After a full uninstall (along with supporting app files) and reinstall, I am still seeing the same behavior.
Here was my original post… Topaz Photo AI v0.10 - #56 by vince.rubiera
MacBook Pro (16-inch, 2019)
2.3 GHz 8-Core Intel Core i9
MacOS Monterey 12.5.1
Lightroom Classic version: 11.5
1 Like
please identify your system fully - make, model, platform, OS level
I am getting a bit confused recently. The vast majority of replies to Topaz Photo AI talk about Lightroom instead which is not only a totally different piece of software. Imagine I would talk about Topaz Photo AI won’t match my expectations to the pizza delivery I got a few minutes ago. 
A lot of folks use LR as their basic photo editor and catalogue software… Topaz installs a plugin to Lightroom that allows a photo that’s in Lightroom to be sent to Topaz’s products for specific editing and then sent back to Lightroom… Pai is NOT a photo editor in the strictest sense… it only performs certain limited specialized tasks… THAT’s why they talk about Lightroom…
2 Likes
Is there a plan to add an option to deactivate autopilot with X1 at the start of loading an image because it’s a waste of time when we use for upscale only and like I said earlier face refinement is too aggressive with large faces.
From the tests I made Photo AI only with enhanced X2/X4 natural/low res model does better outputs than Gigapixel AI alone and normally results should be the same, no ?
I would like to still better use Gigapixel AI standalone so would be great that Photo AI quality upscale could be the same in GAI.
Also the Photoshop plugin doesn’t work when trying to resize X2+++, and only works X1. It should appear on Photoshop like the GAI plugin in the automation menu and not the classic plugin menu where it’s limited to only Denoise/Sharpen.
New today: Ventura beta 7 on M1 MBP, Photo AI 0.10.1
Not new (or fixed) today: The graphics glitches 
(Enhance>Natural only)
Dear Topaz Developers,
it seems as if you do not want to give Photo AI any functions where the user can really determine which areas of the photo are to be optimised.
There is no other way I can explain the absolutely pointless implementation of “Refine”.
Everything should happen as automatically as possible, and that’s where you want to place Photo AI, right?
I understand this approach, but at the moment it doesn’t work, with the majority of my photos I get results with Photo AI that I don’t want at all:
- with Face Recovery, faces that should be deliberately blurred in the background of the image are recovered: please implement a function that allows you to selectively deselect faces, this is a must-have in my opinion.
- with Remove Noise: Please offer the selection from Refine also for Denoise
- And as already written: the automatisms in Refine are completely useless: please give us the brush from Sharpen AI so that we can determine the masks ourselves.
Thanks
Martin
6 Likes
Will the “apply settings to all” function ever come back?
3 Likes
This version seems to produce a red cast compared to previous versions and overwrites meta data (date taken) in the metadata, when used as a LR plug in.
1 Like
Also does the same in stand-alone
I’m also seeing the colour shift in Photo AI 0.10.1 when using it as a plugin from Lightroom
Before:
After:
I’m running the following system:
Windows 10
Ryzen 5900X
32GB Ram
RTX 3080ti
2 Likes
Hi,
Unfortunately having updated to v 0.10 I again am unable to get it working both as standalone and as a plug-in. Few seconds of spinning and then it quits. Issue reopened with the support team.
Grzegorz
I want to share my first picture, Before & After, processed in GigaPixel AI, I was talking about in my Morning’s long post. Face recovery & Noice reduction turned On, with 4x Enlargement and minor, unfinished, processing in newly purchased Luminar Neo.
Sorry, all next files is too big for upload, 12,000 px on the long side, I resized them in PhZ8
Here is another picture of my youngest daughter. Snapshot made with my last Film Camera Nikon N90 in 1994. GP with Face & Noice turn On. It’s still have some unpleasant shadow noise on the bottom of the neck, but is not much GP fault.
Another picture of my wife. Nikon D70, Photo AI 0.6 or 0.9, and only Face recovery ON.
Program removed noice completely, as you can see i a dark part of the image, and Face is well defined.
Last one from Nikon D300, processed in GP with Face & Noice On. Noice reduction made bottom neck wrinkles looking unpleasantly sharp.
All files was sent to GP / Photo AI using Capture One 21 Plug-In option. Right click, Edit with GP or Photo AI, in a second or two opens window, and processing starts.
Every time I am trying to catch sliding down Enlarge 2x, 4x, 6x buttons to press, and to turn off NR switch, when people in a picture, after give time to finish processing, review it or, if need it, readjust Strength slider, and click Apply, that return my photo back to C1 (GP). I already processed hundreds of images with great result.
As one of the reviewer mention, good idea to have option to turn Autopilot off, if you need to use resize option, and keep it until user decided to turn it On, with extra button “Start Processing”, for example, and especially important for program to remember Last Settings, like GP does, that can save time and efforts to repeat it every time.
Autopilot is a great feature, and I use it a lot, but until program can do 100% right on people faces, we need that Turn Off option, please.
Lately, I wait until autopilot finish processing, and after do adjustments, but program already apply NR, and by turning turning that OFF afterwords, might live some artifacts, maybe.
I think Photo AI slightly better even now, and as we all hope it will be better, faster on final release, and we all will enjoy to use it in our photo-editing workflow.
Photo AI, for me personally, earned Status "Program that I Can’t live without "
Memory management seems much better. Gone through my repertoire and no freezing. Tiffs are still only 8 bit, both stand alone and when sent from Capture One. Coming up as 16 bit from Affinity Photo. For what I do the image quality is stunning.
What filmstrip? I am unable to see any way to enable such a feature in PAi.
For what I understand and see, the process order is what the “image quality” suggests
Face Recovery/Denoise/Sharpen/Enhance
We should be able to choose the processes order. If we plan to upscale an image, Enhance should be the very first step as GAI already denoise/sharpen any input. This is why I always get bad results when denoise or/and sharpen are used combined with enhanced whenever autopilot suggests it or manually for further adjustments after enhancement.
I remember I have always asked in the past for this kind of all-in-one program (GAI-DAI-SAI combined),
but if we can’t choose the order we want when combining them it’s useless even more when it’s obviously the wrong order that has been implemented.
In almost all the cases the best results I used to have is with Gigapixel first then Denoise AI or/and Sharpen AI at the end.
Denoise AI and Sharpen AI work best with very high resolution inputs this is why low res or mid range images are necessary to be upscaled first to feed DAI/SAI with high quality input.
Even if we don’t want to upscale, resizing down to the original size at the end will produce better results than doing everything at X1
So in conclusion the better workflow should be the ability to move up/move down any process and at the very end an option to downsize the final image (bicubic, lanczos…)
2 Likes
I’ve been having a similar issue. Did you receive any reply from Topaz as to what is causing this and how to fix it?
I’ve been having a similar issue. Did you receive any reply from Topaz as to what is causing this and how to fix it?
I’ve been having a similar issue. Did you receive any reply from Topaz as to what is causing this and how to fix it?
Open more than one image.
