Im working on a 4 or 5 part ai series right now.first part is kinda over and people i showed said that its so hard to to tell difference between real or ai. Im glad i can fulfill my dreams because of the power of ai. And i need to say thank you the family of curious refuge. You guys make everything easier for me to select which system i should work with. If you guys host a meeting at Istanbul, I will be there for sure (I missed the runway meeting which makes me so sad because Runway is the GOAT right now i.m.o.).
@@peterk3697 I used chatgpt free plan to create prompts for my scenes, midjourney 35 dollars plan (but i used all of my 15 fast hours in just 10 days, i will change my plan), runway unlimited (100dollars), elevenlabs free credits (ı used 4 accounts which equals to 48K credits but i will buy the pro plan) , capcut pro for editing and epidemic sound music library for music and sound effect (its free for 1 month but when i start to publish my videos i will buy the pro plan which is around 30 dollars)
This was one of my older Wonder Studio tests from the best test days. I was able to import all the motion tracking data into blender and create a new camera and place the character into whatever scene I wanted. I still like having full control of my scenes but it's nice to know that some of these AI advancements are starting to make things a bit easier. That being said, I still like having the ability to export the animation and camera track for use in programs like Blender and Unreal Engine.
RunwayML is pricing its self out....right out of the market. Runway's video generate takes about 12 tries to get something remotely like your prompt...an epic waste of time and money.
thanks for mentioning this. After working in MJ for close to a year, I'm intrigued at the developments, but...there's still so much to learn with this. So much that I'd wish leaders would slow down. Nobody is going to use 3D/video in AI commercially with branding in 2025. AI needs to get alpha and text in grips before getting into depth.
@@curiousrefuge Or... don't reward bad behaviour. Especially when product unreliability is the business model. ie. "Pay us more if you want to get an acceptable result." 🤮
I knew the bench turn out weird as in the original image it was already showing. The engine should have extended the bench more, or replace the armrest. A consequence of adding it in to be fancy but not being able to deal with an extended imagine. Perhaps a good tool would generate the larger image from the get go (but kept from user) to prevent this issue.
This is why you shouldnt subscribe to ANYTHING annually, things look different day to day in some case and who even knows what's going to be around and what wont in a year. Keeping up with this right now is crazy
Hi, great videos. Maybe you have an answer to this: What would be the best workflow, to create act one videos with a AI generated voice rather than my own voice?
5:27 While it is mindbogglingly impressive, I can't help to see how the sun did a 180 and suddenly flipped from facing to trailing them. Weird as positioning the light source is more or less a staple of 3D animation.
Well it's not really "Video to 3d scene" if you have to import premade 3d assets. Actual video to 3d scene will probably come along one day but this seems to be only animation. Still pretty cool.
They literally had this on our cell phone for years. LOL You could be a farm animal and many other characters. They would do your facial expressions and head movements and recorded your voice speaking. They even had apps where you could make augmented special effect. LOL.
Just a suggestion for "Audience Favourite" for next year's Halloween Contest. I suggest only allowing votes AFTER all entries are in, as SO MANY votes were cast during the process, meaning later entries didn't get the same exposure. As cool as it was, I highly doubt we get the same winner for audience pick if voting was fair and even, as most votes for it came simply because it was one of the first entries posted.
I think Ideogram beats Midjourney by a long shot. Sincerely MJ have never really stood up for my work cases, which is usable graphics for dev/game projects. For art, maybe holds up, but not my usecase. For usable things especially for text rendering, ideogram beats MJ.
@@curiousrefuge I do mainly social games, in the social casino categories and other similar things. Some trading apps and so forth. Ideogram generates crisper, and clearer graphics for inclusion in projects with less touch-ups required, shortening the production time. MJ as mentioned, have more issues with artifacts,, added distractions and so forth, which makes tidying up an image more time consuming if not near impossible. That's why I say Ideogram is reining supreme as its best at graphics production for game dev. While not production ready, it gets 'beta' quality many times after a few tries, which helps the workflow. So hats the short explanation. Love the content, keep up the good work!
all of this is headed towards node-based modular animation. We'll be building animated motion stacks using modules of animated graphic information. Coding may come into it, compliments of Open AI or...something similar that's easy enough for the powerpoint crowd. Figma won't be impressed.
@@curiousrefuge Thanks. Just think ComfyUI with MJ generation and better interfacing/use of text + 3D. I imagine a UI like this would most likely dominate and instead of using NURBS, we'll have some other version of modeling enhancement since we're not using standard geometry. Or maybe I'm wrong. Maybe that in itself will be a node reference that connects to a prompt.
if there's one director that will never use AI it's Miyazaki! so that comparison is pretty unfortunate... I understand the message regarding giving more film makers a chance to express their ideas but please leave Miyazaki out of this... I don't think we will see a tsunami of master pieces just because people can produce whatever they want.
No we haven’t talked about Runway Gen 2 since gen-3 was released in July. The animation course is about how to create, model, and animate 3d models using ai tools and how to use ai to render using custom models.
You will have to correct yourself again. This isn’t the first AI film. The first one is called’ (I’m sorry but I can say the title due to NDA restrictions), a 140 minutes feature dome with the coproduction of many big players in the AI industry.
Hollywood is gonna love this. Race swapping a character without even needing to cast an actor from a small talent pool and then dealing with their destructive unscripted performances in the media.
Congratulations of filling up the AI Animation course.
Of course I just joined the waiting list for the next AI Animation enrollment. 😊😊
Im working on a 4 or 5 part ai series right now.first part is kinda over and people i showed said that its so hard to to tell difference between real or ai. Im glad i can fulfill my dreams because of the power of ai. And i need to say thank you the family of curious refuge. You guys make everything easier for me to select which system i should work with. If you guys host a meeting at Istanbul, I will be there for sure (I missed the runway meeting which makes me so sad because Runway is the GOAT right now i.m.o.).
which tools do you use? what did you spend?
@@peterk3697 I used chatgpt free plan to create prompts for my scenes, midjourney 35 dollars plan (but i used all of my 15 fast hours in just 10 days, i will change my plan), runway unlimited (100dollars), elevenlabs free credits (ı used 4 accounts which equals to 48K credits but i will buy the pro plan) , capcut pro for editing and epidemic sound music library for music and sound effect (its free for 1 month but when i start to publish my videos i will buy the pro plan which is around 30 dollars)
If we ever a host a meetup there we definitely hope you'll be there! :)
@@curiousrefuge DEFINETLY!
AI has helped me a lot in my work. I am very satisfied
Wow, Need to check out Wonder Dynamics!!! 🎉 Love the quality of your videos.
Awsome! thank for teaching us all
So true a 13:21 about Canva being a real competitor!!! Such a good tool!!! With so many good features
We think so too!
늘 좋은 정보 감사합니다 :)
Thanks for watching!
This was one of my older Wonder Studio tests from the best test days. I was able to import all the motion tracking data into blender and create a new camera and place the character into whatever scene I wanted. I still like having full control of my scenes but it's nice to know that some of these AI advancements are starting to make things a bit easier. That being said, I still like having the ability to export the animation and camera track for use in programs like Blender and Unreal Engine.
RunwayML is pricing its self out....right out of the market. Runway's video generate takes about 12 tries to get something remotely like your prompt...an epic waste of time and money.
Yep... It'll take hundreds or thousands revisions. I didn't know that until I bought the subscription 😭
thanks for mentioning this. After working in MJ for close to a year, I'm intrigued at the developments, but...there's still so much to learn with this. So much that I'd wish leaders would slow down. Nobody is going to use 3D/video in AI commercially with branding in 2025. AI needs to get alpha and text in grips before getting into depth.
Definitely a common critique and we highly recommend getting the unlimited subscription because of the amount of rolls.
@@curiousrefuge Or... don't reward bad behaviour. Especially when product unreliability is the business model. ie. "Pay us more if you want to get an acceptable result." 🤮
I knew the bench turn out weird as in the original image it was already showing. The engine should have extended the bench more, or replace the armrest. A consequence of adding it in to be fancy but not being able to deal with an extended imagine. Perhaps a good tool would generate the larger image from the get go (but kept from user) to prevent this issue.
We're excited to see how this develops over the next 6 months :)
Need to check out Wonder Dynamics!!!
Congratulations of filling up the AI Animation course.
Thanks!
Thank you💯
Thanks for watching!
the reverse shirt cult. kinda looks interesting, will have to check them out
Thanks!
❤❤❤just amazing!!
Thanks for watching!
Thank you
Our pleasure!
This is why you shouldnt subscribe to ANYTHING annually, things look different day to day in some case and who even knows what's going to be around and what wont in a year. Keeping up with this right now is crazy
Great video. thank you!
Appreciate you watching! :)
Hi, great videos. Maybe you have an answer to this: What would be the best workflow, to create act one videos with a AI generated voice rather than my own voice?
We would suggest using sometingl ike ElevenLabs to get your voice file as an .mp3 and then upload it.
5:27 While it is mindbogglingly impressive, I can't help to see how the sun did a 180 and suddenly flipped from facing to trailing them.
Weird as positioning the light source is more or less a staple of 3D animation.
Definitely will have some room for improvement but this is just the beginning :)
잘 보고 갑니다.
Thx for watching!
AI animation tools create movies so different from what directors have in mind that the great majority of them will quit making movies.
Interesting prediction!
Well it's not really "Video to 3d scene" if you have to import premade 3d assets. Actual video to 3d scene will probably come along one day but this seems to be only animation. Still pretty cool.
Fair point!
yes Wonder Studios ,!!! I will pay them
So cool, right?!
@curiousrefuge yes definitely
They literally had this on our cell phone for years. LOL You could be a farm animal and many other characters. They would do your facial expressions and head movements and recorded your voice speaking. They even had apps where you could make augmented special effect. LOL.
True but not quite like these recent tools in quality/fidelity
@@curiousrefuge Well, yes it's better and less work for sure. Still takes hard work to make quality stuff, however. Human's get bored.
Just a suggestion for "Audience Favourite" for next year's Halloween Contest. I suggest only allowing votes AFTER all entries are in, as SO MANY votes were cast during the process, meaning later entries didn't get the same exposure. As cool as it was, I highly doubt we get the same winner for audience pick if voting was fair and even, as most votes for it came simply because it was one of the first entries posted.
That’s a great point. We’ll do that next time. :)
Can't believe particlepanic's "unblind" didn't make the cut. Creepy AF!
So many good submissions!
As a new AI creator, do you recommend making a channel and going for it??
We simply suggest starting slow and working on the things that inspire you.
2:54 The image quality drops a lot.
Thanks for the heads up!
I think Ideogram beats Midjourney by a long shot. Sincerely MJ have never really stood up for my work cases, which is usable graphics for dev/game projects. For art, maybe holds up, but not my usecase. For usable things especially for text rendering, ideogram beats MJ.
Would love to learn more about your use cases!
@@curiousrefuge I do mainly social games, in the social casino categories and other similar things. Some trading apps and so forth. Ideogram generates crisper, and clearer graphics for inclusion in projects with less touch-ups required, shortening the production time. MJ as mentioned, have more issues with artifacts,, added distractions and so forth, which makes tidying up an image more time consuming if not near impossible. That's why I say Ideogram is reining supreme as its best at graphics production for game dev. While not production ready, it gets 'beta' quality many times after a few tries, which helps the workflow. So hats the short explanation. Love the content, keep up the good work!
all of this is headed towards node-based modular animation. We'll be building animated motion stacks using modules of animated graphic information. Coding may come into it, compliments of Open AI or...something similar that's easy enough for the powerpoint crowd. Figma won't be impressed.
Interesting prediction!
@@curiousrefuge Thanks. Just think ComfyUI with MJ generation and better interfacing/use of text + 3D. I imagine a UI like this would most likely dominate and instead of using NURBS, we'll have some other version of modeling enhancement since we're not using standard geometry. Or maybe I'm wrong. Maybe that in itself will be a node reference that connects to a prompt.
Why does all the realism dissapear when the character animates
Probably hard to render out that much detail for each of the frames required for the final animation.
Resolution of the output? Most of these are garbage quality lower then 1080p
It's difficult for the tools to nail down the movements. Definitely a lot of progress needs to be made in this area.
❤❤❤
Thanks for watching!
Can someone tell me what AI 🤖 model can generate consistent characters consistently.
Midjourney and also Leonardo Ai
We like Midjourney
A NU Hollywood is being created in real time.
We agree! :)
Problem is paper 📜 where is paper some website comming saying AI But Not realise paper
Sorry, not sure what you mean?
How to create a dance video for my AI model with expressions and lip syncing?
ComfyUI has a bunch of nodes that can help with this!
4 H100s and a decent machine are very different things hahahaha
True!
if there's one director that will never use AI it's Miyazaki! so that comparison is pretty unfortunate... I understand the message regarding giving more film makers a chance to express their ideas but please leave Miyazaki out of this... I don't think we will see a tsunami of master pieces just because people can produce whatever they want.
Time will tell!
Miyazaki has been anti-technology for a long time. He even hated CGI effects and that is now the industry standard.
it's gonna be so easy to do quality content that it will diminish it's value.
There will be a new tier of 'quality' : )
I don't unerstand you AI Animation course, you promote that every course is updated but you using a runway 2 🙂
No we haven’t talked about Runway Gen 2 since gen-3 was released in July. The animation course is about how to create, model, and animate 3d models using ai tools and how to use ai to render using custom models.
You will have to correct yourself again. This isn’t the first AI film. The first one is called’ (I’m sorry but I can say the title due to NDA restrictions), a 140 minutes feature dome with the coproduction of many big players in the AI industry.
Would love to see evidence of what you are referring to.
Try comparing Mochi1 to Runway in action scenes with people. There is no comparison. Mochi1 does awesome videos, Runway can not do action, full stop.
We'll check it out!
guys, im starting out as an AI enthusiast making similar stuff,
would appreciate your feedback!
Feel free to share your work on our discord!
Hollywood is gonna love this. Race swapping a character without even needing to cast an actor from a small talent pool and then dealing with their destructive unscripted performances in the media.
It will definitely be disruptive!
Can I upload an image of Trump, Biden or Zelensky to do that or it isn't possible at all?
Very often it gets flagged!
First 🎉
Thanks for stopping by :)
This is not the future animation because you still need expensive actors in order to make movies with this process
We believe it is at least the first step :)
ideogram rhymes with video spam (learn to pronounce it correctly)
Thanks!
learn to correctly pronounce Sotheby's
Woopsie Daisy!
super content but !! super boring voice as well
Sorry to hear you feel that way!
@@curiousrefuge you shouldnt be , your content is grade A+, its just to me its that , keep up the great work