For Stable Diffusion, it's a little bit more work because you need to find the combination of the best Model + LORAs, but you can achieve better results than the others.
Im a writter waiting for AI video to video tech to develop. I have written LOTS of scripts for short film, full features and many ideas that even sometimes i end up seeing someone produced and sold to Netflix when i had the idea years before and mine was better. I got born poor, i have no rich friends, but i got born with talent, have won several nation wide art awards by myself learningnfrom the internet. I KNOW I can create amazing art, i just need the luck and AI will give us ALL that luck. No need for expensive cameras, expensive lighting set ups, expensive actors, vfx editors or score writters, rich producers or gate keepers. We will all get that with software. Just amazing. Give a few weeks after video to video AI gets to the point where i record myself and replace the footage correctly with different styles and characters like in ebsynth and youll see art you had never seen before. No more humanity will have to suffer rich people art. We will gain the 99% of art that used to die poor and unknown.
I believe in the future being able to create our dream projects won't be an issue. The real challenge will be to make it stand out from a sea of millions of other works that will no doubt flood the world due to AI. Having the resources and connections to be able to do promotions on social media, getting into film festivals, doing influencer collabs, advertising, networking ,etc. will be a must. This is the struggle a lot of current smalltime content creators are facing. They may have the talent but they can't get people to find them in the first place.
I think the cream will rise to the top. The music industry already went through this with DAWs. SoundCloud and TikTok’s algorithm has found more musicians in recent years than talent scouts. Same will be true for online film distribution. TH-cam and other social platforms will propel otherwise unknown artists and put them in front of the world. And checkwriters will have to take notice.
even better!!! we could get rid of all those writers, production people, maybe even actors. I bet future AI tools will have better ideas than you ever will be able to come up with. At some point, directors won't be needed!! Never ending stream of content to watch!!! super exciting times ahead!!!
I'm the researcher for my company tasked with finding the best tools to streamline our work. Your channel makes my job easy, I just follow your advice. 😎Prodigious work!
The best part is that a person like yourself doesn't need talent anymore! If that troubles you, don't worry, all your uniqueness will be irrevelant soon.
As long as live action productions continue to exist, I guarantee you they will never move on from a location before getting a complete gaussian splatting scan of the set. It will become as expected and standard as recording room tone. Being able to digitally grab any pick up shots and create them fully in post with the actual location is going to change absolutely everything. David Fincher is no doubt creaming his jeans somewhere right now 😂
Except that room tone is old fashioned and going away quickly! Why? ML-based noise reduction and voice isolation! Works better and wastes less time. I'll miss the respectful "60 seconds of silence" demanded by the sound recordist as a memory of a bygone era...
Each time I see a video where they compare Stable diffusion with any other image generator it seems that they do not exactly know how to enter a prompt in stable diffusion
Or what resolution to generate at. That double face is usually from generating larger then what the model was trained to do. If decent SD workflows are used the quality is there.
As always, appreciate the depth of showcasing and live demoing along with the curation of broad range of updates! I'm excited to give the contest a shot, would love to do your course and join the community. It'd be amazing to learn better tips and tools to get the renders of video (currently Pika/Runway) to match the resolution and detail from MJ!
Sounds like this would be great for uploading old movies and improving the CGI, or maybe adding CGI where it wasn't before. I'm thinking something like Mars Attacks Mayberry, where Sheriff Taylor and Opie save the world.
Yes yes yes! We discovered one tool for cartoon animation! Very realistic and we even use it for our cartoon narrator called Sticky! We recently did a video about it and I think animators should try it instead of complex softwares like Blender!
In my humble opinion, the same prompt for different AIs is not a totally fair comparison. I used some AIs, but mostly Blue Willow, and even there the prompts for V2, V3 and V4 need to be adapted to each one. In BW there is a guy, spectral_aidream, who used to do realistic movie stills, using V4 and even with V3. Each tool has its own tricks.
Better prompt understanding/following is far more important than bigger batch sizes. Bigger batch sizes is a waste of my time, and a waste of my generations (and therefore money) when MJ continues struggling with even basic things like specific colours on specific items in a composition. I can get thousands of beautiful images in a row in MJ, but that’s not much use to me if I can’t use 95% of them, because it’s made fundamental errors following a prompt. For me, better prompt following is the highest priority, more so than better image quality, higher resolution, or bigger batch sizes. A standalone app with useful in-painting would be next highest priority.
Great video...thanks! As we all know, Adobe wiped out Figma as a competitor...and has it's sights set on Canva with Adobe Express. Have you had a chance to investigate further their Ai roadmap? I'd be very curious to know how closely their model is following Canva's moving forward. 🤔
We try to showcase the tech that is available to most artists via a normal machine. Because SD is so intense on GPUs and technical we find that it is a bit challenging for non-technical folks to use. Whereas the other tools can be used by anyone fairly easily.
How can someone "reference" a Dalle-3 image in Midjourney? I use Discord and don't see a command for reference. There is describe but I don't think that is what you mean.
The cat eating pizza actually makes some sense because grammaticaly speaking it can also mean a "pizza that eats cat" So perhaps since the expresion "Cat eating pizza" is ambiguous, the AI created ambiguous video.
I, actually, use Leonardo all the time. It has, IMO, a more attractive style than Midjourney. Plus, on the paid version you don't have the same level of censorship as Midjourney.
Very informative, thanks! But could you do a call-out before starting to show the creepy stuff? With AI and everything possible with it, the creep factor can really get up there, lol.
Yes, you could use Midjourney to do it quickly or train your own model in stable diffusion on it. But I’d recommend using Midjourney with potential reference images as a faster workflow.
00:06 The week was historic for AI Film 02:35 AI tools can enhance film production 05:03 Canva introduces new AI-powered tools for image editing and video integration 07:12 Canva introduces AI-powered features for brand voice customization, automatic translation, and quick animation. 09:31 Canva and Adobe Max announced new tools for image editing 11:48 Mid Journey is the top image generation tool with high-quality results. 14:09 Moon Valley is a promising AI video generation tool with smooth movements 16:32 Mid Journey 6 will have better image quality, better language understanding, more resolution, and bigger batch sizes. 18:42 Chat GPT can now perform web searches and provide relevant information. 20:43 AI is being used to solve child abuse cases Crafted by Merlin AI.
Thanks for a great video full of resources! I want to produce a feature film with AI, with identifiable main characters and scenes that can be used from different angles. How far away are we from this?
This looks really cool and, I plan to play with it. I'm just concerned about the potential of future copyright laws allowing companies/corporations who own the code to steal my artistic license because their A.I. "made it".
Dont know which stable diffusion you used but that was the worst example. SDXL could definitely achieve these shots. Just cant be doing the midjourney prompting to get them. Rest of this was pretty solid.
That’s so odd. I used SDXL. Perhaps it’s because I used the exact same prompt for all of them. Maybe SDXL needs negative prompts to pull more quality out?
Your voice is sounding kindof muffled this video - not sure if you're using a different mic or you changed some settings but if you compare it to the last video you can notice the difference.
Man I love your videos thanks a lot with your help. But please use a good mic and not the Enhance Speech from Adobe. it's weird, now everyone has the same voice!
@@curiousrefuge it’s a 22 minute video that based on the title I expected to be all about it the most realistic AI video tool yet, but it’s just a 22 minute long video featuring a bunch of updates about different AI tools from chatbots to image generation/editing, but nothing about video generation. You should’ve just titled it “what’s new in the world of AI”.
Not necessarily - we are constantly using free or freemium tools throughout our projects that we also both love and endorse. We try to be as fair as possible and want to empower people to make the best informed decisions about their tools (whether it's paid or not).
Thanks so much for the feature!! Loved the course and keep up the amazing educational content!
Fantastic Ai insight video, and luv u guys were in the mix in Denmark film peeps!
Sooooo happy you are doing these filmaking AI news! Top notch video
For Stable Diffusion, it's a little bit more work because you need to find the combination of the best Model + LORAs, but you can achieve better results than the others.
Learn to be less lazy and actually design and draw yourself 😅
@@toututu2993 oh, that's cute. If you make a little drawing for me, I'll put it on my fridge.
@@Alain.Robert AI really makes you guys stupid. I don't blame gaming at least you get to use some brain as a gamer with some effort
Preach 🙏🙏
@@Alain.Robertsubbed for the comment 😅😅🎉🎉🎉🎉
Im a writter waiting for AI video to video tech to develop.
I have written LOTS of scripts for short film, full features and many ideas that even sometimes i end up seeing someone produced and sold to Netflix when i had the idea years before and mine was better. I got born poor, i have no rich friends, but i got born with talent, have won several nation wide art awards by myself learningnfrom the internet. I KNOW I can create amazing art, i just need the luck and AI will give us ALL that luck. No need for expensive cameras, expensive lighting set ups, expensive actors, vfx editors or score writters, rich producers or gate keepers. We will all get that with software. Just amazing.
Give a few weeks after video to video AI gets to the point where i record myself and replace the footage correctly with different styles and characters like in ebsynth and youll see art you had never seen before. No more humanity will have to suffer rich people art. We will gain the 99% of art that used to die poor and unknown.
I believe in the future being able to create our dream projects won't be an issue. The real challenge will be to make it stand out from a sea of millions of other works that will no doubt flood the world due to AI. Having the resources and connections to be able to do promotions on social media, getting into film festivals, doing influencer collabs, advertising, networking ,etc. will be a must. This is the struggle a lot of current smalltime content creators are facing. They may have the talent but they can't get people to find them in the first place.
I think the cream will rise to the top. The music industry already went through this with DAWs. SoundCloud and TikTok’s algorithm has found more musicians in recent years than talent scouts. Same will be true for online film distribution. TH-cam and other social platforms will propel otherwise unknown artists and put them in front of the world. And checkwriters will have to take notice.
This is the correct attitude
As long as you’re cool with the script being written by ai too that works.
even better!!! we could get rid of all those writers, production people, maybe even actors. I bet future AI tools will have better ideas than you ever will be able to come up with. At some point, directors won't be needed!! Never ending stream of content to watch!!! super exciting times ahead!!!
I'm the researcher for my company tasked with finding the best tools to streamline our work. Your channel makes my job easy, I just follow your advice. 😎Prodigious work!
Aye! Glad we can be helpful. Take it easy. 🙌
Rerender tech is going to flip the industry upside down
This channel is absolute 🔥You guys always have the absolute best AI filmmaking content. Period.
Thank you so much!
The best part is that a person like yourself doesn't need talent anymore! If that troubles you, don't worry, all your uniqueness will be irrevelant soon.
th-cam.com/users/shorts4lOVk4Xrsw0?si=ZSVat5OXtVtPTYts
Great video, literally all about it, the narration, the info given, the apps discussed, keep up the good work
omg how did I not come across your channel before - finally a fellow filmmaking channel that actually gets AI
Welcome to the club! It’s gonna get reaaalll nerdy here. Haha
th-cam.com/users/shorts4lOVk4Xrsw0?si=ZSVat5OXtVtPTYts
As long as live action productions continue to exist, I guarantee you they will never move on from a location before getting a complete gaussian splatting scan of the set. It will become as expected and standard as recording room tone.
Being able to digitally grab any pick up shots and create them fully in post with the actual location is going to change absolutely everything. David Fincher is no doubt creaming his jeans somewhere right now 😂
Except that room tone is old fashioned and going away quickly! Why? ML-based noise reduction and voice isolation! Works better and wastes less time. I'll miss the respectful "60 seconds of silence" demanded by the sound recordist as a memory of a bygone era...
Glad to see myself in video finally
Awesome as usual. Thanks!
Each time I see a video where they compare Stable diffusion with any other image generator it seems that they do not exactly know how to enter a prompt in stable diffusion
Or what resolution to generate at. That double face is usually from generating larger then what the model was trained to do. If decent SD workflows are used the quality is there.
As always, appreciate the depth of showcasing and live demoing along with the curation of broad range of updates! I'm excited to give the contest a shot, would love to do your course and join the community. It'd be amazing to learn better tips and tools to get the renders of video (currently Pika/Runway) to match the resolution and detail from MJ!
15:14 the casual jab while maintaining professionalism frame😂
:)
Sounds like this would be great for uploading old movies and improving the CGI, or maybe adding CGI where it wasn't before. I'm thinking something like Mars Attacks Mayberry, where Sheriff Taylor and Opie save the world.
Been using kits ai for 2 weeks. Great stuff
This is great. Thank you for all the information here!
just what i was looking for... perfect
Thanks for watching !
Very good video.
Clear explanation. ✅
Coverage of common AI Tools for video and image generation. ✅
Unbelievable. Thanks for this comprehensive overview of emerging AI.
My pleasure!
Greetings, just joined your tribe.. some awesome information, beautifully presented, well done!
I love your videos and channel ❤
Thank you!
You bet!
Yes yes yes! We discovered one tool for cartoon animation! Very realistic and we even use it for our cartoon narrator called Sticky! We recently did a video about it and I think animators should try it instead of complex softwares like Blender!
Thanks for sharing!!
Thanks for featuring our Blueberry AI short film! 🦾
Y’all killed it. Great use of the tech!
In my humble opinion, the same prompt for different AIs is not a totally fair comparison. I used some AIs, but mostly Blue Willow, and even there the prompts for V2, V3 and V4 need to be adapted to each one. In BW there is a guy, spectral_aidream, who used to do realistic movie stills, using V4 and even with V3. Each tool has its own tricks.
Very interesting, keep on the good work :)
Thanks for watching :)
awesome channel👍👍👍👍👍
Thank you 👍
Awesome info
Better prompt understanding/following is far more important than bigger batch sizes. Bigger batch sizes is a waste of my time, and a waste of my generations (and therefore money) when MJ continues struggling with even basic things like specific colours on specific items in a composition. I can get thousands of beautiful images in a row in MJ, but that’s not much use to me if I can’t use 95% of them, because it’s made fundamental errors following a prompt. For me, better prompt following is the highest priority, more so than better image quality, higher resolution, or bigger batch sizes. A standalone app with useful in-painting would be next highest priority.
Yep, Leonardo is amazing
Totally. I use Leonardo in my work. That includes variations of my Leonardo prompts, with excellent results, in DALL-E 2; and now DALL-E 3.
Leonardo AI is built on top stable diffusion
a lot of cool uses for ai of late thats for sure.
What was the most realistic AI video tool yet?
Making something realistic is typically a combination of tools - but for images, as of today, Midjourney does a great job for cinematic realism.
it's very interesting...
Great video...thanks! As we all know, Adobe wiped out Figma as a competitor...and has it's sights set on Canva with Adobe Express. Have you had a chance to investigate further their Ai roadmap? I'd be very curious to know how closely their model is following Canva's moving forward. 🤔
Hey, great content, my man. Mind if I ask why you refrain from involving SD in your demonstration?
We try to showcase the tech that is available to most artists via a normal machine. Because SD is so intense on GPUs and technical we find that it is a bit challenging for non-technical folks to use. Whereas the other tools can be used by anyone fairly easily.
@@curiousrefuge I can follow that rationale. Your insight is just so intertwined and applicable, I couldnt help but wonder.
How can someone "reference" a Dalle-3 image in Midjourney? I use Discord and don't see a command for reference. There is describe but I don't think that is what you mean.
You just upload the image to discord, get the url, then paste the url in before you start typing your Midjourney prompt.
i can see AI has huge potential to create some decent COSMIC HORROR films.
Check out our submissions to last Halloween AI competition :)
The cat eating pizza actually makes some sense because grammaticaly speaking it can also mean a "pizza that eats cat" So perhaps since the expresion "Cat eating pizza" is ambiguous, the AI created ambiguous video.
Ah that's a good point! "Speaking" AI is an art in itself :)
캔바 쓰고 있는데 업데이트라니 ㅎㅎ 좋네요 ㅎ
Algo needs to feed ur vids to subs more
I, actually, use Leonardo all the time. It has, IMO, a more attractive style than Midjourney. Plus, on the paid version you don't have the same level of censorship as Midjourney.
You can’t use repeaters in Leonardo right? So you have to run each prompt by clicking prompt each time?
@@curiousrefuge Unfortunately
Very informative, thanks!
But could you do a call-out before starting to show the creepy stuff? With AI and everything possible with it, the creep factor can really get up there, lol.
Haha good call. It’s a bit jarring for sure.
GORE - CI - ON
Gaussian
Is there any tech that can take a picture of the painting The Scream and have it randomly generate a world based on it?
Yes, you could use Midjourney to do it quickly or train your own model in stable diffusion on it. But I’d recommend using Midjourney with potential reference images as a faster workflow.
00:06 The week was historic for AI Film
02:35 AI tools can enhance film production
05:03 Canva introduces new AI-powered tools for image editing and video integration
07:12 Canva introduces AI-powered features for brand voice customization, automatic translation, and quick animation.
09:31 Canva and Adobe Max announced new tools for image editing
11:48 Mid Journey is the top image generation tool with high-quality results.
14:09 Moon Valley is a promising AI video generation tool with smooth movements
16:32 Mid Journey 6 will have better image quality, better language understanding, more resolution, and bigger batch sizes.
18:42 Chat GPT can now perform web searches and provide relevant information.
20:43 AI is being used to solve child abuse cases
Crafted by Merlin AI.
Thanks for the timestamps, Merlin AI :)
Thanks for a great video full of resources! I want to produce a feature film with AI, with identifiable main characters and scenes that can be used from different angles. How far away are we from this?
This looks really cool and, I plan to play with it. I'm just concerned about the potential of future copyright laws allowing companies/corporations who own the code to steal my artistic license because their A.I. "made it".
Fair point. The near future will be interesting to see how this all plays out!
Anybody got a link for the New Horizons music video please?
Great Video, love it. Myself I am running a rather new channel with AI created music videos. I love to use those new tools ;)
You’ll have to share you music videos with us. We’d love to feature them. :)
VFX will just become so much easy
Yes! A.I. tools are going to help VFX artists and many others speed up their workflow exponentially!
is the cinema red chairs real or a ai background?
That was real!
Leonardo is ok, SeaArt has a lot more options.
Dont know which stable diffusion you used but that was the worst example. SDXL could definitely achieve these shots. Just cant be doing the midjourney prompting to get them.
Rest of this was pretty solid.
SDXL 2.0
Why not 7.3?
That’s so odd. I used SDXL. Perhaps it’s because I used the exact same prompt for all of them. Maybe SDXL needs negative prompts to pull more quality out?
"Essentially"
Your voice is sounding kindof muffled this video - not sure if you're using a different mic or you changed some settings but if you compare it to the last video you can notice the difference.
Oh dear, Cat Pizza! Just as horrifying as all other animal body parts on pizza.
Great video, and great content! Just don't push Adobe Podcast too much to post process your voice... you are in the uncanny valley ;)
The days of simple graphic design jobs are over. You have to stand out nowadays
Man I love your videos thanks a lot with your help. But please use a good mic and not the Enhance Speech from Adobe. it's weird, now everyone has the same voice!
Thanks so much for watching, we appreciate the feedback! :)
How close are we to feeding books to AI and it produces films from them?
Mind: Blown! Any of these tools create x-rated scenes? Asking for a friend.
Can AI change the creative world?
21:02 Nutrition
:)
You're being misleading with the title of the video.
How so?
@@curiousrefuge it’s a 22 minute video that based on the title I expected to be all about it the most realistic AI video tool yet, but it’s just a 22 minute long video featuring a bunch of updates about different AI tools from chatbots to image generation/editing, but nothing about video generation. You should’ve just titled it “what’s new in the world of AI”.
Who is excited by Getty Images new AI? It will only cost $12,780 per image for licensing per year!
You use the word implication in context that makes no sense in some parts?
I pray your thoughts are filled with positivity. May your life be plentiful with prosperity. ❤🔥
Same to you!
curios that only "pay" apps have good reviews.
Not necessarily - we are constantly using free or freemium tools throughout our projects that we also both love and endorse. We try to be as fair as possible and want to empower people to make the best informed decisions about their tools (whether it's paid or not).
what happens when no one is making movies or taking photos anymore and ai keeps recycling itself
It's not, don't worry
Hollywood is already recycling itself with all the movie remakes. Maybe AI would actually do a better job. 😂
Why is this guy using an older model human suit for a video.
6:58
Screw AI !!!
Either way, we appreciate you watching :)
Chat gpt wont brows the internet for me
Be sure your plug-ins are connected. You’ll see a drop-down when you click the gpt-4 button.
Hollywood is nearly dead.
Adobe* spelling
Not real AI.
Bro, are you real?
Yes :)
There's no such thing as AI film. There never will be.
😂😂. Watch this age like Butter
@@Backup439 we'll see
Bro what copium are you huffing to be this deluded. Please tell me this is just sarcasm I’m missing.
@@warframeees8013 we'll see is what I said. Just wait.
Ok grandma
But it's AI -that means that it's bad.
Very cool video BTW your website is down right now for me
Still down for you?