EDIT: This blog has MOVED. You’re on the old, dead blog. Everything (including this post) has been migrated onboard my website. Here’s the direct link:
Update your bookmarks! See you there!
We now continue with our original programming….
First out is a test of a couple of noise reduction plugins; Noise Ninja, and Nik Dfine. I’ve owned both for a while, but I’ve never done a side-by-side comparison before. I have to say the results surprised me!
I’ve been using Noise Ninja as my go-to solution, mostly because I bought it first, and actually went through the entire user’s manual one time (during a long flight). So I know more or less how to use it, although I’m certainly not expert. Nik Dfine is functionally very similar, and even the interface is very similar, so it’s easy to transfer skills from one to the other. But, as you’ll see below, you may not need many skills at all. Here’s an image I made last week:
Click to see it bigger. This was shot on a Canon 5D, with the 70-200 @ 200mm; f/2.8, 1/20th, ISO 1600. In other words, it’s noisy as hell. And while I had the camera braced against a doorjamb, at 1/20th it’s also not the sharpest image I’ve ever made.
Above is a 100% crop of the image. Yikes!
Now, here it is with the “Auto” settings in Noise Ninja applied:
Noise Ninja has profiles for the camera, lens, and exposure settings, and applies them automatically based on the EXIF data. It also analyzes the image to “find” the noise, and applies is algorithms to those areas. (I’ve always wanted to use the word “algorithm.”) But in this case, little has changed.
Here’s the same image, with Nik Dfine applied, also using the “auto” settings:
Whoa. That’s pretty good. Some loss of detail, but it killed the noise, for sure.
I didn’t want to give up on Noise Ninja so quickly, though. So I went back and instead of letting it profile the image automatically, I manually selected the areas that had the worst noise. Then I went into the controls and selected “coarse noise” and upped the Strength and Smoothness sliders, as well as the USM amount. That got me this result:
The Dfine version is certainly lacking in some detail – but with a little tweaking of settings, I got this result:
This I like. I’m not at all an expert with this stuff, so if you’ve got methods you like, hit the comments and let me steal them share ‘em!