Why are photographers interested in infrared photography? I think it mostly boils down to getting an interesting ‘look’. They want to see plants glow, as if they’re covered in heavy frost (it’s called the Wood Effect, after the photographer Robert W. Wood who discovered this in 1910). They also like to see really dark skies and have the magical ability to make atmospheric haze virtually disappear. Freckles can be made to vanish. There are probably a few photographers that are curious to find out if swimming suits really do disappear in IR light.
I’m not willing to sacrifice one of my cameras to get converted into full-time infrared. I’m going to stick with using IR filters instead. I do have to pay the price of really long exposures, which means manual focus and framing before attaching the filter, and it additionally means I need to use the Nikon DK-5 eyepiece cover to stop eyepiece-entering light from ruining my shots.
Do I also have to pay a high financial price for a high-end IR filter to get decent IR photographs? Will bargain filters work okay? Some tests are in order.
I doubt most photographers really care exactly which percent of different light frequencies are involved. Enter cheap IR filters. There’s an explosion of cheap IR filters available, and I couldn’t resist trying a few of them. I soon found out that light frequency transmission standards are pretty much nonexistent when it comes to many IR filter producers. In other words, 720nm and 850nm could be replaced with “fat free” and “non-gmo” with equal accuracy.
On the other hand, what I surprisingly didn’t find were optical aberrations, bad threads, or plastics in the filters I tried. I’m sure these kinds of problems can be found in cheap filters, but I guess I got lucky. I wouldn’t touch a plastic filter, no matter how cheap it is.
I ended up testing IR filters from Hoya, BCI, Neewer, and Zomei. All of these filters use optical glass and have metal mounts. I haven’t seen any significant differences in optical resolution or thread quality with any of these filters. As previously mentioned, the main differences I observed have to do with light spectral response.
I avoid pointing the lens at light sources when shooting IR; I don’t think multi-coating is much of a consideration with infrared filters (the lenses themselves struggle with significant internal IR reflections).
I did find a problem unique to the BCI 850nm filter; it had uneven optical density across the field of view. The filter produced what looked like a series of light-dark concentric circles, centered about the middle of the field of view. It was easy enough to hide these circles using an image editor, because the light-dark pattern wasn’t too severe. Still, I’d avoid recommending the BCI 850 filter because of this defect.
I seem to keep returning to my ancient manual-focus Nikkor prime lenses for infrared photography. These old lenses have the best internal absorption coatings for the (entire) tested infrared region of the spectrum, they have great manual focus scales, and they also have the little infrared focus-shift reference ‘dot’ on them. Exposure and focus is manual anyway, so these old lenses don’t present any disadvantages. Unfortunately, I can’t mount any of the old Nikkor lenses unless they’ve been converted to “AI”, in order that I don’t damage the meter coupling tab (on most of my cameras). My old Nikon D5000 and D60, by the way, don’t have a meter coupling tab on them. This means that I can use the old non-AI Nikkors on them without any problem.
Many lenses will work acceptably for infrared at wide apertures, but they start to develop the nasty white hotspot at f/8 or narrower apertures. The hot spots tend to get worse as you use the longer-wavelength filters. Keep checking this link for updates to lenses (from all lens manufacturers) that are suitable for infrared photography here.
I always have to mention my Nikkor 18-55 DX f/3.5 – 5.6 VR lens. It’s IR performance is very good. Manual focus is a pain with it, however, and it’s not FX.
After I did some testing to determine how much to compensate for the infrared focus shift, I discovered that the little Nikon IR focus scale compensation dot is calibrated for roughly 720nm light. You need to shift focus about twice this amount when using 850nm filters. I put a little dot of white fingernail polish on my manual Nikkor lenses to indicate the 850nm focus shift.
My favorite camera for infrared use is my Nikon D610. Infrared exposure times get really long, so I like to crank up the ISO to minimize those times (up to ISO 2500). The D610 sensor is very forgiving with high ISO’s; for other cameras, I lower the ISO and live with longer exposure times. I have to switch to “Bulb” or "--" (time exposure) and use a remote release for exposures longer than 30 seconds (click to raise the mirror, click to start exposure, click to end exposure).
Hoya R72 Infrared Filter
This is a fairly well-regarded filter and lens company. The R72 purportedly cuts off light wavelengths shorter than 720 nanometers (actually meaning less than 50% transmittance of visible light below 720nm). It has good mechanical and optical quality. This filter isn’t inexpensive; I doubt that many people could call it inferior in any way.
I have no way to measure spectral response, but I believe the Hoya light response claims to be true (see the chart below). The sky ends up being reasonably dark, and green plants have their characteristic glow. You can’t see anything when you look through the filter (human vision ranges from 390 to 700nm).
I’m unable to measure a white balance to get a preset with my cameras when I use this filter, since these modern cameras are a bit too efficient at filtering out IR light. (My old Nikon D50 and D60 could get a white balance preset with the Hoya R72). I can get a decent preset white balance by ‘borrowing’ the white balance from my Neewer 850nm filter (see below).
My main complaint with this Hoya filter is that colors (after getting a neutral white balance) are a little bland. Cranking up the color saturation in post-processing only adds a marginal improvement. The medium-dark sky yields a “tobacco” color after the photographs get a proper white balance. In some circumstances, the ‘look’ of this filter is what I want, but not always.
Hoya R72 filter specifications, courtesy Hoya Optical
The graph above shows why you can’t see through this filter (it only transmits above 700nm). Pure infrared light, even at the lower wavelengths, has very little of what we can perceive as “color”. The lack of a broader spectrum of low-wavelength light translates into dull colors.
You can convert the pictures into black and white, if the weak colors don’t tickle your fancy.
Typical Hoya R72 shot with neutral white balance
Note in the shot above that some of the cactus plants become translucent in infrared light. There are nearly always some visual surprises that you’ll see after a day of shooting IR.
Typical Hoya R72 shot with hue shift for the “blue sky” effect
I have been perfectly content with the ‘look’ of the Hoya R72 for several years, but I think my tastes are starting to change. Can I get a better ‘glow’ from other filters? Are the colors the same with other wavelengths? Read on.
Bear Claw Industries (BCI) Infrared 590nm, 665nm, 720nm, 850nm
These filters are inexpensive, so I figured it would be a low-risk experiment to try them. I got the 82mm thread size, and I use step-up rings to fit my different lenses. The filters came as a kit; normally I would avoid ‘IR’ filters with wavelength cutoffs below 700nm.
I have no complaints about the quality of the optical glass or metal threads on these filters. But…
I knew there was going to be some trouble when I opened up the box and could see right through three out of four of these filters. You’d think that the BCI 720nm filter would look just like the Hoya R72; not even close.
The “infrared” 590, 665, and 720nm filters visually all look like shades of red/orange; this would be expected for the 590 and 665 filters. Orange is 590nm to 620nm, and red is 620nm to about 750nm, according to Wikipedia, although it also says you can’t see beyond 700nm. Hmm…
Methinks BCI must be stretching the truth a bit with their 720nm filter. I couldn’t see any “Wood Effect” whatsoever with photos taken through the 590, 665, and 720nm filters. I just can’t detect any infrared effect. The BCI 590, 665, and 720 nm “infrared” filters are a bust.
I can’t comment on how these filters would perform on an infrared-converted camera, but their filter labels of “Infrared” seem to be false advertising.
The BCI 850nm filter, however, is opaque to human vision. Shots through this filter show up as deep purple, unless you configure a proper camera white balance first. Light transmission is about 4 stops lower than the Hoya R72, so exposures are pretty long.
I really, really like the lighting effect with the BCI 850, although the pictures need to be converted into pure black and white. There isn’t anything you’d call ‘color’ with this filter, so infrared color photography isn’t an option. The sky is rendered super dark, and plant chlorophyll really glows. The bad news, though, is that many of my lenses that seemed fine with the 720nm IR spectrum start getting nasty internal reflections at 850nm. Exposure times are quite long with this filter, typically extending to a few minutes. I don’t dare use really wide apertures with this filter, because the infrared focus shift typically leaves you with out-of-focus shots. My Nikkor 18-55mm VR “kit” lens, which is normally very good at infrared, started to show slight internal reflections at f/11, which only got worse at smaller apertures. These longer 850nm wavelengths work well with very few lenses, compared to the 720nm wavelength filters.
I am completely unable to get a white balance measurement with this filter. Fortunately, I can ‘borrow’ a very usable white balance preset from my Neewer 850nm filter (see below). It still requires conversion into black and white, but at least the camera LCD shows a decent neutral-balanced picture (with a faint blue color cast).
BCI 850nm long-wavelength loses color information
The shot above is typical of the character of light transmitted by the BCI 850nm. The original shot was converted into black and white, to rid the slight color cast. Photos without a preset white balance will look deep purple. This filter is excellent for eliminating atmospheric haze. You lose the “distance” cues in landscapes, since the gradual increase of atmospheric haze is absent.
Be prepared for very long exposure times. You might find this filter handy as a super-neutral-density filter, although only for black and white.
My big complaint with this filter, as I alluded to above, is uneven light transmission across the field of view. The sky can look like it has a monochrome rainbow in it, depending upon the lighting conditions and direction. I can fix this defect in my editing software, but this is the fatal flaw in my BCI 850.
Neewer 850nm Infrared Filter
Here’s where I got a very pleasant surprise. I was expecting this filter to match the results of the deep-infrared BCI 850nm and Zomei 850nm filters. Not even close. The nearest visual equivalent to this Neewer850 would be the Hoya R72, but the colors are quite different (much more saturated). I’m guessing that it should probably be labelled something like “Neewer 700nm Infrared”. This filter is really dirt cheap, and I like the lighting effect it achieves.
I have no complaints about the Neewer filter threads, and I haven’t noticed any optical imperfections, either. It does seem that there is a gradual vignetting toward the edges of this filter, which generally adds to more dramatic skies.
The Neewer850 is visually opaque, but the color results are wild. I’m actually able to get a successful camera white balance measurement with all of my cameras when using this filter! I usually point my camera at green lawn grass to get the white balance preset measurement. My camera histograms show a healthy amount of red, green, and blue in them, which might be why the white balance measurement succeeds.
The exposure is about a stop slower than the Hoya R72, and about 3 stops faster than the BCI850/Zomei850. A typical sunshine exposure is ISO 400, f/8, and 30 seconds for this Neewer850. I like to keep exposure times at 30 seconds or less, since otherwise you need to deal with the hassle of externally-timed exposures. I avoid apertures wider than f/8, due to the uncertainty of focus. Exposures are all over the map with infrared, so you will have to rely heavily on your camera histogram feedback.
I really like the colors as-shot with my measured IR white balance preset from lawn grass. The sky is a great red-orange, compared to the drab tobacco color result from the Hoya R72.
If I want to create the in-vogue “blue sky” color infrared shots, I prefer to use the Viveza 2 plug-in (via Lightroom) and perform a hue shift of roughly -150. This provides essentially the same effect as the “red-blue channel swap” popular with Photoshop. This hue shift only works if the photo has a proper neutral white balance to begin with.
Depending upon the photo, sometimes the effects from processing the shot in the HDR Efex Pro 2 plug-in (via Lightroom) can be really pleasing.
Neewer 850nm, as shot. Nikkor 20mm f/4 at f/8, Nikon D610
The shot above shows the colors I get when the camera uses a white balance obtained from green lawn grass. The orange sky looks much more vivid than the Hoya R72 tobacco color. Foliage has the nice Wood Effect ‘glow’, as well. Note the gradual vignetting of the sky.
Neewer 850nm, hue shift, via the Viveza 2 plug-in in Lightroom
The sky shown above looks more saturated and dramatic than the typical Hoya R72 filter photo. Vegetation takes on the characteristic yellow color. This hue shift is equivalent to the red-blue color channel swap in Photoshop, but Viveza 2 gives much more control over the channel mixing. Addition of extra warmth in post-processing after the hue shift would benefit shots like this.
Neewer 850nm after HDR Efex Pro and Viveza 2
In the shot above, I first used the HDR Efex Pro plug-in via Lightroom. I started with the original neutral white balance shot, but turned it into HDR to get more dramatic lighting and saturation. Next, I took the single-shot HDR and then performed a hue shift inside Viveza 2 to transform the orange/red colors into magenta.
You wouldn’t want a steady diet of this stuff, but the effects that can be obtained from the Neewer 850nm filter combined with HDR Efex Pro and Viveza 2 can be very dramatic. If you want your photos to stand out from the crowd, here’s one way to do it.
Zomei 850nm Infrared Filter
This filter produces visual results that look almost exactly like the BCI850. Unlike that BCI850, however, this filter produces perfectly even light transmission across the field of view. This Zomei is becoming my go-to filter for super dramatic pure black-and-white IR photography.
The exposure times nearly match the BCI850 filter. They’re about 4 stops slower than the Hoya R72.
My camera histograms, when using the preset white balance I borrowed from using my Neewer 850nm filter, typically show a perfectly even balance of red, green, and blue. It’s as if I used a grey card to calibrate this filter. The Hoya R72, for comparison, shows about 80% of the histogram as a red channel response. As I mentioned earlier, I cannot get a direct preset white balance measurement using this filter, so I just borrowed the white balance from the Neewer 850 filter.
If I could only keep a single IR filter out of my collection, this would probably be the one. My only complaint is the long exposure times.
Zomei 850nm, toward the sun
Shooting in the direction of the sun leaves the sky dark, but not nearly as dramatic when compared to shots with the sun behind your back.
Neewer 850 comparison shot
It's astonishing to think that the Neewer 850 shot above has the same specifications as the Zomei 850nm shot. They couldn't be more different!
Zomei 850 away from sun
The sky has really turned dark with the sun behind me. When you're shooting dry vegetation and rock, the subjects look about the same as regular black and white. Plants with chlorophyll, however, turn almost pure white.
Compared to cameras and lenses, trying out some cheap IR filters carries a very low financial risk. Don’t expect scientifically accurate results from these filters, however. If you get nasty hot spots in the middle of your photos, don’t blame the filter; blame the lens. You’ll need to experiment with focus compensation, depending upon which filter is being used (longer wavelengths require focusing much nearer).
Please, please remember to cover your camera eyepiece to avoid washed-out pictures (such as the Nikon DK-5) or else use the built-in eyepiece shutter on the more ‘pro’ camera models. Always take the photos using Raw format, to give yourself sufficient elbow room when you process the shots in your editor.
These kinds of filter experiments can open up a whole new photographic area to explore. There are more and more companies starting to offer IR filters, so go try some of them! Avoid any plastic filters, however.
The range of visual effects that you can achieve with different IR filters is really amazing. If you have an adventurous photographic spirit, this is a great avenue to explore without having to make a significant investment. There’s a kind of dreamy charm that is unique to IR photography.
And no, I don’t know if any of these filters make swimsuits transparent. Those tests will have to wait for another day;) Also, I’ll bet that swimsuit manufacturers got wise to IR photographers a long time ago and switched to opaque fabrics.