23
\$\begingroup\$

I know what an ND filter does. I know what a polarizer does. I also know what two polarizers stacked together and rotated properly do.

So the question: why should I use an ND filter to achieve a darker image at the input, when I can use 2 polarizers instead and rotate them to exactly as dark an image at the input as I want?

Also asked by Julien Gagnet

Is it possible to use two polarised filters to create a variable ND filter?

I was reading that by attaching two polarised filter we could create a variable ND filter.

Has anyone done this? How was this done? Any drawback (colour cast, quality...)? What would be the strength in light filtering of such filter?

Improve this question
\$\endgroup\$
1
  • \$\begingroup\$ ND filters and Variable ND filters are two totally different animals. Since every VND filter I've ever seen or heard of is, in fact, two stacked polarizers they have as much if not more in common with polarizing filters than they do with conventional ND filters. They have a LOT more in common with two stacked polarizers than they do with a conventional ND filter! \$\endgroup\$
    – Michael C
    Commented Jul 13, 2016 at 8:31

9 Answers 9

Reset to default
25
\$\begingroup\$

  • Polarizers are often more expensive than ND filters and you need two of them.

  • Stacking two filters can cause vignetting with wide lenses.

  • You have an extra glass surface with two polarizers which can cause flare and potentially loss of contrast/sharpness.

  • This arrangement can cause colour shift toward yellow (but so some ND filters).

  • Extreme wide angle lenses will exhibit uneven darkening due to the difference in incidence angle across the polarizers.

Improve this answer
\$\endgroup\$
3
  • \$\begingroup\$ And stacking filters will decrease image quality, as it's another piece of glass light has to go through before reaching the sensor \$\endgroup\$
    – t3mujin
    Commented Jun 2, 2011 at 11:47
  • 1
    \$\begingroup\$ Also the effect the polarisers has to create the darkness will not work evenly the wider the angle of your lens gets, in extreme cases causing a cross-like pattern of light and dark \$\endgroup\$
    – Dreamager
    Commented Jun 2, 2011 at 12:01
  • \$\begingroup\$ Even a single polarizing filter may cause massive loss of sharpness on long focal lengths. (maybe more expensive filters somehow avoid that, I am not sure because I only used cheap filters). Also, wouldn't the color shift be neutralized by white balance? \$\endgroup\$
    – Display Name
    Commented Jan 1, 2018 at 6:42
16
\$\begingroup\$

The short answer is that yes, you can do this. Just stack two polarizers, and when you rotate them relative to one another, the transmission will vary.

Make sure that the polarizer in front is either:

  • linear, not circular, as the latter essentially un-polarizes the light as it exits the filter.
  • a reversed CPL (but now the threads won't line up).

The point is that the light exiting the first filter needs to be polarized. The second polarizer will need to be a circular polarizer if you want autofocus.

I just tried this (two CPLs, the one in front reversed) and the attenuation seems to vary between one CPL's worth (1.5-2 stops) to nearly black. I did get a strong purple tint when I was approaching max attenuation.

Be aware that you'll still get all the effects of a polarizer as well.

Improve this answer
\$\endgroup\$
7
\$\begingroup\$

In addition to what @Matt Grum gave as reasons:

A polarizer will reduce reflections, while a Neutral Density filter will not. You might want to include reflections in your images.

A Neutral Density Filter optimally does not change the hue or color of the scene at all.

You may not want to darken the sky, and a circular polarizer turned the correct way will darken and emphasize the sky, while a ND filter will darken the entire scene.

Neutral density filters are offered in many different styles. Graduated Neutral Density filters are probably the most useful, where you only want to darken half of the image or a portion of the image. They also provide different transition amounts, with hard transitions or soft transitions.

ND Filters have various transmittance values. You could potentially have a ND filter that far exceeds the maximum light filtering ability of any two polarizing filters.

Improve this answer
\$\endgroup\$
5
  • 1
    \$\begingroup\$ Your statement that "you could potentially have a ND filter that far exceeds the maximum light filtering ability of any two polarizing filters" is not true, but I won't down-vote your answer, because I appreciate your help. You should fix your answer though. \$\endgroup\$
    – Frantisek
    Commented Jun 2, 2011 at 16:07
  • 4
    \$\begingroup\$ @RiMMER Polarizing filters don't work perfectly, even when turned at 90 degrees to each other some light will pass. You could view a sheet of steel as a very strong ND filter. This filter will far exceed the maximum light filtering ability of any two polarizing filters. \$\endgroup\$
    – Matt Grum
    Commented Jun 2, 2011 at 16:41
  • \$\begingroup\$ @RiMMER - I stand with my original answer :-) \$\endgroup\$
    – dpollitt
    Commented Jun 2, 2011 at 18:40
  • \$\begingroup\$ @RiMMER - I have used both a stacked polarizer, and really strong ND (ND400, allows 0.25% of the incoming light to pass) filters - the ND filter blocks FAR more light. \$\endgroup\$
    – Fake Name
    Commented Jul 1, 2011 at 3:13
  • 1
    \$\begingroup\$ Maybe machine vision polarisers are much more efficient than photographic ones, but two of my linear filters crossed will pass 0.0045% light. Much closer to the steel sheet than the most effective ND filter you can find. This is not a recommendation from my part to use the filter stacking instead of ND filters btw :) \$\endgroup\$
    – Michael Nielsen
    Commented Feb 22, 2013 at 18:25
6
\$\begingroup\$

Normal unpolarized light rays have many different "orientations". A polarizing filter only lets through light with a certain "orientation" and proportionally filters out light rays with different orientations. The further the orientation is away from the orientation of the polarizer, the less light makes it through, up 90 degrees where no light makes it through.

If you stack two polarizing filters at right angles, there is no orientation of light that can make it through both filters so the result is zero light transmission. If you vary the angles (most camera polarizing filters rotate to allow this) so they're not quite at 90 degrees you'll let a very small amount of light through, and thus get an ND effect allowing long exposures ect.

The best thing about doing this is that you can vary the strength of your ND filter. The only thing you need to do is get two filters of the same size and make sure the front most polarizing filter is not a circular polarizer.

A non-circular polarizer doesn't mean it's square! Just that after filtering out all but a certain orientation of polarized light, the filter mixes up the orientations of the light coming out the other side. This is done because polarized light with only one orientation messes with the camera's AF.

The only downside of this is that stacking filters can cause vignetting with wideangle lenses.

Improve this answer
\$\endgroup\$
4
\$\begingroup\$

I just got a Singh-Ray Vari-ND filter for taking longer duration images. It is essentially what you want to do in a commercial setup. The paperwork with it clearly warns of the color effects mentioned above that you may see on some lenses.

Improve this answer
\$\endgroup\$
1
  • \$\begingroup\$ Yea, beware color casts when stacking dense filters. \$\endgroup\$
    – Shizam
    Commented Oct 31, 2010 at 4:29
2
\$\begingroup\$

Speaking from my own experience, I would rather carry different ND filters that combine two polarisers. I stopped using polarisers as I didn't like the way they affect the colours and saturation. I guess two polarisers together will affect the colours even more. Anyway, good luck with your experiment

Improve this answer
\$\endgroup\$
2
\$\begingroup\$

I guess this is an old thread, but I do have some experience with double polarizers. I first came across the idea while attempting to attenuate the IR signal getting to my Cybershot F717's sensor, so that I could take infrared pictures in broad daylight. For that, they do an excellent job.

I have 4 polarizers here, 2 circular and 2 linear, which I have tried in varies combinations; resulting in considerable variation in their abilities to block light, and considerable differences in the hue that they cast on the images (I'm speaking of non-IR here). However, I did find one set that seems to have very little effect on colour.

But there is a difference between using double polarizers and ND filters. The ND filters seem to simply reduce the amount of light getting in; but I sense something different going on with the polarizers, especially near the extinction point, if I can call it that. I notice this especially when taking picture of landscapes with cloudy horizons. The polarizers seem to make the clouds pop out of the sky, more than the ND filters. I can't explain it. Also when one wants to get just the right amount of contrast (in the sky) it's much easier with the double polarizers.

Improve this answer
\$\endgroup\$
1
\$\begingroup\$

A properly rotated polarizer can reduce reflection, but two polarizers with perpendicular axis just reduce the amount of light passes, despite the polarization of the light. This is because light, as a transverse wave, has only two axis of polarization. If both directions are reduced by the same fraction, the result is just a reduction of the overall intensity.

So two such polarizers in theory is identical to a ND filter. But because of the practical reasons mentions by others, you may want to choose just a ND filter.

EDIT: Well, the real reason that light only has two direction of polarization is that photons are massless. If photons are massive, it will have an extra massive polarization. Just so you know~~ :-)

Improve this answer
\$\endgroup\$
0
\$\begingroup\$

My experiment with crossed polarizers (2 cpl, one reversed) has convinced me that they are not the solution I was looking for. Crossed as dark as I could get them worked about to about 16 stops -- so I was shooting 5 minutes at F22 at ISO 200 in daylight. The vertical and horizontal banding was worse than I expected, and the purple color cast was unfixable (by me, at least) in Photoshop.

Improve this answer
\$\endgroup\$

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Not the answer you're looking for? Browse other questions tagged or ask your own question.