And, if I may add a guess to your reply.
I suspect the user who asked why their free year ends in Jy and not Sept has a current annual upgrade plan that ends in Jy 2023… since the freebie is tied to & offered to those who own the 3 programs and an active upgrade plan.
Batch processing is a mess.
When the film strip opens, the first image selected is on the right. Since most of us (in Europe and the Americas anyway) read from left to right, this is counter-intuitive:
And here is an example of where Recover Faces hits the wrong target. Fortunately the two faces in focus did not really need FaceAI:
This morning I just downloaded Topaz Photo AI and gave it a spin. I have been taking digital photographs since the introduction of one of the first digital cameras, Sony’s Mavica that used floppy disks way back in 2004 that produced images of only 640X480. So, I thought I might as well try Photo AI on one of those images to give it the ultimate test for me. I own Gigapixel AI but have had really poor results trying to increase the resolution of such low res images. However, this morning I was really pleased to see Photo AI increase the resolution 4X and still produce an acceptable image. After watching the QuickStart video, I learned how to adjust the levels of noise reduction and sharpening to produce an even better image. Like some of the other users, I, too, think the sharpening tool is too strong in auto mode. After adjusting the strength manually, I was able to restore some of the fine details that were lost. I’ve attached one of my test images (640X480) and the resulting 4X image (2560X1920) with manually reduced sharpening to try to preserve the granulation (gold beading) around the butterfly pendant. (I don’t know why the image size is indicated in the filename as 850X637 since Windows 10 indicates the file size is 2560X1920 just like the image I produced with automode.
You were a bit late then. I got my first digital photo camera in 2002 and it was the very first 5 megapixel photo camera, the Nikon Coolpix 5000. At that time people used one, two or three megapixel digital cameras already. 
I use this software as a standalone program, it is a fantastic program, for me it would be greatly enhanced with a crop tool, it would then really be the 3 tools in one application. As it is, I have to first crop in another application before bringing it into Topaz Photo AI usually as a Tiff, would prefer to use Photo AI with a raw file.
I turned off noise reduction and kept the “recover faces” option turned on. The face in the “before” image has clear and distinct eyelashes. The face in the “after” has blurry eyelashes that appear to be cut short even. Sharpen defaulted to off, but turning it on did not bring back the eyelashes. Enhance resolution also did not help. It appears that the program aggressively decided her face needed to be recovered (I saw nothing wrong with it) and in the process made it worse.
i did test the software and in my case all pictures come off as if i had run trough a photoshop filter to make it look like an old blurry painting. not to mention the auto pilot set up insane settings like scale the image by 228x the original size making it look like a NASA outer space picture. and it was a panda image witch the software could not detect the panda face… so very dissapointed.
Your post might be more useful if you had uploaded the source image…
didnt do cause never process it. but will do.
original
Personally, I think it looks like a pair of underwear
Win 11 Pro desktop PC. PAI commercial rel. 1.0. Standalone vers. Testing Sony ARW raw (straight from camera - no prior manip.). Processor = Auto (I have AMD RX3800 XT & Intel i9 12th gen).
Autopilot (AP) initial run. Then added Face Detect & some sharpening manually.
Another example why manual, user-controlled brush masking needed. Plus, AP ‘saw’ moderate noise in image & it’s severe color noise.
This is the “subject” AP detected (ie, the entire image - corner to corner). Refine didn’t help - either on default or portrait.
AP did not detect the severe noise in the image & only reduced color noise minimally:
I manually turned on Face Detect. PAI didn’t recognize the child. Only the Mom. There was no way I could tell the program to fix the child’s face too… (b/c of the drastic diffs b/wn the cleaned up woman’s face - which looks great after the manual detect - and the quality of the rest of the image, I lowered the Strength on the face detect somewhat):
Saving to a DNG file in Capture One finally works really well
Thanks for the update
Ron L
I used the Sony Pro Mavica back in 1993 for student pictorial photos and other on-campus scenics, etc. Lots of floppy disks, yes. I used to shoot these student pix on B&W film and hand-print each of them…
It looks promising. When using it as a stand alone program I get good results. However when I go to choose to open a file from Windows photo (where I start all my photo rehab) it opens, processes the photo as it is suppose to, then when I go to move my wireless mouse the program closes instantly. Nothing is saved. Could this stability issue please be addressed please?
Hi
I currently have all three applications and my license expires 26th Nov 2022.
NB Upgraded progressively since much earlier versions GAI since v1.1.2 ; SAI since v1.4.2 ; DeNAI since v.2.0.0
They all work as well as needed on the (limited) occasions when I need them but by PC is now quite old.
It’s CPU is a 1st Gen Intel i5 760 Quad Core and has Instruction Set SSE 4.2 and my GPU is:-
Operating System: Windows 10 Version 2009
Graphics Hardware: NVIDIA GeForce GTX 1050 Ti/PCIe/SSE2
OpenGL Driver: 3.3.0 NVIDIA 516.59
CPU RAM: 16375 MB
Video RAM: 4096 MB
Preview Limit: 6132 Pixels
So my question?
As the applications are working well, will Photo AI work equally well even though my CPU does not support AVX ?
And if it does indeed run & work properly on my PC what might I notice with respect to the lack of AVX ‘support’?
TIA 
In this case I believe it didn’t upscale the two faces in front because face recovery is specifically designed for improving smaller/background faces - usually the other filters do a sufficient job on large/foreground faces as-is.
Topaz Photo AI dropped support for certain older hardware, such as CPUs that don’t support AVX. This was done for a variety of reasons, that roughly boil down to making installers smaller, the app easier to maintain and improve, and ensure those using the product have a solid baseline experience.
Unfortunately, if you don’t have a CPU that supports AVX then processing images will not work at all. Since we don’t expect most users to know whether or not their hardware supports AVX, we’re actively working on improving the messaging to the user as early as possible - ideally, before they’ve even installed the program.
Hi Anthony
Well, I am pleased that the 3 applications are running aok for me. 
But as I understand it the majority of your development will now(?) being going into Photo AI?
If that is the case can I clarify something with you.
On “My Products” on account it has always been the case that latest versions get listed as available up until the expiry of the license. With them being downloadable after the expiry date.
Now, as I understand I should be able to download Photo AI free of charge up until the license expiry (if there are ‘dot’ releases before then, they will appear as approriate)?
Therefore, should/when I upgrade my CPU (whole motherboard?) in due course I surmise I will at least be able install Photo AI 1.0.0 (.x) even next year or beyond?
Thanks for the free copy. I tried it on three images I was currently processing. I encountered no program errors, but for me, for now, the results were not as good as what I can get from Denoise/Sharpen/Giga.
Like someone else mentioned, I don’t usually use Standard or Normal settings when I use Denoise and Sharpen, so unless PAI has similar ranges of options available, I would probably keep using the separate apps.
Also, I don’t see a way in PAI to get my favorite view of results, with a single image view and an “Original” button to toggle back and forth between original and processed images. Am I missing something?
