|
Search this group's discussions
|
White Balance / Color Temperature?
|
I shoot RAW exclusively, especially when shooting with my IR10D, but was wondering what you guys prefer to set your white balance to?
I've tried shooting an exposure of the sky and that seems to work ok.
Recently, I've just been setting a Kelvin white point manually in the 10D to 2000°. It gets as close to black and white as possible so when I view the image on the LCD after shooting, it's pretty close.
I've also thought about creating my own fake white balance test shot syntetically in Photoshop and then moving it onto my memory card. In the camera I would pick that photo for Custom White balance. I wonder if though this technique you can simulate color temperatures below 2k or above 10k or have a wider hue setting.
I have most often just been picking the color temperature manuall during RAW conversion using the dropper tool either on tree bark, neutral wood or the asphalt on a road. Seems to work OK.
Thoughts? Experiences?
Posted at 5:39AM, 18 September 2005 PDT
(permalink)
|
|
There are a lot choices for setting white balance with infrared and they can lead to several of different kinds of effects. Fortuneately shooting raw gives you a lot of choices after the shot.
Most of the infrared color gradients that can be captured on a CCD or CMOS sensors occur between 700 nanometers and 800 nanometers. This is one reason that the Hoya R72 and Wratten 89B filters are so popular. At longer wavelengths, characterized by the #87 & #87C filters, the RGB filters on the digital focal plane behave as IR pass filters at these longer wavelengths. But each RGB channel has a slightly different neutral density, that leads to a distinctive midtone color hue. ...
Opps, got called away to lunch so I'll continue or edit this later.
Originally posted 81 months ago.
(permalink)
starfish235 edited this topic 81 months ago.
|
|
I haven't done nearly enough IR work to decide on a "best" method, but I do know that if I'm doing a series of shots and have the wherewithall to remember to shoot white balance, I have great success shooting a full-frame grass, leaf, or shrubbery shot and using that as the custom white balance reference.
I'd show some examples, but last time I did remember to was a while ago and a dead harddrive took those shots with it to the grave.
I look forward to Starfish's comment, it sounds very promising already :)
Posted 81 months ago.
(permalink)
|
|
|
Like boodoo, I like to set the white balance for the #89B filter (Hoya 72 equivalent) using full frame green vegetation. This give trees, grass, and folage white glowing highlights. I like this because these are typically the brightest areas. Using full frame vegetation results in a gain pattern where the red and green channels have about the same dynamic range, but the blue channel has about 65% of the dynamic range of the other two channels. This setting leaves a reasonable amount of separtion between the red and blue channels for switching them to achieve a nice false color IR effect. It also yields warmer mid-tones if you decide not to do any channel switching.
I don't like to set the Infrared white balance for the #89B filter using a bright white object. This typically results in a red channel with less dynamic range than the blue and green channels. ... The the most intense IR highlights (whites) tend to look like blown out cyan. This isn't a real good situation. You can get the red channel gain boosted back up ahead of the blue and green by chosing an incadescent or daylight white balance setting if you're shooting in raw. I've also had the white balance move as the sun rises ... I've had a good white point setting at ~9:00 AM, but the white point moved to blown out cyan by 10:00 AM.
Using Blue sky to set the white balance tends to gives the truest highlights. All 3 color channels tend to have similiar dynamic ranges. However, the mid-tones typically seem a little colder and the sky tends to be fairly neutral.
Clouds are probably one of the most variable subjects you can use to set white balance. The sun lite side behaves like a white wall, and the shadows are like a neutral grey card. Using the dark underside of a cloud tends to give the very neutral midtones. Foliage tends to look a little bluish because the blue channel gain has been pushed to the limit.
You can change the midtone hue using levels or curves adjustment to set the midtone by sampling a different area. Sampling a slightly different white point can also change the midtone hue when all 3 channels have similiar dynamic range in the IR.
One other technique that I've used is to set the white balance with a #87 or #87C filter and then shoot the 89B (Hoya 72 equivalent). See the color IR shot of Mary Magdalen's grave for an example of this technique.
The leaves still have the white IR glow. All of the channels have fairly uniform gain deep in the IR Spectrum and the natural distribution of IR colors in the scene, between ~700 to 800 nanometers, creates a different response in each channel. This technique seems to yield a slight orange bias with the #89B filter, but auto-levels or auto-color renders fairly natural looking midtones.
Setting the white balance for the #87 & 87C filters isn't very sensitive ... almost any method you choose works about the same.
Originally posted 81 months ago.
(permalink)
starfish235 edited this topic 81 months ago.
|
|
|
Wow! Starfish, this is far more information that I expected! I'll definitely try setting the white balance on foliage. I can't use different IR filters for setting balance since my filter is permanently installed in front of my 10D's AF and CMOS sensor.
I also want to try putting a range of synthetic WB samples on the card to see what happens...
Posted 81 months ago.
(permalink)
|
|
|
iso100, I don't see any reason why you can't use other IR filters with your IR-10D. One of the nice things about the ~720 nanometer IR filters is that they have a very flat transmission curve in the IR Spectrum (~90% transmittance or -0.15 EV). This means you can add another IR filter with a longer IR cutoff wavelength in series with your built in IR filter and get a uniform response curve for the added filter with only a 0.15 EV neutral density bias ... That's less than 1/6 of an f-stop.
You don't have to add additional IR filters in series, but you can.
Originally posted 81 months ago.
(permalink)
starfish235 edited this topic 81 months ago.
|
Would you like to comment?
Sign up for a free account, or sign in (if you're already a member).
|
|
Digital Infrared /
Discuss
Feed – Subscribe to Digital Infrared discussion threads
