If you’re doing michelangelo tier stuff, I could see a need for this line art doer. If you’re making a comic or animation, eh… there’s more ground to cover that you can get called lazy for. I dunno, I guess it might depend on what you’re doing. Also, there’s a thread asking for the no AI rule to be dropped on the krita subreddit. I hope not, because last thing we need is to be like artstation and flooded with AI generated art that was two second to make.
I understand your pessimism, but let me remind you that krita is an official part of the KDE project. It has a name behind it, so the odds of the devs doing something really bad is lower than some literal who github project.
Am I less of an artist because I use the colorize mask tool?
To quote my previous post : “Unlike generating flats, which is a tedious busywork job that needs zero creative input, lineart is a fundamentally expressive task that requires you to think, feel, and train yourself to use your tool as an extension of yourself. It’s also, unlike the example of flats, a major component of the finished image, especially if there is no color intended to go in the work.”
It’s such a ridiculously bad apples-oranges comparison to make. Some other guy up there also was like “but what about transform tool or bezier curve tool isn’t the computer is do that???” and even worse someone was comparing being a digital artist at all to this, like raising concerns of an algorithm completely doing a task that benefits from and really is a major part of the creative process is equivalent to a hard core traditionalist gatekeeping you from the title of real artist for not drawing trad on paper you pressed yourself with charcoal you made with your own two hands or whatever.
The problem isn’t automation itself, it is what is being automated. Also, in a lot of these whatabouts I’ve seen raised in this thread like screentones or color adjustments or whatever the artist is still doing things manually and their input is the deciding factor to the largest degree anyhow. Not at all like feeding their art into a black box and seeing what comes out the other end. The differences shouldn’t need to be explained, honestly.
That’s the thing, as it was described this tool is not a feed into black box and see what comes out. It’ll follow the sketch very closely. It’s not supposed to make stuff up. If there is nothing there nothing will be done. If the lines are stiff, stiff will be the output. It’ll not prettify or correct mistakes.
Thanks for the links!
Looks like the presentation is little bit biased, considering the line art tool at the same level than Stable Diffusion and other similar AI stuff…
Grum999
@raghukamath @sooz @AhabGreybeard maybe the topic should be passed in “slow mode” as it start to become a hot mess
Grum999
I just checked: the first file I saved on Krita is from June 26th. I jumped the ship because of Krita’s public stance against genAI, so now I’m here expressing my concern, even though I know this is not technically genAI. I’m sure there are many other artists who just started using this software for the same reason.
I’m just going to add that if krita went full stable diffusion, the KDE project would never recover from the damage control. It be handing GNOME their userbase on a silver platter. That’s why krita I see, won’t ever go full AI. At least, I strongly assume they won’t.
I’ve said this already, this is the KDE project we’re talking about here. Been around since the late 90s. We absolutely can trust them since they have a proven track record of caring about the end users.
That would still be true for for the fast line art tool as it is proposed and the examples given. It can’t produce something from nothing.
And imagining Krita would pull a full Adobe and proclaim “all your art are belong to us” you could bet there would be a “clean” fork available in minutes. This is the advantage of Open Source.
This kind of feature just leaves the door open for more scams, I’ve already been scammed by people taking my sketchs and making lineart and then coloring by AI.
I always promoted Krita as the possible future of art tools in the same way I thought of blender but man if they are just gonna jump into the AI Wagon I guess boiling the planet and burning your userbase is worth it.
From what I read, the way this works is that you feed your sketch in and see how it interprets it into lineart. It is doing the step of inking your pencils 100% without your input and out of your control. If you hand off your pencils to someone else for inking you are not in control of the inking process, their interpretation of your work is theirs, not yours. That is what is happening except it’s an algorithm tracing over your lines instead of another human being.
inb4 someone accuses me of saying that having your art inked means it erases all of your credit in the overall piece and it is now owned by the inker, no. Bad. Shame on you. That’s not what I mean at all and you know it. But just because you provided the prompt doesn’t mean you did the work or controlled the process of interpretation to ink. Yet another one of those things that should be obvious but we’re having to explain it anyways, smh dadblast it.
If you cut out a part of your drawing and transform it, you are the one who moved it and placed it and decided where it should go exactly. Not the computer.
When you hand off a sketch to an algorithm and it automagically interprets which lines are important and which ones are not and inks it for you, you have been completely divorced from the decision making process. The computer is making those decisions, in the web of its inscrutable neural network, not you.
I still have to decide if the result is good enough or not, have to provide a very clean sketch and possibly have to clean up the lineart too. I mean, it’s hard to tell how good this thing is even going to be at this point. When I read that just a few dozen examples are actually used to train the network, I assume it probably can’t be that good. The feature as it’s currently presented seems to be just a slightly better version than the auto vector functionality Inkscape has and people sometimes use to make outlines from sketches.
This thread is now in slow mode. Please remember the topic is about this new tool, not AI in general.
If you’d like to discuss AI, there is a thread dedicated to that:
AI related meta thread
There is now a new topic to submit your sketch and lineart. Call for donation of artworks for the Fast Line Art project - #9 by YRH
Even before generative AI anybody could take sketch that’s publically posted and do his own lineart on top and then post as personal work. The problem is people on the internet more than any tool.
And before that they could do it with mud and a cave wall… dude is about making it so hyper accesible for people to scam artist. Yes it happened before but now anyone can do it, before you needed some skill to actually do good lineart.
It doesn’t excuse that you could to it before if suddenly you make it so you can get it in a click instead of having to at the very least learn the basics of linework.
Many of us know how to draw, in fact I have no problem to provide several examples of my lineart to help the development of this tool, it would be very useful for my work, I love to draw but I am not in a situation where I can take the time I want to do the commissions.
As I’ve already said before this is about reducing the amount of harm that it can cause. We’ve already seen so many examples of AI technologies coming out and people using them for malicious purposes. (copainter, the “speedpaint” creators) Why add another? Especially for a program I believe puts artists first.
Also as an aside I’m glad to see more people speaking out about this, I thought I was going crazy ngl