Two linear polarisers or two circular polarisers - to make a variable ND

[GrahamH]

I remember from school days using sheets of polarisers. I assume they were linear). Rotating one 90 degrees to the other cuts off most of the light (effectively all).
I was wondering if two linear polarisers could be used as a variable neutral density (ND)?
Would it work with two circular polarisers?
One of each?

Anyone tried?

A CPL recently broke and thinking of ways to resurrect it.
Graham

[Ken MT]

You will need to use circular polarisers if you are using a SLR with through the lens autofocus and metering. The reason for this is that both of these systems use semi-silvered mirrors to siphon off some of the light coming though the lens. If that light is linearly polarized it renders either the metering or the autofocus ineffective.

[Mark von Kanel]

Graham variable ND filters are normally made of 1 circular and 1 liner filter i think, have a read of this article:

http://digital-photography-school.co...density-filter

[GrahamH]

Thanks Mark.
Can save me a couple of hundred with this.
My original CPL (Hoya) has separated the screw on scetion and the polariser section. I can glue the two together and find an el-cheapo linear as suggested.
Graham

[revi]

Note that you have to use a linear and a circular polariser, with the circular closest to the lens/body. Two circular polarisers won't give you a variable ND filter (check up on what comes out of a circular polariser and on how they work...). With two linear polarisers, otoh, you'll have trouble with your AF and light metering. And putting the linear between circular and lens gives you the worst of both worlds: no vari effect and metering/AF problems.

I seem to remember older posts here on this subject, where it was said that such constructions (like cheap vari-ND filters) can give colour casts on part of the scene. Such colour casts are difficult or impossible to correct.

[Colin Southern]

I've heard the question asked a few times, but to date, nobody had come back and said that it worked well.

I suspect that there are a few "gotchas" we don't know about.

[jcuknz]

Well don't waste you time with two cpl's becuase it doesn't work and a pair of cpl's will only give you a three stop reduction whereas the CPL plus LPL starts at 3 stops and gives you more/less depending on how you think of it.
Trouble is that LPLs are rather more expensive than CPLs, why I for one have yet to investigate a LPL to see if the theory works in practice

[revi]

A circular polariser as used in photography is a combination of a linear polariser and a quarter-wave plate (see http://en.wikipedia.org/wiki/Polariz...lar_polarizers). So if you combine a linear polariser and a circular polariser, the incoming light 'sees' (in order) the 1st linear polariser, the 2nd linear polariser and the quarter-wave plate.
Note that putting the linear polariser behind the circular polariser isn't very useful either...

What might cause problems is that the quarter-wave plate is made of birefringent material, which means that its effect is different for different wavelengths (=colours!). I'm not sure what the effect in practice
would be, but given that different parts of a scene have different colours and different polarisations of the light, I can imagine that the combination gives colour shifts that vary over a scene in a way that's hard to predict and hard to correct.

[Colin Southern]

In practice, I just use my Singh-Ray Vari-NDs and get on with making pictures!

[GrahamH]

In practice, I just use my Singh-Ray Vari-NDs and get on with making pictures!

Exactly my point - from Adorama $390 (plus tax, shipping whatever). Dead CPL plus cheapo linear (costing just a few bucks).
Rather save my money for something else (like food).
Graham

[revi]

Well, Colin is a professional photographer (i.e. he gets at least part of his income from photography) so investing a
few hundred to assure good results and save on PP time looks like a good move.

And there are other vari-ND filters that are cheaper (I don't use any, so no experience with them);
http://www.learningdslrvideo.com/var...lter-shootout/ lists 6 of them (including the Singh-Ray)

Not sure the combination of a CPL and a linear polariser would be cheaper than a vari-ND by the way. No to mention
that the cheapest filters might give more glare problems (as you add at least 4 interfaces in the optical path)

@Colin: does the Singh-Ray also work as a polarising filter?

[Colin Southern]

Exactly my point - from Adorama $390 (plus tax, shipping whatever). Dead CPL plus cheapo linear (costing just a few bucks).
Rather save my money for something else (like food).
Graham

I take your point, but this is something I've seen quite often, where ...

1. People think the Singh-Ray is too expensive - so they don't buy one - and thus don't get the images that owning one would have allowed them to.

2. People think the Singh-Ray is too expensive - so they buy a cheaper one - and then start posting about colour shifts and interference patterns.

3. People think the Singh-Ray is too expensive - so they try to make their own - but (a) it doesn't work and (b) if they'd put the time they spent trying into something that produced a $$$ return (eg overtime at work etc) then in the time they wasted, they'd have enough $$$ to buy the Singh-Ray.

So although the above rationale won't apply to everyone, I think that for many it's a classic case of trying to save $$$ actually costs more (in $$$ terms or image terms or both) in the long run.

Personally, I own two of them (77mm & 82mm) - they're extremely high quality - and I know I've got them for the rest of my life. I've had then for many years now and I've long since forgotten about what they cost initially.

[Colin Southern]

Well, Colin is a professional photographer (i.e. he gets at least part of his income from photography) so investing a
few hundred to assure good results and save on PP time looks like a good move.

You can't EASILY replicate an ND in PP - you'd have to stack a zillion shots and average them - and that's a major PITA.

And there are other vari-ND filters that are cheaper (I don't use any, so no experience with them);
http://www.learningdslrvideo.com/var...lter-shootout/ lists 6 of them (including the Singh-Ray)

I haven't looked at the market recently, but I've heard of problems with the likes of the lightfader. I think a lot of people just think the Singh-Ray is over-priced and thought they could do it a lot cheaper - only to discover it wasn't as easy as they first thought.

@Colin: does the Singh-Ray also work as a polarising filter?

No - but they do make a couple of Vari-ND derivatives with a polariser.

[revi]

You can't EASILY replicate an ND in PP - you'd have to stack a zillion shots and average them - and that's a major PITA.

Sorry,I wasn't clear here. I tried to say that it's worth investing in good quality to avoid spending time correcting the errors from a cheaper alternative,
I didn't mean to imply to you would use PP to imitate the effect.

Remco

[Inkanyezi]

The "variable ND" consists of two circular polarisers with their linear sides facing each other, i.e. the QW plate of the first facing front and the last one facing the lens. Most of them are not ND, but close to ND, and when set to very high density present clearly visible and uncorrectable colour shifts. Those filters are properly designated as "fade filters" and used for fade-in or out when shooting movie.

The reason for using two circular polarisers is that polarisation of the incoming light is unwanted, as well as polarisation of the outgoing light. They were made for movie in the same fashion also before the event of electronic cameras, as the beam splitters of most zooms for movie also polarised the light.

It is quite simple to make such a fader with two circular polarisers, mounting one in the normal direction in a regular filter holder, and flipping the glass of the other one so that the circular side faces front, then screwing it onto the one that cannot be moved. The caveat of course is that it might be less useful for other purposes than a fader, as you may expect colour shifts when it is set to high densities.

Before digital or Pellix, a "variable ND" would not need polarisation broken at the last surface, but still needed that incoming light would not be polarized, so there might have been such filters with only one circular face (quarter-wave plate). The polarising vari-ND has no QW filter at its front end.

The colour shifts, as I understand it is in part a matter of polariser quality but also an unavoidable physical property. A very high quality polariser would display less colour shift than a lower quality one. When polarisers are combined to very high density, such colour shifts are unpredictable and may be severe.