Product Direction in 2024

I would now return to the general introductory topic of the “Roadmap.” What if Topaz also shared something about its ideas? Topaz Photo AI released two versions of 2.4.x, not met with universal enthusiasm according to the posts here. I too went back to TPAI 2.3.2. I’ve always wished I didn’t have to constantly look for other alternatives, the Topaz AI tools (even before the redirect to TPAI) looked promising (for my needs). I used Denoise AI and Sharpen AI, sometimes Gigapixel AI. The first two tools mentioned remained without further development. A pity. Yesterday I received an offer again for the upgraded Gigapixel v. 7.0.4, for $59 (I have the old version 6.3.3). But why buy it if I have no idea if it would end up as both individual DAI and SAI tools? For the money? The promised “master branch” of TPAI has taken a strange turn, it’s being patched, and what’s next? I’m definitely not going to pay such high amounts for the unpredictable development of TPAI. Sure it suits some, that’s fine, but not me, starting with the changed UI and ending with the bugs. I would expect improvements from development and updates. Does Topaz even have its own “Roadmap” or is it just trying to extract something from users willing to contribute here? Even the TPAI was pretty good, for my modest needs for sure. For example, I would like to be able to edit only parts of the photos (for example, the selected “subject,” adjust brightness and contrast for it in addition to sharpness, and blur its background, etc.). Instead, I watch holes being patched with other holes. It’s really not worth the money to me. I have to say again: It’s a pity, but what can I do – probably look for an alternative, without knowing more about the Topaz intentions (and reliability to maintain a chosen direction).

Looking forward very much to improved quality models. That is my #1 concern. It’s still not possible to upscale a somewhat pixelated 480p video to a crisp 1080p and not have everything looked smoothed over.

I use Both Daily too But HDR in Video would be nice too.HDR 10 Basically Has 2 Values.

1 Like

Can you please add the model to the export name?

NOW: …enhance
BETTER: …enhance-HQ (or standard or lowrez…)


I love the tenets you described here. I use Photo AI regularly and just upgraded to v3. Completely worth the price!

Hi, I too am frustrated by the competitions that have a total ban of the use of AI. My CC and others distinguish between use of AI for image improvement and image and/or object creation. The use of AI is still new. I am young enough to remember when some competitions banned digital images. That quickly changed. Use of stock or any object in an image that is not made by the maker has always been banned in most competitions. AI generated images are treated in the same manner. As a practical matter we are in a transition period, where software publishers must use the term “AI” to remain in business. Most image makers have little interest in competitions. They just want nice looking images. Makers in that class run the gamut from those who simply want an image showing the family at the local zoo, event photographers, portrait makers, product photographers, and/or fine artists.

We will just have to wait, hopefully not too long, until competition organizers figure out a fair solution, keeping in mind that there are some who like winning so much that they cheat at solitaire.

1 Like

Hi, I have stopped using Photo AI in 2022 for lack of quality of the adjustments. However, seeing the recent results and the focus of your roadmap, I have restarted and am already more happy with some results. Since I am mainly doing wildlife photography and tend to photograph close to the ground, I am frequently facing issues with heat haze. The camera shows the eyes in focus, the resulting picture shows the green focus point, however, the result is not tack.
So, my primarily wish would be a particular emphasis on improvement of the eyes and eye-area. I did try the latest Photo-AI version on such heat-haze issues, however, the resulting eyes, or in particular the light reflection in the upper part of the eyes appears rectangular and extending beyond the eye after sharpening and slightly denoising.

Would be great to see this further prefected :wink:

I am very interested in the natural photorealistic result of AI as opposed to fantastical. I don’t want to have it perceived immediately as an AI construct of my image.


Hi Eric, one frustration that I have as a professional is that on a big project, if running into a snag I can’t get someone from Topaz on the phone. Time is money and the time it takes going back and forth through email on a professional job is frustrating. For example, I spent two days on an artifact problem that occurred using Photo AI and we had to start from scratch 3 times.

You asked what things we’d like to see included: A big one that just came up would be a way to use AI to stretch a picture without incurring artifacts. For example, I have a seascape that is a 6 x 4 format and I needed to stretch it to 8 x 4. When using photoshop and then Photo AI to sharpen I got artifacts in the picture. It would be so nice to be able to stretch a picture and then sharpen all within Photo AI. Also, FYI, when I had to go back to a photo that was already upsized to noise reduce using a mask, the masking was incredibly slow and the circle to make the mask bounced all over the screen. This is on a Mac Studio Pro with the fastest processor and 64gb ram.

Great Idea - something I would be happy to do so you have my support with this suggestion.

still there should be a CHOICE. A switch between photo realistic and fantastical. Or a slider from 1-10 or something like that.

Does PhotoAI solve upscaling better than GigaPixel AI? This quote implies that it does.

To support this goal, please consider adding functionality similar to Luminar 4.x DeHaze. Here’s an example (original on left, processed on right; no other processing than DeHaze). There’s some color shift in this example, which I don’t like, but OTOH I have over-applied this (and could correct the shift). It is a useful feature and does in produce natural and photorealistic results. Look at the center - the near trees - for what the feature works best for. Of course, fog and haze can be photorealistic but what we sometimes want to do, I think, is create a photo that represents an better reality - a more striking or interesting one - than we see in the unprocessed photo.

The second shot shows the clarity this feature adds to the stream.


Please add basic functions of video and photo editing like “Flip”, “Levels”, and “Color Correction”

i.e. I must render an entire segment and load it up in Adobe Premier Pro just to flip, rotate, etc. Basic stuff.
I must be missing something basic because often I am unable to get a smooth output if I attempt to slow-mo a video with unusual frame rates, or incorrect frame rates. Could this be simplified? Perhaps a method to flag drop-frame to non-drop