Kodak Ektar 100 - red color cast

I would like to ask for some advise or ideas about how to get rid of the red color cast present in many scanned Kodak Ektar 35mm scans.
Here is my workflow:
- Kodak Ektar 100 35mm developed in a standard lab
- scanned by me on Nikon 5000 ED (tiff, 16bit, 2000 ppi)

Maybe half of my scans show red color cast due to natural color cast. i.e. red walls for example. Using the red hue sliders in LR or PS will modify the entire image, so for example what is supposed to be red will also lose it red color.
I tried also Viveza to do local red reduction, but it is very difficult to get correct skin tones.

So what I am asking if any of you have a good technique to solve this problem?
Thanks
Miklos

Miklos, I don't understand; you say red "cast" but then continue with another issue of local correction. To my knowledge, the term cast is used in global context (even effect in all parts of the image) and corrected globally...???

If a red wall (or something like that) does reflect/shed red light over the skin of your subject (only/locally) then that's a lighting issue, not film/scanning problem. You'll have to adjust locally indeed - or may just opt to avoid tricky/mixed lighting conditions. (Or, use an appropriate color correction filter - if there's such a filter, which will cast its own color to other parts of the image, I mean the parts which aren't lit by the reflected red/reddish light...)

To rule out the possibility of me being completely misunderstanding you, maybe you can elaborate a little more (maybe with sample images?) letting us better understand your exact problem.

Regards,
Loris.

P.S. I don't know the palette of Ektar 100 BTW, maybe it's a very saturated (especially red biased maybe?) film; you may try other films with more neutral palettes.

Back when I was shooting color negative (years ago) Fuji NPS-160 was wonderful with a very good / accurate / neutral color rendition. (I don't even know if they still produce that film though...)

Miklos, how are you scanning these? Using the Nikon software or something else? If using Lightroom to process the scan, I would try doing a manual white balance using the eyedropper.

I'm scanning Ektar 100 in 35mm and 120 formats on an Epson V750 without issue.

I use Silverfast which offers a Negafix profile for Ektar 100. On most scans, I don't have any color issues. On severely underexposed scans, I sometimes have to use the Color Compensation tool set to 20%-80% to reduce a slight magenta cast. This happens rarely, though.

Ok, here it is. Apparently they have a new / equivalent product now...

BTW, the Ektar 100 page says "Ultra-vivid color". That hints to me that it's probably not an ideal film for skin tones - in my book... (BUT please note that I haven't used it before and I'm just interpreting what I'm reading in the product page!)

Loris Medici said:

To rule out the possibility of me being completely misunderstanding you, maybe you can elaborate a little more (maybe with sample images?) letting us better understand your exact problem.

Regards,
Loris.

Click to expand...

Hi Loris,
yes, I meant by local, that I just want to fix the skin tones and not the whole image.
I know that I could fix the color cast by using filters, but my style of photography is not exactly suitable for that. There are many situations where I don't have the time really analyze the scene and decide that there is a red wall so I quickly have to pull out a compensating filter.
As for the characteristics of Ektar, it is a beautiful film under right conditions. Skin tones are warm, saturated to the right degree.

My whole question is more about post-processing. What works fine for digital raw images, does not work for scanned film. At least under those extreme situations. The problem I might be having is that I am trying to apply the same color correction rules which works perfectly on digital raw files. So I was wondering if some one knew a "scanned color negative" specific workflow in terms of color correction. And again, I am emphasizing, this is for images that have a color cast because of the conditions where they were taken.

tgphoto said:

Miklos, how are you scanning these? Using the Nikon software or something else? If using Lightroom to process the scan, I would try doing a manual white balance using the eyedropper.

I'm scanning Ektar 100 in 35mm and 120 formats on an Epson V750 without issue.

I use Silverfast which offers a Negafix profile for Ektar 100. On most scans, I don't have any color issues. On severely underexposed scans, I sometimes have to use the Color Compensation tool set to 20%-80% to reduce a slight magenta cast. This happens rarely, though.

Click to expand...

Yes I use the Nikon software. And that is not the issue (I know many people are criticizing it). When the lighting is neutral (daylight) the colors are perfect straight from the scanner. As I explained in my previous post, the problem is when there is a natural color cast because most of the light is reflected by say a red wall.
Using the eyedropper toll in LR makes the image even more warm and red.
I will post some examples later tonight.

miklo

try using the hue control on the red and the yellow.

I did find your workflow description to be so minimal as to be difficult to know.

I use Nikonscan on an LS-4000 my flow is something like this:

http://cjeastwd.blogspot.com/2009/10/quick-negative-scan-tutorial.html

it could well be that the ektar responds a little differently to the other negatives (based on some conversations I've had, I have as yet not had a go with it, but as its my only other choice for 4x5 I reckon its only 3 sheets before this box of Fuji is finished)

pellicle said:

miklo

try using the hue control on the red and the yellow.

I did find your workflow description to be so minimal as to be difficult to know.

I use Nikonscan on an LS-4000 my flow is something like this:

http://cjeastwd.blogspot.com/2009/10/quick-negative-scan-tutorial.html

it could well be that the ektar responds a little differently to the other negatives (based on some conversations I've had, I have as yet not had a go with it, but as its my only other choice for 4x5 I reckon its only 3 sheets before this box of Fuji is finished)

Click to expand...

That's what I am talking about! Can't wait to get home and try our your technique.
My workflow is more simple, but I am willing to do something new to see if it works better.
Thanks for the link to your workflow!
Miklos

If you can use the "Curves" adjustment panel, that would be the best way to make the adjustment. You can do just about anything with the Curves panel: Color balance, brightness, contrast, white point and black point. It just takes time to get the hang of it.

If you can't wrap your brain around the Curves panel, the next best thing would be the "Levels" adjustment. It's easier and it's a lot more visual.

With the Curves and the Levels panels you can do just about anything that the other panels can do.

Whatever method you prefer to use, balancing skin tones should not be a difficult job to do with Photoshop. However, I agree with the others. If you can get the balance right when you scan the image you won't have to do so much fooling around with Photoshop.

I don't have any technical advice beyond what you have been given. I have a PS plugin from OnOne software called PhotoTune3 that is very useful for color correcting which contains a feature called Skin Tune. I believe they have a trial download which you could try to see if that will solve your problem. Also Silverfast Ai Studio scanning software has a color cast correcting feature.

worker, if you have not got around to the hue command panel I I suggest you give it a whirl. The levels are gross adjustment, curves next then hue adjustment to trim up how red-yellow, yellow-green and blue-cyan sit with respect to each other.

try also adjusting the lightness of them as by adjusting that you can remove some obvious chroma noise in some areas.

miklosphoto said:

Thanks for the link to your workflow!
Miklos

Click to expand...

your welcome.

that's not a fully hand holding sort of tutorial, but points in the right direction.

happy to answer any queries you may have about it.

Miklos,

As a side note: If you're going to scan as positive and then invert in Photoshop, then calibrating / profiling your scanner with an IT-8/7.1 target may help in eliminating casts from the scanner. (Usually resulting in too warm tones - because scanner lightsources are often blue or green/blue biased...) As a bonus, profiled scanners give more contrast, I mean better separation in the shadows of the negative. (= Highlights in the positive image.) At least that's my experience with Epson flatbed.

Loris

Loris Medici said:

If you're going to scan as positive and then invert in Photoshop, then calibrating / profiling your scanner with an IT-8/7.1 target may help in eliminating casts from the scanner.

Click to expand...

that's intriguing ... I will have to follow that up ... but as my scanner is en-route to Australia it will have to wait till after I'm there with it!

I think that may help in doing less post-processing (or doing it in a milder / more subtle manner) indeed. (Especially for the less experienced.) OTOH, will it have a strong effect remains - my transparency scanning experience says it should make a perceptible difference in the unmanipulated scan...

pellicle said:

worker, if you have not got around to the hue command panel I suggest you give it a whirl. The levels are gross adjustment, curves next then hue adjustment to trim up how red-yellow, yellow-green and blue-cyan sit with respect to each other.

try also adjusting the lightness of them as by adjusting that you can remove some obvious chroma noise in some areas.

Click to expand...

Yes, I do use the Hue adjustment but only after everything else. However, I find that getting a good scan and importing it via Camera Raw can take care of the majority of problems. Once imported correctly, I usually add a Curves adjustment. Once I get that far, I have already had three chances to get the image looking the way I want it. I only do other adjustments like Hue as either a "touch up" or for some effect I am experimenting with.

I started trying to adjust images with Brightness/Contrast adjustments and the Color Balance control but never got good results. Then somebody showed me the Curves adjustment but I never was able to get used to using it but I did learn to use the Levels adjustment. That allowed me to get some good results but, when I got CS4 and started learning how to use it, I discovered how the Curves adjustment works and I was able to get my head around it.

Since getting CS4 and learning how to use it, I have gone from taking anywhere from 15 minutes to an hour to fuss with an image and still not getting it right to only needing a few minutes to get it to look just the way I want it.

The three main things that have gotten me this far are:

1) Getting good scanning software and learning how to use it. (VueScan, in my case.)

2) Learning the Camera Raw import process.

3) Learning how to use Curves.

The Hue adjustment is important but, for me, I need to use it less when I get the other three steps done correctly from the beginning.

(And, oh yes... I went through that period where I juiced up the "Vibrance" setting on every picture I made, too! But I got that out of my system. )

Hi

Worker 11811 said:

Yes, I do use the Hue adjustment but only after everything else.

Click to expand...

yep, that's where I was meaning to use it.

The three main things that have gotten me this far are:

1) Getting good scanning software and learning how to use it. (VueScan, in my case.)

2) Learning the Camera Raw import process.

3) Learning how to use Curves.

The Hue adjustment is important but, for me, I need to use it less when I get the other three steps done correctly from the beginning.

Click to expand...

yep ... that's what I was meaning. I only nip the red up or down a couple of points and then the same with the yellow and then the cyan ... seems to be consistent within a film so I suspect its how my scanner responds to various films.

I know what you mean about scan + adjustment times. I was the same back when I started too.

I'm sorry, but can you clarify what you mean by "Learning the Camera Raw import process"? What process?

pellicle said:

your welcome.

that's not a fully hand holding sort of tutorial, but points in the right direction.

happy to answer any queries you may have about it.

Click to expand...

I tried last night your technique of scanning negative as positive and cut all the ranges where there is no data in all three colors.
It gave me much better results. I must say the color cast issue is not gone but becomes much easier to manage in Photoshop.
I saw that some one was commenting on your blog that it did not work for Ektar, I must say it worked just fine for me.

Unfortunately, everything comes at a cost. Scanning takes much more time this way. What I used to do normally, I just load the entire strip of the uncut film into Nikon 5000 ED with the film adapter attached to it. Then run the scan for the whole roll and post process in LR. The scanning took me one hour and I did not even had to sit at the computer.

With your technique I would need to adjust the curves for each individual frame before scanning it. It will result in spending much more time in front of the computer.

What I will do probably, will mix my workflow with you workflow in a way that I still run the scan first for the whole roll of film as negative so I can quickly have a proof in LR. Then I just go back to those images where I see a problem or which are the absolute keepers so I can maximize their quality.

Anyway, I was glad and thankful to get the very useful tutorial from you. Thanks again!

tgphoto said:

I'm sorry, but can you clarify what you mean by "Learning the Camera Raw import process"? What process?

Click to expand...

As I am sure you know, many digital cameras today can export uncompressed image data directly from memory instead of compressing the data into JPEG files. This data is known, generically, as "Camera Raw." (I say "generically" because every camera manufacturer has their own standards for Camera Raw. Camera Raw for Canon cameras vs. Nikon cameras might not be the same.)

Consequently, there must be an intermediate step to import Camera Raw images into Photoshop or other applications. However the benefit of this is that you have MUCH more control over the image as it is being imported into Photoshop.

You have exposure, brightness, contrast, color temperature, tint and several other controls at your disposal that give you a wide range of control. There are other features of Camera Raw import that let you do things like pre-crop the image before going into Photoshop and methods of making localized color adjustments or effects.

On top of all this, you get the added bonus that all these adjustments are done "non-destructively." If, at any time you decide to, you can go back to the original Camera Raw import stage and change one or more settings to readjust the image to your liking.

"Learning the Camera Raw import process" refers to understanding the controls of the Camera Raw import dialog and how to use them. It takes time to learn but the user will benefit greatly from the effort. Properly learned and executed, it can improve your work by orders of magnitude in a very short time. That is part of the reason I said that the time it takes me to produce a satisfactory image decreased from a matter of hours to mere minutes. I learned about Camera Raw and how it is used.

If you have "good" scanner software like SilverFast or VueScan you can create scans output in Camera Raw format (Or better yet, in "Digital Negative" format which is basically an open source form of Camera Raw. )

So, yes, you are adding in another step to your scanning workflow by doing this but you are gaining more control over your image and, hopefully, improving your work. Done properly, the amount of extra time it takes you to go through the Camera Raw process will be far less than you would spend tinkering, otherwise.

I suggest, that, if you don't already use Camera Raw to import images (digital camera or otherwise) into Photoshop you should take a little time to investigate it. I think you'll like what you see.

And, to bring this back around to the topic sentence, I think that the color cast problem would be minimized or even eliminated all together if you used Camera Raw to import your scanner data into Photoshop.

Hi

miklosphoto said:

It gave me much better results. I must say the color cast issue is not gone but becomes much easier to manage in Photoshop.
...
Unfortunately, everything comes at a cost. Scanning takes much more time this way.

Click to expand...

yes, what I did was to observe where is the film base and where is the darkest point for each of the channels and then save that in my nikon scan as a film type. Then I just load that and I know I'm capturing the whole range properly of the film. So under exposed or over exposed it always works

saves me heaps of time

where this falls down is that the thumbnail mode for the SA-21 does not like this, and I need to take time to align these.

PS: for batch scans of things I have been using the Epson, and only use the Nikon for stuff I wish to take care with ... but I agree that it is hardly fast. Perhaps something similar can be achieved with vuescan. Have you read this by Erik Krause

For C41 I have another super-advanced workflow only applicable to Nikon
scanners: Do normal advanced workflow, then in color tab look at the
film base color values. Take the largest one and divide by the next
one. Take the result as a multiplier for the corresponding channel
analog gain value. Do the same for the remaining channel.

If f.e. your channels have film base color values: Red 0.9, Green 0.6
and Blue 0.5 the resulting analog gain values will be Red 1.0, Green
1.5 and Blue 1.8. Unlock film base color, do another preview and lock
again. The film base color values on color tab should be all more or
less the same. If they differ you can repeat the steps.

This way you get a pretty neutral negative. Now you can increase all
(locked) film base color values to 1.0 and scan the whole roll of film
with these settings. In my experience this neutralization of film base
color by different channel exposure works much better than the
mathematical one and you get almost clipping-free image data.

Click to expand...