Table of Contents

Examples

This page contains various examples from the 3 cameras.

Lenses

The Nikon D70-IR and the Fuji S3 Pro-UVIR both used the Nikon 18-200 f/3.5-5.6 G ED-IF AF-S VR DX lens, a good performer in infrared.

Then Canon EOS 5D-IR used the Canon EF 24-105mm f/4L IS USM lens.

Image stabilization was turned off for both lenses.

Raw-file conversion

All frames were converted from RAW to 16-bit TIF using Adobe Photoshop Camera Raw, in order to compare “apples to apples”. Photoshop is definitely not my preferred RAW converter (see Raw File Converters), but it was used to eliminate the raw-file converter as a variable. White balance was arbitrarily set using the chimney towards the upper left, since it is large enough, with the same lighting in every photo.

Some of the images (the S3 ones in particular), were out-of-gamut for the sRGB color space. This is not an Bad Thing (and in fact might be indicative of the S3’s wide dynamic range). The originals were pulled up to 0.8 stop in the red channel (or in one case the blue channel) in order to avoid “blowing out” the red channel when converted to sRGB color space, which is notoriously bad in distinguishing saturated reds. This was done because only one browser, Mac OS X Safari, knows how to correctly display images in color spaces other than sRGB. Please see Web Browser Color Display.

The S3 Pro-UVIR images were converted from RAF (raw) files as 12 megapixel images, not the default 6-megapixel, because the extended dynamic range feature of the S3 was utilized.

After conversion to 16-bit TIF, the following post-processing was done, all in 16-bit mode:

1. Each frame was converted to Lab mode.
2. Each frame was “Auto Leveled” in Photoshop to equalize brightness (except for the “Venus Filter” frame ).
3. The 5D-IR and S3-UVIR frames were down-sampled to a width of 3008 for the large images (native D70-IR resolution width), 600 for the medium images, and 300 for the small images. These were done using Photoshop Bicubic Sharper.
4. Photoshop Unsharp Mask of {15, 50, 0} on the luminance (“L”) channel.
5. Photoshop Unsharp Mask of {150, 0.3, 0} on the luminance (“L”) channel.
6. The Lab image was converted to the “sRGB IEC61977-2.1” color space, black point compensation off, Perceptual Rendering, Relative Colorimetric.

The images were saved at Quality=40 for the 600-pixel-wide images, and Quality=High for the 300-pixel versions.

Image comparisons—infrared

Image comparisons—visible light

Neither the 5D-IR nor the D70-IR are capable of visible light shooting, so there is no comparison to be made. I did shoot the Fuji S3 Pro UVIR both unfiltered and with two filters which ought to allow visible light shooting: the light-green B+W 489 and the more exotic B+W 486 UV-IR Cut 489 (see the B+W Filter Handbook for details).

For the Fuji S3 Pro UVIR, I used the same Photoshop CS2 settings shown below to facilitate comparison with some exposure compensation as needed (up to 0.8 stop “pull”).

The four visible-light images are shown below:

Fuji S3 Pro UVIR—no filtration

Without filtration, the S3 Pro UVIR renders images with a very strong color cast. There is roughly as much infrared light (perhaps more) than as visible light.

Fuji S3 Pro UVIR—B+W 489 light cyan

The B+W 480 light cyan filter is designed to cut off infrared, but its spectral transmission curve slopes off gradually, so it still lets too much infrared through.

Fuji S3 Pro UVIR—B+W 486

The rather expensive B+W 486 interference filter is designed to eliminate both infrared and ultraviolet with a sharp cutoff. It uses special coatings to do so, and is noticeably more effective than the B+W 489 above.

Applying Photoshop “Auto Levels” yields an image which is not too far off from a normal camera. I made no attempt to fine-tune the white balance, but using a gray card or MacBeth Color Checker card might allow good results, though images are unlikely to remain as true as an unmodified Fuji S3 Pro color camera.

When considering the B+W 489 filter, please note that a cyan color cast will increasingly appear in the corners of the image as the angle of view increases. This becomes a problem with the Nikon 12-24mm/f4 DX (or equivalent wide-angle lenses from Canon). Cleaning the 486 filter is also problematic because its filtration derives from the properties of the coatings on the glass, not the glass itself.

Image comparisons—ultraviolet

Quartz lenses

Everyday lenses attenuate ultraviolet light severely, though lenses with very few elements and/or single coatings will let a small amount of ultraviolet light pass. To assess ultraviolet response, one needs a quartz lens. These are are and very expensive, and one was not available for this test. Here are a few candidates if ultraviolet photography appeals to you:

105mm/f4.5 UV-Micro-Nikkor

This lens is very rare, and can only be obtained used. Stephen Gandy at CameraQuest.com has offered them from time to time. Used ones sell for about $4500, but they are extremely rare. A complete lens comes with Nikon’s special UV filter, which mounts on the front of the lens, and can be flipped to the side for composing/viewing, then flipped back into place.

Coastal Optics 105mm/f4.5 UV macro

Apparently this lens, made once you order it, because delivery time is quoted at 6-10 weeks. About $4500.

ukaoptics.com UV1054B

Comments from Mike Caruso at UKA Optics:

This lens was designed for 416mm working distance, but can be used for general shooting.

The price of the UV1054b lens is $1570. The filter thread size is 49mm. It is not apochromatic. It is a triplet designed for 250mm wavelength and all surfaces are AR coated.

Nikon 35mm/f2.5 Series “E” Nikkor

According to Bjorn Rorslett at naturfotograf.com, a cheap, though less effective alternative is to remove the front and rear coatings from a 35mm/2.5 Series E Nikkor (eg polish them off). I had no quartz lens, nor did I choose to ruin my 35mm/f2.5 by removing its coatings.

Ultraviolet filters

In addition to a quartz lens, an exotic filter is mandatory—the filter must block all visible light, and infrared as well. Bjorn Rorslett at naturfotograf.com reports that the Baader “Venus Filter” is the best filter he’s yet found for ultraviolet photography. I purchased one through alpineastro.com [transmission curve].

Ultraviolet crude test

I shot a few frames with a multicoated 35mm/f2.5 Series “E” Nikkor while holding the diminutive “Venus Filter” against the front of the 35mm/f2.5 Nikkor (it fully covers the front element). A 15-second exposure in full sun was required, and it’s unclear whether the response is entirely ultraviolet, or whether it also has some infrared in it. The 15-second exposure explains the blurred trees.

Click to see a table with S3-UVIR medium-sized versions.

The individual red/green/blue color channels from the “Venus Filter” frame show that something unusual is going on, but whether it is pure ultraviolet, or ultraviolet contaminated with other wavelengths is something beyond my expertise to judge with any certainty.

Fuji S3 Pro red/green/blue channels from image above [LARGER IMAGES]
Red	Green	Blue

High resolution examples

If you are interested in a version of this article with high-resolution examples (all images sized at 3008 wide, the D70 resolution), do send an email as I am considering making it available at a reasonable price. The issue is bandwidth; the article would be 50MB were I to include high-quality images for all the example shots, which would quickly swamp my server and end up costing a great deal of money for bandwidth should I offer indiscriminate downloading. Please note that I will be out of contact from September 14-27 on a photographic trip to Norway.

A paid article is planned for the future that will include a head-to-head comparison of the Canon EOS 5D-IR and Nikon D200-IR (and perhaps the Nikon D70-IR) in the usual high-quality diglloyd.com fashion (it will probably be around 200MB). Please keep an eye on the Paid Reviews page.

Main Page | Examples | Usability | Hyper-Utility | Conclusions