I tested it with a variety of photographs. The worse the photo, the better you'll like it. I took some really bad low resolution photos directly from the web just to test it and was amazed how much they were improved just by batch processing. If you have tons of decades old photos that you plan on digitizing, It would be worth it just for that.
Great in depth video and tutorial. I’ve been looking for a competent versatile program to use for my media, seeing your vid just made my verdict of choice easy. Thanks
Good overview of the product. I actually think I'll probably hold off another year for any "AI' software like this. I currently have the Topaz sharpening plug in which does just fine for my needs. I suspect that another year of development will bring far better software, so I'll save my money for the time being.
Thanks for the overview. After looking at the results of the various "filters" I really don't see this software being worth $200. Even though I am still working with an old version of Photoshop (CS6) and using three old Topaz filters (actually "antique" in computer years); DeNoise, Clarity and very occasionally, Simplfy I do not see anything that I would call "amazing" with this Topaz AI. I can do pretty much the same things using this old version of PS. There's a lot more I could say but will refrain.
I tried a few times and the processing time was so long. It took multiple minutes to even remove one small thing, and I’m using a fully loaded 16 inch MacBook Pro that is 1 year old. It’s just awful.
I was mainly interested in the image remove function. I ended up subscribing to Canva Pro ($15 monthly) that does this in seconds, under a tab called "image grab" and another called "image erase." Hope this helps. This was a great vid as well, so thank for posting.
It’s for one year, but you do own the software. Updates are included for that year, but after that you’ll have to pay to keep getting updates. You can still use it once your subscription ends.
Throwing in my experience so far with Photo AI 3 and my 40+ year old scanned photos, both b&w and color slide. I have Sharpen AI and Denoise AI already, just bought the Photo AI. I am finding that the Face Recovery from Photo AI is fantastic on these old pictures scanned both from prints and from neg or slide. You just won't want to spend the time picking out faces in the other two softwares to obtain the same result, even if it's possible. NB, you can deselect faces that Face Recovery finds if you want. Sometimes it's too extreme and the face is way too out of focus in the background to attempt to resuscitate. So it's nice to be able to pick and choose which faces to let the AI work on. I am finding that sometimes I need to use Denoise or Sharpen after using Photo AI in order to smooth out the whole image, OR to pick some parts that need Denoising/Sharpening using the brush. One criticism of Photo AI 3 is that the brush is incredibly slow in comparison to using a brush in Denoise or Sharpen. Don't know why it is so damn sticky yet! But its stickiness is very off-putting, hence using all three softwares on these old photos to obtain an acceptably sharp and denoised image. I am VERY grateful to Topaz Labs for these softwares when it comes to historic images!
The face of the man lost a lot of definition. I think the original looked much better even with the noise. I tried the free demo and thought it did a horrible job.
I agree. The man lost definition and Photo AI turned his skin waxy. The second example, with the landscape, was unsalvageable, at least with this software. The Photo AI took the photo from "blurry" to "dizzying."
Sounds like not worth $200 to upgrade for the year, since Lightroom’s noise reduction and sharpening is improving (and it only costs me $120 for the year for LR and PS) there is less need to have both
TIme to process 1 image is ultra fast. But good luck batching 20+ images. The more you batch the slower it gets (from 1s. for an image to more than 1 minute for an image). Topaz WAS good. Now it's shit. (EDIT: latest version Topaz Photo AI 3.0.2 fixed that issue and one can now batch process any amount of image without slow down, aka GPU memory is, at last ! properly refreshed)