My Not Nearly Complete, But Rather Entertaining, Circular Polarizer Filter Article

Published September 15, 2017

So, a while back I wrote a not quite complete article on UV filters. To do that, I had to buy new testing equipment and learn to test filters. This was not what I wanted to do when I grew up. But somebody has to do it, and I did get to buy new toys.

More importantly, Tyler (Who handles the purchasing) asked me why, many years ago, I chose the Circular Polarizing filters that Lensrentals stocked. A better person than me would have confessed that I’ve never known the first thing about Circular Polarizers; that I just bought the most expensive to be our ‘best’ and the cheapest to be our ‘basic.’ But instead, I just said, “Well, we should do some scientific-type testing and a more thorough evaluation now.”

Like a metaphor for my life, the results ended up being the opposite of what I expected. I thought if we found tons of differences testing simple clear and UV filters, there would be many more differences in more complex polarizing filters. So today, instead of showing you amazing differences between the various brands, I’ll just save you some money on your next CP filter purchase. That should work out for both of us: you save money, I get a shorter blog post.

I should mention our methodology has improved somewhat since we did our first filter article. I used a red laser to measure transmission then, and several people made the very reasonable suggestion that a green laser might be a better choice, being in the middle of the spectrum and all. Other people said I should get a spectrometer and measure the entire spectrum. So I did both of those things.

This should give you some hints about me as a person. If someone asks me to test a lens at a different aperture, I have a screaming fit about testing taking up a few hours of my time. Someone suggests I spend $15,000 on new equipment, and I’m like, “Yeah, great idea, that would be cool to have.”

Finally, I had no intention of testing every polarizing filter on the market. I did what I usually do; bought the ones B&H Photo had in stock in 77mm size. So we’re going to compare, in no particular order, except alphabetical, with the current price for a 77mm:

  • B&W XS-Pro High-Transmission Circular Polarizer MRC-Nano         $102
  • Heliopan Circular Polarizer                                                                          $200
  • Marumi EXUS Circular Polarizer Filter                                                     $140
  • Sigma Water Repellent Circular Polarizing Filter                                   $150
  • Tiffen Ultra Pol Circular Polarizing Filter                                                 $103
  • Zeiss T* Circular Polarizing Filter                                                               $180

So, About the Polarizing Part

If you’re thinking about buying a circular polarizing filter, you probably want to know which ones polarize the best and which ones the worst, right? I know I did. Now we could have just gone outside and taken pictures in the bright sun and said this one’s good and that one’s bad. But we never do anything simple when we can complicate the crap out of it.

So what we did was take our laser transmission set up and modified it a bit. Primarily, this shines a laser into a power meter and gets a reading. Then we can stick mostly transparent stuff in the beam and see how much it reduces the power reaching the meter. Excellent lasers are almost entirely polarized, but I have a budget, and that budget didn’t include (much to my sorrow) $10,000 for the lab-grade, steel-melting laser I wanted. I got a little 5-mwatt green (530 nm) diode laser.

It’s sort of polarized. So we shined that laser through two sheets of polarizing film, each of which have a 1,000:1 polarizing extinction ratio. So basically, the light that passed through the film was really, really polarized. Then we put the filter in the beam of polarized light to see how much light it let through in the open position. For now, we’ll just call it ‘most of the light.’ Finally, we moved the lens to the polarizing position, which should have blocked all of our polarizing light if the filters were really effective.

Circular Polarizing Filters Test Circular Polarizing Filters Test

Here’s where I expected to put a table showing how efficient the polarizers were. Instead, I’ll just tell you that all of the polarizers we tested blocked all of the light, within our capabilities of measurement. They were all at least 99.9% efficient at doing their job, polarizing light. There’s something you don’t see very often; a photography product that completely does what it says it will do.

Also About the Glass

When we tested UV filters, we found several for which the glass wasn’t flat. We repeated the test on all of the CP filters, and all passed with flying colors, so I won’t bore you with repeating it. Again, my expectation was with two pieces of glass we might see more, not less, bad glass. But no, they all passed just fine. That may be that better glass is used in CP filters, or the polarizing effect evens things out. I don’t know. But they all passed just fine.

What About Light Transmission?

If you’ve ever used a CP filter you are probably aware that even when not polarizing it absorbs some light. You should be aware of this because, well, it’s darker when you look through it. We thought it would be worthwhile to see how much light it does absorb.

Why? Well, partly because we had that laser transmission bench already set up, but mostly because two of these filters claim to let more than 99% of light through. The Marumi claims it lets 99.4% of light through, and the B&W High transmission says it passes through 99.5%.

No, I’m not a rocket scientist, I’m just a regular scientist. But high-tech scientific principles tell me that since everything looks a bit dark when looking through these CP filters, it’s unlikely that more than 99% of the light is passing through the filter. But maybe that’s just me being cynical. Or maybe there are alternative facts that say darker isn’t the same as less light.

Anyway, since I was rather fired up and my BS meter was pegged at full maximum, we removed the linear polarizers and measured absolute transmission for each filter in the non-polarizing position. If you remember, when we tested clear filters the best let 99% of light through, the worst was down around 90%.

The transmission results for CP filters were:


Now, the Marumi and B&W are nowhere close to 99% transmission, but I will admit that they did indeed have higher transmissions than the others.

Some, probably most, people don’t care about how much ND effect their circular polarizer has, and if they do they may well not want the higher transmission variety, they’d prefer a bit more light blocking. After all, if you need a circular polarizer, you probably are shooting where there is lots of bright sunlight. But the takeaway message is that higher transmission filters do tend to give more transmission. Just not as much as is claimed.

Let There Be Spectrometry

And so, in the days after the first article, the people spoke as one and said, “You show us but one wavelength of light, yet there are as many wavelengths as there are fish in the sea. Give us spectrometry, that we may see the effect on all manner of wavelengths, each unto its own kind. And make the graphs brightly colored.” 

It’s taken several weeks for us to get things calibrated and running, but this post seems a good place to start using our new spectrometer. We know that some polarizers give a bit of color cast, especially when polarizing, so we thought it would be interesting to see look at their transmission spectra.

We looked at transmission both in the open and polarizing position and did not see any changes with these polarizers. I’m told there are some that do have a color change with polarizing. I’m only showing you one spectrometry report for each filter, to keep this short post short. Also, don’t put much stock in the absolute transmission between the filters. We weren’t testing for absolute transmission since we’d already done it; we just wanted to look at the curves.

The High-Transmission CPs

The Marumi and B&W filters have similar transmissions and very similar spectra. Both tend to have some UV filtering activity and drop off a little bit at the blue end of the spectrum.


Olaf Optical Testing, 2017


Olaf Optical Testing, 2017

The Standard CPs

These all have a stronger Neutral Density effect than the first two, and their spectra are different, too. The Sigma, Zeiss, and Heliopan filters are all very similar with a bit more transmission at the blue end of the spectrum and a bit less in the red-yellow range. The Tiffen has a similar pattern, although maybe a tiny bit more of a green peak.


Olaf Optical Testing, 2017


Zeiss T*

Olaf Optical Testing, 2017


Heliopan Digital

Olaf Optical Testing, 2017




Olaf Optical Testing, 2017


So What Did We Discover Today?

Well, several things, one of which is really useful. So I’ll get that one out of the way first, and then let this post just steadily deteriorate. If you are buying a circular polarizing filter because you want some circular polarizing, it doesn’t seem to matter much which one you choose; they all polarize like gangbusters. So I saved you some money today.

The second point, one which I’ve been told before I did all this testing, is set the white balance after you put the CP filter on, not before. Because CP filters will have a color cast. Or just shoot in raw and fix it later, which is what we mostly do anyway.

There is a third point, and it’s a painful one:

Once Again, Roger Lets Technology Triumph over Common Sense

I didn’t want to test filters; I really didn’t. But people wanted me to. So I chewed up my testing equipment budget to buy laser transmission stuff and an optical spectrometer, spent a few weeks getting everything calibrated and establishing norms, and then a couple of days testing these CP filters. I did this in clear violation of Roger’s Third Law: No Good Deed Goes Unpunished.

After I was done, I told Aaron I had just documented that CP filters had different light transmission percentages and different color casts. And that high transmission filters had one look, and it was different than regular CP filters, which all were really similar. Because I was proud that my investment in time and money had paid off.

Aaron took the filters from me, put them on a piece of paper, took this picture with his cell phone, and said, “Yeah, you’re right.”



Oh, and BTW – I know what you’re thinking. Yes, I’m going to publish spectrometry reports on the clear and UV filters we tested in the last article. Next week, I promise. I need a few days to recover my pride.

Roger Cicala and Aaron Closz


September, 2017

Addendum: MTF testing

Several people made the very pertinent comment that they would expect some effect in sharpness and contrast. FWIW I took a moderate telephoto, very sharp lens (Batis 135mm f/2.8) and MTF tested it first with no filter, then with a B&W Multi-coated clear filter, then 4 of the CP polarizing filters we tested above.

The clear filter made absolutely no MTF difference.

All tested polarizers (B&W, Marumi, Zeiss, Tiffen Ultra, and Helipan) caused a slight decrease in MTF at high frequencies. There was no detectable change at 10 and 20 lp/mm. At 30 lp/mm there was a consistent 1-2% drop, at 50 lp it was about 4%. All of these CP filters were very similar, I could detect no difference between them.

This is a quick, off-the-cuff check. I’ll look at things in more depth when I have time. But the bottom line is fine detail in photos is affected a bit. I don’t find that the least bit surprising.

I’ve also ordered a couple of $40 filters and we’ll see how those compare.

Author: Roger Cicala

I’m Roger and I am the founder of Lensrentals.com. Hailed as one of the optic nerds here, I enjoy shooting collimated light through 30X microscope objectives in my spare time. When I do take real pictures I like using something different: a Medium format, or Pentax K1, or a Sony RX1R.

Posted in Equipment
  • TomDibble

    “CPL requires two pieces” – only kinda sorta.

    CPL requires two *films*, but those to not necessarily need to be applied to two separate pieces of glass (most basic case would be each film applied to opposite sides of a single pane of glass). My guess is that the film is applied on the “inner” surface of the glass panes which are then sandwiched together, keeping the relatively expensive and fragile films from being damaged, but an actual engineer at Hoya etc might be able to give a more technical view of it.

    Also, while my cheapo CPL appears to have two panes of glass (it is fairly thick), both panes rotate at the same time (quick/easy test: put a finger smudge on the “back” side (the “quarter wave” plate) and elsewhere on the “front” side (the “polarizing” plate) then rotate the thread mount; the finger smudges never differ in their relative positions).

    I third the request for a similar test on el cheapo filters.

  • The back of a neoprene mouse pad works well too: press the filter ring down into the pad, and rotate the lens counter clockwise. It won’t help with stuck CPLs though.

  • HenWin

    My son is a journalist–with a LARGE online news organization–and it sickens me to hear when not just him, but everyone in the “newer” generations have suddenly started starting sentences with what I would call “illegal” words. (It’s ’cause I don’t know what those kinds of words are called, but I know bad American when I hear it.

    As far as “y’all”, as I noted to Joshua, below, the proper beginning to the phrase would be “All y’all”, the plural form of “y’all”!

    I’ll continue to fight my losing battle w/ trying to keep American from “progressing” by using words that belittle their original meaning! 🙂

    The newest overused word, by the way, IMHO, is “adorable”. I think it has taken the place of “awesome”, which as people should know, usually isn’t.

    OK, I’ve wasted enough time. 🙂

  • HenWin

    I’m not from the south per sé…I’m from DC and while it’s close, it ain’t there. However, my momma was from Chattanooga and I know that you are using “y’all” incorrectly. “Y’all” is singular. “ALL y’all” is plural! 🙂

  • LeFred

    On a white paper, an Hoya HD polarizing filter looks almost like a clear filter compared to a Marumi. Maybe worth a look?

  • Dave Hachey

    I agree with Brandon. For professional quality work you need to go with commercial products, partly for much better specs and performance, but also for maintenance. I use to run a mass spectrometry facility with 20+ instruments running upwards of $500K. One service call would damn near kill my budget, so I did most maintenance myself (actually a techician did the work). The DIY projects I mentioned above are fun teaching tools, but I’d never use t hem in the lab.

  • J.C. Overgaard

    Wesley, you’re confusing linear polarizers with circular. CPL requires two pieces. Google it. I think most modern cameras handle linear polarizing filters just fine, but they screwed up autofocus on older cameras and that’s why CPLs were developed.

    I agree that another test with some cheap models would be nice, and some linear polarizers, too.

  • That just made my morning. It’s way cool and I totally did not know that.

  • Wesley Hetrick

    Even better, take a crayon or marker and draw a line on the front and back of a CP filter, now spin it.

  • Wesley Hetrick

    Ever had a pair of polarized sunglasses? Sorry, grab the threaded portion of a CP filter while touching the back glass, now spin the front ring and you will see that the glass with your finger on it in the back is moving. See, one piece of glass.

  • Brian Barwick

    I have found that using a surgical glove has always worked.

  • Joshua Koerner

    As someone from the south who has also studies languages fairly extensively, indo-european ones at least, y’all is actually grammatically correct! Not that you probably needed me to say that to know it 😉

    Also, you just let us know what we need to tell you to buy for your next tests, and I’m sure someone will be happy to oblige 😉

  • Brandon Dube

    I do not think a lulzbot could produce an optomechanical quality print. Ignoring requirements like stiffness and rigidity, optical tolerances are ~1 micron on scale; most cheaper 3D printers come in closer to 50um in tolerance.

    The long term cost is a lot higher, but I would ultimately rather going the all aluminum and brass route from vendors like thorlabs, newport, opto-sigma, etc. With those you buy a guarantee and warranty, DIY you do not.

    For the spectral transmission bench things like drift are very important. For the laser based one, things are so simple the requirements are not stringent.

  • Brandon Dube

    A CP filter must have two pieces of material, it is impossible to produce one that varies the angle it selects otherwise.

  • Brandon Dube

    I do not get that impression — the bulk of the transmission curve comes from the waveplate and polarizer, not the glass the polarizer is on or the coating. There are two clear types polarizing coatings/mechanisms used, B&W-style, or Zeiss-style. The B&W style matches the thin film linear polarizers sold by thorlabs or other vendors in characteristics, the Zeiss-style ones, I have no guess what they are.

  • Brandon Dube

    The spectral transmission curves are for randomly (or unpolarized) light. The single-percent transmission numbers are for “perfectly” (~1M:1) polarized light. There are two classes — murami-style, and zeiss-style, with around 40-45% transmission and ~35% transmission, respectively, for unpolarized light.

  • Hendrik Müller

    This is one fantastic article, i so like your amazingly cynical sense of humor, thanks for writing it….

    Only one serious question to be sure: Transmission was measured with perfectly polarized light, other than Aron with the smartphone? im i right, a polarizer with 100% transmission (measured by you) would have 50% transmission with unpolarized light?

  • Christopher Morgan

    I don’t see where you accounted for the quantum effects of light and polarized lenses. https://www.youtube.com/watch?v=zcqZHYo7ONs

    Now go forth and buy more testing equipment.

  • Claudia Muster

    I suspect that B&W and Marumi designed their blue-attenuating curves by purpose. It enhances the contrast in landscape photography.

  • Thanks for the info Roger! I assume the “Also About the Glass” refers to sharpness across the image frame? I ask because I spent a bit of money on a Heliopan CPL a while back, but sold it after I saw how distorted the edges got compared to the B+W that I replaced it with.

  • Wesley Hetrick

    I wish you would have tested say a $30 Hoya filter. Let’s see what you get with going really cheap. Also, every CP filter I have is just 1 piece of glass. Was that a mistake or do these expensive ones have an extra piece of glass?

  • Dave Hachey

    Oh yes, I know. I’ve been using spectrometers and spectrophotometers of various types for more than 50 years. Dig around on YouTube and you’ll find some interesting stuff. Next up a DIY Raman spectrometer. I have no idea what I’d use it for though. Cheers guys…

  • Steven Dean

    Thanks for the response Brandon. I’ve got a $1250 Lulzbot Mini that prints totally adequately to make optomechanics. You’d have to buy some hardware to go with the printed plastic (screws, springs, etc.) but you could make that $200 mount for about $5 once you had a printer and some filament. Even a $400 printer should be able to do it. I’m sure ya’ll end up needing some kind of custom bracket from time to time; and that’s where a 3D printer really excels. There are also a ton of third party places that will print for you for a nominal fee if you’d rather not deal with having a printer around the office. The big advantage of the $200 Thorlabs mount is that it requires no time on your end to set up and will not drift very much over time; but I don’t think either of those are very important considerations for a young, financially bootstrapped research company.

    Expanding and re-focusing the beam should be about $100 worth of optics using a Galilean beam expander. Cheaper if you don’t go with AR coated optics.

  • Brian F Leighty

    Sure understand that from a scientific standpoint you can’t show that. Just pointing out that there are reasons to get more exspensive filters

  • Brandon Dube

    An IR spectrometer is only like another $5k and I think we have spare ports on the integrating sphere 🙂

  • Brandon Dube

    The laser-based bench is obviously always polarized (1M:1 or more with the sheet polarizers, ??? without). The spectral results are with unpolarized light from a SLS201L source.

    If TL wants to give us a discount, that’d be lovely 🙂

  • Brandon Dube

    It’s a spectrophotometer-based transmission b ench for both imaging and nonimaging optics. It’s a lot more than a dispersive element and a light source 🙂

  • Dave Hachey

    Now I get the need to buy new toys for the lab, but geez Roger, $15K for a spectrometer! You over paid by about $14,900. There’s a whole subculture on Youtube that builds these things out of spare parts. I built something similar about 30 years ago as a science project with one of my kids (I should probably do the same with my grandson). Good article though…


  • Dave

    Nice analysis! One thing I’d point out though is you should be careful with how you discuss polarizer transmission. If you measure the transmission of a polarized laser beam which is aligned to the polarizer, then you’re measuring polarization along the transmission axis of the polarzer (which is what I think the filter manufacturers specify). However if you just look through the filter with your eyes, it will always look dark because you’re seeing the total transmission of unpolarized light, which can only be as high as 50% for an ideal polarizer. So even though some manufacturers specify the transmission above 90% their filters will still darken the image at least 1 stop.

    I noticed in your spectrometer graphs the absolute transmission is much lower, I’m guessing that’s why the absolute transmission at each wavelength is much lower (all under 50%), did you use an unpolarized light source here?

    Still though the spectrometer graphs are fascinating! Those differences in the curves will absolutely result in a color cast. I’d love to your spectrometer setup…do you have a true spectrophotometer or a spectrometer?

  • Brandon Dube


    Transmission is determined by absorption in the materials that make up the filter (which for a polarizer is very nonnegligible) and reflections. Unfortunately, in this case the absorption related part dominates.

    We would need a scatterometer to measure the reflections off of the rear surface of the filter. That’s an $80,000 instrument, which we cannot afford.

Follow on Feedly