Calling analytical: more polarizer mysteries

As I understand the difference between linear and circular polariser is that the first will prevent AF from working properly. I never had the chance to compare them but I would expect that they both should work the same. I got B&W and Hoya Pro polarisers (both circular), I'll check them over the weekend and see if they display same behavior pattern.

Think of a LINEAR polarizer as fence posts. Light moves in waves, but unlike ocean waves, light waves can be tilted sideways in various directions. (I’ll call this orientation or alignment). Waves that are aligned in the correct direction can pass right through those fence posts. Light that is aligned in the wrong direction does not pass through the fence posts because the waves cannot fit through the fence posts.
(example image: http://www.edbergphoto.com/pages/Tip-polarizers.html)

Normally light waves are in random orientations, and any polarizer blocks a portion of these waves, making the light appear fainter. However, when light bounces off of a reflective surface, such as water or glass, a lot of the waves line up in the same orientation, causing glare. By using a linear polarizer in the correct orientation, you can block many of these reflected light waves and reduce the glare. A polarizer can also be used to make the sky appear bluer for these same reasons.

A circular polarizer works much the same as a linear polarizer, but after the light passes through those fence posts, the polarizer gives the light a “twist” of sorts. Because they have “fence posts,” they should be equally good at blocking glare.

Afterwards, however, LINEARLY polarized light continues aligned in the same direction.
The waves in CIRCULARLY polarized light spin or spiral as they move forward. So the waves spiral in orientation as they move. (ex drawing: http://www.air-stream.org.au/Polarization)
Remember, this is the orientation of the waves of light. It’s not something you can see with your eye. Your eye just senses the brightness and the color (frequency) of the light. It can’t sense orientation of the waves.

Why this matters to cameras. To correctly autofocus/autoexpose, some cameras use a beam splitter. It splits the light into 2 paths. But a beam splitter may not work correctly with light that is linearly polarized. It could end up blocking some or all of that light, depending on how the light and the beam splitter are oriented compared to each other. Remember, the linearly polarized light is all oriented in the same direction, and that direction is not changing. Beam splitters, however, work fine with circularly polarized light because that light has a changing / spiraling orientation. If you’re willing to manually expose/focus, I believe you should be fine if you use a linear polarizer with a modern camera, although I haven’t tried this myself.

To answer your other questions:
I'm not sure why your CPL doesn't work by eye when it does work with this camera. This doesn't make sense. Maybe you aren't looking at a object with enough glare to strongly show the polarization effect.

Most circular polarizers are listed at "circular polarizer" or CPL, while anything listed as "polarizer" is most likely linear).

To test your filters to see if they are actually circular / linear polarizers try the following:
Hold a filter in front of you, one in each hand, pointed at a bright reflective surface, and twist the filter. You should be able to see the brightness change.
Now stack two filters, one in each hand. Gradually twist them around, one by one.
1. If you can turn them into positions where very little light passes through, they're either (a) both linear filters, or (b) the furthest filter is linear and the filter nearest to your eye is circular.
2. Switch the positions of both filters and repeat. If you can only get it dark in either #1 or #2, then you have one linear and one circular. If you can get it dark in both #1 and #2, you have two linear.

Stacking a linear and a circular can only become dark if the linear is furthest from your eye and the circular is nearest your eye. Use this info to figure out which is which.

I'm guessing that you might get a slight green/purple tint if the polarizer blocks certain colors of light unevenly, or if it confuses the camera's exposure/white balance sensor.
You can stack (linear or linear/circular) polarizers to create a "neutral density" filter where you can adjust the brightness. However, when I tried this, I found a purplish haze that I didn't like. It also stuck out enough that it blocked part of the frame.

The answer was a bit technical. I hope this helps. You can also find plenty on the web if you search.


Edited by Micholand - 03 February 2012 at 19:53

The circular polarizer (as defined for camera filters) is a linear polarizer with a quarter wave plate on the back that "unpolarizes" enough light to assure that the linear AF sensors always get some light. It does not have to produce perfectly circularly polarized light.   The image effect depends only on the selective dimming, not on whether the light is polarized when it hits the sensor, so both types have the same image effect.     

As mentioned already the purpose is to avoid AF problems. No effect on the main sensor whether digital or film. But this needed only when the polarization is least favorably aligned to the AF sensor. (By "sensor" I mean each individual line in the AF sensor pattern.) With more recent camera designs that have oross sensors, multiple sensor locations and likely more sensitive sensors, the circular filter is probably unnecessary. In any case I have never had a problem with Sony or Minolta DSLRs using linear polarizers, including old Minolta polarizers.

The Minolta and B+W filters many look different for a couple reasons. The spacing of the filter lines will affect the darkening. The Minolta may be more closely spaced and thus polarize more completely. (Less light gets through off the intended polarizer direction.) You can crudely compare by seeing which requires the greater exposure adjustment from no filter.      

The color effect is likely a "thin film" effect of the quarter wave plate (which is likely a coating of a specific composition and thickness) or of other coatings on the filter. The differences could be from degree of polarization, whether the quarter wave plate is a surface coating or a sandwiched between two glass elements, nature of front coating, etc.

But the B+W should produce some effect with rotation if it is a polarizer. Look at some uniform source like a sheet of paper through both filters at once. If there is no darkening when they are at right angles, then the B+W is not a polarizer.       

Edited by analytical - 03 February 2012 at 20:06

Whoa! Lots of (interesting) information on how polarizers work, but not really answering my question. Sorry if it got hidden in the OP.

So, let me isolate it here:

1. No polarized eyeglasses. Would not matter, because all else is equal, only the filters changed.
2. Strong reflection on a floor tile and on a wooden desk.
3. Viewing through both filters under equal conditions, I get a darkening effect from the Minota and no darkening whatsoever (not even slightly) from the B+W. However, the B+W adds a greenish and then purple tinge.
4. On the camera, both work as expected, with completely eliminating the reflection at the maximum setting. No color tinge visible with the B+W there.

So, again the question: why is the B+W not reducing the reflection in scenario 3, while the Minolta filter does.


To follow up on this suggestion:

ifreedman wrote: To test your filters to see if they are actually circular / linear polarizers try the following: Hold a filter in front of you, one in each hand, pointed at a bright reflective surface, and twist the filter. You should be able to see the brightness change. Now stack two filters, one in each hand. Gradually twist them around, one by one. 1. If you can turn them into positions where very little light passes through, they're either (a) both linear filters, or (b) the furthest filter is linear and the filter nearest to your eye is circular. 2. Switch the positions of both filters and repeat. If you can only get it dark in either #1 or #2, then you have one linear and one circular. If you can get it dark in both #1 and #2, you have two linear. Stacking a linear and a circular can only become dark if the linear is furthest from your eye and the circular is nearest your eye. Use this info to figure out which is which.

If this test is reliable, then my B+W would be a linear polarizer. However, it works well on all my lenses and DSLRs as far as AF is concerned and, as outlined above, also for the polarizing purpose, and it clearly says "C-POL" engraved on it.

If anyone reading this has one of the B+W CPLs, would you mind doing the eye visual test as outlined in my #2&3 and report? Thanks.

Hi Arno,

I have a 55mm B&W CPL (non-MRC) and Minolta 62mm and 77mm CPLs. I'll give it a look over the weekend. The Minoltas as I remember are AC. Its a mystery to me as to why you are seeing a difference. Coatings? Maybe a German engineering marvel? The Schott Glass B&W uses is reputed to be superior to Japanese glass.

Regards,

Kelly

analytical wrote: In the hand held tests are you holding the filter so you on the camera side of the B+W? If you have it reversed, then if the quarter wave plate is efficient, it would circularize the incoming polarization from the reflection before it reached the linear polarizer layer.

BINGO!

Wow, I had no idea that these filters could be reversed, and that is what I had consistently done. Once I looked through it like the camera does, it worked as it should, perfectly eliminating the reflection.

The test described by ifreedman also assumes that the cpl has camera side toward you.

Again, BINGO!

With the proper position, it now shows that the old Minolta pol is a linear one, and the B+W comes out as a circular.

Okay, that answers everything for me. Thank you so much.

Excellent point Analytical. Good catch w/ the reversed filters! Glad you figured this out.