Adobe Firefly Tutorial. BEST AI UI YET!
Vložit
- čas přidán 24. 03. 2023
- I have access to new Adobe Firefly! In this tutorial I'll show you how the tool works and what you can do with it.
Support me on Patreon to get access to unique perks! / sebastiankamph
Chat with me in our community discord: / discord
Control Lights in Stable Diffusion
• Control Light in AI Im...
LIVE Pose in Stable Diffusion
• LIVE Pose in Stable Di...
My workflow to Perfect Images
• Revealing my Workflow ...
ControlNet tutorial and install guide
• NEW ControlNet for Sta...
Ultimate Stable diffusion guide
• Stable diffusion tutor...
The Rise of AI Art: A Creative Revolution
• The Rise of AI Art - A...
7 Secrets to writing with ChatGPT (Don't tell your boss!)
• 7 Secrets in ChatGPT (...
Ultimate Animation guide in Stable diffusion
• Stable diffusion anima...
Dreambooth tutorial for Stable diffusion
• Dreambooth tutorial fo...
5 tricks you're not using
• Top 5 Stable diffusion...
Avoid these 7 mistakes
• Don't make these 7 mis...
How to ChatGPT. ChatGPT explained:
• How to ChatGPT? Chat G...
How to fix live render preview:
• Stable diffusion gui m...
I love your videos! They are always so well-made and engaging. I can't wait to see what you come up with next.
Thank you kindly, that means a lot! 😊
I'm blown away by the quality of Firefly. That's impressive.
Very impressive for a first stab at AI generation. BTW, Why can’t you trust an artist? Because they are shady… and a little sketchy.. and they’ll try to frame you.
Indeed! And they managed to release it in a format that's actually easy to use, compared to everyone else 🌟
But please, DRAW your own conclusions.
Good one bud🤣
😂
@@sebastiankamph this is the dad joke we needed at the intro
SO STOKED for this!! Waiting for my request to be accepted (you’d think those of us paying $$ monthly for years would get immediate access)
Hang in there. I got access the other day and I’m same as you- subscriber for years.
For sure! I've been told they are sending out mass-invites.
Oooh looks great!!!!!
Love it already.
As it seems to be trained with adobe stock photos, it really shines there. I had good results for product photography, shirt designs and background patterns. It is a really good tool for creating stockphoto like content, but at the moment, it is not detailed enough. But it is a very good first step, as long as you don't want to recreate anything from popculture, like marvel superheroes. And the banned word list is really strange sometimes.
Text and a transparent background alone makes it the best player in town.
The Text Effect is cute and the Text to Image is underwhelming and looks like more like Midjourney v3. The UI and tools are clearly Firefly's strength.
WOW. impressive and looks great
Thank you Sabastian😮
FYI that "dramatic lighting" in MJ is what is called "lowkey or low-key" photography.
Great video!
Letters generating are impressive but other features are just midjourney with extra steps.
So jealous you queue jumper :)
Hopefully the rest of us will be getting in soon.
the letter feature is the new word art :P
We'll see it everywhere, just like flickering SD animations 😅
That Level of detail in the letters was quite fascinating. Each letter felt like a little scene within itself, like the splash one actually kind of looked like little 3d scenes of water splashing within a container, while simultaneously feeling like framed portraits of an underwater scene.
The text thing is just nice for showcase, its actual real world use will be absolutely minimal.
Other features are indeed exciting.
Wait until you see it everywhere.
@@Oscaragious Oh god I really hope I don't.
Must say that thing you consider best thing in Firefly - constructing correct words and sentences , are existing for more of decade in 3DStudio Max (probably in free Blender too) .
Yes , you need to manually make textures for that letters instead of AI , but you can use free pictures of any material (for example fur, grass , leaves ...) , use them as texture and have same effect .
You can also inprint or project them on any object and follow contours of that object (or person if you make tatoos) .
Main selling point of Firefly is ease of installation and big company support which is important for people that already work in Photoshop , but do not know much else about computers .
BTW , love your videos although do not have strong enough video card to run Stable Diffusion (GeForce GTX 1050 with only 2Gb) .
Cheers 😀
thanks for the video. But what about the "Graphic" option / content type?
Whelp. That's it. Throwing Adobe Firefly expertise onto my resume now because if they're going to expect 10 years experience in things that were out for only 5, I can say a 15 minute video is enough experience for something that's not officially out yet.
Entry level jobs in 2025:
Required: 10 years experience in Adobe Firefly.
Does that mean I can say I'm literally an expert now then after a day?
@@sebastiankamph Considering your expertise in everything else, yes actually. lol
I can't afford anything adobe except adobe reader.
Hahhaha#
DeepFloyd AI will generate text to image with letters now. And it’s open source.
It still needs a lot of work to catch up to mid-journey, in the realism and quality of the generated images. But that's just my opinion, after using both.
Nicee.. can you change the font type for the font to image feature?
Yes! So far it's just some preset fonts, but the functionality is there.
will they have a free version when its released?
Could we do the letters with controlnet?
Can't test as I'm trying to fix automatic1111 on my machine
Yes, I made a video on that. czcams.com/video/1jBnGx72mpg/video.html
@@sebastiankamph Awesome! I just tried to do candies on the letter A, and it's not as good as Firefly. Thanks for your getting back to me.
what does it mean trained on their stock, at what point, after it has been trained on the stolen stock it seems XD
i can not joint the beta show me pls thanks
12:50 Buffalo Bill has entered the chat.
This is quite exciting! I'll wait a bit for their AI to become better and for them to implement it into Photoshop and then I may have to pay for Photoshop for the first time in my life! (Let's be honest, their software is way too expensive for broke hobbyists and paid subscriptions end up costing way too much in the long run.)
10 bucks a month is a pretty solid deal imho.
So the fireflies just gave up and build their own Ai App?! , Ellie is going to be so pissed
😂
👋
Also this is the only AI program I would consider using. I am a digital painter and while SD and Midjourney might be more impressive they are very exploitative of actual artists and photographers. I would even say parasitic considering how many popular prompts are the names of artists. Sorry but to actually be a tool for artists these things need to function like tools and not an attempt by tech corporations to take over entire industries. You should not be able to input an artist's style into a generator and have it use their art fed into it without their consent. (and before everyone gets mad at me with "artists reference too", software programs are not people, these things aren't even actual AI they're just advanced algorithms with images fed into them)
The hands didn’t scare me in this video like most other human gen and teeth just weird looking.
I hate having to use text prompts for everything and so do many artists. It is the main thing that leads to feeling a disconnect between yourself and the software and feeling like you lack control over details. If they want us to buy that these are tools and not a replacement they need to find a way to integrate AI generation into a UI more like traditional digital painting software.
ok
I could counter that by saying I hate spending years learning to master a software and be at it's mercy it doesn't change the UI too much so I have to learn it over again. Times change, move with them, or be left behind. You're stuck in your ways and want familiarity, I get it, but the world doesn't work like that.
Within a year I fully expect integrated vocal prompts. I disagree with your "traditional" stand. EVERYTHING evolves, good or bad, that's a personal opinion. Tradition is something which means "dated". AI is going to take over, you go with the ride, or get left behind at the gate. Firefly is way too censored, but it's Adobe, what can you expect. Won't be long until there is a "firefly" model for stable difussion.
I do understand there will be those who will stick by adobe, that's fine. The very fact Adobe are doing this (e.g they follow the money) tells You how much influence AI has. See like people downloading movies in the early 2000s. The movie companies tried to sue, etc they tried everything to stop the downloads, not for the reason you think, they didn't want CHANGE. Change happened, and now, low and behold, people DOWNLOAD (stream) movies daily. So torrent technology was way ahead of the movie industry, they want to control everything, but didn't know how to control this.
Now these companies realise, adopting instead of resisting is far easier all round. A company called fantasy ai (I think) is trying to bully the market, but because of how models work, they may indeed be digging their own grace with their approach. The AI community is big, it's friendly, but you don't want to rub them the wrong way. I fully expect fantasy ai to disapear or change their tact pretty quick or be blown into oblivion.
@@marksutherland774 No I am not stuck in my ways, I am not impressed by what is in fact massively inferior software in terms of customization potential. Having to use prompts exclusively is extremely limiting in terms of the amount of control I have over the final product.
The AI community is not "friendly", they are extremely greedy and amoral run of the mill venture capitalists and defensive of the massive amounts of art theft they commit. Plus they WANT you to be scared of their tech (see that completely insincere "we need to pause AI development" publicity stunt letter) and "change or get left in the dust" is a line they tout to try and make you feel powerless to try and regulate them. Acting like they're a force of nature that can't be criticized or regulated is giving them way too much power, sorry. They're not gods.
AI art tools are not going away but THEY are going to have to adapt if they are going to survive the massive amount of critique and pushback they're facing from artists and the public at large right now. One of those criticisms being the weird, niche interface they currently have. If they want to win over artists maybe stop trying to replace the creative process itself and structure these tools in such a way that they don't exist to stunt the entire industry (as "prompt engineer" is a completely unmarketable skill since everyone who knows how to do it, which takes a day or so max, is inherently completely interchangeable with everyone else).
BTW, torrenting is essentially dead right now, and streaming is not torrenting or downloading because it doesn't give you a copy of the work you're viewing. Media piracy is a shell of itself (not gloating over this, I used to pirate everything). It took a while, but the corporations won.
@@choptop81 You don't get it. Take torrenting, yep I agree it was illegal but it forced a change for the better, for the consumer. £15-20 a dvd/blu ray, it costs pennies to make, say 75p all in, of course you have the cost of making the film and distribution. Of course they'd then skin you for the "special edition", "directors cut" etc etc
What torrenting did was effectively introduce a method of delivery, we call it streaming now where the distibution cost is very small compared to delivering dvd's across the world.
You may well be right, the corporation won in the end, BUT only because they took the model of torrenting on board, but they lost billions while they desperately clinged to the old delivery models.
You could stream torrents, e.g watch it as it downloaded. it's where streaming came from, napster was one that did it.
Here's a quote from the wikipedia page about napster "Napster paved the way for streaming media services and transformed music"
You say AI community is greedy, well let's counter that. Artists need inspiration, they usually take that from the style or combining styles of other artists. Yes, some will genuinely come up with unique styles, it's rare. Look at anime, take samdoesart, he literally changed the style of the eyes and a couple of very minor touches, it's anime, he didn't create anime. So artists are greedy in trying to keep everything to themselves, fearing technology, it's a pointless agruement of course, either for or against.
These tools, and firefly being for the "corporate" world, are not designed to "win over " artists, they couldn't care less about artists. You don't use it, fine, there will literally be thousands to your one who will find AI art tools miles easier to grasp than the software you use. Are you more skilled, absolutely, but, who really cares if you are, very few.
Models is where the artist can create their style and get paid for it. Charge a pounds for a model of your style, fair play. I pay for various online AI tools, I have stable difussion installed locally and plenty models. There is nothing to be scared of the tech, it's there to be used and, if you learn to use properly, will enhance your work, doesn't need to replace how you do it but learning it will enhance how you already do it.
In the end it's each to their own. The AI "PAUSE" letter is more aimed at the likes of chatgpt, be assured the government will be using a version wit uptodate date not mid 2021.
So let's look at it simple terms, creating art took years to learn to produce quality images, now it can literally take 5 minutes. So instead of being critical, you could create a model of your own style, keep it to yourself then use it for inspiration to create a manual version of something it has produced for you. For this you need to install stable difussion, not hard, create the model, not hard on a site like astria.
And yes I know torrenting is dead, it was an example of how , what you could deem "illegal" software was actually way ahead of it's time and paved the way to modern streaming, which napster, a torrent software, did.
No I didn't use chatgpt to write this (it would have done a better job) and good luck with whatever you're doing.
@@marksutherland774 Whatever man. If you don't think AI devs are greedy you're either flat out deluded or an absolutely pathetic corporate simp. Or both. I hope you're aware that your suggestion of artists monetizing their style for image generation models would send the devs of SD and Midjourney into conniption fits.
Software algorithms =/= humans. Directly feeding someone's art into an averaging algorithm is not comparable to looking at it and taking creative inspiration from it. Originality isn't even that important as much as it reflecting your tastes and importantly the labor you put into it. Greg Rutkowski and Artgerm aren't terribly original stylistically and their stolen art is still some of the most popular prompts.
I am not blanket anti-AI art and that's clearly obvious from my comments so I don't even know who you think you're arguing with. What I'm saying is this stuff IS going to get regulated, possibly severely, because it is actively anti-artist at this point.
Release date?
Unknown!
Not ever paying for Adobe subscription services. It looks OK but nothing special. Plus it's censored to hell.
SD Automatic 4 life.
Is this free to use
The beta is free for now. Have no idea about their pricing structure for the future.
1 where was your last video?
This one: czcams.com/video/TInvekF6NRw/video.html
Wow , this is even better the my imagination , ppl will now stop to tell me "this is not art" finally
That's what happens when models are trained with HIGH QUALITY stuff.
No dad joke 😳😩😬
No dad jokes??😠
ps. Great beta review, though👌
I asked the alpaca if he wanted to go on an adventure. He replied: Alpaca my bags and let's go!
@@sebastiankamph 🤗
0:43 - Incorrect. They can all do perfect text like Firefly. Just use a mask like Firefly does.
There's no good automated tool for that with live editing. I have a tutorial on text in Stable diffusion but the workflow is not nearly as good as this.
12:40 the girl Is italian. Look her left hand.
By the way, I'm afraid it will cost an arm and a leg per generation, just like Adobe stock.
Bro it's basically stable diffusion but you have to pay for it and with so little customization 😢😢😢 it sucks o rather use stable diffusion
12:40 Italian Nazi
Remember people, never support ai arts with your money or praise as they already destroying what is great about humanity which is "creations" and quality craft with thoughts and life knowledge that brought us happiness. AI art doesn't have any clue or idea for our inspiration and greatness as much as a living human itself.
What corporations and ai programmers have done is not okay at all. It's getting worse and worse of how they can just insult us all with these ai arts techs that will make our entertainment media worse.
its meant to be a tool, not actually replace digital artists, yes u can generate a prompt in any style, and itll look nice, but for now the output is quite random, adobe programs have been evolving over the years making it easier for a designer to do their job, this is just another tool that can help us speed up things, ive been a designer for 18 years, and dont think itll replace us.
@@visualdosage234 All I see these ai arts thing never being featured in an art programs. But then I'm not that artist that will find it useful to be honest.
what does dripping shit look like?
i didn't like it
very nice. Have you heard of Bard? From google
I have, made a video on it when it was announced. Haven't tested it.