Finally! Adobe Denoise brings Photoshop / Lightroom NR into the modern era

Adobe Denoise made its debut for Camera Raw, Lightroom CC and Lightroom Classic in the company’s April 2023 update, as shown here in the What’s New dialog from Lightroom CC.

Throughout its 16-year existence, Adobe’s Photoshop Lightroom has impressed us in many ways, but there’s one area in which it has trailed the competition to a frustrating degree. While it tends to beat its rivals in terms of camera support, processing performance and in the sheer breadth of tools and adjustments on offer, when it comes to noise reduction processing Adobe was long ago bested by alternatives like DxO’s PRIME / DeepPRIME and Topaz Labs’ Denoise AI engines.

And when we say ‘long ago,’ that’s no exaggeration. DxO’s PRIME engine debuted a decade ago with Optics Pro 9, and Topaz Denoise has been around even longer. We must confess to having been more than a little surprised by just how long Adobe allowed that situation to continue, but last month it finally offered a rebuttal, bringing its noise processing back from the prehistoric era.

The new Adobe Denoise engine made its debut not just in Lightroom Classic v12.3, but also in Lightroom CC v6.3 and Photoshop’s Camera Raw v15.3, providing vastly more capable denoising to subscribers of any of these three apps. In this article, we’re going to take a look at how it compares on some out-of-camera JPEGs to give a sense of the scope for improvement on offer.

Whether you’re using Camera Raw or Lightroom, Denoise is accessed in one of two ways: Either with the Enhance option in the right-click menu, or through a dedicated Denoise button.

An extra step in your processing workflow

Regardless of whether you’re using Camera Raw, Lightroom CC, or Lightroom Classic, the basic procedure for using the new Denoise algorithm is much the same. You select the image(s) you wish to denoise – batch processing is possible – and then click the Denoise button, which you’ll find in the Detail section alongside the sharpening and still-existing noise reduction sliders. (Alternatively, you can right-click on the images or their thumbnails and select ‘Enhance.’)

A dialog box will open in which you’re shown a roughly 700×700-pixel crop from your image and are given a single slider on which to choose the percentage strength of the denoising effect to apply. The preview crop, which you can reposition anywhere you like within the image frame, will be updated to reflect your chosen strength level after a few seconds. When you’ve dialed in the strength you like, click ‘Enhance,’ and Adobe Denoise will do its thing – while you grab yourself a coffee or three.

The only user control over Denoise is a single strength slider for each batch of images processed. Note that using Denoise precludes your also using the Enhance Detail or Super Resolution tools.

Processing takes a while, to say the least

Suffice it to say that this step can take quite a long time. A batch of just ten Raw images, totaling just 407MB and ranging from 20 to 60-megapixel resolution, took over two hours to process on a mid-range Dell XPS-15 9570 laptop.

Once processing completes, Adobe will then spit out the result as a new processed DNG Raw file, much as already happens with its Super Resolution tool.

While it’s good news that the process is compatible with batch processing, using it in a batch does require that you use the same noise reduction level for all the images in that batch. And of course whether used singly or in batches, it also inserts another (long) step in your workflow.

It’s also worth noting that Adobe Denoise can’t be combined with the Raw Details or Super Resolution filters with which it shares a processing dialog. Not only can’t you choose more than one effect at a time in the dialog box, you also cannot run Raw Details or Super Resolution later on the new de-noised DNG, or vice versa. The ‘Enhance’ effects represent an either/or proposition.

A very worthwhile improvement over in-camera JPEGs

With all of that out of the way, let’s look at some results from Adobe Denoise on out-of-camera JPEGs to see how the algorithm performs. The crops below all come from the three images that are shown above in their full-size out-of-camera JPEG versions.

All examples here were made at the default 50% strength setting, so there’s scope to further increase or decrease the effect to your tastes. But as we’ll see, even at the default there’s a big improvement to be had.

Test 1: Dog (Panasonic GH5 II, ISO 8000)

Out-of-camera JPEG	Adobe Denoise

We’ll kick things off with an ISO 8000 shot from our Panasonic Lumix GH5 II review. Straight away it’s pretty clear that Adobe Denoise has made a significant improvement over the out-of-camera JPEG. The denoised version is much crisper, with the improvement in detail especially noticeable in the finer hairs on the left which are pretty much textureless mush in the camera’s version.

Also worth noting are the dog’s eyeball as well as the surrounding area. Here, the color has been restored where it’s largely lost in the camera’s version, and the image feels more lifelike as a result. The improved resolution can also be noticed in the reflections, especially in the blown-out highlight at upper right.

Out-of-camera JPEG	Adobe Denoise

It’s also interesting to note the changes in defocused areas of the image. A huge amount of detail has been restored to the dog’s fur again here. Even though it remains well out of focus, individual hairs can be made out that were, once more, totally lost in the out-of-camera version.

With that said, the bokeh does look a little busier in parts of the image as a result, and depending on your artistic goals this might not entirely desirable. Note the stray hair near the bottom right corner of the crop, which looks a bit haloed after the noise reduction.

Overall, though, the Denoised image looks noticeably better across the board, and while some very fine noise remains, it has a much tighter, less clumpy pattern to it which is quite pleasing to the eye.

Test 2: Portrait (Sony a7 IV, ISO 12,800)

Out-of-camera JPEG	Adobe Denoise

Ramping up the sensitivity to ISO 12,800, let’s take a look at a shot from our Sony a7 IV review. Again, there’s significantly more detail to be found in the Adobe Denoised version when compared to the out-of-camera JPEG.

The folds in the eyelid look especially improved from the rather blotchy-looking original, and we can also see significantly more detail in the finer eyebrow and eyelash hairs, as well as the pores of the subject’s skin.

Note also the improved saturation throughout, as well as the crisper lines in the glasses frames, which have also picked up some very fine scratches. (Or perhaps it’s a texture in the frames themselves.)

Out-of-camera JPEG	Adobe Denoise

We can also see a less-attractive side effect where the algorithm reaches the limits of its capabilities, though. Here, it does a very impressive job of restoring the fine thread details in the fabric, but at certain points this detail falls off very suddenly. That leaves some areas that are now more conspicuously devoid of detail.

This is where we’d like to see the algorithm’s intelligence kick in to recognize the thread pattern and attempt to recreate some detail which clearly just doesn’t exist in the original Raw image. As yet, it seems unable to do so.

But with all of that said, it has nicely restored the shading from the shadow, which is quite blotchy in the original. When not pixel peeping as we are here, the fabric still looks fairly natural.

Test 3: Bird (Canon EOS R8, ISO 16,000)

Out-of-camera JPEG	Adobe Denoise

Our final comparison is, perhaps, the most impressive of the bunch. Here we’re looking at a shot of a female cardinal from our in-progress Canon EOS R8 review. Seen up close, the bird’s feathers are largely devoid of detail in the out-of-camera JPEG, and the eyes and beak are quite soft too.

Seen out-of-context in this crop, you could believe the bird was slightly out of focus. In fact the shot was crisply focused in just the right spot, which becomes clear when Adobe Denoise brings vastly more of its detail to life. The colors also feel more natural, with the beak in particular a more realistic tone than out-of-camera, where it actually felt a little oversaturated.

Out-of-camera JPEG	Adobe Denoise

The improvement in the foliage detail is no less noticeable, with detail in the fine hairs on the leaves restored that was mostly lost in the original. There’s also noticeably more contrast in the fine texture of the branch in the foreground.

But again, notice how the two bokeh balls in front of the branch – perhaps some fine rain droplets – have become harder-edged as well. This has made them a little more distracting as a result.

In this shot they’re still not very noticeable, but if you’re trying to achieve really smooth bokeh you might want to either avoid using Adobe Denoise or layer the image with its original so you can paint the effect back out in the defocused areas.

Conclusion

Overall, we’re pretty impressed with the job Adobe Denoise has done here. Of course, we’ve yet to directly compare it with its rivals, something we’re certainly intending to do in the near future. Subjectively, though, we’d say that Adobe users may now be able to forego third-party tools like DxO PureRAW or PhotoLab and Topaz Denoise AI in favor of the first-party solution.

That said, while those rival tools are already quite intensive in terms of processor power and rendering time, Adobe Denoise feels decidedly more so. (That’s something else we’ll test in detail in our coming comparison article, of course.) It does feel like there’s some work to be done on Adobe’s part to optimize its algorithms, and until it does so you’ll likely want to be judicious in your use of Denoise, or ready to wait a fair while to see your results.

We’re very happy indeed to see this new tool unleashed for Photoshop and Lightroom users alike.

But with that complaint aside, we’re very happy indeed to see this new tool unleashed for Photoshop and Lightroom users alike. If you’re already a subscriber, you’ve gained access to them at no extra cost, and they have the potential to bring significant improvements to your high ISO shots.

And speaking as a well-seasoned reviewer, it’s nice not to have to keep calling out Adobe in this one area, as we’ve been doing for quite some years. It seems the company is now finally in the same ballpark where its competitors have been playing for a long time. If our voice was heard we’re happy, and we’re sure many of Adobe’s low-light-loving customers are feeling a similar sense of relief right now!