“Ai Kitbashing with Tim” needs to be a legit series on the channel! I’ve tried pretty much all of your process kitbashes and have advanced the quality of my output each time.
0:44 confession.... Rogue One is the FIRST Star Wars film I ever watched in it's entirety (after decades on earth). This came after I watched The Mandalorian and got intrigued. Was so impressed after Rogue One, I ended up watching every non animated Star Wars film ever made!
How cool! That’s an experience I’m quite jealous of. My experience of Star Wars sadly has been one of diminishing returns. It was my entire childhood though, so idk who I’d even be without it haha
@@dsoprano13incredible, it is the only non-animated among star wars movie/series i tried three/four times to watch and never arrived past the first half an hour. My preferred is the empire strikes back. And i'm there since the beginning in 1977
His hand motion with 3 fingers!!!! And so blurry! Tim, you are so bright! Open AI should be all over you for you to test their SORA and be an ambassador! I am sure you could create a good film with SORA
As I delve into storytelling and video experimentation, VideoGPT seamlessly integrates into my process, enhancing my content with its subtle touch of sophistication.
I think this is neat and all, if content creation is all you care about. To me though, it feels like letting a chess engine play for you. The fun to me of animating is watching it all come together.
Thank you so much! There’s a lot to still fine tune here, but I thought it’d be great to get the ideas out there for others to play with and add to! I love this wild era of AI video!
Excellent to hear! Still thinking of a few other details to add in as well! We're all on the front lines here, and it's really exciting to see all the crazy directions we're taking things in!
absolutely!! and i love how you getting out of the box with the use of the tools, and not only that, you learn the new ones so fast and then teach others that's amazing, big respect🤓🤝 @@TheoreticallyMedia
Nice job, Tim! Your film reminds me of those classic cars in Cuba that are built out of anything they can get their hands on for cheap or free. I think it's great that you're trying so many tools. Eventually I imagine that some tools will simply shine and be the preferred ones for this kind of work. Anyway, bravo!
Haha. You ever hear that Johnny Cash song “One Piece at a Time”- where he’s working at the Cadillac factory and steals one part every day, but it takes him 20 years to assemble the car!
Small correction on Tarkin in Rogue One... It was not a deep fake, it was a full CG face replacement made by ILM. To my best knowledge the first time deepfake was used in Star Wars, was for young Luke Skywalker in "The Mandelorian" and later in an episode of "Book of Boba Fett".
Very valid point! And ha, yeah-- that Luke didn't work out too well the first time around. He did look much better in Book of Boba though! (didn't they hire the guy that did the better version?)
@@TheoreticallyMedia The reports to what degree RU-vidr and master of Deepfakes Shamook was involved vary between "He was adiving" to "He did it himself". But he definitelly helped ILM to work out a better version in Book of Boba Fett.
Thanks so much! Still a lot of fine tuning with this method, but I wanted to get it out there to see what others might do with it. I love this free-for-all era of kitbashing! So many opportunities for everyone!
Enjoyed the video, since trying out Viggle I've been contemplating a similar workflow. Using Facefusion and then Topaz with Iris over the top solved the face, but ive yet to find a better method to improve the body.
Kitbashing is THE Best! Uncanny Harry had a pretty good quip about it: In 6 months our Kitbashing will be incorporated into the platforms-- so for now, we're all leading the charge and showing the way!
Good question! I think that it is established in your reference photo. I’ll test it out with some heavy film noir shadows on a character to see what happens.
I am actually going to have to watch this twice. Maybe a third time. I think I will try something with this method but I don't have all the software mentioned or alternatives. Do you know if I can do the lighting corrections and stuff in Wondershare?
@@TheoreticallyMedia i will find out soon enough! Haha I won't complain if not though. I like the idea of this workflow for concept films or concept trailers.
Is there a reason why you didn't use the Speech to Speech in ElevenLabs using your own voice? I use that all of the time to get what I need and your tone in the recording does a great job in effecting the output voice.
I did try with my own voice, but it was still coming out garbled. My have been because I was yelling? It was the first 11labs major fail I’ve had in awhile. Surprised me!
Ever see his Epic Cheat Day meals? I love looking at those. Sadly, I just want to eat them without doing the 6 months of workouts! Haha, I’ll just take the food please!
I’ll try it out. Hypothetically, you might end up with the same character? Since Viggle only allows one reference. But what happens if we put in a reference with 2 people? Might break, but I’ll give it a shot.
Haha, basically yes. I beta test for a few of them as well. But, yeah-- somehow I've turned all of this into a job, so I've got to stay on top of them all!
Very good clarification-- I called it a "deep fake" as that is sort of what the media reported it as, but you are totally right, it is more akin to a CGI character that was mocapped. That said, it is funny how random YT'ers (and not so random ones like Corridor Crew) have done the same with better results!
@@TheoreticallyMedia Didn't mean to be all "Ackshually..." about it. 😂 Had deepfakes been more of a thing back then, it would have looked a hell of a lot better! As it stands, he's right out of the uncanny valley, with only a few shots looking somewhat passable.
@TheoreticallyMedia It's not my cup of tea. I'm definitely willing to wait a couple / few years for SORA quality text to video to come to the little people like me......especially when we get that quality in Open Source!!! 👍
@TheoreticallyMedia I didn't jump on board text to image until I saw sufficiently high quality being regularly generated by some of the tools. Time is precious and I feel it's a waste of MY time to play around with tools that are too cumbersome or not sufficiently good quality producing tools. That's just me. I was blown away by SORA...... and that's the standard I seek in order to get heavily involved (time and $$ wise) with an AI video tool.
@TheoreticallyMedia I definitely find there's a good selection of sufficiently good quality text to image tools available now and that's why I've spent a lot of time and $ on software tools subscriptions and hardware tools (like an HP Workstation with a lot of GPU under the hood). Once I get access to RELIABLE and cost effective SORA quality tools, I'm all in on AI video...... trust me on that.
Man, I had to see that 3000 times while editing, was kicking myself every time! That said, I’ll say that the overall point is that the name seems to work more as a placeholder to get Kaiber to create a somewhat consistent character. Not making excuses, apparently I cannot smell what he is cooking!
Haha. Not gonna lie, I often think that as well in the middle of one of these. But I think when you look at it as animation, or like, heavy VFX/Green screen films, there’s some time saving (and cost) here. Especially if you’re a “garage filmmaker”
@@TheoreticallyMedia I can see that. I'm also pretty sure consistent characters in Sora is going to be in a real distant future. Like years lol Even if the technology allows it, Hollywood will not. If you think the strike was ugly from the script writers, imagine how bad the actors strike will be!
Tim I’m shocked you never talked to me for any content. I make way more content than these guys, I also helped Pika train data and was the head moderator and artist for them. We talked a bunch..
@@TheoreticallyMedia oh, wow! I didn't realize you were a Head Mod, Mikey. You were so seldom present, I'm surprised. And, a Data Scientist too? Wow. I clicked a lotta thumbs. Can I put that on my resume?
This is what literally everybody does to create anything, even if it’s just subconsciously and it never makes it out of their head before it becomes what they’re trying to create. It’s also what people do when they use temp music on movie edits before the composer comes in to replace the music, very often replacing the music with something very similar. I do understand what you mean, but this is basically just a higher Fidelity version of what people have been doing for decades, if not the entirety of human history
@@ChristopherCopeland The thing is. AI is not inspired like humans are, because AI does not like anything. It sees trends in the data its fed and outputs an approximate results of what something a human made might look like.
