Control EVERYTHING with Multiple Motion Brushes in Runway Gen-2

[@chariots8x230]

16:03 It’s cool, although I’m surprised it took about 2 hours to create such a short trailer. I thought it would take much less time, but maybe the technology isn’t there yet to speed things up to that level.

[@TokenizedAI]

It takes 2 hours because you spent 70% of the time waiting for the animations to generate. And you need to generate 3-5x as many variations as you actually end up using because many aren't usable. That's just how the workflow is right now. Unless you compromise on a lot of things.

[@jackiewatkins1830]

Man, I'm a long time traditional artist(graphite 13 years) just nowxentering the digital space and I'm being blown away by all this... organizing my productivity has been an issue bit it's exciting to see what's possible as I've been playing with this stuff good jo!

[@TokenizedAI]

If you embrace the tech, you'll have an absolute blast as an artist. All of a sudden, you can do stuff all by yourself which previously required help.

[@Beanskiiii]

@@TokenizedAIno they scammed everyone who paid $1000. Open AI Sora is already miles better

[@hakosho2003]

In the case of motion, I prefer Pika Art, RunAway cannot generate sudden or "speed" movements. In Pika Art, if you remove the background of the character, it is easy to generate someone running or walking without deforming the legs. The same does not apply to RunAway

[@TokenizedAI]

Pika is great at generating from a straight text prompt but animating an existing image (and that's what most of us need) still isn't better.

[@TokenizedAI]

I know what you mean but Pika lacks in ways that are more relevant to what I'm usually working on. It's always going to be a trade-off, unless you want to subscribe to 20 different services.

[@WilliamSheck]

Great Job, Christian. (Another) new era dawns. Honestly, one year ago I thought we were five years out on this frontier. Things are moving at Warp speed.

[@TokenizedAI]

Yup, I agree. And I think more surprises are coming.

[@shadrackkiban4716]

Man this is so eye-opening I’ve tried using pika , but as you said it keeps creating weird motions , this brush tool is definitely a game changer I can feel my creative juices flowing just by watching this video

[@HansJoachimNolte]

i like your honesty. unlike many other authors, you state in the video itself that the ai doesn't really work perfectly. others write exaggerated headlines to generate visitors. i don't think the quality of the video itself is really great. you can see in every scene that it's AI. i think it will take a little more time until the results are perfect. but: your video was definitely enough for inspiration...;-)

[@TokenizedAI]

Of course, if you compare the quality to a studio production, it's not there yet. But if we think about where we were 18 months ago, this is still amazing. And what a lot of people forget is that there are so many aspects of the pre-production phase of films where this level of quality is absolutely sufficient because it's not the actual final production. But it's still 1000x cheaper than the alternative.

[@joselitosering]

right on! that trailer was bad ass!

[@The-Spondy-School]

Nice work, Chris. I've started using Runway just before their 5-brush rollout. I agree in what a game changer it is. I keep hoping there will be a gradual lengthening of their 4 & 8 second limits on film -- yet as you demonstrated, by having a sequence of related video clips -- albethem short - it's possible to stitch them all together in a compelling way by using the right background music. Again, nice job.

[@TokenizedAI]

Yeah, the issue with the extended clips is that they all get wonky. Personally, I like the shorter clips anyway. At least for trailers because otherwise everything is too slow.

[@The-Spondy-School]

@@TokenizedAI Agree -- that's been a huge part of my education -- getting ok with a more rapid pace of informing and getting informed. Look forward to seeing how you develop your film school project.

[@RLeaguer_Saint]

I always enjoy your videos, and learn a lot - this one was no different. However, you convinced me that I'm not ready for Runway yet. It's one thing to produce a trailer, where everything is disjointed by design, because you don't want to give away the plot, so there really is no plot. And for that, the multi-brush tool in Runway is very impressive. But for anything more, like if you wanted to link images into a meaningful story, it would be an enormous amount of work to produce anything of any length and quality. It's an extremely impressive and cool effect that Runway have pulled off with AI here, but, it seems to me, it's a long way from troubling film designers and graphic artists at this point, where you would need much finer grained control that the slight panning and zooming that is availble here. That said, I really enjoyed watching you play with it, and as always, I learned a lot, so thank you for the fascinating video. I think by the time it becomes truly useful (and I'm sure they'll get there), it will become so complicated that it'll be another tool that professionals have in their toolbox, but likely out of reach of most of us. Time will tell...

[@TokenizedAI]

Oh yeah, for a full story it's incredibly challenging. I'm working on something that I'd call a mini-pilot. Can't share details but it's really tough. Let's not even talk about dialogue. We need to hire someone extra who only does the mouth movement. Plus, it's illustration, so imperfections are as obvious.

[@storytellerHD]

Oh this is definitely inspiring on so many levels. You're crushing it. I'm impressed by how you're considering the story elements that engage the senses, i.e. the sound track. The money and resources it would take to create a "like for like" studio version of this trailer would be exponential, not to mention the time. Speaking of time...can you please comment on the average rendering speed for each clip? We have a very good feel of what to expect from MidJourney for stills (paid pro version) but still trying understand what Runway output times would yield. We're hearing that it's really, really slow. We can't validate those claims. I'm also hearing that their support is ultra slow to non-responsive. Again, these are just posts on various forums. Before we subscribe to the Runway Unlimited or Enterprise version we'd like to know more. What say you, Christian? How has the render time fared for you?

[@TokenizedAI]

Sorry, I missed this earlier. Yes, rendering is slow. About 1-2 minutes per clip. But you can render up to 5 at the same time and it's not significantly slower than if you just render a single one. Since you need to keep working on new images all the time, you can juggle back and forth. It's not ideal but I didn't find it too problematic.

[@morgangitonga]

Nice tutorial...an eye opener too. Thanks for sharing

[@LL-vd3qi]

WOW! Great instructional info and your final video was awesome! Thanks for taking the time & effort to post this for us! 🙏👍

[@ianmeizoso2072]

Amazing! When you say the face kept deforming, but you didn't have enough time to fix it... What would you have done to fix it with more time?

[@TokenizedAI]

I just would have re-generated it until I get something decent. Sometimes that's all it takes. 😅

[@johnleighdesigns]

its great stuff - a different workflow compared to traditional shooting footage oneself or creating video from stock footage or a mix - I can see fantastic opportunities to use AI video generation for creating animated style comic books, storyboards, bedtime stories, trying ideas and concepts for proofs of concepts - ive been trying a few approaches and thinking about combining traditional parallax image/card cutouts in after effects and/or blender along with some AI video footage too

[@gigboideaslab]

Your videos are always inspiring, Chris

[@TokenizedAI]

Glad you like them!

[@pennyclaringbull3060]

This was great, exactly the information I have been looking for.

[@NeYapayBuZeka]

magnefisent, but 95$ is expensive dou you recommed more cheaper video ai from runway?

[@TokenizedAI]

I only use Runway at the moment. Don't have time for 10 different tools, I'm afraid. I can justify the cost because I use it professionally.

[@GadielGarciaLopez-oz3il]

The Teaser was SO GOOD!!! But I have a question: What's the website where you create dose images?

[@TokenizedAI]

Midjourney

[@ART-ificial]

Great video, is motion brush not available in Runway anymore or is it only visible in paid plans?

[@unicornplus4444]

thanku sir , m watching from a very remote village in india ..... m interested in ai generated video and i find ur channel very informative .......

[@Shiva..]

Amazing workflow Christian! Which video editor did you use in making this?

[@TokenizedAI]

I use DaVinci Resolve.

[@Shiva..]

@@TokenizedAI Thanks!

[@thewebstylist]

So Amazing! And great to see you back our amazing Ai brotha and teacher 100

[@TokenizedAI]

Thanks a ton!

[@thewebstylist]

We miss you Christian!!

[@TokenizedAI]

Too much on my plate right now 😭

[@marilynlucas5128]

Christian, you're very talented. Keep it up.

[@TokenizedAI]

Thank you 🙏🏻😊

[@MrRoopu]

Amazing 🎉 it’s really wonderful

[@metamediag]

Which Video Editor do you use?

[@dolphinseven]

Awesome! Can't wait for the cinema release!! :)

[@TokenizedAI]

Wel'll see 😆

[@urniurl]

That's great. Curious as to how many credits approximately you used? 5 rolls per shot on average x 3 secs each would be my guess.

[@TokenizedAI]

Some more, some less. But I wouldn't say as many as 5 per clip. Maybe 3-4 on average. I compromised on some of the clips. But since I have an unlimited plan, it doesn't really matter anyway.

[@gamedev2022]

Great work, looks really nice

[@TokenizedAI]

Thank you 🙂

[@MicahJohns]

Very cool... thanks for making this.

[@TokenizedAI]

You bet!

[@BigSaaSsolutions]

It was one of the coolest i've seen

[@TokenizedAI]

Thanks!

[@morpheus2573]

Great work. Very inspiring. 🎥🤩🍿

[@TokenizedAI]

Thank you 🙌

[@markcandlin2635]

That's Great thanks Chris

[@TokenizedAI]

Thank you!

[@news-oko-plus]

Great job ❤

[@TokenizedAI]

Thank you! 😄

[@ppovoa1]

really cool teaser.

[@TokenizedAI]

Thank you 🙂

[@Empoweringlessons]

Great trailer. Great video

[@peacefusion]

What do I think? If I was a hollywood suit I would mail you the greenlight to make this movie now.

[@TokenizedAI]

🤣🙌🏻 Let's go!

[@yodowhatnext9742]

That was awesome 👌

[@TokenizedAI]

Glad you think so!

[@galvanizingtechniques1111]

That was cool man..

[@ArtSolace]

sorry i'm pretty new to this. can you do this with the free number of tokkens or is a paid subscription necessary to really play satisfactorily? how much does it actually cost?

[@TokenizedAI]

It would be unrealistic without a paid plan, I'm afraid. And if you want to create A LOT then the unlimited plan is the only real option. Anything below that only makes sense for very selective use. But if someone uses this professionally, the unlimited is totally worth it (despite the cost).

[@marilynlucas5128]

Have you heard of magic animate? You could use it to make the human move in any way you want. It will be nice to see what you create with runway gen 2 + magic animate and maybe deep fake face swap. You could make a blockbuster!

[@LouvoresPauloRicardo]

excelent as usual

[@manfrombkk]

Keep on prompting.. ALOT !1

[@TokenizedAI]

Every day I'm prompting.... 😅

[@iXzenoS]

The problem I have with the current AI video tech is that it's usually 5s or less. It's difficult to do something solid with such super short clips. Like what if I want to create a nature or city scenery video that is 30min or 1 hour long? Is there any good way to extend or piece together these 5s clips to make something of decent quality that is, say, 1 hour long?

[@TokenizedAI]

Hmm, not sure I agree. There are tricks you can use to extend your clips without actually extending them in Runway. I also think we tend to underestimate how often full-length movies cut scenes. You need to work with different camera angles. I'll admit that it's not easy, since you need to generate a lot of images and clips. But it's definitely possible.

[@jimberry7865]

It's cool what you can do with the current gamut of AI video tools, but I don't see much real-world functionality with them yet. It seems like we are at least 5 years out from this being something that could be used to assemble a real movie -- even at 1080p. The quality is noticeably sub-par to anything a professional studio would produce. It also seems like AI for video should reside locally in a tool like Adobe Premiere w/ Media Encoder. I can't see how it would be cost effective for a company to host the generation of a series of 10 minute video segments (with a bunch of "re-rolls" like in mid-journey) for people in mass. I would think it would be more cost effective to use local compute, similar to when you render a movie in premiere and you tax your graphics card(s) with the task. Plus having it inside your video editor would allow you to assemble it as-you-go into a timeline, rather than having to coordinate between tools.

[@TokenizedAI]

Actually, I disagree because I'm personally involved in several projects that are real world use cases. You're focusing too much on the actual final production. Don't underestimate how much time and money can be saved with these tools in the pre-production phase and even before an idea is even green-lit. Stories and concepts are being pitched with AI-generated concept trailers where the final quality is secondary, but the mere fact that you can show them, increases the likelihood of success tenfold. I agree that the quality is still subpar. But everyone always underestimates how quickly tech evolves. Remember where we were only 18 months ago?

[@jimberry7865]

I guess your right there are good use cases, concepts, etc. I'm just so used to the quality of midjourney that when it comes to video, the difference stands out. AI has already disrupted stock photography but stock video is not really being affected yet. A red-letter day will be when you can narrate a scene into the tool with character and environment consistency (complete with wardrobe crafting) and act like a director on set. For that, you'll need a "ChatGPT" type interface where you can have an ongoing conversation with the tool to groom the output. It should spit out low-res mockups (a step above storyboards) that you can tweak, and then when it's good, do a full upscale. Prompt Nirvana : Scene Six: [Jerry Smith] walks down the sidewalk at 3pm looking up at the windows of [Mike's apartment]. [Mike johnson] opens the window and yells out, "Hey you want to come up for a beer?" [Jerry smith] pauses and thinks for 3 seconds and says, "sure I'll be up in minute." The scene cuts. --Wardrobe Jerry05 Mike06 --length 00:40 @@TokenizedAI

[@Kemetrical]

Great video!!!

[@TokenizedAI]

Glad you liked it!

[@mosiarmstrong]

That was dope.

[@TokenizedAI]

Thanks 🙂

[@christianclavet]

I would not categorize that as AI "Filmmaking" at all. It could be used as a concept art animation only. Ai as a LOT more to do before these movements and images are realistic. It the next blockbuster would go out on theaters made by this, I would never go. This is only good to make concepts for now and nothing more. Calling that a "AI Movie" is deceptive. But for what's its doing (concept shots), it's really good.

[@TokenizedAI]

Patience....young Padawan. Patience. Yes, these are just concepts. Give it some time. I think we'll all be surprised how quickly it's going to happen.

[@pushismano77]

amazing😍

[@TokenizedAI]

Thanks 🔥

[@ruihmartins]

Briliant.

[@DigitalAscensionArt]

Let’s goooo!!

[@TokenizedAI]

🚀🔥

[@j5646]

Amazing does not do this justice.

[@kresimirzex2720]

Just so everyone know he payed 1000$ for one year plan. If u don't do that ur always lacking credits. So its not cheap.😊

[@TokenizedAI]

Actually, this is false. I'm on a monthly plan, so I can cancel anytime I like. No way in hell I'm paying $1k in advance for something that might be obsolete in 2 months. But yes, you need an unlimited plan to really get the most out of it. But for anyone who uses it professionally (not just as a hobby!), it's a very reasonable expense. I have less useful things that I need which cost more.

[@reginahelenasantos8988]

yeahhhh !!!!!!!!!!!!!!!!

[@HugoVacacionesBailes]

Now, I want to watch the full movie, not just the trailer, ha ha ha...

[@TokenizedAI]

🤣 Maybe one day...

[@learnitbetter2365]

WOW!

[@TokenizedAI]

😊

[@sissycarolina4863]

Finally when I get a two minute video done where it looks half decent is $100 in credits. NOT WORTH IT.

[@TheGoodContent37]

-Clicked on video due to title aligning with my needs and wants -Saw this boring old looking guy, thought a little less of the video, he doesn't look like any AI creator I know -The guy took AN ETERNITY to show me what results would be achieved if I decided to watch the entire video so I thought even less of the video -Got frustrated because he talked a lot without giving me anything of use so I hovered over the progress bar to see if at some point he was going to show the freaking results so I can judge if I should spend my awefully limited life span on his video or ignore him for being lame and talentless -Found out the end result at the end of the video (Of course, always at the end, so damn obvious) -Got freaking impressed into oblivion by the end result, holy shhhhhhheeet! This old boring guy has talent, he is only a noob at youtubing So, I'll subscribe, check past videos, see if there's more good stuff there, like, comment etc. This is a hidden gem. Mainly because this guy doesn't know how to engage properly from the beginning. A solid sub for sure.

[@TokenizedAI]

I guess it depends on everyone's own priorities. 😅 Here's the thing. There's no point in me pretending to be someone that I'm not. If I tried to be more "engaging", as you say, it would come across as super inauthentic and it would be much more difficult for me to record these videos at all. So I prefer to focus on delivering substance rather than superficial fluff. The people who subscribe to my channel come to learn rather than looking for entertainment 😁 If that means that I get less subs, then I'm fine with that. There's a niche for everything. After all, 70k subs gotta come from somewhere.

[@TheGoodContent37]

@@TokenizedAI That's ok, this stubborn attitude some artists have is the difference between 70k subs and 300k subs in other AI KZbin channels with the same life span. That's why this is a hidden gem that more likely than not will keep being hidden, almost no one will share it to keep that edge over others. Sadly these channels end up burning out in a couple years but everyone will draw as much juice from them as possible before that happens. Of course no one admits this and many would even lie to your face swearing this isn't true but I just like brutal honesty, it might help you get out that toxic mindset that you shouldn't adapt to achieve more. Just showing the end result in the beginning instead of the end would be good enough to get better results but to each their own of course. Keep up the good work until is not feasible anymore 😀

[@TokenizedAI]

It's not stubbornness at all. The difference is that you are talking about people who mostly do YT full-time. YT is not my primary thing and I don't really want it to be that anyway. So, I guess that's what counts, no? I'm curious why you call it a "toxic mindset". PS: I didn't show the end result in the beginning this time...as a test. This was 100% international. I've done a similar video before where I did the opposite. So I wanted to see how viewers behave in each case. PPS: Also not saying that I won't take your feedback to heart. Trust me, I will. I'm just providing some context on why it is the way it is.

[@TheGoodContent37]

​@@TokenizedAIWe artists tend to have those type of toxic ideas that limit our growth. Like purist ideas or concepts that would somehow "protect" some abstract invented fantasy like value that we asume defines us. Like not "selling out" or refusing to adapt or conform to the world even a little bit. You know, that romantic idea of the poor suffering artist that died alone and unknown. Those are toxic ideas that limit the artists to get the opportunities and resources to grow and create even more. I'm not talking about opening an Only Fans account and selling naked pictures, I'm talking about learning to conform to the algorythm, to act a little, to pander to audiences, things like that that artists with those toxic ideas hate and think are below them, because they think they are gods or better than everyone else or that "THEY" are the ones that have moral values or what not. I just hate seeing this same ol' tale on repeat again and again. Artists that have those toxic ideas and seeing them disappear from this capitalist planet leaving the world without what they could have done if they weren't that freaking stubborn and uptight. While the most disgusting, low quality low effort talentless shit thrives, because they do sell themselves, oh boy sure they do, and they do it good. The rats partying when the mouse is gone, or doesn't want to play. But hey, you do you, these are only simple personal opinions. It's awesome that you tested that preview at the end so you can see the results. Maybe that works out better outside of common sense and evidence. Check the data. Hopefully that can help you get more success so you don't burn out that quick when life catches up to you and keep giving us the good stuff. And who knows, maybe a strike of luck can occur if you are consistent and the algorythm gods could favor you out of nowhere, people win the lottery all the time. I'll be liking and watching either way.

[@milestrombley1466]

Freaky, but cool!

[@TokenizedAI]

😆

[@LouisGedo]

👋

[@klaurcschwackerberg1880]

These days if I hear the word mind-blowing , I know it is time to skip 😅🤣

[@TokenizedAI]

I could say "exciting" but it honestly doesn't do it justice.

[@benugdsen75]

This has potential, but the anatomy of humans is still unusable. People with no legs, one leg, walking backwards… hands with nine fingers, weird faces…. It’s still not possible to use it for film productions. Big sceneries as spaceships flying over a city skyline works… But for character work.. No. Not yet.