Several years ago, my mind was blown by the amazing quality that could be produced by the Topaz products. There was simply nothing out that could even compete. Well, there was nothing else out there until Adobe decided to dip their toes into the ever larger lake called AI.
For this grand showdown, I chose an image that does not give either Topaz or Lightroom and advantage. It is an image that has fairly strong noise, but also has a sharp eye on the subject. The test itself is actually quite simple. Run both of them with default settings and see which produces the better image without changing any other settings or tweaking afterward. It will be judged on overall image quality, time to produce the new image and, of course, what does the noise look like?
About the Test Subject
Below, I have an image of a male eastern bluebird returning from the nest. This image has a sharp eye and face, but it's body is leaning toward the camera. This leaning produces and out of focus foreground. I was interested to see how it would handle the out of focus areas. Would it attempt to over-apply the noise control to this area? The background of this image is also quite noisy. Will they mask the image properly to avoid too much noise control on the subject?
Image Metadata:
Sony a7R III
Sony FE 200-600 mm at 306 mm
ISO: 3200
Lighting Conditions: Morning Sun
Shutter: 1/6400th of a second
Lightroom Version: 12.4
Topaz AI Version: 1.5
TEST 1: The Background Denoise
In this test, I focus on just the background. Will both denoise products knock out the noise in the background and does it look natural? The shot below is the image at 200%. Here we can see the noise very well and see what needs to be removed.
Running the Test and Results
Running both AI tools does take time. The AI engine has to analyze every pixel of your image and make choices on what needs to be done. However, Lightroom took so long that I actually went to the internet, surfed and got a coffee. This is a new feature, so I am going to give them a break on that.
Lightroom Results - the noise is reduced to a very high degree and gives the image a very nature noise "profile." (200% View)
Topaz AI Results - Topaz seems to go for very clean and crisp, rather than keep some of the noise. It did remove more noise. (200% View)
Test 2: Overall Image Quality
One of the issues with denoise solutions is that they are notorious for simply smoothing out the entire image. However, this is what AI is supposed to know how to handle, and why they have an edge over traditional processing techniques. Let's see how they do.
Lightroom Image Quality - A bit of softness in the subject's eye and face, but overall quite good. (100%)
Topaz Image Quality - The image is clearly sharper, but not by much.
Conclusion
This test was an apples to apples comparison of a moderately noisy image with a sharp subject. For me, the winner is Topaz by a somewhat narrow margin. The reason behind this win, is that Topaz has better noise to sharpness algorithms than Lightroom. The second reason is speed. Lightroom was unacceptable in this area. I am running no junk PC either. This has fast, 6 core Xeon processors (overclocked) , with a 6 gig Nvida graphics card (overclocked) and 32 gigs of RAM.
As long as we are on the topic of performance, the Topaz software uses your GPU to make it even faster. I noticed very little GPU usage from Lightroom.
The remaining noise (after correction) in Lightroom seemed more natural, while topaz is clearly going for clean and noise free. This is a subjective topic for sure. Which one do you like better?
For those who own Lightroom and you are doing this type of image work, then I think Lightroom will work with a little sharpening after the fact. If you want to step up to Topaz, you have a much stronger AI toolset than just denoise. The reality is, that Topaz combines all of these into Topaz Photo AI now. Meaning, Topaz it is faster and easier to use for professional photographers , who are batch processing images.