Good Times with Bad Filters

Posted by

OK. First and foremost this is a fun post. It is not episode 362 of "Should you put a UV filter on your lens". Some people use them. Some don't. There's not enough bandwidth to ever end that argument.

But here at Lensrentals, we have a ton of filters. We have some really good, very expensive filters. We have some OK, middle of the road filters. And because some customers, uhm, happen to return a very cheap filter in place of the one they were sent, we've obtained some crappy filters. Brand names aren't necessary. If it cost $22 in 77mm size, it's a crappy filter.

Anyway, one of the techs has to clean all those filters, make sure the threads are OK, and test them out. Honestly nobody likes to do it, so it gets put off until we need some filters or there's just nothing else to do. So the other day Kenny is cleaning filters and testing the threads by mounting them one in front of the other until he made a nice mountain of 50 UV filters.

50 UV filters, cleaned and neatly stacked.

Not being the kind of people to let well enough alone, we decided to mount them to a 5D Mk II and 300 f4 we had handy and take a few pictures.

The well protected lens.

And of course see if the filters affected image quality. See if you can tell which images was shot with the 50 UV filters, and which without:

Shot of the building across the parking lot without filters (above) and with 50 UV filters (below). The one with the filters is actually better than I expected.

Of course there's a lot of vignetting and haloing on the full size image:

Compared to no filters

Roger, do you have anything constructive to say, or are you just wasting blog space again?

Yes, actually I do. Fifty filters stacked is pretty ridiculous. But in that stack of 50 filters, as I said, there are some very good ones and some very bad ones. Lets compare a stack of each, shall we?

First, I had Kenny put the worst filters on the top of the stack (all were nonbrand, or brands we know are cheap and bad) and take a picture of the stack at an angle. All were freshly cleaned and if you look straight through them reasonably clear. Like a filter should be. But if you stack them and try to take an angled picture through several layers of them, the results were ugly.

View through a half dozen cheap filters stacked on top of each other. Try counting the filter rings inside the stack.

Yes, I know they don't look clean in the image, but every one of those filters was freshly cleaned, and checked under a light. And if you look straight through them they were pretty clear. Looking at an angle tends to show you the weaknesses of a filter much better than looking straight through it. And remember: most of the light rays coming into the lens are coming in at an angle, not heading directly to the sensor in a straight line.

Now lets compare the stack with the expensive, top of the line filters (B&W, Heliopan, etc.) stacked the same way.

Stack of expensive UV filters one atop the other.

Hmmm. I'm starting to think there might be a difference here. But the proof is in the pudding. Lets modify our original experiment to something only slightly ridiculous. Instead of shooting through 50 filters, lets take the shot through 5 top of the line filters and another through 5 bottom of the line filters.

Here's a 100% crop of a bumper sticker across the parking lot shot first with no filter, second with 5 stacked high end UV filters, and then with 5 stacked low grade UV filters.

100% crops of a bumper sticker shot through no filters, 5 stacked good UV filters, and 5 stacked cheap UV filters.

Now stacking 5 filters doesn't have a ton of real world implications. Most people rarely stack two. But it is a fun demonstration that there really is a difference between good filters and cheap filters.

The good filters do a remarkable job: 5 stacked filters means 10 air-glass interfaces before the light even gets to the lens. That there's only a little bit of image quality loss through all those filters is pretty impressive. This crop is from the center of the image, there's more degradation to the sides, but still, it's an impressive performance. And certainly lends credit to the idea that a high quality, multicoated UV filter has little effect on image quality.

Five bad filters, though, is another thing entirely. I'm completely aware, for those of you who are going to feel the need to point out the obvious, that nobody shoots with 5 UV filters. And I understand that one cheap UV filter wouldn't have nearly as bad an effect on image quality as 5 of them. But I don't think you can disagree that the good (and expensive, I know) filters have much less effect on image quality than the cheap filters.

BTW - before anyone asks, I avoided name brands of cheaper filters for a reason: many filter manufacturers make both pretty good, and pretty bad filters. You can tell the difference by the price or by reading carefully about the number of coatings, etc. A Tihoya $29 "high quality" filter is not the same as a Tihoya $79 "Professional" filter. This wasn't meant to be a filter review, just a fun demonstration of the obvious.

Roger Cicala


June, 2011

214 Responses to “Good Times with Bad Filters”

Dave Steinberg said:

Great post. Please don't ever leave well enough alone!

Greg said:

If I were you, I'd publish this stacked filter technique fast to make sure you get credit for it. Someone is going to love this "fresh new look".

LensRentals Employee

Roger Cicala said:

Yep, we'll call something like "The Dream Machine" and advertise the "soft romantic look" it gives. Plus think of the variety: don't want a really dreamy look? Just take off 10 or 20 filters.

Joe Sankey said:

Hilarious. I can imagine the staff giggling like kids as this was getting mounted on the camera.

I think more equipment reviews should err on the side of ridiculous - it's a lot more entertaining to read. (Let me qualify that... I think they should *intentionally* err towards the ridiculous. Way too many do that on accident as it is. Much less entertaining.)

Heidi said:

It's not the glass but it's the air in between the filters that's contributing more to image degradation. Humid Tennessee air this time of the year, trapped in between filters, exacerbates the deterioration by refracting the light rays. Next time, try creating a vacuum in between each stacked filter and you shall see a marked improvement in all stacked filter examples. :-D :-D :-D

Pete Templin said:

Wait, you didn't extend the hood on the 300/4! That should block some of the spurious light rays coming in at an angle, right? ;)

Anthony said:

My only problem with this post is that I don't have access to that many filters to play with the "soft romantic look" myself. :P
Thanks for the laugh at seeing that stack mounted!

Heidi said:

So... when can we expect to see stack of 50 filters available for rental?

Carl said:

Cute post Roger. The vignetting in the first shot is more due to the stack of 50 acting as a long hood, don't you think? I'm sure Crye-Leik realtors didn't mind you posting their name, haha. As for creating a vaccum seal between filters, that would be a bit more trouble, and also would increase the spacing between them...but I'm sure Roger will do it someday! It could be an episode of "mythbusters" or something.

I wonder if anyone has compared the material for the actual glass used in filters? I know fluorite, or otherwise synthesized glass to mimic it, is supposedly some of the highest quality glass used to make lens elements. Would it be an advantage for making a thin, flat filter element? You would think they should use something exotic, if the price is to be in the $300 range for some filters...Seems like the companies have a very high profit margin on filters...sort of like cabling in high end hi-fi.

It would be more interesting to compare a few of the high end filters, and maybe also compare the difference between a "UV Haze" filter, and a standard UV filter. I own a Kenko UV Haze filter (now discontinued) that performs unbelievably well, even though as a brand, they are sometimes spat on by those who like to spend a lot of money (Singh-Ray comes to mind).

I also own a Heliopan Circular Polarizer, so in that instance I decided to spend more. But I wonder how it would compare to other circ pols, especially the Hoya and the new Kenko line, that claim they have a finer polarizing film, which allows more light in (I see that as an advantage where circ pols are concerned. Shutter speeds get slowed horribly by mine.) I also think it's a shame that LINEAR polarizers interfere with the metering of digital cameras, because...if you have ever worn a decent pair of polarized sunglasses, the linear effect has more profound influence on objects with highlights on them...as you rotate the sunglasses, or your head...than a circular polarizer does...at least in my opinion. Anyway, I DO UNDERSTAND the need to not devolve into a brand debate here...even though I wound up mentioning brands anyway!

Andy said:

You guys should seriously make "Stack of 50 Varying-Quality Filters (for that "soft, romantic look")" as a rental item to see if anyone bites.

At any rate, this blog post is reason 362 of why I love Lensrentals.

John said:

Kenny deserves a raise. He apparently is a creative person who has great potential. He is able to do more than just put paperclips into a chain. Test him and give him something really boring to do to see his potential. You might get another bolg topic.
Besides that, it is an imformative blog that'll make me think more about 'to filter or not to filter'.

Zeno said:

For a fun demonstration of the incredibly non-obvious, do the same trick with a stack of linear polarizers, each rotated just a little relative to the next. If the first and last filters are rotated by 90 degrees from each other, it's the same as a pair of crossed polarizers and you shouldn't get any light through, right?

But you will get light through, and you'll need quantum mechanics to explain why.

Try it!

alek said:

Ditto first comment from Dave - yet another great post ... funny, interesting, informative, and funny - makes a good point in an easy to see manner ... and methodology is solid.

Nice job Kenny & Roger!

Lee Derrickson said:

A picture is better than a thousand words. This article says more about cheap filters than any other way you could put it.

Groovy said:

Could you please repeat this with an UWA, I think this technique is great but I'm a tiny bit concerned about vignetting! Keep it up!!!

JSturr said:

That's why I only shoot with an Iphone - filters... heh..

John H. Maw said:

I'm sure you would get the same result order no matter how the test was carried out, but the extent of the image degradation depends very much on the lighting conditions. It would be interesting to see a night scene including street lights or similar within the frame. I have found that this is a situation where even one good filter can make quite a difference.

LensRentals Employee

Roger Cicala said:

Actually I think that article, from a source I respect greatly, only tests the amount of UV light transmission. It does test for flare, which is important and overall light transmission. But it makes no test regarding degredation of image quality - simply evaluates the amount of UV blocked. UV blocking was quite important back in film days, but Is not the reason those shooting digital use UV filters today.

I'm not a tester but if I was looking at a test of filter quality, I'd be a lot more interested in effect on image quality than the amount of UV light blocked. Obviously there are people for whom UV blocking ability really is important. Anyway, I wasn't trying to make the point that you have to buy the most expensive filter in the world, simply that bad filters are bad filters.

JD said:

Zoltan, unfortunately, that page doesn't deal with any actual sample prices, so it's hard to tell what's going on in terms of actual cost.

But that does bring up a problem I have with all this, is that picking out a filter is tedious because every brand has its designations on multicoating, and it's tough to know how good they are anyway.

Carl said:

John Maw. I have shot night scenes with street lights, with my Heliopan Circ Pol filter on. Even though it has 8 layers of coating on each side, there are still faint ghost image reflections of the lights. In my opinion, when shooting night scenes which include bright lights...or even when shooting something as bright as the moon at night, it is best to not use any filter at all. My experience was the same when using just the UV Haze filter. However, if the scene does NOT include very bright lights in the shot, then I don't see why a multicoated filter wouldn't work ok. I have even stacked both filters for night shots of the sky, for "star trails", without any noticeable "ghost image" reflection of the stars. I have also tried Milky Way shots with the Circ Pol filter, since the sky in my area in summer, is never crystal clear, and there is always some light pollution from a nearby town. It works ok, but it holds back so much light, that I wind up having to use an ISO that is too high (with a wide angle zoom at f/2.8). So I have found the best way, is just to either wait for a crystal clear, very dark sky, or else travel to where there is one...and shoot without a filter. But again, for shooting city lights at night, it is also best to shoot without a filter.

Steven said:

Not sure if anyone is missing the very obvious, that the UV is only there to save the $1,500 lens, Reminds me of Murphy's law, that the $1 million dollar computer will blow to protect the 10 cent fuse.

Lucy said:

The fuzzy bad filter shots look kinda like Instagram. and there's PLENTY of people downloading that stoopid app.

Shocking said:

Not Trolling but you use the expression "the proof is in the pudding" which should really be "the proof of the pudding is in the eating" You won't find any proof ( test ) of the quality of your pudding until you eat it. Or stick it on your lens.

Jon said:

Fun test. If you want to tackle another hot topic, perhaps you can tell us your method of cleaning filters and lenses.

Jon said:

D'oh! I just noticed the blog post on lens cleaning. A prior entry in your search engine didn't produce any results.

Walter Underwood said:

Use a lens hood instead. If you've ever dropped a lens with a filter on it, you already know that the shards of broken filter don't "protect" the lens. And fingerprints don't hurt a lens, you just clean them off.

A lens hood can only improve a lens, a UV filter can only hurt it.

Scott said:

Fun and ridiculous yes, but still interesting. Intrapolate five filters down to one and it's easy to see that a single crappy filter must be having some negative effect on your images.

Erik said:

At a certain large NY camera store, the basic HOYA UV in 77mm size cost around $50 while the no-name is $23. So the generic advice "If it cost $22 in 77mm size, it’s a crappy filter” is not contradicted by the lenstip article.

louis said:

Hi if i was you i would give this Guinness so you can claim your medal from Guinness book of records just for the hell of it before someone steals your limelight. Its all fun lets not get serious.

Jimmy said:

Hey, great blog! The idea of putting 50 filters on a lens might seem ridiculous, but then - hey - is there a better way to show that one has to invest in best filters possible if having them on the lens is what they want, and only best results count? Many thanks for the "test" :)

Jes said:

Ditto first comment from Dave – yet another great post … funny, interesting, informative, and funny – makes a good point in an easy to see manner … and methodology is solid.

Rob said:

I've used UV filters as lens protectors. Is there an alternative? I've never heard anything bad about UV for photos (wasn't it a film issue?). Can we get clear lens protectors? Minimal filtration?

This test shows it's not an issue using cheap filters because the degradation isn't visible to the naked eye.

Herman said:

I always dreamt of owning a looooong lens...
Now I know how to achieve it!

Brainiac said:

This test is great, but it does not give answer all questions.

Please, repeat it and check the effect on image quality if you stack 50 lens caps.

Nik Caduzo said:

I understand what was the point, but I do not believe that such exaggeration can achieve the real goal. I would like to see the same image but with only one (preferably good quality) UV filter.

Bob B. said:

OK...OK...this was a REALLY fun read...plus...I decided to replace all of the $10 "T" brand filters on all of my lens last month with B&W MRC Brass Pro filters. Now I have two formats and 9 lenses to cover and that is a LOT of mulla for... in effect..ummm nothing! LOL. I ordered from Hong Kong for substantial savings...but it was still painful! I noticed an extreme quality difference all the way around when the filters arrived..the brass rings are solid and non-binding to the lens and it was more than obvious to me that the glass and the MRC coating were head above the "T" brand. Lens rental...THANKS FOR VALIDATING MY PHOTOGRAPHIC INSTINCTS!!!!!

@ID7 said:

I'm off out to buy 200 filters. Great article, thanks for erm.... doing it.... My project for the weekend!

Zoltan said:

Anyone paying $80-$100+ for a filter is only kidding themselves and making filter makers very rich.
I personally have used the Hoya HMC and Marumi MC filters and did not notice any IQ degrading with them at all. I could recommend them as good filters that are fairly priced.

Yes avoid the ultra cheapie uncoated ones by all means, yes the Hoya green range is poor and a few others. But don't stand here and tell me b+w filters are worth the money because I've tried them they are not.

SN said:


Mr S. Tyru said:

Guys - I've spotted the problem, it's not the 'add ons' to the right of that lens that are the problem, its the 'add-on' to the left of the lens

(cue Nikon v Canon debate)

Daf said:

Ha ha - fun test.

I once read somewhere and now stick to it - Why spend hundreds if not thousands on an expensive Pro lens, and then stick a piece of crap filter in front of it....
I now get the best I can within budget (but will shop around and try eBay)

Lee Duguid said:

Great post, thankfully I've never been a fan of UV filters....now I have a reason to hate them :)

Scott said:

Just use a lens cap for "protection"...Or a lens hood....A lot cheaper than expensive filters which are not worth the ectra expense. As for protecting the lens, any drop can damage more than just the front element, get yourself a good insurance policy on your gear...

JeffT said:

I think you're making the inverse of the mistake that engineers made with the O-rings on the space shuttle - ignoring the effect of stacked tolerances.

For example, if a "good" filter lets through 99.5% of the light with perfect fidelity and causes 1% distortion, stacking 5 of them will result in 0.995^5 = 97.5% of light getting through with about 2.5% distortion.

Now if you look at a "cheap" filter that is only very slightly less "good", let's say 97.5% (2% more distortion than the "good" filter), stack 5 of them and the net effect is 0.975^5 = 88.1%... a very significant difference with 5, but a very insignificant difference with just one.

In reality, the effects are more subjective, but the math is the same. Stacking filters magnifies the imperfections exponentially, not linearly, and so looking at the effects of multiple filters can be very misleading when you're trying to evaluate a single filter versus another single filter for price vs. performance.

Marc said:

No filters seems to have the best result. Is it possible to rent no filters?

Marty4650 said:

Lensbaby should market this concept. A set of 50 cheap filters that are stacked for "creative photography"....

Someone would buy it.

Ashley Pomeroy said:

This is great fun - I have a mental image of Count Von Count from Sesame Street putting all the filters together. "Von filter! Ha, ha ha! Two filters! Ha ha ha!", and so on for an hour and a half. The shot of the B+W 77mm filters is iconic, you should put that up as a big poster in your shop.

"Now I have two formats and 9 lenses to cover" - I tend to buy a few large filters and a bunch of stepping rings, which is doubly handy because, with a full-frame camera, some of my lenses vignette a tiny bit if I use a matched filter. E.g. my Tamron 28-75mm vignettes with 67mm filters at 28mm, but not with a 77mm filter mounted with a stepping ring. Stepping rings are cheaper than filters.

Carl said:

The replies to this blog seem to have gotten out of hand last night. Apparently a bunch of people from Europe (or nightowls in the USA) think this is twitter or something...

Leave a Reply