I just ordered a linear polarizer.

IF the type of camera we own is a DSLR, then we are used to exchangeable lenses and lens-attachments. One type of attachment that has always been available since the days of SLRs, is a polarizing filter of some sort.

But a DSLR differs from the older, film-SLRs, also, in having an auto-focus system which is physical, and which fails to work with linearly-polarized light, aka plane-polarized light. And so where ‘in the old days’, we used to buy linear polarizers, these days, we’d buy circular polarizers, for use with DSLR cameras. In the pursuit of photography, they do everything a linear polarizer used to do, and never prevent the auto-focus from working.

This also has as practical consequence, that it will be difficult in most localities, to find a photographic retail store, that still sells linear polarizers. Usually, such stores only sell the so-called Cir-Pols. Yet, because the linear polarizer has always been technically easier to design and manufacture, than a cir-pol, it only stands to reason, that the linear variety can still be bought somewhere. It’s just that retail stores need to pay their rent, and frequently can’t make a profit out of stocking those.

One place where one can certainly still buy those is on-line, which is where I just ordered one. And I did so, because I wanted to test the hypothesis of This earlier posting. In other words, where I’ve run into a self-imposed question which I cannot find the answer to, just by thinking about it, I now intend to find the answer, using the Scientific Method.

(Updated 02/21/2018 : )

The subject is mildly irritating, that research organizations exist, which have much-deeper pockets than I do, and who can easily afford to carry out the simple experiment, which I now intend to carry out. But, I can not find a direct answer to my question easily, which those research organizations could have published. Yet, because in this blog, I’m also the person who proposed a hypothesis, it still isn’t far from correct, that I should be the person to test that.

I should add, that if what the reader is used to, is a smart-phone camera, where, within the preview window, he can tap on any part of the picture, and where the camera then brings that part of the picture into focus, it also seems logical that such a camera has a different type of auto-focus, from what the DSLRs had. What seems to make the most sense here, is that the software takes a sequence of measurements, with the physical focus set to different distances. And at each setting, the software can also compute, how high the spatial frequencies in one sub-rectangle of the picture become. The software then computes, which focus-setting gave the maximum sharpness, and quickly backtracks to a computed setting, that gives the best sharpness.

If this type of software-based auto-focus was assumed, then to put a linear polarizer in front of it would still work just fine. Only, I don’t generally see a lot of cell-phone cameras, that accept standard attachments. Those are usually self-contained.

The direction of plane-polarized light which a polarizer should transmit, to give the best results in photography, is quite consistent. Because glare tends to be polarized parallel to the horizon, the best setting is usually to cut this glare, and to transmit plane-polarized light oriented perpendicularly to the horizon. Yet, in the case of cell-phone cameras, the designers know that the camera will be held in landscape-orientation, or in portrait-orientation, so that to install a polarizer in one fixed orientation, will generally be a mistake.

Further, there is a trick which polarizer can be used for, which is to photograph somebody or something which is behind a glass window, but to minimize the appearance of reflections, due to the window. And in that case, the photographer needs to filter out the light, plane-polarized parallel to the surface of the glass, which in turn involves rotating the polarizer to some odd angle, on a mechanical ring, that attaches it to a camera.

The photographer can usually see in his viewfinder, at what position he has obtained the effect he wishes to obtain, and to have some sort of handle on the ring of the polarizer itself, to denote the plane of polarization it will transmit, is usually not necessary in practice.


There’s an observation to add, about a side-effect, if a circular polarizer is designed in one specific way. The logical way in which those designed for photography work, is as if first to plane-polarize the light, and then, to circularly polarize the emerging beam.

If the polarizer is designed to do this in two discrete stages, then the second stage would be a birefringent substance, with its Axes at a 45⁰ angle to that of the linear polarizer. The electrostatic wave along one axis would be made to phase-shift 90⁰ with respect to the electrostatic wave along the perpendicular axis, of the birefringent, so that circularly-polarized light emerges in principle.

The fact to note about this, is that it will only succeed fully, at one ideal wavelength. Due to whatever thickness of the birefringent chosen, non-ideal wavelengths of light, will also be phase-shifted by something other than 90⁰, in a completely predictable way. The result of this will be, that light not at the ideal wavelength will only emerge elliptically-polarized. This can be represented according to the hypothesis linked to above, as a considerably smaller quantity of photons being of one intrinsic handedness, than are of the other intrinsic handedness.

The elliptical nature can be minimized, because the band of visible wavelengths is less than an octave wide. I.e., The shortest wavelengths we can see (‘violet’) , are just longer, than half of twice the longest visible wavelength (deep red). And those violet wavelengths emerge just as a visual anomaly, beside deep blue. Well the ideal wavelength can be chosen, as the shortest wavelength multiplied by the square root of two. That way, the ratio by which the wavelength of each end of the visible spectrum, is wrong, will be the same, and the error can be minimized.

There is a reason why nobody would care about a circular polarizer producing, largely, elliptically-polarized light, when working with photography: It will not stop the auto-focus from working, and it will also not seem to alter the spectral composition – i.e., the overall coloration – of the image.

But when dealing with Physics, circularly-polarized light is supposed to be circular, not elliptical.

(Edit 02/21/2018 : )

The polarizer arrived yesterday, and I carried out the experiment it was intended for. One observation which I left out of the ‘lab-report’, because it would have created a reference to more topics, was that the coloration of the light coming out of the (second) linear polarizer did in fact change, as it was rotated with respect to the first. At one angle, the light was decided golden, then yellow, then gray, then turquoise, then sky-blue. The earlier text in this posting explains why this could happen.

I only found that this made it harder to inspect visually, casually, whether the average intensity did in fact correspond to ‘50%’. I did not get the impression anyhow, that the average intensity corresponded to ‘25%’, which I would have noticed easily, as being darker than 50%.



Print Friendly, PDF & Email

Leave a Reply

Your email address will not be published. Required fields are marked *

You may use these HTML tags and attributes: <a href="" title=""> <abbr title=""> <acronym title=""> <b> <blockquote cite=""> <cite> <code> <del datetime=""> <em> <i> <q cite=""> <strike> <strong>