Moho 14. And The Rise Of AI.

[striker2311]

if Moho wanted to beat Adobe. Number 1: Add more features that I mention. Number 2: get recognition by sponsoring content creators into signing a contract. to use Moho, not Adobe Animate or other software. and have a description on their video of what software they're using. so people would see and use that software. and sometimes mention the software in their social media. do not always mention software in social media or else people would get annoyed and never use the software. and do not sponsor them in their video like "ThIS VideO iT's SpOnSOR bY MOhO" People would get annoyed and never use that software, and it would get a bad reputation. or maybe add some extra other AI features. if Moho added this feature this will make animation way more easy. or someone could make a script about this. AI auto shadow lighting etc. I hope Moho 14 added these features. btw this it's just my opinion.

Believe me, it all sounds well and good, this all AI business, and yeah I'd love to see AI integrated in MOHO like after detection the structure of body add lightning and all or like automating similar actions by the user and use those actions for every other rig once done for one rig, keep learning from the mistakes and all, but I'm afraid that would or could take away the ability of creativity from the users or artist. It would take away the ability to think about the various factors like how should a head turn look like and all (just a example). Though yeah for faster processing it could be helpful.

For number 2 getting recognition can come, right. I mean we don't need to force people by signing a contract to use only moho and not using any other software, this is limiting their working freedom and choosing moho for whole process, where artist uses a combination of softwares to get the result that they want. It could also show how desperately we want others to recognise us. It could be good but can backfire in long run if something doesn't go right.

So yeah, believe me adding AI is so good in MOHO if it really makes sense to add a particular feature through AI rather just throwing AI into the software. and I believe moho 14 will have a lot of better features as I have confidence in people who are developing moho from all these years.

[JoelMayer]

Okay here's a common misconception about Klaus (i'm not gonna comment on the whole AI can of worms):

There was nothing really automatic about Klaus' lighting engine. There was a human that used a custom made tool to add light and shadows via vector shapes in different colors and blending modes. The tool was just convenient enough to easily animate those lighting/shading patterns over time (as long as the character stayed in the same perspectice). It was basically the same principle as layering multiple Multiply/Screen layers in After Effects using Vector shapes. As you can see, the still had to animate these shapes by hand for each key. It's really not Rocket Science, you could do this in AE, it's just more convenient for them doing it in their own tool.



So while Klaus looks great, it's not really any new, ground breaking technology and has absolutely nothing to do with AI.

People always pretend like there was a drawing going into some magic piece of software and came out fully lit on the other end. Nothing could be further from the truth, it was just an added step to the whole process and was all hand made. The only difference was the custom built software that made it more stremlined for the artists. There's no magic button.

[JoelMayer]

Harmony has a system like that but still needs a normal map to function properly (otherwise you could also just the ol' "inner shadow" layer style trick).

Best way to get good results: put in the **** work.

[lucasfranca]

Okay here's a common misconception about Klaus (i'm not gonna comment on the whole AI can of worms):

There was nothing really automatic about Klaus' lighting engine. There was a human that used a custom made tool to add light and shadows via vector shapes in different colors and blending modes. The tool was just convenient enough to easily animate those lighting/shading patterns over time (as long as the character stayed in the same perspectice). It was basically the same principle as layering multiple Multiply/Screen layers in After Effects using Vector shapes. As you can see, the still had to animate these shapes by hand for each key. It's really not Rocket Science, you could do this in AE, it's just more convenient for them doing it in their own tool.



So while Klaus looks great, it's not really any new, ground breaking technology and has absolutely nothing to do with AI.

People always pretend like there was a drawing going into some magic piece of software and came out fully lit on the other end. Nothing could be further from the truth, it was just an added step to the whole process and was all hand made. The only difference was the custom built software that made it more stremlined for the artists. There's no magic button.

Yes, just that. Everyone wants the button with auto shading.

But looking at this video, everything can be done in Moho if we only have two things: feathering mask and a realtime preview of masks and blending modes.

The feathering can be solved well with a blurred stroke in Moho and using the line width tool. For the blurred stroke, if we had a smarter blur on the vector stroke (not an image brush), with the border transparent and the center opaque (to avoid processing maybe a tool to turn off the feather to the center of the shape), an interpolation of this in a few steps (not the soft edge in the style palette effect) would do just fine. This should also be WYSIWYG (What You See Is What You Get), or in real time on stage and also solve the issue of interpolation between points in relation to the line width, solving the bumps on the line in some cases.

SUMMARY: I see that Moho can do this feat, yes, and soon. And if it did, a simple demonstration of a process would bring thousands of new users to Moho.

[Lukas]

Best way to get good results: put in the **** work.

Well, in the near future AI will definitely put in a lot of the work for you.

I mean... https://twitter.com/_akhaliq/status/1659424744490377217

AI is moving so fast, don't forget how bad it was 1 year ago. This is not gonna slow down. Creative business is already being affected.

Shading in Klaus might have been done by hand, but plenty of effects like that are going to be handled by AI tools in the future.

[Yuri Sekulov]

I agree with the opinion that creative freedom should not be taken away by adding ready-made tools, for example - to add shadows or lights later, you can use already created free applications based on artificial intelligence / EbSynth /. I think for future versions of moho it would be nice to add to the control function the force of impact on the separated points, in relation to the bone. As for example /Spine/
Another possibility is to have more tools for modeling and texturing simple 3D objects and the lighting of imported ones.
I miss the ability to draw in dimensions other than x or y. as for example with the 2d option of a blender
And finally as a classic animator who is forced to work with like Spine, just because it exports a Jason file. Don't you think that such a possibility/jason file extension export/ will give a lot of potential to promote Moho in other areas that use 2D animation such as animation for games and applications.

[hayasidist]

...
And finally as a classic animator who is forced to work with like Spine, just because it exports a Jason file. Don't you think that such a possibility/jason file extension export/ will give a lot of potential to promote Moho in other areas that use 2D animation such as animation for games and applications.

Moho uses JSON.

BUT many people see "This program uses JSON" as meaning "and so its output can be used in any other program that uses JSON". That's simply not the case.

JSON is little more than a definition of how to use punctuation characters such as [] {} :, "
What you put between the punctuation marks could be made to follow a standard - defining a standard vocabulary, syntax and semantics is clearly achievable --- and some might even suggest that there are already candidates out there. Spine might be one, Lottie perhaps thinks it could be. Trouble is they're all quite different in how they describe basic stuff such as layers, skeletons, points, colours and effects, transitions... So whilst it's technically feasible to define a set of things that could be exported from Moho to another program, there's often not a clear / direct correspondence with other programs (and vice versa).

This goes into a more in-depth consideration of the file differences between Lottie (JSON) and Moho (JSON) https://www.lostmarble.com/forum/viewtopic.php?t=35050 [I haven't had an in-depth look at Spine]
and this gives you some starting point for considering if an exporter / converter is viable: https://www.lostmarble.com/forum/viewto ... 16#p210416 [again, focus is Lottie - Spine's JSON is here: http://esotericsoftware.com/spine-json-format]

That's not to say it's impossible for a future Moho to have options such as "save as JSON for Spine" and "save as JSON for Lottie" ... just that (IMO) it's a load of work when the dev team's effort is (still IMO) better used elsewhere...

[Breinmeester]

Best way to get good results: put in the **** work.

AI isn't going to go away and it's definitely going to affect our industry. Believe me, I understand having a purist view on the craft, but I just don't think that's a sustainable attitude. It would be wise for Moho to consider where AI can help keep up with the competition. And I think that in that regard regenerative lighting makes a lot of sense.

[slowtiger]

I'd really like to discuss AI, what it can and what it cannot do, how it works and how it scams artists, who finances it and who profits from it and who not - but only with folks who have a decent understanding of animation in terms of real, years-long experience in production.

That said, I think the team at Lostmarble knows pretty well how to market their software. Any excited "but you HAVE to do XY now!" is, at best, just annoying.

[striker2311]

Best way to get good results: put in the **** work.

Well, in the near future AI will definitely put in a lot of the work for you.

I mean... https://twitter.com/_akhaliq/status/1659424744490377217

AI is moving so fast, don't forget how bad it was 1 year ago. This is not gonna slow down. Creative business is already being affected.

Shading in Klaus might have been done by hand, but plenty of effects like that are going to be handled by AI tools in the future.

Don't forget about adobe firefly btw in case. and not only that roblox it's developing an AI that can make games just by typing a word. https://www.youtube.com/watch?v=e2w0mBheCQ4 I do think AI it's going to bring us good news and bad news. the good news it's AI it's going to make your life easier. The bad news it's AI might replace your job etc etc.

I don't believe it will be like, going to replace jobs of artists (in some cases it might but). AI art will be just an another form of art and a platform to showcase one's idea, just like it is for photography, traditional art, digital art, animation. AI is just going to be another way like AI animation and AI art, etc like that

[opeth]

Something that anyone can easily do is worth nothing. Although there are filters that automatically do what can be done in photoshop, it cannot replace the work done by the photoshop master. People do, ai just puts something out by copying what people do, basing it on, mixing it up.You have unlimited imagination, drawing skills and animation techniques. Ai can only imitate you

[Breinmeester]

Something that anyone can easily do is worth nothing.

I think that's a generalization that's not always true. Reading is a good example. While (almost) anyone can do it, reading a lot is a telling sign of intellect.
That said, if it's true in this case, then that's bad news for all of us. It would mean that the skills and competence we've spent years to master will soon become worthless because anyone can produce similar results faster, cheaper and without investing lots of time and money. So while your comment has a dismissive tone, I feel it only shows the importance of AI's arrival in our field.

I'd really like to discuss AI,... but only with folks who have a decent understanding of animation in terms of real, years-long experience in production.

That's a bit elitist, wouldn't you agree? If anyone I think up-and-comers should have a bigger say. Established professionals know the worth of their skill set and already have the connections and channels to put them to use. Animation students today fear their skill set will become absolete pretty soon and with good reason. To them, the threat is real.

That said, I think the team at Lostmarble knows pretty well how to market their software. Any excited "but you HAVE to do XY now!" is, at best, just annoying.

I too have a lot of faith in Mike and Victor and the fact that they know their user base and markets. However, I don't see why it's annoying to explore how new techniques can be incorporated in a well functioning piece of software. Isn't the nature of software development that the software should develop? I can disclose that before Victor and Mike bought back Moho from Smith Micro our team was seriously looking to redesign the pipeline with alternatives for Moho, simply because the development of the software was stagnant and therefore its future unreliable.

Anyways, we've had this discussion before when I was still an animation student, some 20 years ago. Back then CGI entered the field and opinions were mixed. Some people said 2D was dying or already dead, others said it doesn't matter how a film is made as long as it tells a good story well. While the latter group was obviously right, it is undeniable that if all other factors were the same, the general audience would choose a CGI film over a 2D one and still does. On the other hand, CGI brought a whole lot of opportunities to the industry, for 2D as well.
The difference with AI is that for CGI you still needed a lot of people with a skill set that overlapped the skill set of 2D artists greatly. And you needed to invest in expensive hardware, so it was only an option for people who knew what they were doing. AI is available to everyone and it's cheap to produce. Personally I think that's a good thing. Think of all those people that were potentially just as skilled and talented as Mozart that we haven't heard about simply because they didn't grow up in a family where they had access to a piano. Democratizing the production of art means its quality level will eventually rise. But it also means competition from people who haven't put in the years of effort to get to where we are today. After all, the audience doesn't care how it's made, as long as it tells a good story well. And if the audience doesn't care, broadcasters, producers, distributors and the like definitely won't care if it was made by a studio or an adolescent in his parent's basement, as long as they can sell it to that audience.

The whole "AI took our jobs!" discussion aside, I think the most important thing to take notice of in terms of impact on our industry is how AI will set the bar, just like CGI raised the bar 20+ years ago. If cinematic lighting becomes a standard because AI can create it as an afterthought, producers are going to want it in their productions. If other software has things like those readily available in the application, studios might want to switch to that software. It's certainly not make-or-break yet, but it wouldn't hurt to explore what sort of possibilities there are that make sense for Moho specifically. Again regenerative lighting seems like a sound and achievable option.

[slowtiger]

That's a bit elitist, wouldn't you agree?

Not at all. I've seen my fair share of new technologies appear and disappear over the decades, and had many a discussion about it. There's always that certain kind of people who, while not the least experienced in the field, urge everyone to embrace the hot new thing, to the point of insulting any sceptic, and make a fuzz about "using" whatever is the new thing, but have not the slightest idea about how to make things work in practice, let alone know about the inner workings of stuff. I don't waste my time with these any more. Show me people who really use stuff, and I'm happy to discuss with them.

"AI" is a useless term anyway since it covers everything and their deceased uncle since the 70's, but doesn't really define a certain technique. The newest iteration plunders images off the internet and stuffs them into a database, a bunch of poor humans has to tag each of those because software is still lousy in detecting what's really in an image - remember the first "automatic" porn filters? Which relied on recognizing "cylindrical pink shapes" to determine whether some image contains nudity? What it can do now automatically and with some success is recognizing differently coloured areas in images and isolate and index those. Now a text prompt searches that database for matching tags, putting the resulting images in a pool from which the software further constructs new images. If it's animated, like the newest one featured in https://www.cartoonbrew.com/tools/stabl ... 29044.html, it produces some kind of replacement animation which we know since Yellow Submarine (1968, the start to Pepperland sequence) and Frank Film (Frank Mouris 1973.) It works completely in 2D and has no spatial understanding (the head turns are abysmally bad). The results are a flickering mess, you could as well use random animated GIFs as textures on 3D characters (remember the "scramble suit" in Linklater's A Scanner Darkly?. It can't create any movement of its own, only uses video reference to replace the moving parts with what it dugs up from its database. It's just a big copy&paste machine, and the results are almost always tasteless kitsch. That's no surprise since a big bunch of it's "training data" (an euphemistic term for "we ignore copyright") comes from readily available sources like DeviantArt. (No surprise that the text equivalent, ChatGPT, writes bad prose, since it harvested tons of fan fiction.)

If you really want to introduce some progress into animation workflows, give me a local database on my own machine and the software to automatically index it. Give me hierarchical filters (that's "prompt" in IT speak) so I can easily have useable source pools. Give me access to the weighting algorithms and let me define my own. Give me discrete layers of colour and brightness information, of shadows and lighting directions, of texture orientation and depth mapping. Give me adjustable smoothing in space and time as well. Give me playback controls with an API to have them scripted and animated. Give me an animation software module which can put all this neatly into their own layers so I can use it only in parts where I want it. And then you still have to pray for a story and a script which will put all this to good use, where the whole copy&paste aesthetics is relevant for the story and makes sense.

[synthsin75]

Seems like a very odd thread to be a new member's only posts on the forum. No introduction, no questions, no shared work. Just some agenda.

Talk about AI in animation just sounds so much like wanting a "make animation" button. Quality will always take work.

[Hoptoad]

I want an AI that will alert me whenever I put a thumb on the wrong side of a hand.