Photography: Hard Work
Now I’m no clueless amateur with Photoshop but the best artisan is always limited by the tools available. In my case, that tool is a Kodak Easyshare CX6200. Now Kodak’s consumer point and shoots aren’t exactly the greatest snappers and the Easyshare series are fairly close to the bottom. It’s a creaky old 2.0 megapixel (1600×1200) disaster.
The objective was to capture noctilucent clouds: Night time photography. This is all but impossible on a point and shoot, they simply don’t have the exposure or the sensitivity. The longest exposure it can do is half a second and the sensor is noisy as hell. A crop of a shot from the session is to the right, but brightened up a bit since this blog has a white background.
It’s a foul mess of noise and hot pixels, and while you can indeed see the clouds, it’s hardly a great image and there’s only so much you can do in Photoshop. Garbage in, garbage out. There had to be some way to give Photoshop more information about the image.
There is. Using a small mini-tripod (that’s why the hole at the bottom of your camera has threads), I placed the camera on the ground and fired off 18 shots, then covered the lens and did another. It’s imperative that the camera does not move and your subject is absolutely still.
In Photoshop, you then load the 19th (dark) image and subtract it from all the others using the Apply Image function. This removes junk added by the sensor such as hot pixels. It’s called darkfield subtraction and commonly used by amateur astronomers for exactly the same purpose.
The next step is to add all these images together. This CAN be done in Photoshop but it’d probably take forever. I use a piece of freeware called RegiStax 5 for it which has an incredibly awful interface. It adds all the images together and averages them, a process known as stacking. Noise, which is random from frame to frame, does not survive the averaging, but the detail of the image reinforces from frame to frame.
You then drop it into Photoshop, pull the curves around a bit and you get an image far better than any one shot could have been with much less noise.
