Hey, mate, thanks very much. I love the global RU-vid community. I'm a long way from Hollywood here in Yorkshire, England, so it makes my chances of getting my movie made even slimmer if I rely on the traditional Hollywood route. AI is my best hope for filmmakers like you and I :-)
For London landmarks you could use Luma's image to video. That would give you a solid base to work from. I've recently been trying Luma but I didn't bother looking at the text to video option and went straight to image to video. The results are impressive half of the time. 😄
Thanks, Tasha, I'd be really interested to see what you do with it. Do you have connections with any willing actors that you could collaborate with for scene tests?
@@HaydnRushworth-Filmmaker I'm sure I could find *somebody* but I have altogether different types of scenes in mind of stuff that's already been shot... You might see what I'm talking about in my next piece depending on which route I go with it 😊
Buddy I think you are right but I think it's actually 5 years maybe 3 years. The progress over the last year showed me that by next year ai Video will be huge.
For your shots where you want to change the background, you could try green screen. Then you could either get the AI to produce the background or you could ask it to keep the green screen and replace it with a 3D model - although that will require Blender, interior models, and doing your own match moving (which is actually not that hard to do).
I did try one example with the green screen, but I need to run more tests to get it to work effectively. Really interested in your take on using 3D models. I'm entirely Mac-based these days, so whenever I come across an approach that requires a PC I tend to tap out at that point. That said, I think you're on to a really important angle. I've been increasingly thinking that the most controllable future for AI Filmmaking would actually be some kind of AI-enhanced blend of game engine technology and AI creativity. Ultimately, I'm hoping for a tool that is simple enough for a screenwriter to learn to use.
I think the male viewers is who the algo pitches your content to, based on your own status. I get the same thing, and the majority are 50 and over. It assumed like unto like. Those were great clips you got with DOMO, especially once you added the music. They worked very nicely. Also, a possible project short of a film could be an audiofile, uh, screen show equivalent, meaning using Midjourney to generated images. I would watch something like that because I like still images a lot.
Thanks very much re the music and Domo clip. The first time I watched it back with music added I got goosebumps and realised video-to-video really is the most likely way (at this stage) to tell the most human stories. In terms of the visual audio book, that's the next thing I'm going to work on. I'll start with the opening scene and try to versions... a narrated one, like an audiobook reading, and a dramatised one without the narration. Although, having said that, the movie itself opens with a voiceover.
@@HaydnRushworth-Filmmaker I'm wanting to do better lip-syncing for my vids. Runway is OK for talking, but not so much for singing. I'm ready to try video to video. There seem to be a couple possibilities I'm now investigating. This is probably something you will crack.
I'm in my 20s but feel like we are on the same wave in terms of thinking, not using youtube but writing to document my process... Keep pushing, we are gonna figure it out.
Hey, thanks very much, I really appreciate the encouragement. For what it's worth, you'll still feel like you're in your 20's when you're in your 50's... trust me, it gets really frustrating. :-) :-)
I've tried the free plan, and while it doesn't do much, Luma did produce some interesting results. I see some potential, but I'm approaching it from a different perspective because I've been heavily involved in prompting to create images. So, instead of giving Luma a script, I try to describe what I want it to do, just like I would with images. I'm not sure what works best. I have no background in filmmaking; my hobby was creative writing (poetry), and now AI Art.❤
Hey, so sorry for the delayed reply (3 weeks *doh!*)... To be honest, I'm prompting in videos in very similar ways to the prompts I'm using for the stills images, but I add camera movement, motion and action references in the prompts. It's still a dark art trying to prompt effectively, and different tools will produce quite different results with exactly the same prompt, so you almost need to pick just a handful of tools and try to learn their particular nuanced boundaries, but it still early days, and I maintain that it's better that we stumble around like toddlers in these early days rather than wait until the tools are "production-ready".
Awesome video. I am trying to get into filmaking and I have been building youtube channels. I started with MIDJOURNEY pics and story telling, now im getting into AI video + MJ. Thanks this helps me narrow it down... I also understand why your wanting video-to-video... I have been doing image to video
Thanks for the complements and encouragement. I think the more serious you are about brining human stories to life, the more you realise you’ll need real actors and video to video.
@@HaydnRushworth-FilmmakerSet upon your shotlist, create your storyboard with precise generated finalized stills (these become your final shots), animate. Then you’re ready for a rough cut. It’s golden. Just don’t rely on the prompts alone for the animation step. Give it exactly what you want visually and then it can handle the rest.
@@HaydnRushworth-FilmmakerAnd one final note, it will infer the action from the still. If you have a character holding an axe, it will chop, and if it’s holding a laser rifle, it will fire lasers. If you have someone with a shocked expression they will gasp, and if you have their eyes closed, it will open them. Give the stills the “momentum” you expect in the shot, and it will infer the context. It’s not perfect, but it’s genuinely good once you give it that baseline.
Thanks very much, I really appreciate the feedback (genuinely). It's a risky thing to "build" a channel without completely trend-jacking, but it means I get to connect with an audience where we have things in common :-)
You might try writing a story for the AI vids. You would have to mess with the apps enough to know what will most likely happen, from there you can get a feel for it that will release more creative freedom. Maybe a story about doing an AI vid. Some of those crazy AI vids naturally lend to a comedy.
I am building an A.i. series and I feel your pain when it comes to getting an image I can use to change to a video that I can use. A.i. right now is a dreadful mess. Lip sync videos are so lifeless and it's extremely hard to make anything work. Continue your journey and maybe one day we will both get our projects completed.
Good Advice - I'm in the process of trying to finish my first film and have hopes that AI might help in certain scenes, but nor relying on it to do the whole thing. Like your suggestions about using Ai more around the Social Media content, makes sense.
Thanks very much. It occurred to me yesterday that we really need storytellers and AI filmmakers to collaborate over great, short stories just to push the envelope with AI video. I suspect animation and claymation effects are the lowest hanging fruit, but at least it will get filmmakers and storytellers exercising their AI muscles. In terms of the social media idea, I haven't seen anybody else doing it yet, but in theory, it has great potential for building buzz with a target audience. :-)
Gen 3 doesn't allow you to upload an image yet, so relying on prompts is risky especially when you're spending credits. Luma lets you set a start and end frame, so there is some element of control there (as long as your images are good from the start). My go-to now is actually Kling AI, which also allows you to control the camera direction. :)
I've heard good things about Kling, but I hear the wait list is still a challenge. I agree entirely about the cost of credits, it really is a major consideration, especially since the cost of "failed" results are entirely shouldered by the creator.
An AI-generated video is an abomination. It has all the same issues as AI images but adds them to every individual frame, each of which seems to act independently. There are no rules they follow; there's no sense of physics. Body parts shift and morph, sometimes disappearing altogether. Solid objects can pass through solid objects at any time, resulting in a hellish dreamscape that's impossible to take seriously. I think that people talking it up are delusional.
I agree entirely that this is largely the state of AI at the moment. As a solo screenwriter living in Yorkshire, England, I'm such a long way from the production possibilities of Hollywood that I'm facing two options: a) Give up on ever seeing my screenplay turned into something watchable. b) Cling on the the faintest hope of any tool that finally allows me to create a visual version of my Screenplay. I turned it into a eBook to make it more digestible for a non-screenwriter audience, but I didn't write the story to be written, I wrote it to be watched. It's true, though, I really am an absolute dreamer, living in my own little dreamworld, naively, optimistically following a dream :-)
Just an opinion but if I were you I would dump AI and go back to traditional filming. And if you spent that long on a screenplay I would dump that too.
If nobody tried out new things then no progress would be made. For a vanity project (no disrespect meant - I've done plenty of vanity/self-published work of my own), why not try going a non-traditional route? No-one is risking huge sums of money and we all get to learn something. There's nothing wrong with your opinion, but it really doesn't match up with the author's stated objective of using *AI* to turn his script into a film.
The interesting thing is that I loved the entire creative process of writing that Screenplay. It really is such an enjoyable story that it was incredibly satisfying and rewarding to not just see it come together but to feel it coming together as well. I'm not exaggerating when I say that I regularly laughed and cried whilst I wrote the screenplay, and the process of turning it into a visual version is also incredibly enjoyable. Giving it up because it hasn't "succeeded" yet would be like giving up golf because you haven't won the US open yet.
Thanks very much, that sums it up really well for me. The costs and risks of trying AI are so low that we have little to lose but an awful lot to learn.
@@HaydnRushworth-Filmmaker I realise I may come across as cruel but it's a tough business we are in. My first novel was 250,000 words and I thought it was brilliant. It took me eight years to realise that the agents were right. It was crap. I suppose you've heard of a film called, Tangerine made in 2015? The film was shot with three iPhone 5S smartphones. It made its world premiere, at the Sundance Film Festival. It won a host of accolades and no AI at all. One really cannot better reality. The director is now highly regarded. He got his big break. I can only wish the same for you.
