Long exposure colour cast

[Stagecoach]

Had my first chance yesterday to go out and trial a 10 + 6 stop ND, my objective to be able to achieve 5 minute exposures in brilliant afternoon sunshine.

Again I came across 'colour casts' and a possible 'softening' of close subjects which I had encountered before on random shots but can't put my finger on the cause.

Here's an initial shot used to determine exposure, no pp. The sun is roughly 90 deg to the left;

No 1


Here's the shot with 16 stops ND fitted, no pp;

No 2


Further info;
a) The lens was the 16 to 35 f/4 Nikon but I have encountered this before playing with a different lens.
b) A skylight filter is fitted but I have previously eliminated this as the cause.
c) Other shots taken yesterday with 16 stops ND, 300s exp with the sun directly behind or in front do not show any cast at all.
d) I have encountered this colour cast previously using a single ND.
e) I'm 99% sure the viewfinder was shut for the example above.
f) I take great care to ensure the filters are aligned correctly to maintain the light seal.

I'm coming to the conclusion this is related to the light angle, so I may need to order the appropriate lens hood kit. I'll also do some further controlled testing to try and nail it. Any other ideas or experiences with this would be appreciated.

But the good news is that things are recoverable.

No 3

[Polar01]

Did you cover the viewfinder? Myself I use a woollen watch keepers cap, which also works well to keep the sun off my and and warm in winter.

Cheers: Allan

[Stagecoach]

Did you cover the viewfinder? Myself I use a woollen watch keepers cap, which also works well to keep the sun off my and and warm in winter.

Cheers: Allan

Mentioned in e) above Alan. For this one I'm 99% sure it was shut (the D800 has a shutter that can be closed over the viewfinder).

There were a couple of 5 min exposures I took where I had forgotten to close it and these were almost white exposures.

[Manfred M]

Which 10 stop filter are you using. The Lee Big Stopper is known to have this blue issue. I can’t find it right now, but saw this on the Lee website.

[DanK]

Some NDs, even expensive ones, have color casts. Take a shot of a uniform color, like a wall, first without the filter and then with, and you’ll see whether yours does. One reason I settled on the ones I have (Marumi 3 and 6 stop, Hoya 10 stop) is that they are quite neutral.


Sent from my iPad using Tapatalk

[Polar01]

For me that is the light leak colour that I get when I forget to cover over the viewfinder. I also use the Lee big stopper and have at times doubled them up thus using 2 big stoppers now there is a colour cast, usually do not worry that much as my vision is usually B&W.

Cheers: Allan

[Stagecoach]

Which 10 stop filter are you using. The Lee Big Stopper is known to have this blue issue. I can’t find it right now, but saw this on the Lee website.

I'm using the LEE system Manfred. Yes they do tend to make the overall global scene it bit bluer which is very easily addressed but with the purple you see here it is localised.

Here's an example of the blue global cast you get, 10 + 6 stops, 300s.

SOOC JPEG



SOOC JPEG





Edited, incorrect long exposure loaded, replaced with SOOC JPEG

[Stagecoach]

Some NDs, even expensive ones, have color casts. Take a shot of a uniform color, like a wall, first without the filter and then with, and you’ll see whether yours does. One reason I settled on the ones I have (Marumi 3 and 6 stop, Hoya 10 stop) is that they are quite neutral.


Sent from my iPad using Tapatalk

But I can use these Dan in certain directions, same set up and exposure time and there is absolutely no purple localised casts anywhere.

[Stagecoach]

For me that is the light leak colour that I get when I forget to cover over the viewfinder. I also use the Lee big stopper and have at times doubled them up thus using 2 big stoppers now there is a colour cast, usually do not worry that much as my vision is usually B&W.

Cheers: Allan

That's good info Alan and supports that it is localised areas affected. I wonder if the viewfinder shutter is not sealing 100% and some light coming from a certain angle is getting through.

I can try and replicate when I get it and then put additional covering over the viewfinder and compare.

[Manfred M]

The image with the magenta cast seems to be a lot softer than the other images. Are you catching some flare? The Lee filters (as other brands of square / rectangular filters) are uncoated and subject to flare.

[Polar01]

Something just came to me, there is a rubber gasket on the Lee Big Stopper, the filter needs to be in the first slot and positioned so that the gasket makes a good seal at the top holder. I am going to say that in your image #2, the filter may have been positioned a little low so, thus creating a small light leak. It is very important that the Big Stopper is in the 1st slot (closest to holder) and in the correct position so that it makes a light proof seal. If I think back I believe I had something like this happen so since than I always as a matter of habit make sure that I have positioned the filter so that the gasket makes a good seal.

Cheers: Allan

[Polar01]

Another thought, test it out on say a 30 sec exposure, use the Big Stopper and position it a little low in the holder to create a light leak on purpose to see if it is similar to you #2 image. Better to take 30 seconds that 5 minutes to try out the theory.

Cheers again: Allan

[Stagecoach]

The image with the magenta cast seems to be a lot softer than the other images. Are you catching some flare? The Lee filters (as other brands of square / rectangular filters) are uncoated and subject to flare.

Yes, the softness of the foreground rocks in the magenta cast image is puzzling Manfred and certainly not evident in the image in post 7. Possibly camera movement as two tripod legs were a few inches in the water, but, the sea wall on the horizon looks sharp.

It's also evident in another image which did not have a magenta cast and was pointing at the light.

It's certainly not my focusing and I can't guarantee it was not camera movement as the ground around there is broken coral over sand.

[Stagecoach]

Something just came to me, there is a rubber gasket on the Lee Big Stopper, the filter needs to be in the first slot and positioned so that the gasket makes a good seal at the top holder. I am going to say that in your image #2, the filter may have been positioned a little low so, thus creating a small light leak. It is very important that the Big Stopper is in the 1st slot (closest to holder) and in the correct position so that it makes a light proof seal. If I think back I believe I had something like this happen so since than I always as a matter of habit make sure that I have positioned the filter so that the gasket makes a good seal.

Cheers: Allan

Another thought, test it out on say a 30 sec exposure, use the Big Stopper and position it a little low in the holder to create a light leak on purpose to see if it is similar to you #2 image. Better to take 30 seconds that 5 minutes to try out the theory.

Cheers again: Allan

I missed these two Alan,

I'm pretty certain as mentioned in f) above that alignment and sealing of the filters was spot on. My practice is to remove the whole unit and place it very carefully down each time between shots.

I'll undertake further tests tomorrow, especially the 'misaligned filter' as you mention to try and nail this.

Some further investigation has revealed that the purple/magenta areas are caused by light entry, either via the filters OR the viewfinder. It is suggested, via the filter they show as vertical areas and via the viewfinder as horizontal areas.

Whatever, each is easy to solve but it will be useful to know exactly the cause and its affect.

[xpatUSA]

Any other ideas or experiences with this would be appreciated.

A bit of GIMP analysis:

Hue-map:



There's some kind of split down the middle (arrowed). With this hue-map opened in RawTherapee, the affected area has virtual no green, a good bit of red, 100% blue and an HSV saturation of 100% . . not good but I have no idea why.

Blue channel center-area histogram:



Pretty skinny compared with -

Blue channel lower left histogram:



Food for thought . . .

[Stagecoach]

I can enlarge that image Ted, the original RAW, original JPEG and downloaded JPEG from CiC to the maximums in both ACR and PS and there is absolutely no indication of anything unusual happening as shown arrowed in your example.

[xpatUSA]

I can enlarge that image Ted, the original RAW, original JPEG and downloaded JPEG from CiC to the maximums in both ACR and PS and there is absolutely no indication of anything unusual happening as shown arrowed in your example.

Grahame, I did try to emphasize that the split was in the hue-map image. So opening the original, etc., is irrelevant to what I wrote, sorry.

Seems obvious to me that you did not create a hue-map from from "the original RAW, original JPEG and downloaded JPEG".

Here's a hue-map from a completely different app:



The "split", although it may be a red herring, is still there.

[Stagecoach]

The "split", although it may be a red herring, is still there.

I only looked and assessed the images as I would in the "real world" Ted

But, the good news is I have just completed a controlled test series of 17 long exposure images with some pretty conclusive results that I think explain what happened to that image and in addition how to recognise the cause/source of purple/magenta "areas" with long exposures using square filters, and it's got circles and light angles involved

I just need to download the images and prepare for posting.

[xpatUSA]

I only looked and assessed the images as I would in the "real world" Ted

But, the good news is I have just completed a controlled test series of 17 long exposure images with some pretty conclusive results that I think explain what happened to that image and in addition how to recognise the cause/source of purple/magenta "areas" with long exposures using square filters, and it's got circles and light angles involved

I just need to download the images and prepare for posting.

Looking forward to it, Grahame,

I can only hope it will visible in my unreal world . .

[Stagecoach]

I undertook a set of controlled tests today to help determine the cause of the purple areas on 'some' long exposure images I have taken using the LEE ND Square filter system.

On the web I had come across talk of purple areas being caused by light entry via the filters and the viewfinder eyepiece and wanted to nail this down further and have come up with the following;

1 LIGHT ENTRY VIA THE FILTER/s.

This could be caused by filter alignment or the seals. My filters are new and the seals are perfect so I was only able the test alignment.

No 1.1 : A single 10 stop was misaligned to the extent where the foam seal is visible before a purple area was evident on a 30s Exp.


No 1.2 : A 10 + 6 stop were both misaligned to where a very small light gap was visible and a purple area was evident on a 300s Exp.


Conclusion, major misalignment is needed before any affect and in each result the area is clearly circular also with a softening of the image.

Edit added ............. ( I can simulate filter seal leakage)

No 1.3 : A 10 + 6 ND with the 10 reversed so there is no seal between it and the holder. During the 300s exp I shone torchlight into the small light gap now between filter and holder.



Conclusion, no purple evident but still that circular pattern.

2 LIGHT ENTRY VIA THE VIEWFINDER

Light entry through the viewfinder can be caused by it being left open (a leaf shutter on my camera which could possibly stick) or not covered adequately.

No 2.1 : A 30s Exp with the viewfinder closed.



No 2.2 : A 30s Exp with the viewfinder open and strong torchlight shone in from an angle of around 30 deg to the viewfinder plane. (the same result from above, below, left and right)


No 2.3 : A 30s Exp with the viewfinder open and strong torchlight shone directly into the viewfinder.


Conclusion, with the viewfinder open the affect of light entry causing the purple areas and softening very much depends upon the light direction.

Final overall conclusion, the example image in post 1 was due to ME not closing the viewfinder or it sticking.

With the viewfinder closed in these tests I could not achieve anything by shining bright light into it from any direction to support sealing leakage.

Grahame