Quick Feedback on Photo AI from Giga/DeNoise/Sharpen AI user

In short: very disappointing compared to the individual apps.

AutoPilot results beyond horrendous – very plastic skin. Refining it manually took 10 times longer than it does with the individual apps, with far worse results, eg: artefacts such as striping in the noise reduction (on skin), blurred patches on textiles (that were sharp in the original; doesn’t understand the difference between weaving and digital voice); a strange reversal of effects on Preview when size set to ‘Fit’ (background came out very ‘plastic blur’ with the subject still very noisy*); and OH. SO. SLOW!

This is horrid guys. Not a patch on the separate apps. I wasn’t bothered by that until I read in one of the comments that all development is going into Photo AI rather than updating Giga/DeNoise/Sharpen. So now Now I’m worried. I’m all for fewer but better updates, but is Photo AI supposed to be the ‘upgrade’ for the other 3 apps that haven’t been getting upgrades? I’ll be very annoyed if this is all I see before my upgrade subscription expires in Oct - and I’ll think twice about paying $99 for the next year’s updates.

Topaz, please can you offer some reassurance on this situation?

  • Update: I’m a ridiculously slow, inaccurate typist. In the (quite some time) it took me to type this, the image preview quietly updated, so the subject now looks very plastic, and the other artefacts described remain. (When I started writing it had finished rendering (according to progress indicator) and this is not a slow computer, it’s v high spec incl. 64GB RAM).

I agree, I Purchased Photo AI because I too, also understood that the individual apps would no longer be updated. My initial test results are very disappointing as the results from photo AI are not as good as the noise and sharpen AI. I would have expected at this point, Photo AI results should have been equal to or better than the individual apps. Topaz, it is imperative that you bring Photo AI up to the current level of the individual apps.


I have reached out to both of you through email. I want to figure out how the results in Topaz Photo AI compare to the individual applications so we can improve the application to have similar or better results than the individual applications.

1 Like

I also have to agree. As posted here just a few minutes ago in another thread.
You also had a perfectly good UI in the Gigapixel products allowing mass-editing of models and values, I have no idea why you threw that away for Photo AI and replaced it with a basic film strip (which loads forever when faced with many images).

What I would have expected from Photo AI:

  • Based on the same UI as the existing (good) products, but being able to combine Sharpen/Enhance/Denoise for the first time there (so I don’t have to use 3 products every time)
  • And -just optionally- (if the user wants to) being able to use the autopilot to suggest settings, which one can correct/overwrite for all selected images at once (in the proven UI)
  • and some CLI functionality to make life easier

But what we actually got was:

  • Always forced Autopilot, many times the results are not as good as in the old single products.
    Also not happy with the default plastic face look many images get. I understand that Barbie is currently the film to watch, but that doesn’t mean, that I want my pictures to look like that :wink: I had to set face to 20% from the default 80% to get some sort of OK results…
  • Also no way to at least disable specific models (to force autopilot to so something right)
  • New UI with basic film strip, which doesn’t allow batch operations (applying models/values for more than one image) and is extremely slow when faced with lots of images (thousands)
  • Unsuable batch features, see here
    • An ‘apply to all’ feature, which is useless as it also copies the individually detected values for the first image over to all other images (and not just selected models)
    • A CLI which only does the same thing one can already do in the UI (run every image through autopilot)
  • A remaining Blurry Patches issue, which doesn’t seem to be addressable

I very much hope that future road maps address such issues. The last few months I can’t say that anything has been improved in that direction. All I read about is updated models, keyboard shortcuts, camera support and so on…
If that continues I will also not renew (as many others, from what I have read in various threads).

The product itself has great potential. But the way it is now (compared to the old products) that potential is -sadly- not being used.

P.S. The way Photo AI stands now, one wishes that Topaz Labs would at least continue to update the old products. But sadly it looks like it’s the end of the road for the old AI programs. Even if the new king in town is not yet ready to follow in the footsteps…


Hi Lingyu, thanks, great to hear from you, I’ve responded. TBH I’m amazed that you aren’t already aware of the differences though - it’s immediately obvious.

I (like many here it seems!) am full of praise for Giga/Sharpen/DeNoise. Hope Topaz hasn’t really dumped those excellent products for Photo AI(?) as many comments suggest…


@Jetband @donreynolds Thanks to both of you for sharing your insights. Looks like Lingyu has started email threads with you both! Hopefully that helps us continue improving the app for you!