Lightroom Classic is catching up. AI-based masking has been around for a long time in programs like ON1 Photo Raw but was only recently released in Lightroom. Now, Adobe has moved on to noise reduction. Does it stand up to scrutiny?
There is a wide range of AI-based noise reduction programs available; Topaz DeNoise AI, ON1 NoNoise AI, and DxO Deep Prime all perform much better than Adobe’s terrible noise reduction results. Those apps work well as plugins for Lightroom and Photoshop. In the case of DXO PhotoLab 6 and ON1 Photo Raw, they have also integrated AI noise reduction into their own raw development and asset management software. Even camera brands’ proprietary software offers their own AI noise reduction; OM Workspace, which is growing in popularity with OM System users, has built-in AI noise reduction.
All those programs work fine. I’ve run multiple tests in the past, running the same raw file through each file, and each of them does the job too well. Some are slightly better than others, but those differences have more to do with speed and usability than noise reduction performance.
I just tested Lightroom’s new AI-driven noise reduction feature along with other programs to see how well it performs compared to other tools.
Worry about monopoly
Since many people use these AI-based programs as plugins for Lightoom and Photoshop, there is concern that Adobe is creating a monopoly because these third-party plugins will become redundant.
However, there’s more to it than noise reduction. Other standalone programs, such as DxO Photolab 6, ON1 Photo Raw and Capture One, are highly appreciated by users because they produce very different development results that many users prefer.
Meanwhile, Topaz Denoise, not an integral part of the raw development and asset management suite, isn’t the only tool in that company’s arsenal. Furthermore, it will still be a popular plugin with Capture One owners even if it becomes redundant for Lightroom users; Capture One currently does not have AI-based noise reduction.
Stable closing after the horse has latched?
Sorry for mixing up metaphors, but Adobe’s horse not only ran away from the stable, but people stopped watching the race it was supposed to run.
My first DSLR, I was able to use up to ISO 400 without noise. With my second, it jumped to ISO 800. I changed to mirrorless Micro Four Thirds, and that number doubled again to ISO 1,600, then ISO 3,200. I can use my OM-1 at ISO 6,400 or even 12,800 without thinking about applying noise reduction.
Even raising the ball by two stops, noise was not an issue for me. Furthermore, with my latest camera, I can shoot long exposures without image noise.
Modern cameras perform better than I need them to, and I suspect the same is true for many others. Advances in sensor technology mean that noise reduction is becoming unnecessary for photographers.
Adobe should have introduced this feature three years ago. Still, it will be a welcome addition for those using Lightroom, especially if they use an older camera and shoot moving subjects in low light.
Lightroom’s Denoise Performance
There are a few problems I found. First, I was disappointed that the preview image didn’t fill the entire screen, something that happened with ON1 and Topaz, although that’s a complaint I’ve heard with Deep Prime. Worse still, Lightroom’s preview is not clear so it is difficult to judge the final result after processing. All other programs I have used provide a clear and accurate preview of the final image.
Second, I have an old computer; I built it about eight years ago. However, it has a 3.4GHz quad-core Radeon R7 processor, 32GB of RAM, and a solid-state drive. Although it’s old, it runs pretty fast for most tasks. ON1 applies noise reduction almost immediately, DxO PhotoLab 6 Deep Prime takes a little longer, Topaz Denoise takes longer but they still work in seconds. Lightroom took 14 minutes to denoise some images. Those with newer machines will find this time significantly reduced, but it shows that Adobe’s noise reduction tool is slower than other tools I use. Photographers with older cameras might think twice before using it often.
My tests were taken with different ISO settings, conditions, and exposures, so I’ll have plenty of photos to try. I am using an image to illustrate my findings. It was shot at ISO 6,400, 1/12,800 sec, at f/5. I used aperture priority and dialed at -1 EV.
Before denoising, I first zoomed in 100% in Lightroom and then cropped it to similar sizes, then I applied Lightroom’s Denoise at the default values.
At first, I turned off sharpening because I’ve always found Lightroom’s algorithm too aggressive. But I think the results of the AI look a bit soft. So I increased the sharpness to 40, but the result still looks too muddy. By increasing the sharpness to maximum and increasing the radius to 1.5, I was able to achieve a more satisfying result, but it is still far from perfect.
This is the same image that has been removed with other programs.
The difference in these results is seen at 100% and is small. I repeat that no AI program produces bad results.
Pushing noise reduction to its limits
I’m eager to push the software further. So I found the noisiest photo I have in my catalog. It was a photo I took for the sole purpose of testing noise reduction software over four years ago, long before I started writing for Fstoppers.
It was a 1,800-second exposure shot taken at night at the conclusion of an extended exposure photography session with the Olympus OM-D E-M1 Mark II, which has had two replacement generations since.
Before taking this photo, I did a long exposure for over an hour. As a result, the sensor heats up and there is a lot of color noise as well as a lot of hot pixels.
The lens I use is an old Four-Thirds lens, which shows more distortion at the edges and is less sharp than the newer Micro Four-Thirds lens I’m using. This will be a challenging test for any software. Lightroom’s traditional Noise Reduction has almost no effect on color noise and hot pixels on this image, so I’d love to see how it works.
Lightroom Classic AI Denoise did a good job with noise reduction. On close inspection, it removed more hot pixel artifacts than other programs, but nowhere near all of them. There are many white spots left on the image that I can see when taking a pixel peek. It also does a good job of removing color noise, even if, again, it makes the image look a bit muddy.
Lightroom isn’t as good as DxO Photolab 6 at correcting lens distortion, and it loses some shadow detail.
This is a harsh test. No program can turn this very poor quality image into a wall hook. But it did give an idea of what could be achieved. Likewise, with all of these tests, I took a pixel peek, zooming in at 100%, something that is rarely needed.
Quick talk about Super Resolution
The Denoise function also comes with a Super Resolution option that allows you to increase the number of pixels in an image. So far, I haven’t tested that and that will be the subject of a whole other article.
What I like and what can be improved
Lightroom’s Denoise didn’t do its job well. This new feature is a welcome addition for dedicated users of Adobe, especially if they own older cameras that are more prone to noise than newer models. It will also please those who are in the habit of shooting at high ISO settings.
I especially like that the denoised image is a DNG file. However, this does not mean that one cannot synchronize noise reduction across multiple images.
The thing to remember is that this is just one component of the entire package. Lightroom is an amazing, arguably the best, catalog-based asset management program out there. It also does a pretty good job of developing raw images, and it has a great user interface for doing that.
However, while its raw development results are not bad, they are not always the best either. Of course, this is a subjective opinion and you may disagree with it. The inclusion of AI-driven noise reduction makes the show better than before. However, I found the development and noise reduction results of DxO Photolab 6 better in most cases. But I’m not sure if the difference is big enough to entice satisfied Lightroom users away from it.
But for those already invested in other programs, this new feature is unlikely to entice them back to Lightroom. For example, over the past few months I’ve been applying DxO Photolab 6 to my workflow and haven’t been able to switch back to Adobe because it has AI Noise Reduction.
Lightroom Classic’s noise reduction is very slow compared to other programs, but this may be less of an issue for those with new turbochargers.
In short, it will be a welcome addition for those dedicated Adobe users, but the question remains whether it is necessary. On the following cropped image, taken at ISO 2,500, can you tell which photos have noise reduction applied and which ones?