Ad
Ad
Ad
Pages: [1] 2 3 4 »   Bottom of Page
Print
Author Topic: Colormanaging scanned negatives - How to do best ?  (Read 6814 times)
Christoph C. Feldhaim
Sr. Member
****
Offline Offline

Posts: 2508


There is no rule! No - wait ...


« on: February 18, 2012, 04:40:14 PM »
ReplyReply

I want to profile the films I am using and wonder what would be the best method.
Which role plays having a better target ?
I have only a Colorchecker passport - doesn't work creating DNG profiles for TIFFs, since these can't be assigned in LR.

When I scan I am using a simple standard procedure to get positives from my negatives - no adjustments - just a repeatable set of adjustments using Silverfasts "Negafix" function and infrared based dust and scratch removal. So - getting a positive from a negative I have standardized.

I'd like to create an ICC profile I could just assign in PS to have better colors after getting the positive file from the scanned negative.
Just for fun I tried that using Argyll CMS and my colorchecker,
but the profile was crap (The target was taken in overcast weather and not totally protected from reflections).

I think I pretty much understood the basics of colormanagement and
would like to have an as simple as possible workflow to create profiles for my scanned negative films,
but I don't have experience with creating profiles.

Ideas that don't cost tons of money?

Should I use a proper IT8 or Q60 target?
Should use a specific light when taking the image of the target like D50?

Thanks
~Chris
Logged

Mark D Segal
Contributor
Sr. Member
*
Offline Offline

Posts: 6818


WWW
« Reply #1 on: February 18, 2012, 04:49:42 PM »
ReplyReply

Hi Christophe,

I did an extensive tutorial on this subject published on this website here: http://www.luminous-landscape.com/techniques/scanning_colour_negatives_raw_or_not.shtml

It predates SilverFast 8 but the principles are the same and you can automate some aspects of it. Don't know whether this is the kind of thing you want, but have a look. Basically I've come to the conclusion that customizing NegaFix is the easiest and best way of doing this.

Cheers,

Mark
Logged

Mark D Segal (formerly MarkDS)
Author: "Scanning Workflows with SilverFast 8....." http://www.luminous-landscape.com/reviews/film/scanning_workflows_with_silverfast_8.shtml
Christoph C. Feldhaim
Sr. Member
****
Offline Offline

Posts: 2508


There is no rule! No - wait ...


« Reply #2 on: February 18, 2012, 05:11:41 PM »
ReplyReply

I have read and understood your article - awesome tutorial by the way...

I have one additional question:
How exactly do you set your CMS settings in SF ?

My impression is, they apply the scanner specific profile as a means to get correct input data, but have no solution to correctly render ALL colors by getting film specific.
Of course I tweak and try to optimize the look of an image with the various tools, including the orange mask method you describe.
But how about rendition of the different colors by different films?
ProPhoto does not work in SF because it seems, that there is not a real conversion to ProPhoto but just an assignment.
When I assign the native scanner profile created by the Auto IT8 process my colors become much better.
Adobe RGB works only better, I believe, because its gamut is much nearer to the native scanner gamut.

So what I do is to use the scanner profile for output and I convert (as opposed to assign) to ProPhoto in PS.

What is not solved at this point is the different color rendition of the various films.
I'd like to have the opportunity to eliminate that from the beginning by using a specially tailored profile for each film type.

Logged

Mark D Segal
Contributor
Sr. Member
*
Offline Offline

Posts: 6818


WWW
« Reply #3 on: February 18, 2012, 05:33:08 PM »
ReplyReply

Thanks Christophe, glad you enjoyed the tutorial.

Scanner profiles don't assure accuracy of colour rendition for colour negatives because the film can't really be profiled with icc profiles in the same way positive transparency film/scanner combinations can be. That is why they developed Negafix. The fact is, that you need to use a separate Negafix profile for each colour negative film type you are scanning. And this goes not only for the type, but also for the ISO sensitivity within each type, insofar as - for example - a Kodak Gold 200 will not render colour the same way as a Kodak Gold 400. If you are using SF Ai Studio, there is an Expert Mode which allows you a great deal of customization potential to fine-tune the Negafix profile very correctly.

The most workable combination of settings in CMS I've found is to use your scanner profile as the Input profile, and use Adobe RGB(98) for the Colour Working Space for the reason you mention. OR, you can scan with no profile embedded by chosing one of the HDR modes (and going into CMS and putting <none> everywhere, in General Preferences making) sure your Gamma is 2.2 and check the box "for HDR output" if you are making an HDR scan). Scan with Negafix working, then when you open the image in Photoshop you will get the missing profile warning. At this point I would Assign the scanner profile and Convert to the Working Space - selecting ARG(98) in this instance. 
Logged

Mark D Segal (formerly MarkDS)
Author: "Scanning Workflows with SilverFast 8....." http://www.luminous-landscape.com/reviews/film/scanning_workflows_with_silverfast_8.shtml
Schewe
Sr. Member
****
Offline Offline

Posts: 5415


WWW
« Reply #4 on: February 18, 2012, 06:09:42 PM »
ReplyReply

Scanner profiles don't assure accuracy of colour rendition for colour negatives because the film can't really be profiled with icc profiles in the same way positive transparency film/scanner combinations can be.

Mark is correct...E-6 chrome processing tended to be very, very consistent (unless the lab screwed up) so an Ektachrome Q60 was a really good way of profiling the E-6 rendering. C-41 color neg processing tended to be all over the map and stuff varied greatly by film type. Even with the same film type, film processed in one lab would not be at all the same colorimetrically as another lab. That's why you can't really create an ICC profile for scanning negs.
Logged
Christoph C. Feldhaim
Sr. Member
****
Offline Offline

Posts: 2508


There is no rule! No - wait ...


« Reply #5 on: February 18, 2012, 06:23:17 PM »
ReplyReply

Mark is correct...E-6 chrome processing tended to be very, very consistent (unless the lab screwed up) so an Ektachrome Q60 was a really good way of profiling the E-6 rendering. C-41 color neg processing tended to be all over the map and stuff varied greatly by film type. Even with the same film type, film processed in one lab would not be at all the same colorimetrically as another lab. That's why you can't really create an ICC profile for scanning negs.

I had hoped to just assign a film specific tailored ICC profile once SF has produced a positive from the negative using Negafix.
This should work, I thought, since the file is a positive at that stage.
But if negative film is so variant it surely doesn't make sense.

I just wonder if the orange masks vary as well - or would it be sufficient to have one mask file per film type so I could automate the process in PS?
Because if they don't vary but the emulsion does any idea about consistent color with negative film would be futile.
Or I would need to shoot a greycard / Colorchecker in D50 (or other consistent) light on every film I expose.
Logged

Schewe
Sr. Member
****
Offline Offline

Posts: 5415


WWW
« Reply #6 on: February 18, 2012, 06:25:44 PM »
ReplyReply

I just wonder if the orange masks vary as well

You betcha...and that's one of the big problems trying to scan various emulsions. And again even that could vary lab by lab or over time. There's just no good way to create an ICC profile for neg film that I'm aware of....
Logged
Mark D Segal
Contributor
Sr. Member
*
Offline Offline

Posts: 6818


WWW
« Reply #7 on: February 18, 2012, 06:28:59 PM »
ReplyReply

On top of which not all "orange masks" are "orange" :-)
Logged

Mark D Segal (formerly MarkDS)
Author: "Scanning Workflows with SilverFast 8....." http://www.luminous-landscape.com/reviews/film/scanning_workflows_with_silverfast_8.shtml
Christoph C. Feldhaim
Sr. Member
****
Offline Offline

Posts: 2508


There is no rule! No - wait ...


« Reply #8 on: February 18, 2012, 06:32:30 PM »
ReplyReply

If I'd process a negative the way Mark describes by subtracting the orange mask of that specific roll of film - would an ICC profile make sense after this step?

I imagine taking
an image of an IT8 target,
make a raw scan,
subtract the orange mask from the scan and
use that file (now a positive with orangemask removed) for creation of an ICC profile
which I'd subsequently apply to all other images after they have passed the procedure??
Logged

Mark D Segal
Contributor
Sr. Member
*
Offline Offline

Posts: 6818


WWW
« Reply #9 on: February 18, 2012, 06:39:52 PM »
ReplyReply

Why not just do it the easy way of using Negafix to get your positive and embed ARGB(98) as the colour working space? If you've "tuned" your Negafix profile for the batch of film you're scanning, it should look the same in Photoshop as it does in the SilverFast preview converted with the right Negafix profile. And if you need to tweak the Negafix profile for the specific film batch, you have the option in Ai Studio of saving it as a new NegaFix Profile (a custom one) which you can load every time you need to convert an image from that same film batch. By film batch I mean same vendor, type, ISO, film processor and processing date. Once you do that you don't need to worry about IT8 targets and all the rest of it. LaserSoft Imaging made it nice and easy for us, and as I demo'd in that tutorial - all kinds of ways to skin the cat, but none of them do it any better and they are all more complexified.
Logged

Mark D Segal (formerly MarkDS)
Author: "Scanning Workflows with SilverFast 8....." http://www.luminous-landscape.com/reviews/film/scanning_workflows_with_silverfast_8.shtml
Christoph C. Feldhaim
Sr. Member
****
Offline Offline

Posts: 2508


There is no rule! No - wait ...


« Reply #10 on: February 18, 2012, 06:45:50 PM »
ReplyReply

I wished Lasersoft would allow us to shoot an IT8 target and use this to create a specific negafix profile so we could have an easy way to adapt.
Though technically somewhat capable I'm not really the type for twiddleing with curves at a computer monitor.
However - the answers helped me a lot since I'm now safe from unnecessary work.

Thanks and cheers
~Chris
Logged

Schewe
Sr. Member
****
Offline Offline

Posts: 5415


WWW
« Reply #11 on: February 18, 2012, 06:56:55 PM »
ReplyReply

If I'd process a negative the way Mark describes by subtracting the orange mask of that specific roll of film - would an ICC profile make sense after this step?

Nope...not in my experience...ICC profiles just don't work for neg film. Sorry...your much better off taking Mark's advice and do the correct setup for Lasersoft and get it as good as you can when you scan–knowing the addition corrections will prolly be required in Photoshop-particularly local corrections since a scan app can't really do local corrections...
Logged
Mark D Segal
Contributor
Sr. Member
*
Offline Offline

Posts: 6818


WWW
« Reply #12 on: February 18, 2012, 07:01:42 PM »
ReplyReply

Hi Christoph,

There's nothing LSI is preventing you from doing. The problem is you can't create an ICC profile from the negative of an IT8 target, because the reference values in the target file are all based on positive values of the patches. An ICC profile is designed to characterize your scanner's response to rendition of IT8 patches in positive mode by comparing the colour values of the patches the device renders compared with the correct - positive - values in the target's reference file. I'd love to see how the math in a profile creation program would cope with the disconnect between positive reference file colour values and negative colour values for the same patches from the scanner. Wow, that would be fun! Perhaps a new branch of abstract photography. :-)

Getting back to those Negafix profiles - if you use Curves for adjusting your images in Photoshop, the principle is the same in the NegaFix Expert Mode; yes, it can be a bit "fiddly"; but most of the time most people don't need to get into that.

And yes - always a good idea to avoid needless work :-)

Cheers,

Mark
Logged

Mark D Segal (formerly MarkDS)
Author: "Scanning Workflows with SilverFast 8....." http://www.luminous-landscape.com/reviews/film/scanning_workflows_with_silverfast_8.shtml
Mark D Segal
Contributor
Sr. Member
*
Offline Offline

Posts: 6818


WWW
« Reply #13 on: February 18, 2012, 07:03:57 PM »
ReplyReply



Jeff,

SilverFast 8 can do a lot of local corrections.
Logged

Mark D Segal (formerly MarkDS)
Author: "Scanning Workflows with SilverFast 8....." http://www.luminous-landscape.com/reviews/film/scanning_workflows_with_silverfast_8.shtml
Christoph C. Feldhaim
Sr. Member
****
Offline Offline

Posts: 2508


There is no rule! No - wait ...


« Reply #14 on: February 18, 2012, 07:46:23 PM »
ReplyReply

Maybe just for the fun of it I'll try to get hold of an IT8 target and try that thing above -
more for learning and seeing how I'll bump into a wall.
At least it will be educative.
When I asked the Lasersoft guys about using IT8 targets they took it as an idea -
so who knows with what they will come up with in the future ..

Cheers
~Chris
Logged

Mark D Segal
Contributor
Sr. Member
*
Offline Offline

Posts: 6818


WWW
« Reply #15 on: February 18, 2012, 07:51:36 PM »
ReplyReply

Keep us posted what you discover!
Logged

Mark D Segal (formerly MarkDS)
Author: "Scanning Workflows with SilverFast 8....." http://www.luminous-landscape.com/reviews/film/scanning_workflows_with_silverfast_8.shtml
MonsterBaby
Full Member
***
Offline Offline

Posts: 133


« Reply #16 on: February 19, 2012, 03:03:33 AM »
ReplyReply

maybe a lil crazy...

but couldnt you take a picture of an reflective it8 target with a certain film (shade outdoor whatever you wanna set up) and then have it developed..

scan it.. use a standard negafix curve for the orange mask so you get a positive.. save it without ICC..

THEN create an ICC profile from the target... !

and there goes your ICC workflow...?

scan.. use standard negfix.. assign new ICC profile.. ??

this way you wouldnt have to fine tune the negfix but just use it for the standard mask
Logged
Christoph C. Feldhaim
Sr. Member
****
Offline Offline

Posts: 2508


There is no rule! No - wait ...


« Reply #17 on: February 19, 2012, 04:11:26 AM »
ReplyReply

maybe a lil crazy...

but couldnt you take a picture of an reflective it8 target with a certain film (shade outdoor whatever you wanna set up) and then have it developed..

scan it.. use a standard negafix curve for the orange mask so you get a positive.. save it without ICC..

THEN create an ICC profile from the target... !

and there goes your ICC workflow...?

scan.. use standard negfix.. assign new ICC profile.. ??

this way you wouldnt have to fine tune the negfix but just use it for the standard mask

Exactly that is the experiment I'm going to do as outlined above.  Wink
I just found out, I can scan in linear gamma without any curves applied.
And the orange mask removal can also be standardized - So I believe its possible.
If it works in the real world only the experiment will show.
« Last Edit: February 19, 2012, 04:13:15 AM by Christoph C. Feldhaim » Logged

Mark D Segal
Contributor
Sr. Member
*
Offline Offline

Posts: 6818


WWW
« Reply #18 on: February 19, 2012, 07:52:21 AM »
ReplyReply

maybe a lil crazy...

..............

scan it.. use a standard negafix curve for the orange mask so you get a positive.. save it without ICC..

THEN create an ICC profile from the target... !


As Jeff and I have been trying to explain, there is no such thing as a "standard negafix curve" because "orange masks" are all over the place, varying in hue, saturation and lightness depending on the vendor, type, ISO, batch, processor and processing date. But to begin with don't confuse orange mask detection with colour inversion; these are separate issues; orange mask detection is one part of the operation, then the colours need to be inverted - properly. Also, on the mechanical side of it, your photograph of the IT8 target would need to produce a result dimensioned exactly to what a profile creation application could read correctly. Let's see what Christoph comes up with, but I'd be surprised if it turned out to be a more practical workflow than simply using a (perhaps customized) NegaFix profile in SilverFast.
Logged

Mark D Segal (formerly MarkDS)
Author: "Scanning Workflows with SilverFast 8....." http://www.luminous-landscape.com/reviews/film/scanning_workflows_with_silverfast_8.shtml
Christoph C. Feldhaim
Sr. Member
****
Offline Offline

Posts: 2508


There is no rule! No - wait ...


« Reply #19 on: February 19, 2012, 08:44:06 AM »
ReplyReply

As Jeff and I have been trying to explain, there is no such thing as a "standard negafix curve" because "orange masks" are all over the place, varying in hue, saturation and lightness depending on the vendor, type, ISO, batch, processor and processing date. But to begin with don't confuse orange mask detection with colour inversion; these are separate issues; orange mask detection is one part of the operation, then the colours need to be inverted - properly. Also, on the mechanical side of it, your photograph of the IT8 target would need to produce a result dimensioned exactly to what a profile creation application could read correctly. Let's see what Christoph comes up with, but I'd be surprised if it turned out to be a more practical workflow than simply using a (perhaps customized) NegaFix profile in SilverFast.

And the funny thing is:
Totally controlled and calibrated colors aren't necessarily beautiful, maybe not even THE prerequisite to get beautiful color rendition.
Postprocessing skills, a good calibrated monitor and a good eye are most likely much more important ....
Logged

Pages: [1] 2 3 4 »   Top of Page
Print
Jump to:  

Ad
Ad
Ad