This week we have made a few small improvements to the Standard Enhance model. This should help fix various issues like occasional blurriness in bright regions, make it more robust to JPEG compression, and have more natural faces. Alongside this we have made some other improvements around the app including increasing thumbnail generation performance, updating close warning dialog logic, updated tooltips on thumbnails, and standardized some menu items, text, and buttons.
On the Windows side, we’ve also introduced patch updates. This should make updates from 1.3.0 to subsequent releases much faster (~90% reduction in download size). These updates will only work from versions that are compatible (1.3.0 for now) and should happen automatically if available when you use the in-app update flow. You can still download the full installer from the downloads page or the thread if needed.
A full change log including other fixes can be found below.
Please give us any feedback or report issues with this release. We’ll be updating TPAI regularly to address those pieces of feedback and issue reports. If there’s a specific image you’d like us to see, you can send it us at this dropbox link.
We’ve recently added the ability to post RAW files under 65MB in the forum. If you are okay with your image being posted in this thread publicly you are free to do so. You are also free to go through our dropbox if you do not want to post publicly.
Improved Standard Enhance model
– Fixed blurriness that can occasionally happen in bright highlighted regions
– More robust to strong JPEG compression
– More natural faces
Added preference for enabling legacy Sharpen models
– Can be found under General tab in Preferences
Improved thumbnail hover tooltip
Changed logic for showing close image warning
Changed “Technical Support” Help menu item to “Support”, changed to link to our new Docs platform
Standardized done button when refining mask, selecting faces, or cropping
Added error message when dragging and dropping unsupported files
Fixed some large images not generating thumbnails
Fixed trackpad scrolling behavior in the update dialog
Fixed inconsistent capitalization of “RAW”
Fixed BMP not being importable from open dialog on Windows
Fixed blank space in export dialog caused by scrolling while exporting
Known issues:
HDR or Monochrome DNGs which previously did not open may have issues when saving as a DNG
– To get around this you can save as JPEG, PNG, or TIFF which shouldn’t have the same color issues
Launched 1.3.1 Ps (2023) Plugin with no issues - used a tree then an Egret image (unprocessed .tiff I sent to Cameron via Dropbox the other day).
The tree image looked pretty natural to me - retained its photographic quality, but
sharpened up. It didn’t look “snowy” or over-processed by Sharpen, to me.
For the Egret image, I Got same issue as with 1.3.0 - really odd. The processed A.P. preview looks much worse than the unprocessed view in Split View. Either serious noise or weird overall noise-looking artifacts being added in the processing. A.P. only turned on Sharpen.
I turned on Denoise manually. That got rid of all the noise-looking “stuff” (politicians aren’t the only ones who can use quote marks …).
Then I experimented by manually trying the various Sharpen sliders (current & legacy) at various amounts. To me, the egret definitely looked better using the Motion Blur slider set. And, this time (unlike the other day) when I adjusted the sharpening in PAI it didn’t seem to bring back that awful noise that appeared after the A.P. initial run.
What I did see was that the ground under the strolling bird looked like crap (and I don’t mean from the bird), b/c it was over-processed (or the wrong model to process it). If I wasn’t caught up in my Summit right now I’d have masked just the bird. I presume that would have handled that. Then, after saving back to Ps I could have re-launched back in to PAI and done the inverse (if only I had an Apply button & an inverse Mask option ).
It seems to vary from release to release (and maybe image to image - I haven’t made a formal study of it), but in this 1.3.1 rel. I felt there was a much less dramatic pixel hop back on the Ps layer for the returned, processed image when the eye icon was toggled off/on.
Thank ou for all the updates, you peeps are doing well trying to keep this working for everyone and Adobe should take a learning to this kind of customer support!
I really love the facial recovery! It makes old, soft images look like they were just shot today (below: Detail crop of a group from a medium format negative scan):
•I do wonder why some obvious faces in groups are not automatically activated even after having been chosen by PAI (reported for 1.3.0 but I saw it again in 1.3.1 with the guy in front here).
•The face recovery area is looking square again, sometimes (not always) cutting across hair. Can this mask be set to always be more oval in shape, since that’s the way heads usually are?
•Even with enhancement on, non-facial details don’t seem to be improving such as we had with Gigapixel (which was magical in its capabilities). I did try PAI without facial recovery on and wasn’t seeing Gigapixel-quality detail recovery on anything, at least on the settings shown:
I did go and try GP 6.3.3 and wasn’t seeing it there either today. But in the past I was always blogging about how miraculous the recovered details were.
An example of 1.2.9 cleaning up non-facial details shirts, background) along with faces, same type of film source:
I really don’t understand why Topaz fires out updates excessively fast in the last Time, instead of intensively testing a beta first before releasing anything…
“Added preference for enabling legacy Sharpen models”:
Could someone please confirm, that the latest release will bring back the old sharpening model from version 1.2.10 (and before) including the options “Lens Blur” and “Motion Blur”? I want to avoid another rollback to a previous version. Thanks.
The new update is really great. The problem I had when saving my image to the back of Lightroom was not working in 1.3.0. It works again here with 1.3.1. The image quality with upscalle seems even better than before. And even the sharpening tool with the right settings is really superior to the old one. I haven’t tested the lens blur and motion blur yet because I don’t have a picture to use them.
The quality of the details is amazing . Scaling x2 plus use of standard sharpen.
But the DNG RAW image output from my Galaxy S7 has already been processed for noise/grain with the new AI-based tool added on the latest version of Lightroom Classic.
Yes, they are there. Adam mentioned them in the Changelog. And, if you look at my screen snips with the image of the egret you’ll see that they are there in the righthand Settings panel. Activate them in the Prefs per Adam’s note.
Does not save preferences, i.e., crop ratio, close image after saving.
EDIT: My error. The program does close image after saving, but it does not close the processing window once it finishes processing one must click to close that window- a useless step. I would suppose that the desitnation directory is already open, out of habit.
It would be nice for the Aspect Ratio to be in the changeable in the Preferences section.
You are absolutely right!
In direct Comparison I see PhotoAI 1.3.1/GigaPixel 6.3.3 are loosing some Details compared to the older Version GigaPixel 6.3.0
Another Bug, which has been recently introduced: The Filter Status Toggle is often not correct, filter is ON, but shows it would be switched off and vice versa…
Unfortunately I can’t express myself in English as well as I would like to!
And another Topic:
The Sharpening-Filter is really bad, it looks in most Cases like “Watercolor”
Ist’s not looking natural anymore… A big Step backwards!
Trying to understand why this app is making faces WORSE with face recovery on vs off. This singular app is just so inferior to the 3 stand alone apps, I don’t understand why this direction is happening. Great to 1 app to do it all, but each of the modules in this 1 app are TERRIBLE compared to the previous 3.
btw, this is the default “autopilot” setting that came up with the sliders/face recovery. I did not place it like that at 80% or turn it on myself. I DID turn it off because it’s awful
Basically one has to shut off pretty much every option here. recover faces, Remove Noise, and Sharpening to get any decent output. The only thing that actually works is the upscaling/Enhance resolution. The other settings are totally useless in pretty much every scenario I’ve tried on various images.
Noise removal? It actually seems to introduce noise where there was none before, or at least none with the setting shut off.
Again, really not clear the purpose of this, and the other settings as they simply do not work. Resolution enhance? def works great… everything else? Total fail to this point