Infrared - Does the sun's angle matter?

[slejhamer]

A question for the IR gurus:

I took some infrared shots at my local botanical garden this past weekend.

When I converted the RAW images to TIFF (using the "Don Ellis Method" of custom white balance selection), I found that:

1. What I expected to be light magenta was orange.

2. What I expected to be icy green/blue was white or gray.

3. Worst of all, there was excessive chromatic distortion in the higher contrast areas (alongside a wooden structure), showing up as yellow and pink highlights. I don't mean a little bit of the G1's "purple fringing" problem, but very noticeable "what the heck is that?!" distortion.

I tried many different points for my white balance selection but could not get a satisfactory result. The unprocessed RAW images had the typical magenta cast and looked "normal" before conversion.

The shots were taken in the very late afternoon, with the sun maybe 15 to 20 degrees above the horizon. I use a G1 with Hoya R72 filter; the R72 was stacked on a Sunpak polarizer and a protective UV(0) filter. I used autobracketing but the different exposure settings did not matter. I have not experienced these problems before when using the same setup, and am most concerned with the chromatic aberrations.

Could the sun's angle have made such a difference? If so, would that imply that the season in which one shoots could also influence the end-result of an IR shot?

Any other thoughts on what I may have done incorrectly or should have done differently? Right now my only options are to convert to grayscale or to diffuse the heck out of the images. (Not a bad effect, but one I don't want to overuse.)

Thanks in advance :) ,

[ken-w]

slejhamer wrote:
A question for the IR gurus:

I tried many different points for my white balance selection but could not get a satisfactory result. The unprocessed RAW images had the typical magenta cast and looked "normal" before conversion.


Hi Mitch - a couple of quick thoughts.

What happens when you do a normal "out of the camera" RAW conversion? (not using Don's technique). If you're seeing a typical magenta cast in the unprocessed raw, a processed TIF using the camera settings indicated by the RAW should come out looking the same. I shoot JPG with IR so I'm not familiar with the issues surrounding IR and RAW.

The fact that autobracketting looked the same likely indicates that the camera was at its lowest f-stop/speed for the mode, in fact likely you were in a slightly underexposed situation (otherwise the + f-stop bracketting should have showed a brighter result).

The chromatic distortion is possibly due to over-saturation. What is the saturation setting on the camera?

Having said all that, I've never done late day IR. There tends to be more shadows and more intense (saturated) light. I wonder if the chromatic distortion might also be due to stacking 3 filters. It would be nice to have some comparsion shots with just the IR on the camera to see if it resulted in the same distortion.

[slejhamer]

Hi Ken. Just tried the "out of the camera" conversion; the magenta cast stayed, as you expected, but so did the chromatic distortion. The yellow was very apparent even before auto-leveling, which also turned up the pink. :(

Re: the autobracketing, I meant that the distortion was there regardless of exposure setting, so it wasn't an exposure issue. The autobracketing did work as it was supposed to; sorry I wasn't clear on that point. All of the other camera settings were "normal"

I will try to get out and shoot some test shots with different filter combinations. But, like I said, I've used this stacked combo before with acceptable results.

Your comment that "there tends to be more shadows and more intense (saturated) light" in the late afternoon gives me a thought: the areas with the worst distortion are along a fence which, because of the sun's angle, had many contrasting shadows. Perhaps the G1 got confused and tried to make color adjustments it ordinarily would not have... It wouldn't be the first time that a G1 has gone slightly crazy. :)

Thanks for your thoughts,

[Don Ellis]

Hi Mitch,

Since I'm not going to be able to help, maybe I'll just post a couple of pictures. :)

But first a word: your question about the sun's angle seems valid because we're talking about reflection. Thinking about my own shots, I'm not sure I've taken any IR photos that early or late in the day.

I now leave all my contrast, saturation and sharpness settings on low, but I don't recall any problems when I used to have them on normal.

In terms of magenta or orange color, I find that I occasionally get an orange RAW rather than magenta -- but that's before conversion.

Here's a recent example when I took the same photo with and without a polarizer (these are the THM files converted to JPG for display here)...

Without a polarizer...
http://www.kleptography.com/dl/crw_2071.jpg

With a polarizer...
http://www.kleptography.com/dl/crw_2072.jpg

I converted both using the same custom white point and then Auto-Contrasted, slight Curves and Sharpened. Here are the results...

Without a polarizer...
http://www.kleptography.com/dl/crw_2071-600ps120.jpg

With a polarizer...
http://www.kleptography.com/dl/crw_2071-600ps120.jpg

As you can see, they're nearly identical. The white point for both was taken about 20% up from the bottom center.

I'm surprised to hear about oranges, yellows and pinks after any sort of conversion. I'm using BreezeBrowser, by the way, but since you've not had this problem before, I wouldn't imagine that it's your conversion software.

Don't really know what to say. If it's convenient, you might consider shooting the same scene when the sun is higher and check the results.

One last thing -- because I use the G1 as a dedicated IR camera, I don't have a UV filter on the Lensmate. I let the R72 serve as both protector and filter.

If there's any way you can post a link to the original RAW file, I would be happy to take a look at it. And the EXIF data might help as well.

Let us know how you get on.

Cheers,

Don

[slejhamer]

Hi Don, thanks for your comments.

I just opened one of the RAW files in Breezebrowser (still on "trial" ) so I could see a full-size preview of the RAW image. The distortion is apparent even before conversion, so it's not the processing software.

Your images with and without polarizer are interesting, but I agree that it does not seem to matter much in the final product. I'm using the cir-pol to reduce glare coming from a pond as well as to darken (relatively) the sky vs. some foliage in the foreground, and the results until now have been favorable.

I definitely need to run more tests to see what works best for me. If I end up with similar results (distortion-wise) I'll link you to one of the RAWs; thanks for offering to take a look.

Question: Do you run your camera settings at "Low" for technical reasons or simply preference (i.e., letting Photoshop do the work instead of the camera?) Also, do you use the low settings on your other camera when you take color photos?

Thanks for taking the time to comment.

[Don Ellis]

slejhamer wrote: Question: Do you run your camera settings at "Low" for technical reasons or simply preference (i.e., letting Photoshop do the work instead of the camera?) Also, do you use the low settings on your other camera when you take color photos?
Hi Mitch,

I like to use the best tool for any given job, so I run all low settings on both the G1 and G2 and change the values, if needed, in Photoshop, which gives me finer control.

Cheers,

Don