Kolari Vision Infrared Camera Anti-Reflection Coating Review

When I got my Nikon D7000 converted into infrared-only, I decided to use the Kolari Vision company. The main reason I chose this company was because of the specifications they gave in regards to their IR filter covering the camera sensor.

Kolari Vision offers (at additional charge) to put an anti-reflection (AR) coating over the surfaces of their glass IR filter that they install in place of the normal (visible-light) camera sensor filter. I was a bit skeptical about this AR coating (it’s adjusted to infrared wavelengths), but I decided to add this option when they converted my camera. They also add materials that are supposed to make this sensor filter easier to clean, if debris gets on it (which it always does).

My camera conversion was with their 590nm sensor filter, but they offer the anti-reflection coating option for all of their conversion wavelength options.

I wanted to mention that I don’t get any money from Kolari Vision, so I have no stake in anybody buying something from them or not. I just have a deep interest in infrared photography.

Kolari Vision claims that un-coated IR sensor glass filters will reflect 7% of the light, while their coated filters reduce this down to 0.4%. They further claim that this reflected light bounces off of the rear of the lens aperture blades, and is a primary contributor to the dreaded hotspot in the middle of your photos (with many lenses). Hotspots, if the lens produces them, always get worse as the lens aperture gets stopped down, and this is the reason why. You can read for yourself about their anti-reflection coating here:

The Life Pixel company, which also does IR camera conversions, doesn’t put anti-reflection coatings on their IR filters over the camera sensor. They claim that they tested this technology, and found that it made light transmission worse and didn’t help reduce hotspots in any way. All modern lenses have multi-coating, and they vastly increase light transmission. It doesn’t make any sense to me when Life Pixel claims that the anti-reflection coating reduces light transmission. They doth protest too much, methinks. Sorry, Shakespeare.

So who’s telling the truth here? I figured that some testing was in order. I can’t prove that the reflected light bounces off of the aperture blades, but I can at least look at the end result of using a coated sensor filter.

For many years, I have done infrared photography by using IR filters on my lenses. I always figured that it got me exactly the same results as an infrared-converted camera, except that the light levels were reduced when using the filters (compared to an IR-converted camera).

I have always known that my Nikkor 50mm f/1.8 AF-D lens got a bad hotspot after stopping it down while doing infrared photography. Because of this, I always avoided using it when shooting infrared. The shot below shows exactly what I’m talking about.

50mm f/1.8 at f/7.1, 30s, ISO 800, Neewer 850nm IR filter

You can clearly see the hotspot in the middle of the shot above, even though the lens was only stopped down to f/7.1. The hotspot was clearly visible, no matter which IR filter I tried on this lens. The hotspot got breathtakingly bad at f/16. This shot uses my own attempt at a white balance preset, and the color wasn’t modified by any post-processing editor.

I could use Lightroom and its “radial filter” to mask the hotspot, but this is definitely a second-rate, band-aid kind of fix.

I decided to re-create this shot with the same lens at the same aperture, the same filter, and the same lighting conditions. While I was at it, I also decided to try a variety of other infrared filters. The main change here is that I’m now doing these tests on my Kolari Vision converted IR camera.

The shot above was done with my Nikon D7100, which has not been converted to infrared. I have seen this same hotspot effect when trying my other cameras with infrared filters, including on my D7000 before I got it converted into infrared-only. My shots would look okay wide-open, but would get ruined after stopping down beyond f/5.6 or so.

In the test shots that follow, I am using the factory-set white balance that Kolari Vision provides, without any further modification by me in an editor. In regular photography, I would do lots of post-processing to alter the colors, including improving the white balance.

50mm f/1.8 at f/7.1, 1/800s, ISO 800, Neewer 850nm IR filter

The shot above was with my Kolari Vision 590nm conversion, but also using the same Neewer 850nm IR filter on the lens that resulted in the nasty hotspot with my D7100. Notice a few differences from the previous photo using the same Neewer IR filter. The exposure went from 30 seconds to 1/800 second (over 14 stops)! The big thing, however, is the total absence of a hotspot. Kolari Vision’s anti-reflection coating appears to have made a huge difference. Also note the loss of color, due to the double-filtering of the light through two visible-light-cutoff filters (590nm and 850nm).

I need to mention that I use Live View with an LCD magnifying viewer to see and focus when I put an IR filter onto a lens. It’s only a minor inconvenience, as long as I’m not trying to follow action. This combination even works in bright sunlight. Hand-held shooting is still possible with this combination of an IR-converted camera and an IR lens filter, although I typically lose about 3 stops of light.

I decided to try some other filter tests at the same lens aperture, to see if the hotspot might show up with filtration changes.

Kolari Vision 590nm sensor filter conversion only. 50mm f/7.1. No hotspot!

Zomei 850nm 50mm f/7.1 with Kolari-converted camera.

The Zomei 850nm filter also causes total loss of color information, and is extremely similar to the Neewer 850nm filter.

Hoya R72 50mm f/7.1 with Kolari-converted camera.

As you can see, the various IR filters on my 50mm lens didn’t make the hotspot appear at f/7.1.

Next, I decided to see if I could coax a hotspot to appear with this lens by stopping it down to f/16. I had no filter on the lens, and only adjusted the exposure by stopping down from the previous f/7.1.

Kolari Vision 590nm only. 50mm f/16. Tiny hint of a hotspot.

Most people wouldn’t even notice it, but at f/16 my 50mm lens shows just the barest trace of a hotspot in the middle of the shot above. This lens is now totally usable for infrared, at any aperture.

Despite what Life Pixel claims, I consider that the hotspot issue is indeed repaired (at least with my lenses) at lens apertures down to about f/11. This may not be a total cure, but I think it’s a huge improvement. When stopping down further, the hotspots are weak enough to largely be ignored. I rarely use f/16 or narrower apertures in my photography anyway, since it ruins resolution due to diffraction.

I suspect that I’d be seeing hotspots if I had the “standard” Kolari Vision 590nm conversion done, which doesn’t include the anti-reflection coating on the sensor filter. I guess I won’t ever find out, since I don’t intend to ever get a camera converted that doesn’t include the AR coating. It’s a cheap investment to make sure your lenses perform as well as possible with infrared photography.

Kolari Vision 590nm only. Adjusted white balance.

I fixed up the white balance in the shot above, using the tree trunk as neutral gray. Notice that there is a bit of a glow from the bushes on the right. This is known as the “Wood Effect”, which comes from infrared heavily reflecting off of the chlorophyll in the leaves. This isn’t a central hotspot, and is generally considered part of the charm of infrared photography. You’ll have to make up your own mind if you like it or hate it.

Most camera lenses demonstrate some of level of “glare” when the shot includes a subject with this heavy infrared glow. The shot above is typical. My Sigma 14-24 f/2.8 Art lens, although awesome with visible light, shows an above-average level of this glare in infrared. With most subjects the glare is ignorable, but sometimes the shots get ruined because of it.

Kolari Vision 590nm only. Red/Blue channel swap.

The shot above shows the more conventional red/blue channel swap from the white-balanced version, using my photo editor. The leaves changed from blue to yellow/orange, and the sky looks a bit more normal. The 590nm IR conversion retains enough of the visible-light spectrum to enable nice colors. It’s easy enough to also convert the shot into black and white, which I end up doing at least half of the time.

I was going to make up a database of my lenses to indicate which ones would work for infrared. After the Kolari Vision 590nm conversion, all of my lenses work with infrared! There are of course some qualifiers in saying “all” of the lenses work; lenses don’t work equally well under all lighting conditions.

My Sigma 70-200 f/2.8 Sport, which I initially determined to be totally unsuitable for infrared (using IR filters on un-converted camera bodies), for instance, now works just fine with most subjects.

Another bonus from getting a camera converted to infrared is that my lenses that don’t permit a filter on them can finally be used to shoot infrared.

Rokinon 8mm fisheye, 590nm infrared, red/blue channel swap

The Rokinon 8mm fisheye doesn’t allow a filter to be attached. It works great in infrared, at least with the AR coating option that Kolari Vision provides. The shot above was at f/11. The reason I used f/11 is because this is one of two lenses I have that cannot focus to infinity in infrared. Stopping down to f/11 gets the lens reasonably sharp at infinity. It doesn’t look quite as sharp as with visible light, but still looks pretty good.

Tokina 11-16mm at 11mm, f/8. 590nm infrared.

I used Silver Efex Pro 2 on the shot above, to convert it into black and white. The 590nm conversion can still provide a nice white-foliage look that people mostly associate with only long-wavelength infrared like 850nm. The distant atmospheric haze was totally eliminated.

This Tokina lens isn’t regarded as good for IR, but it looks fine to me. The main problem with this lens, as with several others, is shooting into very bright IR light (or just outside the field of view). The lens elements show lots of reflections and veiling glare when pointed at bright infrared sources. In shots like the one above, you’d never suspect it has any problems shooting infrared.

590nm infrared with Nikkor 24-70 f/2.8 at 24mm, f/9.

My Nikkor 24-70 f/2.8 AF-S VR also has a poor reputation for IR, even with the Kolari Vision website’s lens database (“bad after f5.6”). It, too, is mostly working fine for me, through about f/11. I have the same caution about shooting into lights, however: don’t do it. Some subjects (especially beyond f/8) will cause an overall central glow, although it’s not quite what you’d call a well-defined hotspot. I used Silver Efex Pro 2 on this shot, too.

The Luminescentphoto web site specifically states that their 720nm infrared-converted Nikon Z6 (didn’t say which company converted it) with this Nikkor 24-70 f/2.8 VR: “Hotspots at all apertures”. They rate it “Poor”. The shot above doesn’t have a smidge of a hotspot, and it’s stopped down to f/9.

Sigma 70-200 f/2.8 with 1.4X teleconverter, 280mm, 590nm IR

The Sigma 70-200 f/2.8 Sport lens isn’t supposed to work with infrared. This lizard obviously didn’t know that fact. Shot at f/4 1/3200s ISO 200. The shot above even throws in the 1.4X Sigma teleconverter. It’s curious that this lizard looks almost exactly the same in both visible and infrared light; most subjects look quite different in infrared.

My Sigma 150-600 Contemporary lens works surprisingly well. I would have thought that all of the glass in that lens would have made it terrible at infrared. Maybe the Kolari Vision AR coatings are working their miracles there, too.

590nm, Sigma 14-24 f/2.8 ART, 1/640s f/6.3 14mm ISO 100

I couldn’t find any positive reviews regarding my Sigma 14-24 f/2.8 ART in regards to infrared. If you keep it pointed away from lights, it can produce fine IR images.

It’s ironic that my cheap Nikkor 18-55 DX VR f/3.5-5.6 G II lens is excellent with infrared. It has better contrast and glare resistance than most of my other lenses do with infrared. It’s a slow lens, but I virtually never need a fast lens for infrared work (I don’t shoot it at night or even at dusk). It’s a pity that it doesn’t go wider than 18mm, though.

Summary

I have to admit that an IR-converted camera produces superior results, compared to using the screw-on IR filters over lenses instead. Many of my lenses that were previously unusable for infrared now work just fine in most (not all) circumstances. My super-wide lenses that don’t even have filter threads on them are suddenly my go-to lenses for infrared, such as my Rokinon 8mm fisheye.

You might have noticed that just about all of the photos in this article were made with lenses that have been reported as either substandard or unusable for infrared. This was done intentionally. I have to question if anybody making those reports has tried to use an AR coated sensor filter.

Again, you will want to avoid shooting into bright lights. Given this limitation, many “unusable” lenses are suddenly usable.

Hand-holding shots instead of multi-minute, tripod-anchored exposures is a real treat. I can always use my 10-stop neutral density filter (most lenses) when I want long exposures.

I think that the Kolari Vision anti-reflection coating on their infrared sensor cover filter makes a huge difference. The cost increase for this camera conversion optional feature is about the same as buying a single good infrared filter. Money well spent, if you like infrared shooting half as much as me.