This award is granted when one of your answers gets 10 likes. Your response really made an impression on the forum and helped move the conversation forward!
You can earn this multiple times
Stenis[Sten-Åke Sändh (Sony, Win 11, PL 6, CO 16, PM Plus 6, XnView)](https://forum.dxo .com/u/Stenis)
(You can click here to see what I wrote. I take that as a proof of that many people would see some kind of effective AI-driven masking tools even in Photolab)
While we are on the subject of masking, seems like a substantial amount of improvements and features for a point release coming in June for ON1 2024.5 In the past I found PL7 noise removal was still better for my Sony camera and lenses for astrophotography
They really have done a lot of work over at ON1 of late. It is one of those Apps that really should be all that is needed but no matter how much I try I cannot get to love the workflow. And my Fuji X images do not look that great out of the box, not to me anyway.
Intelligent Masking which is the topic at hand does not mean that all implementations will have or need to have Adobe T&C’s. So far ON1 AI masking is doing a decent job for me, and that is without the upgrade coming later this month.
Ideally any AI type thing would be local only or with ability to use server option as well if one so chooses. For large models this often requires a lot of hard drive space and computational power, so for that reason it is preferable to have options to use network in some situations or set up local network of rendering machines for in house studio work. Adobe is up there with Open AI, Meta and others who are not into creative industry at all, but data mining and selling. on1 at least seems to be at the moment not in that category. Affinity sold it soul to Canova or whatever is the name of the company so did CaptureOne sold itself to some bankers with DEI mandates. On1 was never really top of the line in terms of quality, and not a mature ecosystem but they had lot of iterations fast and lot of options. Maybe they can fill in the space left behind if they don’t go full Adobe. never go full Adobe.
This remember me my computer always having some cpu activity when topaz products were opened (even when I was not using them).
I wrote a first message to support to ask what was happening.
They responded that they didn’t really know and some stuffes like that.
I wrote them again saying that I really do not want anybody, any process, anything to use MY computer without my consent and that I will dig into this problem.
And by magic, less than a week after that my topaz products did not use cpu when I was not using them, this without any update of the products.
I don’t realy know why was happening in background on my computer, whether it was intentional or not, but at least it was quickly sorted out.
I’d like to add to that, for all the photographers who use these adobe processes : don’t complain when you’re out of work, because agencies will replace you with AI softwares that in a matter of seconds will come up with more proposals than you, and at a ridiculously low price : it’s you who will have given them this opportunity. And those Ai softwares will use your work without any compensation. Ironic, no ?
(And it has already begun).
But this very worrying point has nothing to do with ai masking in photolab.
I think I saw quite a few times people complaining on Topaz forums about licensing subservice running in the background expecting user to be connected to internet to check for license. Because of buggy software, and various other circumstances like being away from internet connection would limit the usability of the purchased software, effectively making it unusable.
I don’t know if they changed that after backlash, but that could be one service running in the background, which I for one , would not appreciate. Its like all the dongle and other methods they tried to combat piracy, which only made people more likley to go for high seas because in some situations with some software, people bought the darn thing and than got bunch of other services they don’t need or want just to do the licensing. So they also downloaded the pirated version which has none of it. That is the level of clown world we see with these companies. I suspect in the case of Adobe, high seas will be full of people after their latest stunt. Which they did before, its just official now.
Remember when Adobe was forcing users to go to subscription and those that didn’t want to and had no need for new version of software, Adobe sent them threatening e-mail saying they could be sued if they don’t upgrade. lol Apparently Adobe was paying some kind of license for old software features, and with no new subscribers they saw it as dead weight, but instead of doing the correct thing, they did the Adobe thing. Its when the first wave of creative migrants went off Adobe.
Yes. Some people not only do not see it, but cheer the process of their own demise.
Technically it does not. . As long as DXO does not go full Adobe. Never go full Adobe.
The OP’s argument was AI is more productive. And I said, its relative term, not absolute. Today I wake up to this article about the rush to use AI so you don’t have to. What could possibly go wrong?
"The 1957 movie 12 Angry Men stars, as the title suggests, 12 men who are on a jury together and must decide the fate of a teenager accused of killing his father.
However, Amazon Freevee users scrolling through the platform’s offering may be forgiven for thinking there are more than 12 jury members as a movie poster for the Academy Award-nominated flick shows 19 characters.
Upon closer inspection it is clear that the 19 men aren’t even real; their faces are smudged, their hands have fingers missing, and the proportions are all wrong. Why? They are all AI-generated."
Using AI to create masks over the areas identified by the shadow clipping & highlight clipping to more easily rectify the exposures to the matching colours and white balance with sliders for fine adjustments. The masked areas would need to be able to be grouped or ungrouped for manual overrides
What I heard from DxO in a webminar (less than 1 year ago I think) about AI masking, is that they can’t talk about it, but if/when (?) they deliver this kind of tool, they want it to be way better than what adobe produce.
Adobes masking is superior to PL masking imho (as is Capture One’s). The AI masking works pretty well and when it does not there are tools available to correct it or do it manually. The fact is AI will never work on everything, hence the ability to intervene if necessary, but it works on so many things it is very useful most of the time.
When I tried it out in Lightroom I started with a photo that should, logically, be easy for the AI but, physically, difficult to mask manually. Namely, a helicopter against an almost uniform blue sky. Whether I told it to select the object or the sky, it missed some, making for a fiddly manual “correction”. Also, given there was only “the object” and “the sky” in the photo, I was somewhat surprised that the two options were not mutually exclusive.
Maybe it’s great for selecting a mountain or a tree, but if I can’t even have it select a plain blue sky, it’s not for me.
Here’s a helicopter against a plain blue sky, using a Control Line in PL to mask just the sky, then change the colour to demonstrate quality of masking…
Exactly. The photo I tried it with was of a helicopter with skids which meant there were multiple “enclosed” sections of sky which is what the AI had trouble with. They were, however, all very blue.
