The ISO capabilities in DSLRs and mirrorless cameras today is pretty amazing. But it’s not perfect. You’re still going to see an increase in noise as you ramp up the ISO, even if it’s not quite as bad as it used to be. But we all love clean, noiseless images, don’t we? Along with the improvements in ISO performance, we’ve also seen big leaps in noise reduction technology. But many of them still aren’t perfect, often eliminating noise at the expense of detail. For some images, you really do need to keep that detail and in this video from Scott Walker at Walks on the Wild Side, we learn three different methods of noise reduction. These methods include simple Lightroom Edits, Topaz Denoise AI and a pretty complex method that’s very involved but “fixes everything”. He also covers a “Pre-noise reduction” process you can do before you bring your photo into 3rd party software.
It’s a very in-depth video, running 36 minutes long, explaining exactly how ISO works, the different types of noise you get in digital images and then his three processes for cleaning up the images.
Lightroom’s noise reduction has come a long way over the years and can clean up noise pretty well these days, but it’s still not perfect at figuring out noise from detail sometimes. There are things you can do to improve it, but there’s usually a compromise. Topaz Denoise AI is an excellent piece of software that, as the name suggests, utilises AI in order to try to produce the best possible result. Such software is typically a one-click solution to noise reduction, and Topaz Denoise AI also has a one-click option, but it also lets you deep dive into the settings to get more control. The most control comes from Scott’s final method, which is a pretty involved process using Photoshop and layers. It incorporates all of the previous techniques, with the same image processed several times with different settings. This is because a given set of settings will apply itself to different parts of the image differently. So, it’s processed multiple times with different settings to produce a range of images, each processed for a different part of the shot. These are then composited together and masked in Photoshop to produce the best overall result from this stack of images. Like I said, in-depth. And the final method shown is about as far from a one-click solution as you can get. But if the image is worth it, it deserves the work! [via FStoppers]
