Topaz Photo 1.5: Can Denoise Max Finally Beat DxO DeepPRIME?
True comparison not from someone trying to plug DxO for 15%
That video you posted does not mention Denoise Max and in fact the video was posted 3 months ago, long before Photo 1.5 with Denoise Max was released.
That’s true I posted it because, it gave a fairer and more balanced comparison between DxO and Topaz and I noticed within the comments for the first video somebody pointed out they only picked the highest subscription tier
Thank you for the explanation. The point of the video I posted was Photo 1.5 and Denoise Max though. I thought maybe you had mistakenly posted the wrong video since it was sort of off topic. If you have something that compares Denoise Max in Photo 1.5 then that would be interesting and on topic.
Here is a new more general review of 1.5.
ROTFLOL…
He’s very entertaining!
For this latest one, we do see instances in his video of incorrect workflows and enhancements being used in ways they were not intended, both in this video and some past ones for this specific user. The blue color shift shown in the video is also not expected behavior, and if the creator reaches out to support we can absolutely help troubleshoot that further.
That said, it is always great to see reviews and comparisons shared, so thanks for posting both threads. It helps the development team better understand the different user cases out there, what users are comparing against, and where expectations may differ from intended workflows.
Thanks again @Tostanica-fezanamu and @AND-E - If you do get any unwanted results with Denoise Max, send in your original and results to support@topazlabs.com and we can share to the development team directly!
1 Like
Topaz probably as a matter of good PR should contact him to let him know his mistakes. If he learns how to use the software properly he may create a new video that is much more positive.
It was already shared internally with the marketing team a while back because of some earlier videos as well that had incorrect information shared and incorrect workflow used.
To be fair though, YouTube creators are also naturally looking for engagement and views, and more dramatic comparisons or reactions tend to perform better on those platforms. At the same time, we still appreciate seeing the feedback, even when critical, because it helps highlight where users may be running into workflow misunderstandings, unexpected behavior, or differences in expectations.
If the creator ever wants to reach out directly with the original files and workflow details, we would absolutely be open to helping test the cases properly and explaining what enhancements would be best suited for them.
3 Likes
This video you posted (it is long, but I finished watching the whole thing) focuses on bird photography. That seems to be common with many of these types of AI denoise comparison videos. I want to see how it works with people, faces, etc. That is where there are often so many problems.
Yes I agree apart from Scott being a Bird Photographer I believe the reason for demonstrating on Birds is because, they can zoom in and look at the fine feather details and the Background for smooth Noise Reduction at the same time which is a good indicator if the Applications Denoise is failing in its job but, yes I totally agree there should be more comparisons with people.
Actually, most of the videos I see from various people so often use birds as the subject for comparing the denoise apps. I guess bird photographers use high ISO a lot and therefore the companies do a lot of their model training on birds. In my case it is people though. AI has more trouble with getting people right, particularly faces, teeth, hands, etc. An app that does well with birds may or may not do well with people.
All so, true, have you had a chance to try the new Face Recovery v3 which is a big improvement over version 2 because the previous version was limited to 512Ă—512 resolution meaning your High-Res images were Downscaled to that resolution first before Face Recovery was applied then upscaled back to the original size again the new version is designed for Higher resolution faces.
No, I haven’t tried it yet. I just got it installed today. I will have to find a photo that is a good candidate for trying it out.
Here is a new one.
Google Gemini Just Destroyed Topaz Image Web? DENOISE MAX TEST
Hi.
Did you read the comments the majority stated alternative views and I believe there’s more to this review than meets the eye meaning untold facts we’re not seeing here’s one examples of an image shot at a stupidly high ISO with Denoise MAX applied
Small Crop from an image shot at ISO 204,800 that’s Two Hundred and Four Thousand Eight Hundred ISO
After using Denoise MAX even the figure is giving the OK sign of approval 
Please highlight and zoom in for true comparison
As you can clearly see If a Picture Paints Two Hundred Thousand ISO Words then, this one is clearly Painting a different story
Here’s a video not a comparison just how far we come in the last few months
Did ask Gemini about it.
"Unfortunately, I cannot process or upscale image files of this size right here in the chat. If you upload an image with a side length of 12,000 pixels, the system will automatically downscale it due to technical limitations, causing fine details to be lost. I also cannot natively export or edit files at this resolution.
For image material of this magnitude (which is quickly reached with high-resolution camera systems or panorama stitches), … "
Gemini is an tool for people with Smartphones.
In almost doing photography for 20 years.
Customers did never come for “Art”, they did always come for making normal images about them.
Doing something right is more challenging than doing something artifical.
Generative denoising do wonders when it works and stays true to the source.
There is a lot of progress to be made here.
To me, the Nano Banana 2 issues are two. It generates “low” resolution files, and looks almost too perfect/too enhanced. The example with the face of the boy looks over the top, like it was painted on the image, if you see what I mean.
