Digital IR - Nikon D70
When I sold my digital SLRcameras in 2003 to buy my first digital SLR, the one element I was most reluctant to let go of was working with 35mm infrared film. Though I had experimented with post-processing digital images to achieve the look of infrared film, I realized that, while it was possible to make some images resemble an infrared image, it was totally impossible to get the same look that a true infrared film gives as it sees light that is invisible to traditional digital cameras. As a result, I resigned myself to setting infrared aside for a while, with the hope that somehow I’d eventually be able to work out a digital solution.
I knew this was possible, as all digital cameras are infrared sensitive because this is an inherent quality of both the CCD and CMOS sensors which they use to capture light. This causes problems, however, with their visible light images. To combat this, camera manufacturers place a hot mirror filter over the sensor, making it difficult to capture infrared light with an unmodified camera. Some cameras do pass enough infrared light to permit images to be made through infrared longpass filters which block available light but pass infrared light when using very long exposures; for my purposes, however, a 2 second to half minute exposure isn’t practical.
After some promising, but ultimately unsatisfactory, experiences with a Sigma SD10, I was still searching for the ideal camera to convert. At this point, I began to look at Nikon cameras which, according to various internet sites (here and here), could be modified to be infrared sensitive by the placement of an infrared longpass filter over the imaging sensor. This had a number of distinct advantages - first, lenses for the Nikon system are varied and widely available (unlike those for the Sigma SD10) and, second, with a filter over the sensor, as opposed to the lens, the image could be seen unimpeded through the viewfinder and easily composed. Other cameras require you to continually add and remove an opaque filter in order to restrict available light from entering the camera, slowing down the creative process and leaving little room for spontaneity.
Camera Conversion
Shortly after learning about the possibilities of converting a Nikon D70 to infrared sensitivity, I approached Nikon Canada about the possibility of having a repair technician modify my camera. While the instructions on-line seem straightforward enough, the idea of taking a new, fully warranted camera and voiding the warranty seemed a little insane. Fortunately, Nikon was willing to carry out the modification if I provided the replacement filter and accepted that the Nikon Warranty would not apply to the modified portion of the camera (the new infrared longpass filter), not an issue for me. This approach had the added advantage of insuring the modification was done in a clean room, preventing any dust between the sensor and its new covering filter.
Finding the replacement filter turned out to be easier than I’d expected. A quick search of the internet turned up dozens of filers, and in the end, I settled on an infrared longpass filter from Edmund Optics. These plastic filters are ideal for blocking visible light, while passing infrared wavelengths. A 2"x2” filter cost me under $7.00 US and provided enough material for a practice cut, as well as the sensor filter itself.
The filter arrived in a little under a week, within a small padded envelope, inside a plastic bag, inside a paper envelope. Five minutes with a sharp craft knife, and I had the longpass filter cut to the measurements required to replace the existing hot mirror filter. With that done, I packaged up the camera, filters and a copy of the website instructions for the modification, and shipped them off to Nikon Canada for conversion.
Focus Correction
When I received the modified D70 back from Nikon Canada, I immediately set it up and made an image. Much like the Sigma SD10, the image was bright red, with some aqua-blue tones (interestingly, an overexposed area on the Sigma was bright yellow, while on the Nikon, these appear a bright cyan). 
After setting a custom white balance on the camera (2000K and a -71 tint), the images appeared as almost black and white on the camera’s LCD screen (though most people’s hair turns out to be the most delicate shade of egg-shell blue)..
The only real problem with the first test images was that the Nikon D70’s autofocus was inaccurate. Infrared light focuses at a closer point than visible light so manual focus correction was required to get a sharp image at larger apertures (all Nikon lenses with a distance focus scale on them have an infrared light focus correction mark). In my initial tests, when I corrected the focus of each image, the final images were perfectly focused, even at the maximum lens apertures.
After spending a couple of days testing the modified camera, I recalled reading on the internet about being able to adjust the focus of a D70 and, after some searching, I turned up the site, and following the instructions. In less than ten minutes, I had modified the AF system to correctly focus the infrared light. The procedure was surprisingly simple and easily reversed if necessary.
The focus correction procedure is detailed below:
1: Preparations
I made the adjustments to the modified D70’s focus with the camera mounted on a tripod, pointing at a brick wall 20m away. I attached the camera via USB cable to a computer, and used Nikon Camera Capture 4.2 software to produce all the focus test images, recording them directly to the laptop. I used an AF Nikkor 50mm f/1.4 lens for the tests, photographing at the largest aperture.
Before starting, I made a reference image so I could know if the focus was getting worse as opposed to better. After auto-focusing on the wall, I made the test exposure and began the modification.
2: Mirror Lock up
It is necessary to hold up the mirror while rotating the AF-mirror adjust screw. Some suggest taping the mirror up while the adjustment are made, but I would be a little hesitant to recommend this, due to the residue the tape may leave and the number of times adjustment may be necessary.
When I modified my camera, I wore a latex glove to hold the mirror, more to avoid dust and debris from entering the camera than because of any concern over fingerprints.
3: Inserting the Key
Once the mirror is out of the way, the two black adjustment screws will be plainly visible on the right side of the mirror box. The one in the back, just in front of the shutter is the AF-mirror adjust screw, which needs to be modified for infrared focus.
A 2mm hex key is carefully inserted into the mirror box and set into the AF-Mirror adjust screw. It is very important that the hex key is inserted very carefully so the shutter, which is very close to the AF-mirror adjust screw, is not damaged.
4: Focus Adjustment
Placing the hex key in the screw is a little tricky, but once seated, it moves quite freely.
Rotating the AF-mirror adjust screw in a clockwise directions shifts the camera’s focus point closer, effectively correcting it for infrared focus. For my specific camera, the amount of adjustment required was exactly equal to the maximum adjustment possible without removing the hex key and making a second turn (about 15-17 degrees).
After each focus adjustment (I made three before finding the perfect angle of adjustment) I made another test image, carefully auto focusing on the wall and comparing the results to the reference image. Once I achieved the correct focus with the 50mm f/1.4 lens at f/1.4, I tested the correction with my other lenses (ranging from the 12-24mm to an older 105mm f/2.5 MF lens, and all of them proved to be quite accurate.
Field Conclusions
After working with the modified camera for a week with the corrected focus, I have made a number of discoveries (some of which confirmed suspicions I had before modifying the focus). The first and most important this is that the focus modification produced a major improvement across the board in image focus and sharpness. Hundreds of images made with the 50mm f/1.4 at f/1.4 and f/2 focused at portrait distances (1-2 metres) were all tack sharp, with the point of focus crisp and contrasty.
The truth, however, is that the focus modification is an improvement but not perfection. While the camera is now perfectly corrected for infrared auto focus with the 50mm f/1.4 lens at portrait distances, the AF correction is less accurate for farther distances at large apertures, and for longer or wide lenses (this I had suspected would be the case, due to the longer or shorter focal length of the lenses and the fact that the modification of the camera’s AF system in effect just modified the length of the AF path by a small percentage).
The good news is that the focus modified camera, while not perfect with all lenses, is closer to being in focus than before the modification. As a result, if longer or shorter lenses are stopped down to middle apertures (f/8-11) the modified AF is much more than accurate enough (much more accurate than the unmodified focus was at the very least).
The notable exception to the above is the Sigma 10-20mm super-wide lens, With the autofocus on, the image with the infrared camera is totally out of focus, but if I pre-focus to 1m, almost regardless of the distance between the camera and the subject, the image is sharp (I find this optically very odd, as it should be important to focus for each image, but I suppose at 10mm, there is so much depth of field, that even at f/4, there is enough depth of field to cover most subject distances when pre-focused).
Another apparent side effect of the infrared conversion is that there is significantly more dust accumulating rapidly on the new sensor cover. My suspicion is that the hot-mirror filter which the manufacturers use to cover the sensor is coated with an anti-static agent, to help repel dust and debris. The solution to this issue is simple, regular cleanings with the American Recorder C02 system. A quick cleaning before each session seems to do the trick. .
Image Post-Processing
More than any other digital image, the results from the infrared modified D70 need to be post-processed before use. As I make all my images in RAW format, it is easy to tailor each image to suit but, even with a standardized batch processing of the images, the results are quite pleasing and easily achieved.
The most important factor in getting a good image with the infrared modified Nikon D70 is starting out with a good white balance. I did this by simply pointing the camera at a green grass and setting a preset white balance (usually measuring about 2000K, -71 tint). Indoors, with both available light or studio flash, I set the white balance off something like white sheets. The main result of this was an increase in the lightness of the blue and green channels (see the histogram examples below), resulting in a more neutral image. Since setting this, most of my images have been quite neutral in tone, with only a slight colour cast to elements in the image.
| Comparison between a conventional and IR converted D70 | |
|---|---|
 |
 |
Made with an unmodified D70 in manual mode, this image was exposed to maximize the tonal range, with the information as close to the right side of the histogram as possible. The preview image (above) looks much too light and seems grossly overexposed. When using RAW file formats with conventional digital cameras, producing an image which is as light as possible without loosing highlight information is the key to maximizing the image quality. This is one of the major differences between working with a digital and film camera. | Made with an infrared converted D70 in manual mode, this image was exposed to insure that as much information as possible was retained in the highlights. The preview image (above) looks much too dark and seems grossly underexposed. There is a pale cyan cast to the leaves in the water (not possible to see on this photo of the screen), indicating the exposure was nearing the threshold. A bolder cyan would indicate blown highlights which would have no tone when converted to in the post-production. The difference between these tones is crucial in determining if the highlights are overexposed or not. |
 |  |
| The resulting histogram on the camera has a large spike towards the right side but nothing actually touches the right edge (which would indicate overexposure). In all but the most extreme cases, this method of exposure for RAW format images will yield the broadest range of tones in the post-processed image. | The resulting histogram on the camera has a large spike on the left side, and a small amount of data indicating the highlight tones (this normally would indicate extreme under-exposure). The Nikon histogram only shows the green channel, so the histogram exposure feedback must be combined with highlight colour in the LCD image to determine the correct exposure. |
 |  |
| In the computer, the RAW file was then post-processed to extend the tonal range out over all the available tones between pure black and true white, resulting in the final image. | In the computer, the RAW file was then post-processed and converted to monochrome, with the tonal range adjusted so that it extends from a pure black to a true white. |
Camera Upgrade
In the spring of 2007, I upgraded the infrared converted 6mp Nikon D70 to a 10mp Nikon D80 camera, which I had converted through Lifepixel, as Nikon Canada declined to convert the camera for me (it is important to be aware that converting a camera to IR through an independent company voids the original camera warranty).