Now this is how I like seeing AI used. It's not a tool that does it for you, it's a tool to help you. I used to spend hours looking for inspiration online. Now I can generate inspiration, or assets to further tweak in my art the way I used to seek out stock photos.
That's how all of us AI artists are doing it. You only see the finish product. Artist don't want to show people the process because they don't want to bore people.
Some times I get worried about art in the future with AI. But its intresting to see how the creaters taste changes everything. I would have chose alot diffrent then what you ended choosing. Goes to show that AI is going to be a very strong tool in the future.
Damn, I love being among the first to discover channels that are def going to be HUGE! This video was amazing. Narration on spot and super interesting. Can't wait for your next vid!
Great video. Looking forward to seeing how to utilize this character for animation sprites, jumping, walking, attacking, falling, etc. The process of generating images one by one until reaching the desired result has left me discouraged. Just when I think I've found the right prompt, the next time something totally different comes up, wrong. It seems to be like playing the lottery. An exercise that requires patience.
Hi! Pixel artist here, wanted to check in on this comment and see if I'm any more replaceable now than I was when this was written. Gimme a reply if you notice this.
@@RobertDrane Hello friend. My results are far from what I want. Far away! Very far away! I feel that AI gives me a starting point. I play with possibilities and collect ideas. At the end of the day, if I need something very specific, I'll need a real artist (or I'll do my own art, since I have no intention of spending money on it from the start). But mind you, this is already the beginning of AI, and there's a lot of advancement out there, but right now there's nothing to worry about.
Wow, the best video for usage of Stable Diffusion I've seen so far. Earned a new subscriber, please keep it up with new stable diffusion designing content
Fantastic work! would love to see you taking this process one step further and taking it into a game engine to see it in action, something you think you'll do in the future?
@@pjgalbraith if you're looking for a quick and dirty way to get something like this up and running, Thomas Brush has a free 2D gamedev toolkit for use in unity. I reckon you'd be able to get something really good up and running in no time using that :)
Using AI to make concept art for future sequels? This NEEDS to be explored Halflife Portal Team fortress Witcher Call of duty Devil may cry Resident evil Any game you can think of Imagine how much development time would be cut at minimal if any, no labor cost whatso ever just to get a solid concept out? The current trend with AI images right now is seemingly turning beloved games into 70s-90s dark fantasy movies. Well my friends i think this is a new trend that could not only give us viewers something nice to see and ponder about. But also could open up the possibility to benefitting the game industry. WHAT OTHER GAMES WOULD YOU LIKE TO SEE MORE OF??
Hey, this is amazing. I'd love to see more videos and tutorials like this. What would also be interesting is taking images from Midjourney and putting them through stable diffusion to render it for background or character models. :)
Incredible video! thank you. Would love to see it broken down into a series detailing the steps further. Still an amazingly instructive. One question, is there a way to get SD to produce an image with transparent background? haven't been able to do that.
So the AI doesn't understand or produce RGBA (i.e. doesn't understand transparency). But there are extensions like this that produce a masked image using a high contrast depth map github.com/Extraltodeus/depthmap2mask#alpha-cropping
Any suggestion on how to animate the character? What would be the easiest inside unity? Thanks! Tried some prompts to split the character into smaller parts to bone animate but no luck :(
Let me ask, if I have 3 in-game item images (guns, armor, medical boxes...) with completely different styles, how can I make Stable Diffusion synchronize the styles? Thank you
Thanks! Can't remember exactly how long it took sorry. I did a 1 hour challenge video on my channel so check that out if you want to see what's possible in that timeframe.
Wow, this is outstanding, thanks for the video! This is the first reason ever to join the dark side of the universe - forced to switch to Nvidia after 10 years using AMD GPUs 😭 Sadly AMD self-hosted solutions are still way too behind CUDA oriented ones, tried here but the quality/speed is way subpar.
Wow this is so good! Was thinking of doing pixel art but since the models are not there yet this might do for now! But how do you make the sprite sheets though?
So... while SD is cool and all, it still looks like more work than just doing without, I guess the artist zeitgeist for 2023 is going to be something like 'use SD for inspiration', but it'll slowly fade as the limitations in control and timesink become more obvious to everyone.
Yeah it was never the magic tool people made out. Although it will improve rapidly. That's why it's important to actually study and understand the capabilities and limitations. If you see some of my more recent videos I show off a hybrid approach.
Yeah there's a lot more that goes into making a game. At the very least this allows for quick prototyping and is a lot better than the crappy placeholder art I've used in the past during development.
Yeah good point that was a struggle. I would probably generate a bunch of separate pieces and create a more detailed composition from them but I wanted to keep the video short and just get the ideas across.
Hey, I think you should try out the depth2map plugin for automatic1111. It really helps with creating alphas that way you won't have to erase your elements as much and can get some cleaner lines
50 years later... man lose the ability to draw without AI help ... sad~ its the same phenomen that ... women nowadays can't look good without putting on makeups..
Now this is how I like seeing AI used. It's not a tool that does it for you, it's a tool to help you. I used to spend hours looking for inspiration online. Now I can generate inspiration, or assets to further tweak in my art the way I used to seek out stock photos.
That's how all of us AI artists are doing it. You only see the finish product. Artist don't want to show people the process because they don't want to bore people.
Some times I get worried about art in the future with AI. But its intresting to see how the creaters taste changes everything. I would have chose alot diffrent then what you ended choosing. Goes to show that AI is going to be a very strong tool in the future.
Damn, I love being among the first to discover channels that are def going to be HUGE! This video was amazing. Narration on spot and super interesting. Can't wait for your next vid!
P.s textual inversion makes it even easier to get generations to look like the style you want
How would I go about this?
The stable diffusion interface I was using from RunPod had a button for this. I'm not sure what project/git repo they use.
WOW! It really shows that if you are already good at something, AI just makes you 1000x better
*1000000000 times
@@progCan 1000000001 times
Great use of stable diffusion! Now for the tricky part: how do you make the characters walk?
U need to creat animation for walking
You can use bones in unity or other programs.
Well this video was a complete mock of my time. This is not Designing a Metroidvania, nor a Game, instead a singular Concept Art for such.
Honestly can't wait to see more tutorials from you!
This is the workflow that I'm looking for, as a technical person I always struggle with Artistic side, this is amazing, Thank You!
this approach is amazing, cant wait to play around with this my self
Your video prove that AI helps designers, not replacing them.
This will make so easy for people who make indie games
Great video. Looking forward to seeing how to utilize this character for animation sprites, jumping, walking, attacking, falling, etc.
The process of generating images one by one until reaching the desired result has left me discouraged. Just when I think I've found the right prompt, the next time something totally different comes up, wrong. It seems to be like playing the lottery. An exercise that requires patience.
Id recommend learning how to paint over stuff, definetely helps a ton, AI does the heavy lifting anyway
Hi! Pixel artist here, wanted to check in on this comment and see if I'm any more replaceable now than I was when this was written. Gimme a reply if you notice this.
@@RobertDrane Hello friend. My results are far from what I want. Far away! Very far away!
I feel that AI gives me a starting point. I play with possibilities and collect ideas.
At the end of the day, if I need something very specific, I'll need a real artist (or I'll do my own art, since I have no intention of spending money on it from the start).
But mind you, this is already the beginning of AI, and there's a lot of advancement out there, but right now there's nothing to worry about.
Wow, the best video for usage of Stable Diffusion I've seen so far. Earned a new subscriber, please keep it up with new stable diffusion designing content
Thanks I've got a lot more content planned! I'm obsessed with Stable Diffusion right now 😂
@@pjgalbraith perfect, looking forward to it. Maybe something with Dreambooth and consistency of characters?
Fantastic work! would love to see you taking this process one step further and taking it into a game engine to see it in action, something you think you'll do in the future?
Yes, please
Thanks! It's been awhile since I've done game dev but I might give it a go to see how well it works.
@@pjgalbraith if you're looking for a quick and dirty way to get something like this up and running, Thomas Brush has a free 2D gamedev toolkit for use in unity. I reckon you'd be able to get something really good up and running in no time using that :)
@@pjgalbraith Unity pleaaaaaaase :)
Wow! Great style and consistency you got out of it😍
Thanks was a fun experiment for sure.
Now THIS was really cool, and goes to show HOW GREAT AI can be... and already is!
Awsome I'd like to see a video teaching all tips and tricks in photoshop, step by step, brushes, workflow, how to illuminate your scenes, etc
that was a great and inspiring video! taught me a lot.
Using AI to make concept art for future sequels?
This NEEDS to be explored
Halflife
Portal
Team fortress
Witcher
Call of duty
Devil may cry
Resident evil
Any game you can think of
Imagine how much development time would be cut at minimal if any, no labor cost whatso ever just to get a solid concept out?
The current trend with AI images right now is seemingly turning beloved games into 70s-90s dark fantasy movies.
Well my friends i think this is a new trend that could not only give us viewers something nice to see and ponder about.
But also could open up the possibility to benefitting the game industry.
WHAT OTHER GAMES WOULD YOU LIKE TO SEE MORE OF??
That's a great idea!
Hey, this is amazing. I'd love to see more videos and tutorials like this. What would also be interesting is taking images from Midjourney and putting them through stable diffusion to render it for background or character models. :)
very very gooood, thaaankss maan✨✨✨✨✨✨✨✨✨✨
Thanks for reminding me of the "Tiled" button. It got buried in newer A1111 versions in the settings.
That was nice to watch!
Incredible video! thank you. Would love to see it broken down into a series detailing the steps further. Still an amazingly instructive. One question, is there a way to get SD to produce an image with transparent background? haven't been able to do that.
So the AI doesn't understand or produce RGBA (i.e. doesn't understand transparency). But there are extensions like this that produce a masked image using a high contrast depth map github.com/Extraltodeus/depthmap2mask#alpha-cropping
How would you create a sprite sheet to animate the character?
One word, beautiful
This is dope work!
Woah, awesome video. I would love to see you going into how you do all the stuff in photoshop, even if it ends up a hour+ video.
Installed, everything works, thanks!
Not so much a game but an image of a single scene of a potential game.
I checked - everything is clean
Any suggestion on how to animate the character? What would be the easiest inside unity? Thanks!
Tried some prompts to split the character into smaller parts to bone animate but no luck :(
I'm not aware of any automated methods at the moment unfortunately. Bone animation should work but might require painting in missing parts manually.
BROTHER, YOU ARE THE BEST!!! You oooh really helped me!! THANK YOU VERY MUCH!
mindbendingly cool. thanks for making this
Let me ask, if I have 3 in-game item images (guns, armor, medical boxes...) with completely different styles, how can I make Stable Diffusion synchronize the styles? Thank you
i would definetely develop a game with this art style, it's so awesome...
btw, how long did you take from zero to the final product?
Thanks! Can't remember exactly how long it took sorry. I did a 1 hour challenge video on my channel so check that out if you want to see what's possible in that timeframe.
What SD Checkpoint were you using?
Wow, this is outstanding, thanks for the video!
This is the first reason ever to join the dark side of the universe - forced to switch to Nvidia after 10 years using AMD GPUs 😭
Sadly AMD self-hosted solutions are still way too behind CUDA oriented ones, tried here but the quality/speed is way subpar.
Wow this is so good! Was thinking of doing pixel art but since the models are not there yet this might do for now! But how do you make the sprite sheets though?
good upload.
Once AI can animate sprites then sure .... However for background elements its certainly possible
Nice tuto
So... while SD is cool and all, it still looks like more work than just doing without, I guess the artist zeitgeist for 2023 is going to be something like 'use SD for inspiration', but it'll slowly fade as the limitations in control and timesink become more obvious to everyone.
Yeah it was never the magic tool people made out. Although it will improve rapidly. That's why it's important to actually study and understand the capabilities and limitations. If you see some of my more recent videos I show off a hybrid approach.
So cool! Thanks!
Good tips. Now if it were possible to connect an AI to Godot (for example) and get it to help with the actual game engine side...
Can midjourney also do this? It's cheaper than buying a computer that could run stable diffusion for me.
thank you very very very much, you help me a lot
what model did you use?
Are there any tricks to animating the sprite and making it look decent?
Yea its just for art-work only. Making the game unique will need a lot of more work unless you just want to do a mod by replacing art works 😁
Yeah there's a lot more that goes into making a game. At the very least this allows for quick prototyping and is a lot better than the crappy placeholder art I've used in the past during development.
What is the website you are using?
How do I become confident enough to do all this since I don't organize folders well to find on time
amazing!
Your music taste is impeccable! I feel so pumped working on my own SD projects while watching yours as mine renders. Any playlist recommendations?
As much as I'd like to take credit the music is all from UA-cam's free music library 😂
@@pjgalbraith No shame! UA-cam royalty free music is tha bomb :)
How do u use the reedit to open and execute de website that have the IA?
What AI did you use?
Looks great although not a fan of the pile of bones bc it's not very cohesive of what bone starts where and where does one end
Could have done a couple single bones to throw in there to make atleast a few of them look like there not fused together in 1 big glob of bone mass
Especially the center piece head could have been singled out and put back in
Yeah good point that was a struggle. I would probably generate a bunch of separate pieces and create a more detailed composition from them but I wanted to keep the video short and just get the ideas across.
I did more detail work in this video ua-cam.com/video/h12JjO9cf20/v-deo.html
That was awesome ❤
By the way just to let you know that your video is recommended by youtube right now kuddos
Hey, I think you should try out the depth2map plugin for automatic1111. It really helps with creating alphas that way you won't have to erase your elements as much and can get some cleaner lines
Yeah I've been working with it lately it's pretty good! It wasn't around when I did this video unfortunately.
That’s so epic 💎 need to buy Nvidia RTX card ASAP 😀
Now I just need something other than a mouse to draw with.
Good quality of video, deserve more views
Thanks appreciate it!
@@pjgalbraith AI is really a future for indie developers. Me and my friend us SD to make placeholders in our Tabletop Games
Nice, but animation?
anyone know how to fix it?
Way to show how to use AI to augment your artistic vision
proving that AI imaging software will not take jobs away from people. Like all tools in the digital world, you have to make them work for you.
for you
I’m amazed how well tNice tutorials is explained. I’ve never heard a lot of these words, but the order and purpose of these nus and buttons
Wow!!!
WIN 11 WORK
you just made one image. a game is so much more...
it seems like you could have saved a lot of time by just making the assets you wanted and not using stable diffusion at all.
I watched tNice tutorials 18 minute video for like 2 hours.
!
not really a game tho, like a poster really
Sorry If I'm not in the hype of the comments but you designed a concept art, not a game.
🤷♂️
Would take less time to draw it on your own, tbh
Says "game designing for free"
Yesssss
.........starts up Photoshop 2 minutes in !
Nooooo😂😂😂😂
It's not required at all, it's just what I'm familiar with, in some of my newer videos I mention free alternatives.
@@pjgalbraith just yanking your chain ;) i use alternatives, got krita gimp an corel draw, no worries. liked the vid though .
@@eelco_de_haan all good gave me a chuckle
Can't wait for them to optimize it further so I can do 512 x512 locally can only do 320x320
And da quality is ass so I'm sticking with dreamstudio and disposable emails
Also other various web apps
50 years later... man lose the ability to draw without AI help ... sad~
its the same phenomen that ... women nowadays can't look good without putting on makeups..
Really nice!
creative comnt...........................Thank you!!!!!!