@Joanna, which version of Topaz are you using?
Mark
The free trial download of the standalone product. I have already bought their sharing and Gigapixel apps.
The video claiming to show its superiority over DeepPRIME is smoke and mirrors, and is dishonest. The only people who will be fooled by it are Topaz fanboys or inexperienced users who don’t see or understand the advantage of fine detail retention and are just happy with a smooth image devoid of noise.
Mark.
4 Likes
I’ve tested Topaz denoising and found that it suppresses detail much more effectively than DeepPRIME.
2 Likes
You make that sound like it’s a good thing. Why would you prefer a package that suppresses detail?
If you meant the opposite, that it retains detail better than PhotoLab, then respectfully, I would have to strongly disagree.
Mark
1 Like
Ah, but Mark is 'Mercun - they don’t do satire 
1 Like
Actually I considered that, but wasn’t certain whether it was sarcasm or a language issue because @platypus is not a native English speaker.
Mark
In our chats, he certainly speaks (and understands) English (including sarcasm) a lot better than I can speak German. I have never been insulted so eloquently
2 Likes
well folks, we do strive for the best, but the hunt for the best (add your poison here) leads to us running after something that will be surpassed by another product doing (add your poison here) even better…within weeks or months.
2 Likes
Hmmm. You’ve given me a goal to strive for! 
Mark
I think in general it is always good to have some competition in the field of state-of-the-art noise reduction algorithms.
While I certainly agree with @mwsilvers and @Shill_Hunter that NR alone without detail retention is easy but useless, Topaz has made some progress in their new Photo AI product. This improvement has not trickled down yet to the standalone Denoise AI but I’m sure there will be an update soon. I keep an updated copy on hand for use with my scanned 35mm negative files which are not RAW. It has always worked nicely on TIFFs, but not so much on RAWs.
Here is a direct comparison of the results from the original RAW(no NR applied), Deep Prime, latest version of Denoise AI, and Photo AI clockwise from the upper left. All NR versions were at their default settings, viewed in FastStone IV at 500% magnification. The photo is @mikemyers version of the Miami Fl. nighttime skyline shot at 12,800 ISO.
Clearly the DNAI version is not as clean as the new PAI version and while DP is a little bit noisier than the PAI version, neither of the two Topaz versions come anywhere close as far as detail retention is concerned. The Topaz products also wash out the colors and contrast to further lower the noise, but it is also clear that Topaz has made some progress in catching up with DP. They’ve gotten the noise down to a very low level and I’m sure that they will next work on the detail retention. IMO DxO needs to spend some time improving DP if they want to stay in the number 1 position.
2 Likes
I completely agree. However, the problem that most of us have with this particular review is that the reviewer is biased in favor of Topaz because he is making money off of the sale of that software and has decided to overlook the loss of fine detail when comparing it to DeepPRIME. This is both misleading and dishonest.
Mark
4 Likes
I have no problem if Topaz improves upon their product and even exceeds what DeepPRIME can do. Competition is necessary for the improvement of products. It is the fact that the reviewer was disingenuous when describing the superiority of Topaz over DeepPRIME that I found most disturbing.
Mark
2 Likes
Mark, I agree completely with you. That so-called “reviewer” is just trying to make money by selling the product. All I was doing was showing the comparison between the NR versions. Yes competition is very good, I was just trying to get DxO fired up about working on improving Deep Prime, in order to stay number 1. It’s not really important, but I like using the number 1 product.
2 Likes
I fully agree that the comparison in the video seems to be biased towards Topaz. When we follow the argumentation in the video and push it to the extreme, the best noise reduction algorithm would be to take the average of the whole image. There is zero noise left (but unfortunately also zero details) 
2 Likes
Not least of which, for me, is the library module. But at least that can be used for free, thank goodness.
Does it really matter what it looks like on the screen? Monitor qualities are so subjective. What really matters is what it looks like in print. YES, I said it print! So you create all the videos you want and there is plenty of editing tools out there but it comes down to is very simple.
- Does the tools fit into your workflow?
- What does it look like on whatever medium you print it on.
This is 2022. Probably far less than one is a thousand photographs actually ever gets printed these days. As a result how an image looks on a monitor matters very much to most people. It certainly matters to me.
Mark
4 Likes