Help me correct color cast (from B+W 10-stop ND filter)

[Heroique]

I’ve heard about the color cast w/ B+W’s 10-stop ND filter, but now I see it!

My main question is what on-location filters might help “correct” for this cast, so I can “get it right” on the film? :confused: (I’m less interested in using PS for correction; I mainly view my slides with an analog projector.)

BTW, these Epson 4990 scans, on my calibrated monitor, re-produce – very closely – what my slides look like when they’re back-lighted w/natural light…

— The first shot is w/o filtration, Astia-100F, 1/30th @ f/22.
— The second shot is also Astia-100F, w/ B+W’s 10-stop ND filter, 34 seconds @ f/22. I think the film suggests exposure/color compensation isn’t necessary for this exposure, and my experience mostly confirms this.

Now, I like warming my daytime shots in conditions like this. Usually, I’ll go w/ an 81-series filter, and sometimes when I want stronger warming, I have an 85c. But this result is, well, “strong & wrong” for my tastes. The slightly hazy sky’s “brownish” cast is especially displeasing to me. (Makes one curious why B+W puts “neutral” in the name of this filter. ;) )

Would CC20 cyan, or CC40 cyan be a good place to start experimenting? Or maybe the 80 or 82 series? Other ways to experiment?

[John T]

Are you sure this is a color flaw of the filter or reciprocity failure of the film. My guess is that the color shift of the film is partially to blame. Astia has much less reciprocity problems than other films, but 22 seconds will affect the color. Since you are willing to experiment, do 2 shots one with the 20 and one with the 40

[Heroique]

Hi John,

Yes, I've used Astia-100F w/o filtration at longer than 20 seconds, and the daytime colors, though slightly different than when taken at quicker exposures, don't change as dramatically as they do here. I don't think I've used over 40 seconds yet...

I should add that the two shots were taken within moments of each other in the same light.

Also Tachi 4x5 w/ Schneider XL 110mm/5.6.

[John T]

The interesting thing about the color of your image was that my B+W was slightly magenta. That, plus the color shift of the film was really difficult to clean up. (Scanning did make that a non-issue later on.)

Your color shift IS pretty strong. Unfortunately, I'm viewing it on a laptop so I don't trust the color on my end. Have you tried to use the cyan filters as viewing filters? It may get you in the ballpark.

[Darin Boville]

I get a shift with dark B+W ND filters using digital--no biggie to correct digitally, of course. But the shift seems is real...

--Darin

[Jim Michael]

Why don't you take a reading in PS from a neutral area (like a cloud) of an image you like and one you don't like, and see what the difference looks like? If the cast is f(t), then determine the slope of the line and calculate the approximate shift per unit of time. That should get you in the ballpark.

[mandoman7]

Going beyond 1 sec. introduces a big variable, if you're trying to nail down the color. Film companies make no promises for color accuracy of daylight films at such long exposures. For testing purposes, I would shoot 2 sheets at close to the same speed before making any assumptions about the effect of added filters.
Having said that, in that past when I used to do a lot of copy work for artists, I came to the conclusion that nearly all filters seem to have a bit of coloring, if you looked closely. Frustrating, but it rewards the worker who has tested his equipment and knows their characteristics, which is as it should be.

[Brian Ellis]

If you have to put a second filter on top of the ND filter to fix this problem (assuming there is another filter that will fix it and assuming it is in fact your ND filter that's causing the problem) I'd get rid of the ND filter. Sticking two filters on a lens generally degrades the image more than I like (just a personal opinion, I know many people do it).

[Eric James]

I brought the filtered picture up in PS and applied a cyan "photo filter"; if the PS filter is an accurate representation of the cyan filter you hope to use, I suspect that you will not be happy (read: horrified) with the result. Do you know someone with a dSLR who would investigate the color cast vs color shift question? FWIW, I've used Astia-F for 20-second exposures and have never seen this sort of color perturbation.

[Heroique]

Thanks for these suggestions to get “in the ballpark.”

Looks like I have some old-fashioned “try & try again” experimentation ahead of me. To start, I think I’ll find some cyan-ish filters, like John says, and hold them behind the transparency — though Eric’s cyan filtration in PS suggests a different color may be better! I may even try Jim’s “f(t)” examination in PS for additional ideas...

Then it’s back to the field – I mean the beach :cool: – to get some more results.

If I didn’t include the yucky sky, the golden-brownish color cast wouldn’t bother me so much. Ideally, I hope I can find just one additional filter to “correct” this color cast & “create” the warm color I do like … at the same time. And, maybe a separate second filter to bring the image closer to Astia-100F’s normal palette.

[Jim Michael]

Although it's been more than a few years since doing a lot of color corrections, the difference in the 2 images looked more blue/yellow to me than cyan.

[mrladewig]

Thanks for these suggestions to get “in the ballpark.”

Looks like I have some old-fashioned “try & try again” experimentation ahead of me. To start, I think I’ll find some cyan-ish filters, like John says, and hold them behind the transparency — though Eric’s cyan filtration in PS suggests a different color may be better! I may even try Jim’s “f(t)” examination in PS for additional ideas...

Then it’s back to the field – I mean the beach :cool: – to get some more results.

If I didn’t include the yucky sky, the golden-brownish color cast wouldn’t bother me so much. Ideally, I hope I can find just one additional filter to “correct” this color cast & “create” the warm color I do like … at the same time. And, maybe a separate second filter to bring the image closer to Astia-100F’s normal palette.

I think the filter is adding magenta. All the notes (from DSLR users) I've seen of problems with the B+W 10 stop have indicated a magenta shift. If so, a green filter may be more effective than a cyan. But in your example it almost appears to be adding a tobacco cast with Astia and I'm not sure how you would want to correct that.

One thing about Astia 100F that may be coming into play is the 4th layer (MCCL) technology which is supposed to help correct white balance under various lighting. Perhaps shooting something else like Provia without the 4th layer tech might help provide a better prognosis.

[Sevo]

If the filter had a constant and uniform wide band cast, they'd simply add green in the production - it must be some more complex flaw.

[mrladewig]

If the filter had a constant and uniform wide band cast, they'd simply add green in the production - it must be some more complex flaw.

That reminded me to go look at the posts I've seen on the B+W filters. Seems these filters are not always cutting IR and the magenta cast in digital cameras has to do with their IR sensitivity. Normally the cut filters in front of the sensor reduces the IR component relative to the visible light sufficiently that the IR isn't noticeable, but when 10 stops visible is cut without a proportionate cut in IR, some shift is developing.

Example discussion (http://www.fredmiranda.com/forum/topic/802380/0?keyword=B%2BW,ND#7400145)

So, it looks like a magenta cast is a dead end for a film user.

I'd still look at the 4th layer function of Astia though. Perhaps it is interacting in an unusual manner with the ND filter. I recommended Provia as a test because it is one of the best films for daylight long exposures. It should have no reciprocity issues at the time you indicated and doesn't employ the 4th layer. From there you should be able to accurately determine what color cast is emerging and set about correcting it.

[Heroique]

I just visited a local expert w/ a color meter, which provided some useful information about this rather un-neutral B+W filter.

According to the Sekonic, the B+W filter caused a 970 K drop in temperature (more red). This was even more than I suspected, and I’d call it significantly un-neutral. This reading also suggests there’s a little more magenta in the actual film than appears in my scanning effort above (and confirms John’s B+W magenta cast in post #4). Even more technically, we had a chart that equated this change w/ a +40 nm shift.

BTW, for comparison I also had a Nikon 2-stop and Tiffen 3-stop ND filter, and we measured these, too. The Nikon produced a 160 K rise in temperature (more blue). The Tiffen produced a 340 K rise (also more blue). These changes, of course, would not be as noticeable to the naked eye (if at all). Much more “neutral,” especially the Nikon.

Then we measured some “correction” filters that might be a good match for the B+W. The B+W’s 970 K drop suggested an 80d filter (which typically raises the temperature about 1300 K or so) would be a good place to start. We had a Lee resin 80d, and the Sekonic said it actually added about 1,000 K…

The Lee 80d – maybe a good filter to start experimenting with, though I think correcting the B+W will be a little more difficult (make that a lot more difficult) than simply adding a common filter like this one.

However, since I like a little warmth, I’d much rather find a single filter that adds less than 970 K (instead of a little more like the 80d) to leave some warmth, and see what happens. Any common filters out there that do this?

[Mike Anderson]

If you have a digital camera, why don't you repeat the experiment with that?

...Mike

[Heroique]

Well, that would be convenient indeed, but I think I'm one of the few in this forum who, irony of ironies, has no digital camera, or even access to one! ;) But I like your idea. If someone else w/ a digital camera can make a similar examination, their results would be quite interesting to hear about, and clarify even more what type of "correction" might be good to experiment with in the field.

[Bob Salomon]

We have never had a complaint with the heavy ND filters from Heliopan regarding a color shift from any user or any dealer.

[brian mcweeney]

Heroique,
For what it's worth, I took both of your posted images into Photoshop to try and make a "simple" color correction. I found that it wasn't just a matter of more/less magenta or blue/cyan. I ended up going into channel mixer and adjusting color amounts in each channel. I got close but never matched your original image. I don't know if this helps any, but it seemed the color shift was across many areas and not a simple single color filtration solution.

[Heroique]

Thanks Brian! That gives me a little more information to go on, esp. in view of my limited PS skills. Now it's time to start some field tries w/ this or that filter, and see what happens. Even if the B+W filter continues to forbid Astia-100F's more normal palette w/ longer exposures, just removing that strange golden-brown tint from the sky would be satisfactory.

(BTW, if you'd like to show your resulting image, that would be fine by me.)

[Brian Ellis]

I just visited a local expert w/ a color meter, which provided some useful information about this rather un-neutral B+W filter.

According to the Sekonic, the B+W filter caused a 970 K drop in temperature (more red). This was even more than I suspected, and I’d call it significantly un-neutral. This reading also suggests there’s a little more magenta in the actual film than appears in my scanning effort above (and confirms John’s B+W magenta cast in post #4). Even more technically, we had a chart that equated this change w/ a +40 nm shift.

BTW, for comparison I also had a Nikon 2-stop and Tiffen 3-stop ND filter, and we measured these, too. The Nikon produced a 160 K rise in temperature (more blue). The Tiffen produced a 340 K rise (also more blue). These changes, of course, would not be as noticeable to the naked eye (if at all). Much more “neutral,” especially the Nikon.

Then we measured some “correction” filters that might be a good match for the B+W. The B+W’s 970 K drop suggested an 80d filter (which typically raises the temperature about 1300 K or so) would be a good place to start. We had a Lee resin 80d, and the Sekonic said it actually added about 1,000 K…

The Lee 80d – maybe a good filter to start experimenting with, though I think correcting the B+W will be a little more difficult (make that a lot more difficult) than simply adding a common filter like this one.

However, since I like a little warmth, I’d much rather find a single filter that adds less than 970 K (instead of a little more like the 80d) to leave some warmth, and see what happens. Any common filters out there that do this?

If it's any help to you, I spent about 30 seconds in Photoshop playing around with the color balance of your Surf2. By adding some blue (+33) and cyan (red minus 20) and fiddling around with the sky I got this result, which on my calibrated monitor is pretty close to your original. Another few minutes of effort would probably duplicate it, except of course for the water that your long exposure changed dramatically. I have no idea whether there's a color balance filter out there that is the approximate equivalent of adding blue and subtracting red (which adds cyan) in Photoshop but maybe this will give you some idea of a direction to try.

[Heroique]

Brian, that’s a super job. Including the water. And especially the sky.

Thank you. :)

I hope I can find a single correction filter that can approximate what you’ve done in PS.

[Mark Sampson]

Maybe you should consider using a Kodak Wratten #96 ND (gel) filter of the appropriate density. The Kodak filter book says they are "for uniform light attenuation throughout the visual spectrum". I've never used any of them with color film but it's worth a try.

[brian mcweeney]

I still feel like there's more than a simple color correction to this problem. Here's what I came up with. First image is the original shot, second is with the ND, and the third is my photoshop work. Channel mixer, slight curve to get snap back, and slight overall green and blue added in color balance. Still not perfect match though.

[dsphotog]

Although not the same...... But could be fun...
You might enjoy shooting the scene at night.

[Heroique]

That’s another fine PS conversion. And like mrladewig’s post #14, your job suggests a touch of green may be part of the solution.

I realize field tests are the only way to be sure, but I want to get the filters that make the most sense for a first try. I’m not sure how to translate the PS work (by both Brians) into on-location filters – but as discussed above, the 80d is one I’m sure to bring.

A couple of others might be, say, a CC cyan and CC green, but I’m curious which strength one might recommend. For example, below is a set I might start with.

1) 80d, 2/3 stop, adjusts color temp +1300 K (or +1000 K if it’s the Lee 80d we tested)
2) CC40 cyan, 2/3 stop, absorbs red (perhaps in place of the 80d)
3) CC20 green, 1/3 stop, absorbs red and blue (w/ 80d, or w/ CC40 cyan)

I hope I don’t have to use more than one filter for correction, but you never know…