I had some exposure problems in the field recently with a couple HiTech 4 stop ND filters so this evening I set up some lights at home, and did a few tests. I shot my Xrite Color Checker and compared the histograms of each shot.
Under "normal" lighting conditions, I got the following exposures:
No Filters 1/80 sec f/4
One 4 stop ND 1/5 sec f/4 (a difference of 4 stops)
Two 4 stop NDs stacked 3.2 sec f/4 (a difference of 8 stops)
So far so good right? But, when I did tests under "low" lighting conditions, I got the following exposures:
No Filters 1 sec f/4
One 4 stop ND 3 sec f/4 (a difference of 1 2/3 stops)
Two 4 stop NDs stacked 60 sec f/4 (a difference of 6 stops)
Isn't this weird? Why, under low light, is a 4 stop ND filter acting like a 1 2/3 stop ND filter, and why, when I stack two of them do they act like a 6 stop ND filter?
These are the exact problems I experienced in the field and what prompted these tests.
On a side note, with the long exposures, like the 60 seconds, there is a really bad magenta colour cast, but I can more or less correct that in post.
Are you covering the eyepiece of your camera when you make the exposure reading with the filters in place? Because stray light going into the eyepiece greatly effects exposure with ND filters. You probably avoided the problem by doing manual exposures and got all the histograms to match eh? If so, then I have no idea what is going on?
Did you check the histograms to see that it was the same exposure? I use a 10 stop and do the math and it usually comes out pretty close, even at sunset conditions.
I never trust what my camera reads, sometimes live view gets a close exposure, but I trust the math, or eyeball it based on the histogram after a test exposure.
Other than that, maybe as the exposure gets longer, light leaking in from the eyepiece can have more of an effect? The color cast does get stronger as the exposures get longer, if I shoot with the 10 stop on cloudy, I get a real warm and rich image. Something I do on purpose many times.
Total guess here on the color cast.... the NDs are probably not exactly the same across all wavelengths of light but are very close. So they probably have an unnoticeable color cast all the time - maybe they let a couple of percent more red light through than other wavelengths, for instance. So at fast shutter speeds, that difference wouldn't be enough to matter or be visible. As you go longer, that small difference starts to add up and becomes noticeable.
Thank you for your replies guys!
Darwin, I used manual exposures and I tried to cover the eye piece to see if it would make a difference but it didn't.
Paul and David, I think you are right about the colour cast not showing up as much in the shorter exposures. That make perfect sense to me.
I just had a thought though, about the problem with the long exposures... The Hitech filters are square, and fit in to slots in my Cokin P holder. Since they are not screwed to the lens directly, and since there is a small gap behind the filters, maybe the light bounces of the back of the filters and in to the camera. Being so dark, they would act as a pretty good mirror.
I'm going to do a few more tests tonight to test this theory. Stay tuned...
Being so dark, they would act as a pretty good mirror.
Not necessarily - color doesn't always imply reflectivity. They are dark glass, which I'd agree would make them pretty reflective. But they are photo filters, which should minimize exactly the kind of thing you're talking about to some extent (most components of a lens and a lot of filters are designed to minimize problematic reflections). Of course, the main reason you hear about coated, multi-coated, diamond coated, etc. glass is to minimize their reflections; And I don't think your filters have that kind of treatment, so it's possible. I'd think that would cause things more like flare or loss of contrast though, than a color cast. (Though I'm, again, guessing.)
I've noticed a weird thing about my B+W ND3.0. This one should actually be a 10-stop-ND but it only darkens the shot by about 6 2/3 stops. That's weird too.
I tested my B+W 106 6 stop screw-in filter. It cuts light by 6 stops both in outdoor overcast conditions and indoors with available light. I'm curious if your theory is correct.
Posted by Tom Nevesely on May 10, 2010 at 11:50 AM
I did a few more tests at home over the weekend, again in normal light and low light. I tried covering the lens and filters with a lightweight (but dark) cloth to prevent light leaks but it wasn't necessary because no matter what I did, the filters worked like they were supposed to.
So now I have no idea what went on. If the problem on occurred once I would be more likely to dismiss it, but it happened to me once in the field and once at home. And now everything is working fine again. I don't know...
I'm not saying this happened to you, Tom, but you reminded me of one of my booboos. I frequently use a handheld meter when working with ND filters or polarizers in challenging light, simply dialing in the filter factors when I meter. It works great and helps overcome lots of problems I encounter when trying to work through that dark viewfinder. Works great that is, until I forget that I had dialed +/- compensation into the camera and failed to remove it.....
Discussion
9 repliesUnder "normal" lighting conditions, I got the following exposures:
No Filters 1/80 sec f/4
One 4 stop ND 1/5 sec f/4 (a difference of 4 stops)
Two 4 stop NDs stacked 3.2 sec f/4 (a difference of 8 stops)
So far so good right? But, when I did tests under "low" lighting conditions, I got the following exposures:
No Filters 1 sec f/4
One 4 stop ND 3 sec f/4 (a difference of 1 2/3 stops)
Two 4 stop NDs stacked 60 sec f/4 (a difference of 6 stops)
Isn't this weird? Why, under low light, is a 4 stop ND filter acting like a 1 2/3 stop ND filter, and why, when I stack two of them do they act like a 6 stop ND filter?
These are the exact problems I experienced in the field and what prompted these tests.
On a side note, with the long exposures, like the 60 seconds, there is a really bad magenta colour cast, but I can more or less correct that in post.
Darwin
I never trust what my camera reads, sometimes live view gets a close exposure, but I trust the math, or eyeball it based on the histogram after a test exposure.
Other than that, maybe as the exposure gets longer, light leaking in from the eyepiece can have more of an effect? The color cast does get stronger as the exposures get longer, if I shoot with the 10 stop on cloudy, I get a real warm and rich image. Something I do on purpose many times.
I see Darwin beat me to all the answers...=)
Darwin, I used manual exposures and I tried to cover the eye piece to see if it would make a difference but it didn't.
Paul and David, I think you are right about the colour cast not showing up as much in the shorter exposures. That make perfect sense to me.
I just had a thought though, about the problem with the long exposures... The Hitech filters are square, and fit in to slots in my Cokin P holder. Since they are not screwed to the lens directly, and since there is a small gap behind the filters, maybe the light bounces of the back of the filters and in to the camera. Being so dark, they would act as a pretty good mirror.
I'm going to do a few more tests tonight to test this theory. Stay tuned...
Not necessarily - color doesn't always imply reflectivity. They are dark glass, which I'd agree would make them pretty reflective. But they are photo filters, which should minimize exactly the kind of thing you're talking about to some extent (most components of a lens and a lot of filters are designed to minimize problematic reflections). Of course, the main reason you hear about coated, multi-coated, diamond coated, etc. glass is to minimize their reflections; And I don't think your filters have that kind of treatment, so it's possible. I'd think that would cause things more like flare or loss of contrast though, than a color cast. (Though I'm, again, guessing.)
So now I have no idea what went on. If the problem on occurred once I would be more likely to dismiss it, but it happened to me once in the field and once at home. And now everything is working fine again. I don't know...