negatives and correct the images in Photoshop just like any other images. This works well for Kodak and Fuji negatives. I've had to use the negative as a positive method for some Agfa negatives which were processed by a lab whose machines seemed to be set up for Kodak and Fuji.

Regards

Raymond McKinley
___________________________________________________________________________

Re: Scanning color negatives
Posted by: "Chris Brown"
Sun Dec 17, 2006 10:52 am (PST)

I appreciate all the informative responses here. I hope y'all don't mind if I dig in with more questions.

John Castronovo wrote:

We use an action that
inverts the image and sets up the layers for us. After a few clicks
which are unique to each negative, we're done except for the fine tuning
which would be needed no matter how you do it.

So when do you assign your profile? Before or after corrections? My other question is what profile do you use and do you use different profiles based on your initial scanning results?

Andrew Rodney wrote:

You know by looking at the image based on this description of the numbers
(all the profile does is provide a scale, a definition of the current fixed
numbers). Do you like the color appearance? If so, tag and move on.

Okay, so the profile applied is relevant only in terms of how the RGB values are interpreted by the ICC profile and how it looks on a profiled monitor. This seem pretty arbitrary, but I am beginning to understand why: a color neg has no definitive color balance.

This brings another question to mind: why not apply this work flow to scanning transparencies? The HCT target is from Fuji Velvia (I think) and Kodak IT8 target is Ektachrome E100 (I think). If I'm scanning a film that's not either of these film types, doesn't the above logic permit me to scan without using a profile then tweak in Photoshop according to the profile I wish to apply?

~ Chris Brown
___________________________________________________________________________

Re: Scanning color negatives
Posted by: "Andrew Rodney"
Sun Dec 17, 2006 11:27 am (PST)

On 12/17/06 10:41 AM, "Chris Brown" wrote:

Okay, so the profile applied is relevant only in terms of how the RGB values
are interpreted by the ICC profile and how it looks on a profiled monitor. This seem pretty arbitrary, but I am beginning to understand why: a color neg has no definitive color balance.

Not quite. The scanner is producing a unique mix RGB of numbers to define whatever color it Œthinks©ˆ you asked for visually or numerically. Those numbers alone don©ˆt define color appearance anymore than a recipe for key lime pie with the ingredients but not the amount of ingredients allow you to define that recipe. ICC profiles simply give the numbers a scale within human vision. The profile defines device behavior which in turn defines its color space and how each color number should appear. R255 is the most numerically saturated red you can define but where does that red fall within human vision? Adobe RGB (1998), sRGB and ChrisScannerRGB all define that max mix of red using the same numbers. The colors are not the same! So by defining where R255 falls in terms of ChrisSannerRGB is necessary. This info is used to produce the preview in your scanning software and in Photoshop as well.

This brings another question to mind: why not apply this work flow to scanning
transparencies? The HCT target is from Fuji Velvia (I think) and Kodak IT8
target is Ektachrome E100 (I think). If I'm scanning a film that's not either of these
film types, doesn't the above logic permit me to scan without using a profile then tweak
in Photoshop according to the profile I wish to apply?

The film has a characteristic look as well as gamut. Ideally you©ˆd build a transparency profile for E6 and Kodachrome using in the case of E6, the widest gamut stock (Velvia which is what Don Hutchinson recommends). You still need a profile to define the device specific numbers that the scanner is producing. Ideally you©ˆd have a profile for each film stock but that©ˆs not necessary if your goal is pleasing color rather than a closer colorimetric match of the original film stock.

The profile is a road map to the numbers you get as well as the preview you see in the scanning software. It gives the numbers a full definition. Otherwise, you only have a partial recipe for color.

Andrew Rodney
___________________________________________________________________________

Re: Scanning color negatives
Posted by: "Richard Wagner"
Sun Dec 17, 2006 2:00 pm (PST)

Chris,

Don't forget that a scanner profile is a device profile - it is not a simple matrix-based profile like a working space (colorspace) profile. My scanner profile made from the 528-patch HCT Velvia target uses a 16-bit table with 25 grid points. Each of the 3 tone curves has 511 data points. It is very specific to my scanner (and Velvia film) - a device profile characterizes the scanner/film - it should not be confused with a working space profile. Fortunately, the differences in dyes between different E6 films are not that great, and the profile works very well for most E6 film. Kodachrome is another matter altogether - applying the custom profile will *not* give good color correction to Kodachrome. Don Hutchison has a good description of how you can make a profile for Kodachrome from the Hutch target and a handful of good Kodachrome slides, but the method is empiric, not colorimetric. <http: //www.hutchcolor.com/PDF/Kodachrome_profiles.pdf>

Sure, you can scan transparencies without a profile, then "tweak" in PS, but why? If we use a custom profile made from the Hutch target, our color is nearly dead-on, and only minor tweaks to levels/curves are needed. I can't imagine trying to color-correct starting from the raw scan without a scanner profile - that would be ugly. Our workflow is to make zero adjustments with the scanner software, batch scan, assign the custom profile, convert to a working space, and make very minor tweaks. I don't think it gets much easier than that.

--Rich Wagner
___________________________________________________________________________

Re: Scanning color negatives
Posted by: "john castronovo"
Sun Dec 17, 2006 4:40 pm (PST)

From: "Chris Brown"

So when do you assign your profile? Before or after corrections? My other question is what profile do you use and do you use different profiles based on your
initial scanning results?

The working profile is in use the minute we're in Photoshop. You can't help that, so you might as well assign it. Whether it should be Adobe or sRGB is a question I'd rather not get into right now. <g>

This brings another question to mind: why not apply this work flow to scanning
transparencies? The HCT target is from Fuji Velvia (I think) and Kodak IT8 target is
Ektachrome E100 (I think). If I'm scanning a film that's not either of these film types,
doesn't the above logic permit me to scan without using a profile then tweak in Photoshop
according to the profile I wish to apply?

This work flow cold be used for transparencies, but it's unnecessary because a good scanner profile will do most of the work for you. The HCT started out on Velvia 50 but some HCT targets have also been made E100G. You can ask for either, but since the Velvia 50 is out of production, there haven't been any made on Velvia recently. I understand that Fuji is going to reintroduce it soon. Personally, I've only used the Fuji targets and we have no problems scanning anything using the resulting profiles so I don't think you need to buy both types. The purpose of the profile is to first characterize the response of the scanner and to a lesser extent the film. Kodachromes use a very different dye set. But none of this will help with a negative unless you want your scan to look just like the neg.

john castronovo
___________________________________________________________________________

Re: Scanning color negatives
Posted by: "J Walton"
Mon Dec 18, 2006 12:06 am (PST)

On 12/17/06, Chris Brown wrote:

I appreciate all the informative responses here. I hope y'all don't mind if I dig in with more
questions.

Not at all!

So when do you assign your profile? Before or after corrections? My other question is what profile do you use and do you use different profiles based on your initial scanning results?

Chris, I think you are confused about the purpose of a profile. It's NOT something you assign to make the image look better. It merely describes the conditions of your scanner scanning that kind of film.

Here's how it would work in a transparency scan (on a properly-profiled scanner):
1. SCAN without any corrections (auto or otherwise) on the scanner.
2. ASSIGN your scanner profile for transparencies. This merely describes the scan you have on your screen, nothing more.
3. CONVERT to your working space (AdobeRGB, sRGB, ProPhotoRGB...)
4. COLOR CORRECT by the numbers, just like Dan taught you.

Here's how it works when scanning a negative (if you're not correcting the image with scanning software):
1. SCAN without any corrections. Since you don't have a negative profile, there's nothing to assign to this scan.
2. OPEN in Photoshop, Invert, and COLOR CORRECT by the numbers, just like Dan taught you. When you finish your corrections and save the file, it will already be in your default working space (AdobeRGB, sRGB, ProPhotoRGB)

Why are the steps different? Because you really can't ICC profile a negative scan - it just doesn't work very well. Since you have nothing to describe those scanning conditions, there's nothing to tag the image with. All you can do is color correct it to look good.

Some have software they like to use to do these corrections. My boss (for whatever reason) likes the Imacom software, others swear by SilverFast (I think that's it). Others like to do the corrections in Photoshop.

Okay, so the profile applied is relevant only in terms of how the RGB values are interpreted by the ICC profile and how it looks on a profiled monitor.

NO. The profile is relevant in that it knows that way your scanner acts with a certain kind of transparency film, and is able to describe the image you just scanned in that context. Does your scanner tend to make things too purple and way too dark? Then your scanner profile, when APPLIED to the raw scan, describes that condition. After being applied it should look like the original transparency on a good monitor.

Now when you CONVERT to an RGB working space, or to a final output profile, you have some meaningful numbers with which to do this conversion. In a sense the profile makes the image yellower and lighter because it knows your scanner makes things purple and dark. (IN A SENSE)

This seem pretty arbitrary, but I
am beginning to understand why: a color neg has no definitive color balance.

I wouldn't put it that way, but I guess that's correct.

This brings another question to mind: why not apply this work flow to scanning
transparencies?

The question is backwards. You use profiles with transparency scans BECAUSE YOU CAN. You correct things on-screen with negative scans because profiles don't work and that's all you're left with.

If we could get negs to profile like trannies, any sane scanner operator would use the profile for negs.

The HCT target is from Fuji Velvia (I think) and Kodak IT8 target is
Ektachrome E100 (I think). If I'm scanning a film that's not either of these film types,
doesn't the above logic permit me to scan without using a profile then tweak in Photoshop
according to the profile I wish to apply?

You are permitted to do whatever you like, of course. Having scanned hundreds of Kodak transparencies after profiling our scanner with the Fuji Hutch target I can say this:

USE THE PROFILE! (Yes, that was shouting!)

I don't know if Tom Thompson will see this post, and if he does he'll have his own thoughts, but I thought I was a pretty good drum scanner operator. I looked at the CMYK numbers the scanner gave me and did my best to make the scan match the original.

I couldn't believe how good the scans turned out after Don Hutchinson got us going with scanner profiles. Yes, different film material had slightly different results, but all of those results were closer than some goofball (me) guessing at what the numbers should be.

I think at this point what will help you the most is to truly understand what profiles do and what they don't do.

-----
J Walton
___________________________________________________________________________

Re: Scanning color negatives
Posted by: "Stephen Marsh"
Mon Dec 18, 2006 4:26 am (PST)

J, as you go on to describe, there is no ICC profile to assign to the neg. This is where Chris has been coming from since the OP. The consensus view appears to be to make the image look right on a characterized monitor so that the *edited* files numbers when coupled with the working space provide a pleasing result (it is a given that the unedited numbers may not result in ideal LAB numbers when coupled with the RGB WS profile). So one can tag it with working RGB, despite working RGB being known to be `technically incorrect' when opening and then edit away so that the tag is now technically correct. Or one could open and choose to not colour manage, which results in the working RGB being assumed behind the scenes and not actively chosen up front by assigning (ensure that the final edited file is saved with the working RGB profile when saving).

Obviously the assigned/assumed working space will have an effect on the raw negative scan, assigning Apple RGB would make the image less saturated and brighter than would assigning or assuming Adobe RGB, which would result in a darker, more saturated image (presuming that the raw neg scan had similar behaviour to sRGB for the example at hand in describing the effect of assigning different profiles).

So I think the real question is "what are the consequences of assigning a working space profile to a neg scan before editing?". This obviously depends on the files numbers and the profile. The only way for Chris to really be happy is try a few and see for himself. Another question could be "Is there a good 'dummy' profile that one can assign to a neg scan?". See the previous answer.

I think these questions are easier to solve once the image has been inverted and the film base removed.

I would agree with you that the textbook purpose of an ICC profile is not something one just assigns at random to make an image look better (or to give the files numbers a better LAB value) – but it depends on which textbook you are reading! <g>

http://tinyurl.com/u9ryy (Fate and the False Profile)

Chris may wish to experiment with assigning various standard or custom made working space profiles to the image, once it has been inverted and the film base removed (if not using an in-scanner solution that provides CM results). The evaluations are visual and via info palette LAB or CMYK readings for key and memory colours etc. Once happy, one would convert to working RGB (or LAB for big moves) and perform the real edits and then save with the working RGB tag.

Sincerely,

Stephen Marsh.