If AI can create more sophisticated file renaming options during image export, then I’m all for AI being added to Photolab
That’s the interesting point: should the programmers come up with one or more solutions in advance, or do I specify my needs to an AI that interactively implements exactly my requirements.
What do you think about it?
#17 AI reflection removal tool
Is it a desirable AI tool for PhotoLab? Perhaps the end of polarising filters .
Ai can’t know what is under water or glass surface if not recorded by sensor.
Maybe you’ll can choose what underwater background you want ?
Harari was asked:
And then the big political and economic question of the 21st century will be what do we need humans for, or at least what do we need so many humans for?
Harari’s answer:
At present the best guess we have is keep them happy with drugs and computer games.
Not much constructive imagination there. The despair of egotistical atheism. It always ends in hedonist materialism. Ultimately, Harari is a nihilist thinker. The rest of us have a world and (other people’s) lives to improve.
AI in PhotoLab does not have to re-imagine my photos. There will be other software better suited for imaginary photos. PhotoLab’s role is to make the best possible image out of our RAW files. Therefore features like:
- AI masking (including green screen)
- AI generative fill (filling in the those black edges after an extreme horizon rotation)
- smooth AI HDR merges
- AI focus depth merges
- AI skin improvement
- AI enhanced resolution
are what I’d like to see. On the other hand:
- Unicorns
- Tidal waves
- Facial reconstruction
- Metamorphosis
- Magic mountains
are areas better left to Midjourney and similar applications. One would take a PhotoLab finished image off to one of the AI image creation tool as a source. That’s the way I see it.
Anything PhotoLab can do to reduce rote work (what else is cutting out a mask), so much the better.
Totally agree with you on this. I thought your list of what AI could offer were all positive additions.
I’d also like to see AI sharpening added to Photolab so I don’t have to send images to Topaz Sharpen AI when I need to (which does do an excellent job fwiw). Export from PL, open in Topaz, apply sharpening, export again - it’s just a faff tbh!
Having just spent hours waiting for Topaz to render some extreme non-sharp images, I am not sure about putting that extra load on the PL infrastructure.
Yesterday, I had to wait 10 minutes for Topaz to enlarge and sharpen an image, just so I could re-evaluate it and have another go. After several attempts, the whole computer restarted with an error saying something about swap files.
What Topaz does is truly amazing, with a layering structure, I have been able to graduate sharpening on images like this…
… where the DoF wasn’t large enough to incorporate the near part of the conveyor belt.
So, I sharpened, just the conveyor belt, from where the DoF ended to the corner but that still left about the half into the corner not quite sharp enough, without over sharpening the rest. So, I was able to apply a second layer to bring that into full focus…
… and, printed to A2, it is truly astounding.
I think trying to integrate such power hungry functionality into PL might just be a step too far for many users. unless, possibly, it could be integrated as a “plug-in” like FilmPack and ViewPoint. At which point, one has to ask if that is any different to passing the file to Topaz.
Not forgetting that Topaz is continually improving at a rate that could be hard for DxO, starting from scratch, to catch up.
As it is, in order to speed up Topaz, I am going to have to upgrade my 2019 Intel MacBook Pro to an M4 - and we already have PL users complaining about export times and upgrade cycles for macOS…
Since Photolab is an raw image developing tool I think just adding AI selection like what Lightroom has would make many people very happy.
Joanna, your point quoted above and your other thoughts are all valid. For the few images I need to pass to Topaz, as much as I’d love a ‘one stop solution’, Topaz does fill this need very well.
I"m late to this discussion.
Here are some things I do beyond Photolab, or in a tedious way within Photolab.
so AI could help by adding new functionality or making existing capabilities easier.
Subject related:
- Subject recognition and selection so I can apply subject selective sharpening. I routinely do this in Topaz in a subsequent step, acting on an exported TIFF. It works great, and keeps getting better. but its such a frequent routine operation for me, and I suspect it could be done even better in Photolab as part of the RAW processing than going over to Topaz and doing it on an exported TIFF. This would save me a step. I do not apply the unsharp mask at all in DXO as I dont want any sharpening until my last step, let the sharpening become exaggerated by some later operation.
- Eye enhancement. this goes with subject recognition. In DXO, I routinely put a local adjustment point on the eye of a bird, and add a touch of upping midtone and touch of increased contrast. (this does less noise amplification than juicing up micro contrast.) It improves the DXO exported image, and works great in subsequent TOPAZ operations to give a better eye.
Clutter removal;
- Night sky. the number of satellite tracks now is exasperating. Automatic recognition, selection, and removal of these would be helpful as an option. These were annoying when taking pictures of the latest comet.
- Telephone wires and poles and fence posts, recognition and removal. I believe Photoshop can do this. I dont want to go into Photoshop, ever.
Finally, were DXO to go into AI sharpening of poor originals, I would want to see them do something better than the rest. Topaz overdoes it. Some way of keeping soft surfaces from looking hard and plasticky would be good.