[Mike Cash]

Just some side-by-side comparisons of shots done with a Marumi linear polarizer on the K20D through a Super Takumar 85/1.9. The shots on the left/top are done with the filter rotated to kill the filter's effect and on the right/bottom with the filter rotated to maximize the filter's effect.

Shot in Av mode with no EV adjustment as JPG files. No post-processing other than combining shots into single files and resizing.

Subjects were chosen for usefulness for polarizer comparison, and not for any compositional merit. (Don't want folks thinking I took these thinking they're good pictures!)

Conclusion:
Linear polarizers work just fine with the K20D and manual focus lenses. There doesn't seem to be any noticeable metering problem. I'm damned glad I paid under $20 for a linear polarizer instead of four times that much for a circular polarizer.

I haven't tested this with an auto-focus lens yet, so have no idea if the linear polarizer affects the auto-focus system or not.

[Rener]

Just some side-by-side comparisons of shots done with a Marumi linear polarizer on the K20D through a Super Takumar 85/1.9. The shots on the left/top are done with the filter rotated to kill the filter's effect and on the right/bottom with the filter rotated to maximize the filter's effect.

Shot in Av mode with no EV adjustment as JPG files. No post-processing other than combining shots into single files and resizing.

Subjects were chosen for usefulness for polarizer comparison, and not for any compositional merit. (Don't want folks thinking I took these thinking they're good pictures!)

Conclusion:
Linear polarizers work just fine with the K20D and manual focus lenses. There doesn't seem to be any noticeable metering problem. I'm damned glad I paid under $20 for a linear polarizer instead of four times that much for a circular polarizer.

I haven't tested this with an auto-focus lens yet, so have no idea if the linear polarizer affects the auto-focus system or not.

hi Mike, very nice test. what is the difference between the linear and circular polarizer. thanks.

cheers,

rene

[ChrisN]

hi Mike, very nice test. what is the difference between the linear and circular polarizer. thanks.

cheers,

rene

From what Mike has said, about $60.

I've seen comments that the older linear polarisers will cause metering problems. I also also use the old linear polariser that's been in my bag for many years, with no problems with either manual-focus or auto-focus lenses.

[stevebrot]

hi Mike, very nice test. what is the difference between the linear and circular polarizer. thanks.

cheers,

rene

That is a simple question with a complicated answer! The short/easy answer is that the polarizing qualities of a linear polarizer work on a single plain in a straight line. These filters are generally made to allow rotation in the filter mount. To use, you mount the filter and rotate the polarizer until you get the effect you desire.

Circular polarizers also polarize the light, but the material is made in such a way that the light is polarized in multiple planes in a circular pattern. (Note: other forum members may feel free to correct my explanation!) The result is that you get most of the benefits of the polarizer without having to rotate the filter.

There apparently are issues with the metering/auto-focus systems on some cameras when a linear polarizer is used. As a result many users have switched to the circular version.

Steve

[bdery]

The information given above by stevebrot is sadly innacurate. To his defense, polarization is one of the most complicated concepts in optics and photonics.

A linear polarizer rejects light polarized along one axis, and transmits light polarized on another axis, erpendicular to the first one. For simplicity's sake, let's keep it simple, and say that light is made of two polarization components, perpendicular to one another. It's a little more complicated than that, but it doesn't need to be, here After light passes through a LP, it is linearly polarized in one direction. If light was ALREADY polarized (this happens, for instance, when an unpolarized source is reflected on a non-metallic surface) then the polarizer will either

1-let the light pass through, if its orientation is matched to that of the polarizer

2-reject the light, if the polarization orientation is not matched.

3-a combination of both, if the orientation is partially matched

A circular polarizer is a linear polarizer. It does the same thing. the difference is that, AFTER the LP component, there's another component, that will spread the light (which is polarized on ONE axis) over the two polarization axes. To give you an image, let's say light before the LP looks like this : + . After the LP it looks like this : = . After the CP it still looks like this : + . So if you send light oriented like this : | to a LP oriented like this =, nothing will pass. If you send light oriented like this : / , some of it will pass (in this case, the horizontal component).

This may be too complicated, or badly explained, I'll be happy to tell you more (or to say things differently) if needed.

Now, about metering... the thing is that the mirrors inside the body polarize the light themselves, and use the fact that light has two polarization orientations, to send light where it's needed for metering, AF, etc. If the mirrors and polarizers were perfect, metering would probably not work at all. But since those components are NOT perfect, some light can still be sent to the parts of the body that needs them. But remember that it CAN be innacurate in some cases (particularly with wide lenses and in low light). Personally, I'd rather pay for a CP without coatings than a LP with coatings.

My two cents.

[mattdm]

Circular polarizers also polarize the light, but the material is made in such a way that the light is polarized in multiple planes in a circular pattern. (Note: other forum members may feel free to correct my explanation!)

I'm not really an expert, but I'll give it a shot.

The result is that you get most of the benefits of the polarizer without having to rotate the filter.

Nah, you still rotate the filter to get the effect how you want it. The circular part doesn't actually refer to the orientation of the filter. Rather, the light waves that get through the filter are regularized, but not in a line: their electromagnetic fields basically define a perfect spiral over time (or, to put it another way, a circle at a fixed point in space).

This is important (in theory, but as the test above demonstrates, apparently not in practice) because the the half-silvered mirror in an SLR (which splits light between the viewfinder and the metering/autofocus sensors) also has a linear-polarizing effect, and if you've ever looked at an LCD screen wearing polarized sunglasses, you know how that goes.
So, there's a kind of filter called a quarter-wave plate that takes linearly polarized light and "twists" it so that it's circularly-polarized instead. As I understand it, a "circular polarizer" is a linear polarizer with a quarter-wave plate behind it. This guarantees that you won't have metering problems.

And that explains why they have the same effect — there's the exact same filtering going on in the first part of the circular version. And it explains why the circular filter is much more expensive — it's got more to it.

I'd heard (and repeated) before that circular polarizers don't have as strong an effect, but I'm not sure now that that's actually true. It may be that in order to compensate for the extra layer in terms of light loss and other optical-quality issues they tend to have a less-strong linear polarizing component. Dunno.

I'd be curious to see a careful test with rotating a linear polarizer carefully throughout 360° — it seems to mean that at many orientations a problem wouldn't be visible, but get it just wrong and you have a blackout.

[Arpe]

I've seen comments that the older linear polarisers will cause metering problems. I also also use the old linear polariser that's been in my bag for many years, with no problems with either manual-focus or auto-focus lenses.

Same here.

[pw-pix]

Linear polarisers cause problems where a beam splitter is used either for auto-focus or for light metering.
I have an old Cokin linear polariser and I used it on my K10D a week or so ago and didn't notice any real problems. I used it all the time on my KM A2 with no issues either.

[stevebrot]

I'm not really an expert, but I'll give it a shot.



Nah, you still rotate the filter to get the effect how you want it. The circular part doesn't actually refer to the orientation of the filter. Rather, the light waves that get through the filter are regularized, but not in a line: their electromagnetic fields basically define a perfect spiral over time (or, to put it another way, a circle at a fixed point in space).

This is important (in theory, but as the test above demonstrates, apparently not in practice) because the the half-silvered mirror in an SLR (which splits light between the viewfinder and the metering/autofocus sensors) also has a linear-polarizing effect, and if you've ever looked at an LCD screen wearing polarized sunglasses, you know how that goes.
So, there's a kind of filter called a quarter-wave plate that takes linearly polarized light and "twists" it so that it's circularly-polarized instead. As I understand it, a "circular polarizer" is a linear polarizer with a quarter-wave plate behind it. This guarantees that you won't have metering problems.

And that explains why they have the same effect — there's the exact same filtering going on in the first part of the circular version. And it explains why the circular filter is much more expensive — it's got more to it.

I'd heard (and repeated) before that circular polarizers don't have as strong an effect, but I'm not sure now that that's actually true. It may be that in order to compensate for the extra layer in terms of light loss and other optical-quality issues they tend to have a less-strong linear polarizing component. Dunno.

I'd be curious to see a careful test with rotating a linear polarizer carefully throughout 360° — it seems to mean that at many orientations a problem wouldn't be visible, but get it just wrong and you have a blackout.

Boy, I was way off-base!

Thanks for the corrections mattdm and bdery!

I have to admit to being almost clueless regarding circular polarizers, but I do know how the linear version works. That is why I put the comment in regarding accuracy. My ignorance regarding the circular polarizer is based on being out of the photo world for about five years during which these items came to market. Having made my confession, I can say that my understanding is similar to the explanation given by mattdm. Having never actually worked with a circular polarizer, I just assumed that the need to rotate the filter was gone. After some consideration, that assumption was pretty silly.

So let me get this straight.

Linear polarizer: Bad for AF systems, works for MF

Circular polarizer: Works for both AF and MF systems

Steve

[mattdm]

Linear polarizer: Bad for AF systems, works for MF

Theeeeeoretically, bad for both, because both rely on the light split by the mirror. But in practice, I haven't heard anyone with a Pentax (or any other brand, but Pentax is where I've been watching) dSLR report real-world problems.

Circular polarizer: Works for both AF and MF systems

Right.

[PePe]

Early this year I had to get a polarizer right away for one particular event. The local shop had run out of circular polarizers of the right size, so I had to go for a linear one.
I have been using it quite a lot on my DA*16-50 and K10D.
No problems whatsoever!

[sewebster]

I think your comparisons aren't really fair... what you probably really want to compare is the shot with the polarizer OFF the lens and the shot with the polarizer set for the best effect. If you just rotate it 90 degrees then you are selectively choosing the polarization that you don't want, the glare, and this shot will be worse than using no filter at all...

[Mike Cash]

I think your comparisons aren't really fair... what you probably really want to compare is the shot with the polarizer OFF the lens and the shot with the polarizer set for the best effect. If you just rotate it 90 degrees then you are selectively choosing the polarization that you don't want, the glare, and this shot will be worse than using no filter at all...

I really bristle at your characterization of the comparisons as not being "fair".

I presented them as nothing more than what they are, and stated quite clearly what they are. And I resent your assumption that what I "probably really want to compare" was something other than what I did compare as it insults my intelligence.

I compared what I wanted and intended to compare. I presented the comparison photos with a description outlining exactly what I did. If they fail to meet what you would prefer them to be, then by all means feel free to do comparisons under the circumstances and for the purposes that you like and then share the results with us.

[sewebster]

I really bristle at your characterization of the comparisons as not being "fair".

I presented them as nothing more than what they are, and stated quite clearly what they are. And I resent your assumption that what I "probably really want to compare" was something other than what I did compare as it insults my intelligence.

I compared what I wanted and intended to compare. I presented the comparison photos with a description outlining exactly what I did. If they fail to meet what you would prefer them to be, then by all means feel free to do comparisons under the circumstances and for the purposes that you like and then share the results with us.

Whoa. I didn't mean my comment in some sort of negative way, so there is no need to be offended. Sorry.

Part of the reason I posted my comment was your choice of the wording "filter rotated to kill the filter's effect." I don't know what you mean by this I guess. I think it would be hard to rotate the polarizer to "kill the effect." You'd have to compare the view through the finder to the view in reality and keep rotating until you got it right. I _thought_ what you did was rotate the filter between the two extreme points: most glare and least glare. In this case, the "most glare" position is not an elimination of the effect of the filter, it is exacerbating the glare problem.

But anyway it seems you already know this and were just doing some linear filter tests on the K20D so that is fine. Didn't mean to be upsetting. Your photos clearly demonstrate the large degree of polarization that is present in reflected light, and that is interesting to see.

[Mike Cash]

The point was that linear polarizers can be used on DSLRs, at least with the equipment and under the circumstances I outlined (K20D and manual focus) with no apparent detriment to metering ability. The conclusion was that for use under such conditions there is no need to lay out four times the cash for a circular polarizer.

I did eyeball the scenes and the glare with the filter 90 degrees from the maximum glare reduction position was similar. Whether having the filter mounted and in that position made the glare worse was irrelevant as it was already so bad that some sort of polarizer would be necessary to salvage a shot from it anyway. The point wasn't whether polarizing filters in general produce an effect or not; that much was a given.