How to handle strong noise

Milford Sound is surely one of the most beautiful places on Earth. This is Mitre Peak at the gates of Milford Sound just before sunrise. Photo by: Jacob Surland, www.caughtinpixels.com

As I have been writing about I have been searching for something I can use as a new style. I have had a feeling of what I was looking for. I have had a few shots at it, without striking a clean note.

But as I was rumbling through older photos I came across this one, which I never processed before and applied one of the Lightroom filters I made last week – and there it was! A clean note. Or at least the beginning of it. I did a little extra post-processing here and there, but not a lot, and I was happy … for about 10 seconds. I zoomed in, and saw that the image was ruined by noise. I had shot it at ISO 10159 (why why WHY!!!). Because I had had the camera on Auto ISO the day before, while driving through a tunnel to get to this wonderful place. And this photo is shot almost in darkness.

It is processed in Lightroom only, except for the Noise reduction. To handle the noise reduction, I started by reducing the size of the image to about half resolution, not half megapixels. I shot this at 36 megapixels, cropped it down to 29.5 megapixels, and what I did was to reduce the resolution from 7360×4004 to 3676×2000 and thus reducing it to 7 megapixels). After I had done that I applied noise reduction to the final result. And this is it.

I used Noiseware for the noise reduction, and I used the Default settings. There are a number of various settings, but I quickly started loosing details, so I stuck with the default filter.

The photo is still noisy, but it actually works quite nice with the colors and style I think. If I had used stronger noise reduction I would have lost so many details, that the photo would have been ruined.

In my Arcanum cohorte one of my fellow apprentices asked about when to do noise reduction. I first tried to do the noise reduction before the resolution reduction, but that gave a terrible result. Then I reduced the size first and then did the noise reduction, and that was a much better result. This is still noisy, but at least it is regular noise.

Leave a Reply