Runway Can Create Characters?! | An AI Animation Tutorial

[@deepworkdesigner]

This is exactly what I did for my AI avatar I use on my channel. Timmy is the 🐐

[@curiousrefuge]

timmy is amazing!

[@richtoonsTV]

I haven't used Runway Frames yet, but I have used this method for animation cases like this. But it's always great to see little pointers here and there to make the process smoother.

[@curiousrefuge]

Frame is off to a really good start, but you won't find as much control as other tools yet.

[@vesicastudio]

Thank You so much, Caleb! Can't wait to play with this!!

[@curiousrefuge]

Any time!

[@AI3EFX]

I need to mess with Runway’s new features more than I do, Kling still seems to still be at the top of the mountain top. Slick video 🍻

[@HarryClipzFilmz]

Same Kling been on a roll though

[@curiousrefuge]

Kling is awesome - Runway has a suite of features though!

[@HarryClipzFilmz]

@ I saw as I was checking it out and I may subscribe to them

[@thehumanaiexperiment]

This is great, thanks, look forward to the course.

[@curiousrefuge]

Glad you liked it!

[@thehumanaiexperiment]

@@curiousrefuge When is your longer course becoming available?

[@AIOneMusic-dj5mu]

always learning something new! Thanks so much!

[@curiousrefuge]

Glad it helped!

[@robinfields6101]

Can you do a video on how you added audio to the video that will match the mouth movements? Thanks

[@curiousrefuge]

Check out Act 1 !

[@jhoansebastianlara]

Amazing! could you tell me how to achieve your "futuristic" background?

[@curiousrefuge]

It's a projector! :)

[@technodramatists]

What is the difference in output between uploading a still image as the reference and adding a video like you did?

[@curiousrefuge]

Output should be the same but it depends on the program you use! Very often the output , regardless, is 720 or 1080 as a base.

[@NGW_Studio]

Also perhaps using a still image as ref will output only the facial/head animation. While using video as ref, will output both face and background animation That's my guess, need to try act one some more🤔

[@SuperDigitalAI]

I think that RUNWAY it seems to be generating the bad versions of videos (not following all the prompt) at purpose, in that way you need to do more generations (more credits / more money need to be expended to get what you want) to get what you want, because the JUMP from not good generation version to the one that you want it's so obvious, they just take it off the part of the prompt with the "hand gestures". In almost ANY other Video Gen App, it follows all the prompt (all the motions/movements), even in the "bad" versions.

[@curiousrefuge]

Interesting theory!

[@SuperDigitalAI]

@@curiousrefuge Yep, at least at this point in time in AI development, I think that the "follow the prompt" as good as it's possible, it's already achieved long time ago by many Video and Image Ai gen apps and we get used to that "oh no no, you have to do more generations till you get it right" thing, but now there is no excuse to not follow the prompt or at least 90% of it with each generation, in my opinion. So, to me ... the not hands doing the gestures in the first 2 generations, it's simply because they remove things from the prompt randomly on the first 2 or 3 generations, after that they really do the whole original prompt generation. It's a "Dark Business" point of view, but I think it's perfectly possible and there is no way to proof if it's happening or not, but we need to remember that business it's business, all the time.

[@danielyanezgarrido]

not at all, runway just has a different prompt structure than the rest. Its not as permisive.

[@SuperDigitalAI]

@danielyanezgarrido ok great. How you know that it's not the case? Just trust in them? Lol

[@danielyanezgarrido]

@@SuperDigitalAI Because Ive been using runway for a while to understand its logic, thats all. Its not perfect but once you understand how it thinks you can get more regular outputs.

[@gocurious7401]

Hi when you are talking are you on a green screen? Or si that a large tv? How do you get your background to look like that? Thanks!!

[@curiousrefuge]

Projector! :)

[@gocurious7401]

@@curiousrefuge Cool thanks! But how? Is the projector in front of you? Not sure how that works?

[@ministryofanimationofficial]

The only good thing about Runway is Act 1, once others offer competing solutions, there will be no need for Runway. Their video gen feels like a whole generation behind competitors like Kling and Hailuo. Their image gen is laughable compared to MJ, Recraft, Dall-E, and their new "Frames" image gen costs $100 a month for what others can do for free.

[@curiousrefuge]

These tools will keep leap frogging each other!

[@ExpatKeith]

The Dor Brothers are killing it with thier AI music vods that are in thioer own way film-making. I am surprised you haven't mentioned them even though thier genre of music may not fit with your brand. They are however getting really good at storytelling and many of the things you talk about.

[@curiousrefuge]

Dor Brothers are excellent. We like their style of adding popular music, quick cuts, and face swapping with famous people.

[@Mediagumbo]

Will there be a deeper dive into this for the next course? Will this be incorporated? Thanks,

[@curiousrefuge]

It's currently baked into AI Filmamking!

[@A.M.-AI.artvideo]

thanks for the advice! great tutorial!

[@curiousrefuge]

Glad you enjoyed this!

[@UnlimitedAIFilmAndMusic]

I wished that Kling ai would run as fast as Runway, everything would be perfect!

[@HarryClipzFilmz]

Use video editing software to up the speed interpolation

[@curiousrefuge]

Hopefully soon!

[@footballskillshighlights8810]

I hope as well. Kling is good

[@heartshinemusic]

I wonder how long it will take before we can also add our own hand gestures (and poses etc.) to the driving video for Act-One (and similar features.) Being able to do our own mo-cap performance + facial expressions + dialogue with one driving source video will be a big leap forward. Just a matter of time now...

[@curiousrefuge]

You can do it already with things like Viggle!

[@heartshinemusic]

@@curiousrefuge Yes, only the quality and realism isn't that good yet, i.e. lip-sync is not up to par with Runway act one. Plus I need a realistic, consistent character.

[@lightlegion_]

Isn’t it amazing how talented you are?

[@curiousrefuge]

We all can be using these same tools :)

[@agnesslovehealz]

Kling has easier work flow to use character reference and background then bring to runway

[@curiousrefuge]

Kling is currently a top 2 tool!

[@LastCallComedy]

We are getting close to having A.I. on the level to do a movie made for Tubi. I guess that aint too bad???😊

[@sunburnfm]

How do you animate in minimax? NVM, I thought you were talking about 'taking' animation.

[@curiousrefuge]

You animate by either text to video or image to video. Thanks!

[@arabianeyes]

this is awesome man, can we use 9:16 ratio for this for instagram reels?

[@curiousrefuge]

You can certainly crop it or use the ratio in Runway

[@dizzypack]

any tool that will use body movement from video you upload ,that would be more usefull ?

[@curiousrefuge]

You can always try Viggle!

[@kiukle]

Does it create 3d models that you can import into your own software? Or just 2d animation videos?

[@curiousrefuge]

2D currently, but there are tools like Meshy where you can create 3d models and then import.

[@kiukle]

@@curiousrefuge Yea I am messing with Meshy and Rodin, Rodin seems to do the basic characters better for what I'm after right now.

[@BaronArtsProd]

this is awesome.

[@curiousrefuge]

Thanks for watching!

[@HarryClipzFilmz]

what are good prompts for hand gestures?

[@curiousrefuge]

For us, we simply said hand gestures, moving hands, etc.

[@HarryClipzFilmz]

@ Thank you. Will use in my next project

[@lachainepassepresentfutur]

You can use Act-one with a single image. ;)

[@curiousrefuge]

True! But we want better motion in our charcter ;) !

[@user-gy4xp2tu6f]

nice 3d render

[@curiousrefuge]

Glad you liked it !

[@alaintanguay4152]

Nice.

[@curiousrefuge]

Glad you enjoyed it!

[@vrynstudios]

Try Kling, it does better too.

[@curiousrefuge]

Kling is awesome!

[@ChezTropic]

❤❤❤❤

[@curiousrefuge]

thanks for watching!

[@punchtalestudio]

runaway is running. Backward

[@curiousrefuge]

Never understimate Runway :)

[@Samt2b]

can you please make Kling AI trending consistent video like which is a man go to car open the cae go inside and drive like consistant

[@curiousrefuge]

We'll see what we can do!

[@TheAiKonnekt]

So, you have to pay $95/month to use this?

[@curiousrefuge]

Yes, but there are many other tools baked in!

[@TheAiKonnekt]

@@curiousrefuge Do you like this better than Kling? I like Kling but dang, it's so slow right now especially with the new feature is just released.

[@SneakysneakFoxxo]

I wonder if they can make some furry avatars or a mix between the two.

[@curiousrefuge]

You can try!

[@SonicPaletteProductions]

Why don't you tell everyone that it takes dozens of times to create a good image?

[@curiousrefuge]

Sorry, we believe most people know it takes multiple generations. But it's true, definitely consider that it's not a "push button once and perfect results" experience. We appreciate you engaging!

[@TheAnimatedSchoolhouse]

We used this process as part of this project: kzbin.info/www/bejne/g6aZeYmMppqCgq8

[@curiousrefuge]

Thx for sharing!

[@excursionfilm6075]

Why would somebody want to “do” this?

[@curiousrefuge]

Creativity, Art, fun, making stories, anything!

[@s.patterson5698]

And THAT'S the main issue i have with these programs....you need several generations to be able to get something useable, but they take your credis for every generation you do.

[@iKINDREAD]

Just use comfy Ai work flow. It’s free and all in one space and far more flexible. These programs are just building interfaces and using the same thing and charging you arm and leg. Look up work flows you can copy and paste into comfy Ui. I was able to do the same thing for free

[@curiousrefuge]

It's true it can take quite a few generations!

[@Mooooty]

Runways Frames is weak sauce.....Runway please stay in your lane and stick to improving your video generator.

[@curiousrefuge]

These tools are still evolving!

[@Saahib_K]

You talking is also ai generated IMO

[@CoolestToysGadgets]

Is runway free?? No

[@curiousrefuge]

No of course not!

[@kevinlittle8956]

Deedle thinks so! kzbin.info/www/bejne/fqu5k31-pcyYe5osi=revSmu89xyKOspO6

[@curiousrefuge]

What do you mean?

[@procrastenabled]

I used a similar process to make the music video for an AI track I remixed...I did the performance, and I'm a big burly middle-aged dude with a beard, and it worked really well, picked up the lipsync through my beard even! kzbin.info/www/bejne/eWKniICmbZKDmNEsi=bxtgCyQr73pgwuMl

[@curiousrefuge]

Very cool!

[@procrastenabled]

@curiousrefuge thanks!

[@ClickArtsDigital]

kzbin.info/www/bejne/bJqrqIeJoJaqp8ksi=zFnoJheJhiUGnPLK I created my first short animation film using just AI generated video.. Thanks! for sharing this feature in runway, i will check it out. It will definitely give realistic voice over for my future projects.

[@curiousrefuge]

Awesome thanks for sahring!

[@ModernGadgetreview]

I had to sit here for three commercials b4 the movie started- TY is getting out of hand

[@curiousrefuge]

True, KZbin is getting harder to see the content!

[@Rick88888888]

More childish A.I. rubbish. Already as it stands video A.I.'s far too often cartoonize serious prompt requests. Just use your own real face!

[@themightyflog]

I like AI but recently it has gotten so boring. People just typing in stuff. Like I loved when folks merged things with AI. Not type in a mid journey can’t control it prompt and then run it through these video AIs that don’t do much. I have not seen folks doing cool stuff with AI like they did a few months ago. And it is all an almost can do this.

[@curiousrefuge]

Check out our gallery and you'll see more than simple demos :)

[@themightyflog]

@@curiousrefuge it’s good for short form artsy cool quick videos. Narrative it can be cool. I don’t see enough people utilizing 3d in connection and it being consistent enough for long form movies. Always a new tool and something always missing that another has. Like there is Runway that does lip sync from a video, stable diffusion can do it as well and then the alibaba lip sync. Each missing what the other has. The alibaba one is the best but can’t be applied to video, runway can apply to a video but not as good looking as the alibaba one, comfy ui gives more control but still flickering. I’m afraid we will get into that situation like Iclone, Daz3d and Metahuman. Iclone goes well with every program but looks ugly, Daz looks great and is super customizable and easy to customize but doesn’t work well with other programs and Metahuman looks great but can’t customize it without such a high degree of expertise and it has been years and they leave it like that. Same thing is happening with AI. Mid journey is getting boring despite the beauty but I can’t get a simple pose.

[@bhavenjain07]

❤❤

[@curiousrefuge]

Thanks for watching!