I am not going for any monthly ransom software, but this looks seriously impressive! GUI looks very neat too.
I quite like how neat the interface looks compare to the cluttered darktable interface which is small (on my 27") and harder to maneuver.
Masking by saliency mapping and object detection isnât new. Couple that with machine learning and we have a compelling package. It is possible in open source as well if someone is willing to put it together.
I mean new as in ânew Lightroom toolâ. I am aware I can do more or less the same in darktable (without the AI, but often MI works better, even though slower). I really like the interface though.
I watched it again on a bigger screen (my mobileâs is super tiny) and I see the image-editor-layer type GUI. It looks like a handy way to view masks at a glance and also allow them to be the basis of new ones. Interface debatable but feature-wise a good thing. I think ART and RT are closer to this than dt, but I havenât been using these 3 in a long timeâŚ
It is a fairly simple picture though, isnât it?
How does the masking work if, for example, you replace the boy with a chicken, or several chickens? Can it cope with subjects that contain holes or is obscured in some way?
Not sure, but very sure they used a well working sample for the demo and also pretty sure that in reality it wonât be always that smooth.
I believe Lightroom is a RAW editor, like darktable and has no such function (I donât use LR, so not sure), thatâs what PS, AP, Gimp can do. And I think the feature is the AI mask selection, what you do after that is pretty much the same as in all the other editors, including dt.
There was already a âselect subjectâ in photoshop tight?
And the topaz mask tool had this for ages I think
At least the ai sharpen and denoising tools from topaz can do this. It detects what is in the picture, and then you can pick âpeopleâ or âskyâ or âtreesâ to mask them for example.
Always good to see it in more tools of course, do om not complaining. But it isnât new or groundbreaking :).