# What variable maximizes color in astrophotography?

I'm doing some research in astrophotography before my trip to Yellowstone and I'm most interested in pulling the color of the galaxies. I will be using a Canon 7D.

What should I be focusing on in order to maximize the color? Will cranking up the ISO bring me to a threshold where I should instead use a lens with a wider aperture?

I have the Canon EFS 10-22mm 3.5-4.5, Canon 17-55mm 2.8, and Canon 50mm 1.4. I have a small budget for renting gear as well. Am I going to be falling short on the equipment to get galaxy shots with color? If so, should I get a prime wide lens to get wider aperture or a full-frame camera that can crank up ISO?

-

As Matt Grum has stated, you'll want to maximize exposure value (EV), but using the longest exposure at the widest aperture you can. Generally speaking, shorter focal lengths support longer exposure times, and when your goal is wide-field photography, focal lengths of 16mm or less are great (so your 10-22 @ 10mm-16mm will be an ideal lens.) To determine the exposure time you'll need to avoid star trails, you can use the "Rule of 600". This simply states:

Divide 600 by the 35mm-equivalent focal length to determine shutter speed.

If you were to shoot at 10mm on the 7D, you would need to adjust to 35mm-equivalent (multiply by 1.62x), and divide 600 by that value:

``````600/(10*1.62) = 600/16.2 = 37 seconds
``````

At 10mm, you could expose the sky for 37 seconds before you start to experience star trails. At 16mm you could expose for about 23 seconds, and at 22mm you could expose for about 17 seconds. At 37 seconds you should be able to get a decent shot with good color, so long as you use the right ISO setting. At all exposure times, I would recommend taking a series of as many photos as you can...dozens, even a hundred. A tool like DeepSkyStacker could then be used to align them, eliminate noise, and stack them into a high resolution (even super resolution) and high color-fidelity final image.

One area where my recommendation differs from Matt's is the ISO setting. I own a 7D myself, and at very high ISO's noise takes on a very different characteristic than at low and medium ISO settings. In particular, ISO 2000, and anything ISO 3200 and above, are really going to eat away at your color fidelity, while ISO 2500 and anything under ISO 2000 won't usually affect color fidelity in any seriously adverse way. It is unnecessary to use that high of an ISO setting with a Canon anyway, as read noise drops to its normative low by ISO 400, between 2.8 and 3.4 electrons worth, vs. 8.6e- at ISO 100 and 4.7e- at ISO 200). The ideal ISO level from a read noise and color fidelity standpoint would actually be ISO 800, where it is only 2.8e- (see SensorGen Canon 7D for details), although with the higher maximum saturation of ISO 400, the 3.3e- read noise won't be a major problem.

You will probably want to tune your ISO setting, which affects the maximum saturation level, to the exposure time. A longer exposure will be more capable of saturating more, so at ISO 400 you could use 10mm and a 37second exposure so a higher max sat. level is useful. For a shorter exposure, such as at 22mm, you might want to bump ISO up to 800 to allow the exposure to saturate faster, and maximize SNR, etc. Keep in mind, the higher the ISO setting/shorter exposure you use, the more the random physical nature of light will affect your shots. At ISO 1600 or higher, with very short exposures of only a few seconds, regardless of whether you manage to saturate some pixels, a lot of pixels will likely never receive any light at all. At this point, taking as many shots as you can and stacking them in post will be the only way to really create a "complete" picture with high saturation and good color fidelity.

-
This answer is a great example of why I love this site... – AndyML Aug 16 '12 at 13:35

The variable you want to maximise is total light captured, i.e. exposure time / f-number squared. That will give you the most information possible to determine colour (or the best colour-signal to colour-noise ratio).

After you've maximised that it's just a case of setting the ISO as high as it will go without clipping the signal, in order to combat read noise.

Depending on what you want you'll just about get by with your gear, at 10mm you ought to be able to pull off a 30 second exposure without noticeable star trails. The 50 f/1.4 will let in more light but you'll have to use shorter exposures to avoid trails.

However if you're just shooting the milky way with no foreground objects then you're fine as you can shoot a load of short exposures with the ISO all the way up and stack them using a software that corrects for sidereal motion to reduce noise.

-

Many of the famous photos of the stars and galaxies (and nearly all of the famous Hubble shots) use false color. The color is not on the plate, its added by an artist or analyst.

-
Just to note, this requires the careful use of filtration and many shots taken under specific filters, managed manually as individual channels, such that each channel can be given its own color before final blending into a single image. You usually need a tracking mount to support false-color astrophotography (and for all the effort, having ideal tools, like a good telescope and a better imager.) – jrista Aug 15 '12 at 17:52
Yeah, and they filter frequencies of the incoming signal, some of which are not in the visible spectrum. They essentially use layers for each set of filtered image. Very cool stuff, way over my head. – Pat Farrell Aug 15 '12 at 22:03