.
What you get here is a brief explanation of human vision and infrared imaging in my own words and from my own experience. Lots of more detailed information and example images can be found on DPFWIW: Digital Photography For What It's Worth.
|
| A very infrared avenue... |
Scientifically speaking, light is electromagnetic radiation ranging in wavelength from 400 to 700 nm (nanometers, 10-9 meters). The human eye has four kinds of receptors for light: three kinds of cone cells for daylight color vision, which allow us to perceive 420 nm as blue, 534 nm as green, and 564 nm as red, and rod cells for night vision which respond to a wider wavelength range, peaking at 498 nm but extending even a little into the proximity of 750 nm ("beyond red"). However, our brains convert the image information from the rod cells to a bluish monochromatic image - this is why faked night scenes in old movies were recorded through a blue pass filter and just don't look exactly real.
Infrared radiation is anything beyond the long-wavelength end of the visual spectrum. Most people think of thermal vision when they read "infrared" but that is not precise. Thermal infrared (TI) - heat radiation from surfaces at "reasonable" temperatures - ranges from 3 µm (3000 nm) to 1000 µm. In between this thermal infrared and visible light, around 700 to 3000
nm, is near infrared which is also called "reflected infrared" because it behaves more like visible light. Optical effects like reflection, refraction, shadows etc. are more pronounced there than thermal effects.
Digital Camera imagers (CCD and CMOS sensors) are designed to be sensitive to visible light, and can distinguish it as red, green and blue by using tiny color filters on each sensor element (the Bayer pattern). But what few people know is, these sensors are also effected by radiation slightly above and below this range. The ultra-violet (UV) sensivity - i.e. the response to "harder" wavelengths of 380 nm and below - is virtually zero (unlike in traditional "chemical photography", where UV contamination can be a problem). In the near infrared range, the sensors are more sensitive and it depends on the chemistry of the camera's color filters how this extra light is interpreted. My Olympus C-4000Z's red and green filters transmit a considerable amount of what I call "immediate infrared" (800 nm), and the blue filters let through at least some "near infrared" (950 nm).
Most cameras have a built-in "hot mirror filter" of varying steepness that prevents (normally) unwanted NI contamination from changing the colors in visible-light-photography. As for Nikon "Coolpix"es, there seems to be a way to remove this filter without destroying the camera, and for Olympus "Camedia"s, some models - luckily including my C-4000Z - only have a relatively weak filter. In some webcams with threaded fix-focus lenses, the filter is a separate piece of glass that can be removed carefully with a small screwdriver. This glass behaves somewhat weird in that it looks reddish against a dark background, but bluish or greenish on a bright surface.
For brevity, I will use the letters N-I-R-G-B-U-V to reference Near and Immediate infrared, Red, Green, Blue, Ultra-violet and Very ultra-violet light, respectively. Far infrared and thermal emission imaging (much longer wavelengths) requires expensive special equipment and is not dealt with here. Color imagery on this site is either NI or RGB. And speaking of near infrared in general (N and I), I will use "nir" like an extra color.
|
| My improvised filter holder on the C-4000Z |
As I said, most digital cameras see both RGB and NI light, but you need to prevent all visible light from entering your camera so it can concentrate on the invisible. This is done using an "infrared filter" (more precisely called an "infrared-only pass filter"), i.e. a filter that exactly blocks visible light and thus looks completely black to the naked eye. The type you will find recommended most often on other "digital infrared" sites - and also the one I bought - is the Hoya R72, sometimes called "IR72" as well. I ordered mine from Foto Koch as I couldn't find it in any of the professional photography shops in town. For what you can do with it, at around EUR 30 it is also relatively inexpensive. My C-4000Z doesn't have a thread for filters on its object lens, so I had to improvise a filter holder made of a carton tube, adhesive tape, silver foil from a cereal packaging and two rubber rings. Sounds strange, but it looks good and is almost perfect.
Taking NI pictures is a bit trickier than normal "point and click" photography. You practically always need a tripod, because exposure times can go to four seconds and beyond. I use a small 12 cm tripod (as seen in the picture to the right) on my bike saddle when I'm on the road, which is quite acceptable unless strong wind shakes my bike. Even in bright sunlight, exposure times usually won't go below 0.5 seconds, so fast-moving objects in the scene can be a problem - but the motion blur also is part of the beauty of IR shots.
Sunshine is even more important than for RGB photography: pictures taken on days with overcast clouds or even rain are almost always blurred and really dull.
|
| My hair isn't blue and this is a black sweater... |
To be precise, nir doesn't "look" at all. You can't see it. What you see in my "NI" images are actually false colors: red and green represent immediate infrared, and blue represents near infrared (a little more distant from red than "immediate"). To get an impression of NI images displayed in "true color", simply press the button labeled "Power" on your monitor :-)
How unedited infrared color pictures look depends on the camera - some will take only grayscale images, some a nearly monochrome greenish. I use one of the "MyMode" custom saved settings in my C-4000Z as the "Infrared mode", with a fixed white balance that emphasizes the blue (N) channel - otherwise the image would be nearly a red monochrome of the "I" channel.
Nir is scattered less than visible (especially blue) light while traveling through the atmosphere, which is responsible for two effects: The sky appears darker, but with an extremely bright sun (most nir light travels straight through the atmosphere from the sun without being "distracted"), and distant objects (mountains etc.) appear a lot closer because of the missing haze in the distance.
Looking at skin in the near infrared range is also astonishing. In blue light (and early black & white chemical photography which was most sensitive in this color), human skin looks awful, because all darker or bluish spots, veins etc. below the surface are very pronounced. Red, and even more, nir light is so "soft" that it bounces off the upper layers instead of penetrating deeper down, so skin will appear clean and pale in nir light. For very aesthetic (although only grayscale) examples of this, visit Infrared Dreams. I guess this effect is also the explanation for the use of red illumination in red-light districts - it simply makes people look more beautiful!
Live foilage (tree leaves), tree bark and grass appear bluish-white (i.e. mostly "N"), just as flowers do, no matter what color they boast in the RGB range. I believe the reason for this is that the "target audience" for flowers are insects (who can see RGBUV) and mammals (RGB), so evolution didn't optimize flowers for flashy behaviour outside that range. As for foilage - "organic solar cells" so to speak -, the sun's radiation has its maximum between red and green (experts call this intermediate color "yellow"), so evolution has optimized the leaves' pigments and energy converters (chlorophyll and carotin) to absorb red and yellow light - they reflect green and nir as some sort of a "light waste" they cannot make use of.
By the way, snow appears more reddish-white (more "I" than "N"), so snow and grass don't look exactly alike.
|
(c) 2004-2007 by dAWiDi@ubahnsound. part of dawidi's widicam.net | 
| Kindly hosted by CONCEPTNET |