They say that it's best to put on a filter on your lens to protect it. It's cheaper and more effective than having camera insurance. There have been hundreds of stories on the internet where multi-thousand dollar lenses were saved from crash/drop by putting cheap UV filters (and/or hoods). However, filters come with a price-- decreased image quality. Adding an additional glass means the lens will get flare more easily. This is especially true when shooting directly into a light source where the light source becomes a flare at the opposite side of the frame. Below is an example of a green flare, thanks to a cheap UV filter, when shooting into the moon:
Adding a filter also means decreased of light going to the camera. A cheap lens today in general will allow 90% of light source going through, and a really high end one will allow 99% going through. However, this is actually quite negligible. 10% is miniscule compared to how much aperture I can open up or how high ISO I can crank up in today's modern DSLRs. Each stop means 50% or 200% difference, so the decreased of 1% to 10% of light is a non-issue, at least for me.
Going back to the issue of image quality -- this is a real concern. Below is a wonderful HDR picture that Ping shot in Alaska with a B+W UV MRC filter. Notice the green flare to the left-bottom corner of the pictures. Personally I think it is fine as it adds some artistry to it... in fact often times people add in fake flare in Photoshop to spice up their pictures.
On the other hand, real flare can come at inopportune moments when you least want them to exist. In my past I never really cared about flare until one day, I came back from from a wedding shoot, and was horrified at a bunch of green flares on people's faces. They're shot using my amazing Sigma 50mm f/1.4 but sadly mis-paired with a really really cheap 77mm Hoya 81A filter! "Why on earth would you use a 81A filter at night" you may ask? Because I didn't have any other 77mm filter at the time, and shooting without a filter while I run around doing event photography is utterly out of the question for me.
The first example below is acceptable because the flare is not interfering with the subject. It just so happened that none of the green flare got on Jennifer, yay!
However, subsequent examples are where I really didn't want any flare to occur, but I did not realize that it was happening at the time since I was too busy shooting and I didn't have time to chimp the LCD. Moral of the story: chimping IS GOOD FOR YOU! Sometimes.
 |  |
All of this of course could have been mitigated by 1) using a flash and blow away all warm natural lighting [thus rendering the color out of whack and will need to turn night pictures into black and white] 2) using a higher quality filter 3) not using a filter 4) not shooting into the light source (candles). On the other hand, I really wanted to capture the candlelight to emphasize that all these shots were made possible with nothing but natural candle-light. Most photographers will just opt for 1) because flash is a lot easier. Capturing natural candlelit subjects usually requires 1600-3200 ISO, 1/30 sec, f/1.4-2.0 (below EV2)... or pretty much at the limits of what steady photographers can hand-held AND pushing the technological limits of today's DSLRs sensors.
In short, not all filters are equal. I have compiled a list of the filters that I've been using below and some examples shots. The first row shows examples of NO FILTER. Subsequent rows show the 77mm Hoya HMC 81A, 77mm B+W 010 UV Haze MRC x1, 77mm Hoya Super HMC UV(0), 72mm Hoya HMC UV(N) and just for kicks a 77mm Hoya NDX8. The second column shows the reflection of the filter and a full-spectrum CFL light source (rated at 94 CRI) where you can look at the colors of the filters. The last 2 columns show an example of shooting into the light source, and the flare on the opposite site of the frame. Are of them were shot using a Nikkor 85mm f/1.4D (portrait lens) on a tripod:
In short, I had a very high expectation of the German made 77mm B+W 010 UV Haze MRC x1 and I was very disappointed with it. People on the internet seem to say that it is great... that it is what German engineering is about. However, my tests show that it simply blows. In fact it is *MORE* expensive than the Japanese made Hoya Super HMC, and performs as badly as the cheap Hoya [regular] HMC. Personally, I've never been impressed with the value/cost ratio of BMW cars vs. Lexus, and this test just made me feel the same about B+W vs. Hoya.
One last note. I didn't cover color range and IR/UV accuracy. I've seen tests that show that the Hoya Super HMC blocks way more IR and UV that it is suppose to thus creating color shifts, and that the B+W MRC has a much more accurate IR/UV block off points. However, without instrumentations I can't validate those tests. If color is a more important issue than flare, perhaps the B+W MRC is a better choice. Trade-offs.
P.S. For more formal results you can visit an external link here:
