Very Early beta access to Topaz Photo AI:Autopilot if you own DeNoise+Sharpen+Gigapixel

Got it. Thanks for the heads up.

It is a very long install process......

Initial (very quick) thoughts.

Even though the image I sent to Photo AI is one I just processed in Sharpen, AI's Auto feature only chose Noise. So, I clicked on Noise, Sharpen and Enhance. Left all other settings alone. I guess that was a mistake as it way over-sharpened. I've never seen output from Sharpen like that no matter what I did to the sliders.

I'm very interested in this product and will participate fully in the beta as I would love to have one product instead of three.

 Photo AI on the right
Photo AI on the right
 
Got it. Thanks for the heads up.

It is a very long install process......

Initial (very quick) thoughts.

Even though the image I sent to Photo AI is one I just processed in Sharpen, AI's Auto feature only chose Noise. So, I clicked on Noise, Sharpen and Enhance. Left all other settings alone. I guess that was a mistake as it way over-sharpened. I've never seen output from Sharpen like that no matter what I did to the sliders.

I'm very interested in this product and will participate fully in the beta as I would love to have one product instead of three.
I'm also trying it. I first used it in standalone mode and it took so long I decided I would never use it. But then I tried the PS plugin and it is much faster, probably no more than the normal Sharpen AI takes. The controls are hard to understand in either case. There is no "Enhance" in the plugin; maybe that just means to "do the processing"? I haven't used it enough to really know what to expect, but in a couple of images it did a better job of overall sharpening than I got with just Sharpen AI. So, yes, a promising start; will be interesting to see how it develops into a product.

****
 
Here's a good example of what it can do. First photo was sharpened in Sharpen AI using its recommended "Motion blur, normal". I also tried it with "Out of focus, normal" but it wasn't as good. The 2nd one is with Photo AI (Plugin version from PS) using Auto mode which suggested only Sharpening. Note the big difference in the dragonfly's tail. The image had been processed from raw using PhotoNinja which applied some pre-sharpening. This might account for some appearance of over-sharpening on the body and the stick he's holding on to. Lots of experimenting to do, but probably too early to worry too much since it will be changing during this intro period, I'm sure.

****

Topaz Sharpen AI
Topaz Sharpen AI

Topaz Photo AI
Topaz Photo AI
 
Last edited:
Downloaded and installed (as someone said, lengthy download even at 150mps). I thought I might as well since I'd picked up the other products on sale)

At first sight rather confusing (I don't know the other products well) but I'll continue to try a few things. Would be good to have all three as a one off final step when needed.
 
Of course, the goal should always be to make the perfect photo without mistakes from a technical point of view.

I have not always succeeded in this over the years. And not every technically unsuccessful photo can be retaken.

Thanks to Topaz AI Sharpen, I've been able to save photos that I otherwise would never have shared because I would have been ashamed of technical errors like shaking or blur.

Of course, these are not masterpieces, but not infrequently important personal memories that I can now share later thanks to Topaz.

I am very looking forward to seeing how Topaz Photo AI is developed further :-)
 
I think that this makes a lot of sense. I think that a shell that can decide how to sharpen, remove noise, and enlarge an image automatically is a great idea. I own all of TopazLabs add-ons for still images and I have always thought that AI could help a lot of their plugins. Today we get that for a lot of their technical adjustments. I would like to see some of that applied to their artistic functions too. Studio 2 AI anyone????
 
Just downloaded and installed. And have only tried it so far on one image. Quite impressed with what I see.

Here is a Lightroom compare with RAW on the left and the result on the right. Much sharper and it has exposed detail not seen in the original.

63a2887bd7fc414186f06254920a7235.jpg

The final result.

ae517de61810495ba70277bfa14061fc.jpg

And on my new PC with the RTX 3090 GPU, it runs fast.
 
If topaz aggressive on improving this, a few more builds, I sure this will make DxO sweats heavily.
Does it do lens corrections? Cause that's one of DxO's strengths. For M43 users, DxO's lens corrections are better than the inbuilt ones.
 
  • Like
Reactions: Lan
If topaz aggressive on improving this, a few more builds, I sure this will make DxO sweats heavily.
Does it do lens corrections? Cause that's one of DxO's strengths. For M43 users, DxO's lens corrections are better than the inbuilt ones.
And, of course, DxO's corrections are measured in the lab. That would need a huge amount of work for Topaz to replicate them.
 
If topaz aggressive on improving this, a few more builds, I sure this will make DxO sweats heavily.
It’s a completely different approach. DXO with Deep Prime is working at the RAW conversion stage and as a result can be quite subtle but also includes optical corrections which are hugely important to a good end result. Topaz is working on images that have already been converted, or started as OOC jpegs. And there’s no reason why the two can’t be used in a complementary workflow.
One of the issues I find especially with Sharpen AI is that if any sharpening has been applied at the initial RAW conversion (or OOC jpegs), there is a tendency to oversharpen, producing “crunchy” images, and also more likely to produce artefacts in out-of-focus or non-subject areas, such as grass, water etc, which then have to be eliminated by masking or using layers in Photoshop.
 
Last edited:
And, of course, DxO's corrections are measured in the lab. That would need a huge amount of work for Topaz to replicate them.
Some software packages use the lensfun database that has correction data on most lenses. ACDSee, On1, Affinity, Aftershot Pro, darktable, etc. use it. Much easier than trying to reinvent it at each company or org. Here is a list of software that uses it (may not be complete):

https://lensfun.github.io/usage/

 
Last edited:
And, of course, DxO's corrections are measured in the lab. That would need a huge amount of work for Topaz to replicate them.
Some software packages use the lensfun database that has correction data on most lenses. ACDSee, On1, Affinity, Aftershot Pro, darktable, etc. use it. Much easier than trying to reinvent it at each company or org. Here is a list of software that uses it (may not be complete):

https://lensfun.github.io/usage/

https://github.com/lensfun/lensfun/releases
Thanks. I didn't know about that :-(

Maybe I'll try the Affinity correction vs. DxO, just to see
 
And, of course, DxO's corrections are measured in the lab. That would need a huge amount of work for Topaz to replicate them.
Some software packages use the lensfun database that has correction data on most lenses. ACDSee, On1, Affinity, Aftershot Pro, darktable, etc. use it. Much easier than trying to reinvent it at each company or org. Here is a list of software that uses it (may not be complete):

https://lensfun.github.io/usage/

https://github.com/lensfun/lensfun/releases
Thanks. I didn't know about that :-(

Maybe I'll try the Affinity correction vs. DxO, just to see
It highly depends on how well the Lensfun profile was created and if a lens reports focus distance or not.

I have a lua script for my Canon EOS 6D that takes vignetting correction samples at almost every f-stop.
 
Last edited:

Keyboard shortcuts

Back
Top