At same its endless competition and you get bored of see the same effect for the thousand time. I wonder why people keep repeating that it’s only get better when you could apply the same reasoning to any tech.
@@hombacom I agree with you, people only know to repeat same effect, but its something new. People are just having their first "talk" with AI. Lets see what happens next year
upvote for miming a song in front of 10K+ views. In all seriousness this was the most informative video about Runway Act one I have found. Thank you very much. Will be back.
Man this is so freaking good. You're the first video I've seen who has actually shown their own examples rather than just what Runway has shown. Do you know if the unlimited plan also covered Act One or will credits still be taken for that separately? I love the idea of this making me a one man army with different characters and takes all driven by me. Amazing.
You can use unlimited on this, I generated 3 separate videos and they take a decent amount of time, probably 5-15 minutes per generation not including errors lol. I did 1 with credits and it was done in like 2 minutes, hope this helps
Thank you so much for this demo! But the one thing you forgot to show us is how much body movement you can get away with. It looks like the software does recognize some body movement, so please explore the limits in your next video. 🙏🏼
Thanks for watching! I used midjourney to create the image, and I believe I used the prompt “realistic photo of an old man reading a book in a old living room”, I hope that helps!
It is impressive how at the narration part on the right side the book slightly shakes as the character is having stronger facial activity, therefore most probably also more bodymovement and therefore more shaking of the book
It's not affecting my career yet, but like you said, it's just the beginning and I think it will. I'm a 3D artist, and my field might soon be outdated. I'm considering switching to game development to keep up, but I'm worried AI will dominate that too, and eventually, we'll all be out of jobs.
I think it will be used more as a tool to help speed up work rather than replacing jobs. I know it can be stressful for a lot of artists, I really don’t want people to lose jobs over ai, hopefully it can be another useful tool in our arsenal.
@@AtomicGains-ql8hg Obviously, with this AI, most customers who need face animation might not choose someone like me to do their 3D work anymore, though not all of them, just the majority. At this point, the AI isn't suitable for everyone since it has many limitations, but as you mentioned, it's improving rapidly. That's the worrying part, seeing the direction it's headed and knowing that CGI might become irrelevant soon. Anyway, this tool isn't something I use; it doesn't relate to what I do as a 3D artist since it's not even 3D.
update: i was in the middle of using runway for a project and got a random popup while i was generating. No email, just a random popup on the site itself saying i can now try Act-One.
Thanks for the comment. The unlimited plan is unlimited, the generations do take a bit longer to generate in the unlimited mode though. With act one you can only do 30 seconds per video at the moment, but you could easily break up your scenes into 30 second blocks and stitch them together. I hope that helps!
This is already amazing, but I hope that they will crack the code of characters with hands or whole bodies. More like combining the facial features with a wire-frame system. (Would be extra cool if you can personally adjust the wire frame for troubleshooting.) For hands to work, they do need to be already in the picture, because I guess it's much harder for the A.I. to create them by itself in the appropriate style of the video. Ultimately, it would be way better to be able to create 3D models of our characters (with more 3D information) that the A.I. then can bring to life.
I hope soon they will do it so that you can import a video where a person is moving or walking or doing something and add the facial animation on top of that instead of a static image
There's also an option to do that with "Advanced Live Portrait", but it has its quirks and the quality isn't nearly as good. But depending on the project, one could alternate between these methods and maybe achieve more than could be done with act one alone.
Its nice, but driving video on a still image have been done months ago, and while this is of higher quality, it's not far more useable. There is already technology to do driving video on top of a moving video. THAT is useful. we need runway to do THAT! Or allow full body motion capture, and not just the face. That would be amazing!! I gave this mode a try and it worked really well on the few images I tried. But the fact that the rest of the image is static is really killing the potential here.
@@mariomills how is masking helping this? If you have a video that is moving, than the animation of the face only was done on the first frame only, where the rest of the video has changed, including the posture of the person talking.
hey there love your video like you are doing something that no one is doing or being able to make our own character work...... please help i cant get my own character to work what are you doing different ?
@@AtomicGains-ql8hg custom character but nothing extreem followed all the rules but still not taking the image i upload. it says " not seeing the face in my upload and its just a pixar character of a old swami from India then i did a smilple character your typical caucaisan young pixar man character and still doesn't see a face. then i look at your video and boom you got everyone on it its amazing..
Once MiniMax get lip sync in it’s a wrap. My character’s arms move and it generates the side of their faces etc, if it’s at a certain angle. I still want to check this out from the video it’s quite dope
question: once we've recorded with our own voice, and used our own custom character, is there a way to change the quality of the voice? so I can ultimately create different people by myself?
Video generation, Motion tracking, video style modifer, image generation, voice changer, music generation... Ok now we have every tools to make a professional movie but every tools have to get put together in one website.
@@kfrfansub Sure. Competition is what keeps companies from gouging users (look at how evil Adobe has ebcome, for example). There are alternatives. I suggest using them whenever possible.
The fact these need driver videos and actors is why this approach will probably be short-lived and soon replaced by image to video + prompts in 12-18 months. It looks like a mish mash of tech from the last 15 years.
I think it's important to see what to expect. With video generators the videos often start off great and then quickly become a mess. This seemed to be flawless over the complete runtime. It's good to know that's possible.
Im still amazed by how far this tech can go. It is on initial stage, remember that guys...
At same its endless competition and you get bored of see the same effect for the thousand time. I wonder why people keep repeating that it’s only get better when you could apply the same reasoning to any tech.
@@hombacom I agree with you, people only know to repeat same effect, but its something new. People are just having their first "talk" with AI. Lets see what happens next year
It’s changing week by week and can get quite overwhelming, but it’s interesting to see how it’s evolving.
upvote for miming a song in front of 10K+ views. In all seriousness this was the most informative video about Runway Act one I have found. Thank you very much. Will be back.
Haha yeah it was quite embarrassing, but It was worth a try. Thanks for watching!
Man this is so freaking good. You're the first video I've seen who has actually shown their own examples rather than just what Runway has shown. Do you know if the unlimited plan also covered Act One or will credits still be taken for that separately? I love the idea of this making me a one man army with different characters and takes all driven by me. Amazing.
I am wondering the same thing cause your gonna burn credits fast to use this for anything. This could honestly replace a lot of vtubing options
i'm going to test it in the next day or 2 and ill respond if no one else answers you lol
You can use unlimited on this, I generated 3 separate videos and they take a decent amount of time, probably 5-15 minutes per generation not including errors lol. I did 1 with credits and it was done in like 2 minutes, hope this helps
@@dankujcreations Thanks. Man I wish these programs just let people use their local hardware.
@ same, probably one day but I think they probably want all the generation data haha
Thank you so much for this demo! But the one thing you forgot to show us is how much body movement you can get away with. It looks like the software does recognize some body movement, so please explore the limits in your next video. 🙏🏼
Thanks for watching! And yes I will definitely be doing more tests in upcoming videos.
Had to pick my jaw up off the floor, this is incredible.
Wow thanks!
Great demo, thank you for saving me the time of doing this myself... and money.
No worries!
Amazing man!
Indeed you're the first testing Act-One; I'll jump rignt into it right now😢
What's missing now: movement of the body :-D Great Video!
Hopefully that's coming in ACT TWO or THREE!
Awesome content!
Can you maybe tell us what the prompt for the old reader on the right was and which image generator you used?
Thanks for watching! I used midjourney to create the image, and I believe I used the prompt “realistic photo of an old man reading a book in a old living room”, I hope that helps!
As always, your test videos are the best on youtube.
It is impressive how at the narration part on the right side the book slightly shakes as the character is having stronger facial activity, therefore most probably also more bodymovement and therefore more shaking of the book
Yes I noticed that too, I’ll be doing a lot more tests in upcoming videos.
That was truly captivating! Thank you! ❤
You are so welcome!
Well Done my friend. Great Tests.
Thank you! Cheers!
comment for algo. Looks stunning, ty for your effort!
Incredible video I must be try it
Go for it!
It's not affecting my career yet, but like you said, it's just the beginning and I think it will. I'm a 3D artist, and my field might soon be outdated. I'm considering switching to game development to keep up, but I'm worried AI will dominate that too, and eventually, we'll all be out of jobs.
I think it will be used more as a tool to help speed up work rather than replacing jobs. I know it can be stressful for a lot of artists, I really don’t want people to lose jobs over ai, hopefully it can be another useful tool in our arsenal.
@@AtomicGains-ql8hg Obviously, with this AI, most customers who need face animation might not choose someone like me to do their 3D work anymore, though not all of them, just the majority. At this point, the AI isn't suitable for everyone since it has many limitations, but as you mentioned, it's improving rapidly. That's the worrying part, seeing the direction it's headed and knowing that CGI might become irrelevant soon. Anyway, this tool isn't something I use; it doesn't relate to what I do as a 3D artist since it's not even 3D.
Amazing! Did you just get a random email? Or are they sending it to content creators?
update: i was in the middle of using runway for a project and got a random popup while i was generating. No email, just a random popup on the site itself saying i can now try Act-One.
Awesome video! Question: Is the unlimited plan really unlimited? Kindly explain if this tool can be used for longer or unlimited creations. Thanks
Thanks for the comment. The unlimited plan is unlimited, the generations do take a bit longer to generate in the unlimited mode though. With act one you can only do 30 seconds per video at the moment, but you could easily break up your scenes into 30 second blocks and stitch them together. I hope that helps!
@@AtomicGains-ql8hg thanks for your prompt reply, very useful indeed.
This is already amazing, but I hope that they will crack the code of characters with hands or whole bodies. More like combining the facial features with a wire-frame system. (Would be extra cool if you can personally adjust the wire frame for troubleshooting.) For hands to work, they do need to be already in the picture, because I guess it's much harder for the A.I. to create them by itself in the appropriate style of the video. Ultimately, it would be way better to be able to create 3D models of our characters (with more 3D information) that the A.I. then can bring to life.
Good Video
Glad you enjoyed
You must be one of the first who has access!!!
So great for artists.
I hope soon they will do it so that you can import a video where a person is moving or walking or doing something and add the facial animation on top of that instead of a static image
That would be great, hopefully they add it soon!
There's also an option to do that with "Advanced Live Portrait", but it has its quirks and the quality isn't nearly as good. But depending on the project, one could alternate between these methods and maybe achieve more than could be done with act one alone.
@@missoats8731 I tried that as well but it really has some Problems qualitywise
Its nice, but driving video on a still image have been done months ago, and while this is of higher quality, it's not far more useable.
There is already technology to do driving video on top of a moving video. THAT is useful. we need runway to do THAT!
Or allow full body motion capture, and not just the face. That would be amazing!!
I gave this mode a try and it worked really well on the few images I tried. But the fact that the rest of the image is static is really killing the potential here.
Easy to resolve with masking
Would this mode fall under unlimited plan?
@@mariomills how is masking helping this? If you have a video that is moving, than the animation of the face only was done on the first frame only, where the rest of the video has changed, including the posture of the person talking.
hey there love your video like you are doing something that no one is doing or being able to make our own character work...... please help i cant get my own character to work what are you doing different ?
Thanks for the comment! What problems are you experiencing with your character? Are you uploading a custom character or using one of Runways?
@@AtomicGains-ql8hg custom character but nothing extreem followed all the rules but still not taking the image i upload. it says " not seeing the face in my upload and its just a pixar character of a old swami from India then i did a smilple character your typical caucaisan young pixar man character and still doesn't see a face. then i look at your video and boom you got everyone on it its amazing..
Is in the Runway Unlimited subscription the Act-One feature unlimited?
Hi, yes it is!
@@AtomicGains-ql8hg tnx bro
Once MiniMax get lip sync in it’s a wrap. My character’s arms move and it generates the side of their faces etc, if it’s at a certain angle. I still want to check this out from the video it’s quite dope
Yeah that would be great, thanks for watching!
question: once we've recorded with our own voice, and used our own custom character, is there a way to change the quality of the voice? so I can ultimately create different people by myself?
You can use a voice changer like eleven labs.
As an animator, I can't wait to be unemployable.
Until AI credits get too expensive and hiring us meat puppets is the more affordable option.
Basically, it s LivePortrait but the paid version 😂
Can it be used in the unlimited Mode?
oh by god wtf the fart omg i missed that the first time lolol
Lol glad you enjoyed it, (not a real fart haha)
I don't get it, does your sample of the reading men, is a picture from MIDJOURNEY ? a fix image ? How come the book is moving a bit...
Hi, yes it is a fixed image, sometimes ‘act one’ will try add subtle movements to the overall image, like in the video of the man reading.
💯💯
Thank you.
Can movement be transferred to a body such as dancing and others?
Not at the moment, hopefully they add it soon.
Does anyone why i get this message "An unexpected error occurred.
Your credits have been automatically refunded." when trying to generate? :(((
okay it was because i was uploading a .mov file! should be .mp4!!! hope it helps someone :)
Thanks for the update!
We're all agree that the best point of this video was the fart, no ? 🤣
Haha yeah, just to confirm…It wasn’t a real fart lol.
6:21 that is ONE HUNDRED PERCENT the villain from incredibles 2 (or the first one)
like!
Thanks!
❤
Thank you!
What was wirh the flying creature in the smartphone footage test? Lol!! Good video otherwise!
Please add a simple 2d character in your testing next time, surprisingly very few of those services work correctly with 2D cartoons
He LITERALLY did this in 3:02 but after the animations happened, It looks 3D. So you already know what to expect. Keep your expectations low.
how to you sync a different audio with your lip sync?
Do you mean how did I change my voice?
@@AtomicGains-ql8hg yes, and how then you lip-sync the other voice... thanks
Video generation, Motion tracking, video style modifer, image generation, voice changer, music generation...
Ok now we have every tools to make a professional movie but every tools have to get put together in one website.
You actually don't want everything on one website because then you have a monopoly.
@@shin-ishikiri-no What about a lot of websites that do all those things ?
@@kfrfansub Sure. Competition is what keeps companies from gouging users (look at how evil Adobe has ebcome, for example).
There are alternatives. I suggest using them whenever possible.
Is it very costly?
But it’s still only the face. Why the body is frozen?
Hopefully they will add body animation soon.
hmm the only thing i do not like is the credit use for it, it should be unlimited in the unlimited tier
reminds me of iClone
I haven’t seen that, I’ll have to check it out.
guys im starting out as an AI enthusiast making similar content
would appreciate your feedback!
I’ll have to check out your content!
Is it free?
8:50 this guy is ALL neck. This is creepy.
Yeah I know! Thanks for watching.
Is it free
Amazing👍🏻
Thanks ✌️
The fact these need driver videos and actors is why this approach will probably be short-lived and soon replaced by image to video + prompts in 12-18 months. It looks like a mish mash of tech from the last 15 years.
We didn't need to see the entire 30 seconds the old man reading the book. We get the idea in 5 seconds and then you can go on to the tutorial
I didn't mind seeing it. I came here to see runway anymations and how well they lipsync.
I think it's important to see what to expect. With video generators the videos often start off great and then quickly become a mess. This seemed to be flawless over the complete runtime. It's good to know that's possible.