Over the past twelve months or so, the race to have the best Noise Reduction software has been heating up.
Recently On1 have entered the fray with their latest offering: NoNoise AI
That’s right, yet another ‘AI’ powered noise reduction application. Before I get into a rant about the use of the ‘AI’ moniker: Lets check out how it performs
Let’s get right into the results. How does well does NoNoise AI do at removing noise from an ISO 8000 file from my Nikon D500?
The below comparison shows 100% crops of the raw file processed in Lightroom with zero noise reduction vs the same file run through NoNoise before processing. Impressive!
It is pretty clear in the above example that NoNoise has eliminated most, of the noise and has not sacrificed any of the fine feather detail in the process.
One of the most complained about aspects of AI powered noise reduction tools is that they take ages to process each file. This doesn’t appear to be the case with NoNoise AI.
On my 2018 MacBook Pro with Radeon 570 graphics, NoNoise processed the above 20 Megapixel file in less than two seconds. As a comparison, Topaz DeNoise took 17 seconds and DxO PhotoLab took 28 seconds for the same file.
I haven’t yet had a chance to test how all three applications go on my speedy Desktop machine, but I’ll be sure to add the results when I can.
Now let’s take a look at how the results from NoNoise compares to the competition.
For the below comparisons I put the original RAW file through a workflow to optimise the quality of the results in each program. I didn’t just press the ‘auto’ button, but made sure to tweak each program to get the best results I could.
The NoNoise photo was processed as a RAW file in NoNoise, saved as a .dng file and then processed in Lightroom to get the finished photo.
NoNoise vs Topaz DeNoise AI
To create the photo in DeNoise I processed the file in Lightroom with zero noise reduction or sharpening applied, then opened the file in PhotoShop and ran DeNoise AI with manual settings to get the best result I could.
You can see that noise has been significantly reduced in both files, and they both retain an excellent level of detail. However, if you look closely, the background of the DeNoise AI image shows an ugly ‘mottled’ effect.
I have written an article on how to remove this problem in DeNoise, but the NoNoise result is clearly better
NoNoise vs DxO DeepPrime
To get the DxO Image I opened the RAW file in DxO PureRAW, applied DeepPRIME and Optical Corrections before exporting it as a .dng file. I then processed the .dng file in Lightroom.
You probably wouldn’t notice the difference in a print, and in fact I prefer to leave a tiny bit of noise in the background to give it a slight texture – but in terms of the amount of noise removed – NoNoise is the winner here, and you can always dial back the NoNoise settings if you want to leave some noise in.
NoNoise RAW vs TIFF
If you try and open a non-RAW file in NoNoise AI you will get a warning telling you that the noise reduction is not optimised for rendered files such as JPEG or TIFF.
I thought I would test out just how different the two file types are.
As it turns out…. very! Using NoNoise AI on a non-RAW file results in much worse results than using a raw file. Look closely at the background and feathers around the eye… the TIFF image shows much less detail and the background is quite mottled.
At this point in time, I would only consider using NoNoise AI on RAW files. As they mention in the warning, that feature is currently in beta and I’m sure they will improve it over time.
Download The Free Trial
If you want to check out NoNoise AI for yourself then you can grab the 14-day free trial at the link below. The trial comes with zero limitations and any files you process during that time won’t be watermarked or otherwise comrpomised
If you decide to purchase the software after you’ve tried it out, it will cost you US$69.99.