Using Crossed Polarizers for a Variable ND Filter

Strong neutral density filters are great for landscape photography when you want to allow motion blur to occur. But they can be expensive and hard to find.
Lately I’ve been exploring using longer exposures in my landscape photography to produce motion blur in waves, trees, clouds and flowing water. I’ve been testing a number of heavy neutral density filters (more about this in some other articles). For those who do not shoot this way, let me explain. Some movements, such as a waterfall or water cascading over rocks in a stream may require a shutter speed of 1/2 a second or longer to create that pleasing blur that you often see. Things like turning the ocean flat or making clouds into streaks in the sky can require exposures in the 30 second to two minute range, or longer and so on. Such exposures can’t be obtained simply by stopping the camera lens right down when the light is bright ad we cannot always return when it is late in the day. Also using a small aperture forces a large depth of field, which you may not actually want. The solution is to use a strong neutral density filter to cut the light levels so that you can use a combination of ISO, aperture and shutter speed that you want to achieve the effect you desire.

In researching all this I found the Singh Ray Vari-ND filter, that goes from 2 to 8 stops of exposure effect as you rotate the filter. Sadly it is a fairly expensive filter and only available in 77 and 82mm sizes. So I set out to work out how this was done. The hint was on their site when it said that you could use it with other filters but not polarizers. Photographers have known about the two-polarizer effect for ages and I guessed that this was what they were doing. Indeed it is used in photo-microscopy to photograph certain types of specimen against a nice black background. A polarizer only allows through light whose rays are vibrating in a certain direction. It blocks others. Put two polarizers together and rotate them so their transmission directions are 90 degrees apart and all the light is blocked. In practice it massively attenuates the light but may allow a little through because they are not perfect.

So I grabbed two polarizers I had and tried it. No result. The reason was because they were circular polarizers. A circular polarizer has a normal linear polarizer at the front and a rotating plate closest to the camera that redistributes the ray vibration directions so that the light is effectively no longer polarized. This is done because linear polarized light can interfere with some camera autofocus and exposure systems. Two circular polarizers didn’t work because the light was “de-polarized” by the first one.

Next step was to buy two linear polarizers and sure enough they work. But this could cause problems with some cameras. So the ideal answer is this: stack a linear polarizer with a circular polarizer, just make sure the circular polarizer is the one closest to the camera lens. The first, linear polarizer will polarize the light. The second, circular polarizer, will use its linear polarizer component to create the variable neutral density effect and then its “de-polarizer” will prepare the light before it goes into the camera lens. To make life hopefully easier I used the same brand filter for all the tests, Hoya.

Using two polarizers for variable ND filter

So I now had a working arrangement that would function on my camera with no dramas. The issue now was to test by shooting and see how it went.

Without the polarizers, the exposure was f6.3 and 1/320 second. With both polarizers attached and set for minimum absorption the exposure was f6.3 and 1/40 second, giving us a 3 stop reduction to start with. I taped the back polarizer so it could not rotate and then started turning and shooting as the filter combination got darker. As you can see from the shots below, all made at the same f6.3 f-stop, we went from a 3-stop reduction to a 10+ stop reduction in light.

No filter, 1/320 second
Using two polarizers for variable ND filter

Minimum 3 stops, 1/40 sec
Using two polarizers for variable ND filter

Approx. 4 stops, 1/30 sec
Using two polarizers for variable ND filter

5 stops, 1/10 sec
Using two polarizers for variable ND filter

7 stops, 1/2 sec
Using two polarizers for variable ND filter

8 stops, 1 sec
Using two polarizers for variable ND filter

9 stops, 2 sec
Using two polarizers for variable ND filter

9+ stops, 2.5 sec
Using two polarizers for variable ND filter

10+ stops, 5 sec
Using two polarizers for variable ND filter

10++ stops, 6 sec
Using two polarizers for variable ND filter

You will notice that at 9 stops there is a color shift starting to occur. By 10 stops this is extreme. So what I found was that these polarizer’s are not perfect, which makes sense because Hoya would not have chosen the polarizing material with this application in mind. The filters are obviously not blocking across all visible wavelengths evenly when crossed and is letting through more at the blue/violet end. Singh Ray have obviously chosen polarizing material to make their filters that are not as strong polarizer’s (hence the two stop minimum) and only get to eight stops when at 90 degrees with no color shift. Incidentally I also tried this on my IR converted camera and it did not work really there, but more on this in a separate article.

The Singh Ray Vari-ND is a perfect solution for those using wideangle lenses, since the single filter thickness minimizes vignetting
issues. But you can also work quite well by stacking two polarizers as I have, especially if you use larger filters
and a step-up ring or work primarily with only moderate wide to telephoto lenses. The exact result you get will depend on the polarizers you use and you may well find no color shift at the most extreme end. Some may be concerned about the effect on image quality. The answer is that if you use good filters and they are clean the effect will be minimal. Given the cost of polarizing filters is not high I would recommend dedicating two filters to this if you like this effect. That way you need not separate the filters in the field and thus can keep in the inner surfaces spotless.

Now as you rotate the filters from minimum position you will get little effect to start with and much greater effect with small movements as you move closer to 90 degrees. You can see this on the graphs below, but especially on the second one calibrated in stops. These graphs are for the ideal polarizer.

Using two polarizers for variable ND filter

Using two polarizers for variable ND filter

What is really amazing, as the shots below show, is that if you set a moderate reduction when you then rotate the combination you can vary the polarization effect on, say, the sky or reflections. This is an additional creative bonus.

Using two polarizers for variable ND filter

So I now have a new creative possibility in my camera bag when I go out to shoot, a variable neutral density filter combination that also gives me polarizer effects.