(For more in-depth editing, check out the second installment: Aurora Post Processing 2: Raw Therapee)
I think a lot of people out there take some great shots of the northern lights, only to blow the final product in their editing suite, whatever it may be. Everyone’s philosophy differs on the editing side of things, so in that respect this is all opinion. I like to keep the editing simple, with as few steps as possible. After a lot of reading and trial and error, here is what I’ve found works for me. . .
What I want to show you now is how I deal with the two main killers of a good aurora photograph: the white balance and noise.
Why bother? Well, your digital camera doesn’t have a firm grasp on what exactly ‘white’ is, not like your eyes do. Different light sources will change what the white reference point is, which is why most cameras have pre-set white balance settings like: daylight, cloudy, shade, incandescent, flourescent, etc. So essentially, the wrong white balance will mess up tone, color, and temperature. Many cameras will allow you to manually set your white balance by giving it some reference. This works fine for me in the daylight, but I haven’t been very successful with this at night. My solution to this is to not even bother trying – shoot in RAW and worry about white balance in post-processing. The camera’s white balance modes only affects jpegs, which are awful. (Actually I shoot RAW + basic jpeg for quick previewing later. I never use the jpeg for anything else.)
I’ve seen tons of Aurora images on the web, in calendars, on stock sites (selling for hundreds or thousands of dollars!) that do not look anything like what the aurora looks like . . . ever. The problem almost always boils down to white balance. Here is an example of one of my pictures with two “wrong” color schemes.
The leftmost picture uses a pre-set mercury lamp white balance, and the second a white fluorescent white balance. These are the two extremes that I see very often, one makes the aurora look way too warm and in the other its nearly blue. The sky is kind of a giveaway too – neither looks like the color of the sky I’ve seen at night. I’ve seen a lot of pictures where, instead of fixing the white balance, the photographer increases the contrast to make the sky dark enough that you can’t see the color in it.
So then, what setting gives you the “correct” white balance? I’m not sure if there is really a “correct” answer to that question. At this point it is easiest to rely on memory. You were there, you saw it, what did it look like? This is more challenging than you may think it is. In my brain there is definitely an acceptable range that will give a realistic, attractive image. If you want to make attempts at calibrating your aurora photos to the 557.7 nanometer green line from atomic oxygen, then be my guest. I will read your article because it would be interesting and then I would keep on making my best guess for what represents what I actually saw. I’ll talk more about that in the processing section below.
Noise is yuck. When shooting aurora, it’s nice to be able to shoot at a high ISO setting so that you can decrease exposure time. Although with a lot of cameras this means you get a lot of extra noise. When I upgraded to the Nikon D7000, the noise issue became much easier to deal with. The sensor is much better than on the D90, even at high ISO settings. However, I still want to do my best to eliminate noise. The camera has “noise reduction” software, but this is limited, it doesn’t work as well as I want, and it only applies to jpeg (which is awful).
I’m a huge advocate of open source, so I use GIMP for my editing instead of Photoshop. GIMP doesn’t have a built-in way to read or edit RAW (.NEF) files so for that I use UFRAW. Once UFRAW is linked to GIMP it’s all like a single package so there is no need to export or import, I just open the file through GIMP.
Like I said before, I like to keep my edits simple, and how I handle white balance and noise is the epitome of simple. In UFRAW, the white balance is handled with a simple drop-down menu on the 1st tab (outlined red):
With my camera’s color profile I typically have good luck using either the Direct Sunlight or Day White Fluorescent white balance settings. You may have to play around to find what works for you. It helps to have your camera’s icc profile loaded into UFRAW (check out color management), but it is not necessary.
Now that I have the color where I want it, it’s time to get a handle on the noise. Luckily the D7000 is not very noisy, and it is 16.2 megapixels. High megapixel cameras are not necessary, but it does allow you to get away with more noise reduction in post-processing.
Zoomed in you can now see the grainy ‘noise’ that I want to try to eliminate. To do this I just use the Denoise slider bar on the same UFRAW tab.
I just use the AHD Interpolation because it seems to work well and give consistent results. Here I’ve taken the slider to 400, which is pretty far. You will lose detail using noise reduction, but even here I haven’t lost much. One thing to look out for, especially when using noise reduction on lower megapixel images is a kind of striping (smooth gradient banding) effect. This was a problem I had with the D90, and I could only use Denoise up to about 150 before this would start to happen:
Here you can see the banding effect in the aurora from too much noise reduction. Especially apparent are the rings that occur around the stars. You’ll have to play around to find a ‘sweet spot’ where you have decent noise elimination without too many artifacts from using this algorithm. This was a different picture taken with the D90.
Other than the fact that I adjusted the exposure in UFRAW (the slider at the top – increased by 0.35 stops), these are the only two things I did to edit this photo. So, I feel successful at keeping it simple and realistic anyway. To finish it up I hit ok on UFRAW, sending the photo to GIMP where I save the file as both a tif and jpeg (booo!) for uploading to the internet. I should note that once in GIMP there are additional tools that can be used for color, exposure, and noise reduction, but I’ll get to those eventually. This is just a start!
If you found this helpful, I do like coffee and beer!