Adobe Firefly Masterclass: Get into Gen AI with Howard Pinsky
ฝัง
- เผยแพร่เมื่อ 16 พ.ค. 2024
- Join Howard Pinsky for a deep dive into Adobe Firefly! Learn about our text-to-image generations, Generative Fill, Recolor, and much more!
Howard Pinsky is a Senior Creative Cloud Evangelist at Adobe: www.behance.net/howardpinsky
Music by Chillhop Music: chillhop.ffm.to/creatorcred
-----
Join us LIVE on Behance: www.behance.net/adobelive
00:00 Start
1:58 Starting work in Adobe Firefly
4:37 Prompting Adobe Firefly
9:40 Pushing Firefly to an advanced direction with prompting
12:17 Generative Fill in Firefly
13:22 Removing part of the generated image
22:52 Text to Image in Adobe Express Beta
24:53 Photoshop Beta: Creative Cloud Desktop App
27:33 Content Aware Fill vs. Generative Fill
35:04 Cleaning up an image
41:29 Restoring an old photograph
42:39 Spot Healing Brush
44:36 Generative Expand
49:45 Selecting channels in Adobe Photoshop
52:04 Generative Fill the pickles
53:26 Generative Expand for videos in Adobe Photoshop Beta
The Photo Restoration and Generative Expand options and processes are incredible.
Excellent video! I will definitively use these tools in my work! Thanks
The one thing I've noticed with Firefly: adding modifications to a gen, either in the prompt or styles/materials, will often use the existing compositions as a base. I'm often seeing the same forms & compositions repeated over multiple generations. If you want to completely refresh the image from scratch, you have to hit Refresh a couple more times. It would be great if Firefly came with a generate option for either 'Refresh' or 'New Seed', especially moving forward when each generation will eat up your credits.
My goodness you knocked this Masterclass out the park! I typically don't watch these Live videos because I find the pace lacking. However this was so concisely done and you covered a ton of ground. Thank you for your effort!
Glad you enjoyed it!
I love the interior design potential this AI has! Mind blown!
Thank you, a lot of technique and inspiration! Best regards from Germany
this is a great video! thank youuuuu!!!
Thank you for the tips! Watching from Malta 🇲🇹
I just love these videos. It’s giving me new ideas to take into consideration when editing a photo. The abilities are endless!
That's amazing! So glad you're enjoying the videos
Amazing video. I got a great intro to the tools and learned so much!
I wonder whether there is a way to use the GenFill to simply remove lipstick without changing the shape of the lips. Over all, I wish the GenFill function would have more options to control the command, as it mostly feels rather clumsy. Especially when performing subtle tasks.
Absolutely amazing!!! Thanks!
Regarding using the 40% mask trick/grey layer trick to create the oil painting look (and I've also seen it used for retouching faces etc.), is there a way to use that trick or something similar but with Firefly and no Photoship?
Absolutely Amazing!
awesome!! I love the video feature of expanding the canvas.
We love Adobe.
I feel like the Text Effect is an old school thing that was dug out of the attic. There are better options out there. The text to vector is pretty cool for logos and branding.
A great option with generative fill would be to be able to freeze certain areas that you do not want changed like with liquify (unless this is already possible and I cannot find it). Great tutorial. I usually enjoy your videos. This tech is wild indeed.
You can do that with masks
BIG THX for sharing this , little struggle but in the end I made it! ;)
What's the best way to download/upscale these generated images for commercial purpose?
Hey, first thing: thanks for great video, it helps a lot to grasp techniques to use ff efficient way.
Now QUESTION for a million: once generations will become limited to specific number, will each iteration consume the limit of credits even while I didn’t yet got the image that I need?
Or a single credit is for the moment when I finally got what I need?
Of everything ive see so far from Adobe, Photoshop is the best so far.
Watching from Atlanta GA
Thank you!
Question: In Photoshop, does leaving the 3+ different options of Generative Fill in each of the layers add a ton to the overall memory and size of the working file?
Yes, it’s better when you choose preferable option to delete the rest.
thank you
I used now and i explore.
Hello, It's excellent and very nice video to teach easily for every one.
How would we use this during live video calls such as zoom meetings, etc to create custom ai backgrounds for live video?
Thanks alot :)
BOOPcorn... classic Pinsky!!
B O O P
From Nigeria
Could i use this for physical product rendering
🎉witam za duzo pani gada pozdrawiam .
Is anything real? Idk. Lollllllll Love playing with this program. And it’s just the beginning isn’t it? I feel like if I don’t jump on the bandwagon now, it will become so overwhelming I’ll be totally lost. Helpful video! Thanks!
great tool
Time to play catchup!
Do you have to buy all these programs? Can Canva do it all at one price
Only Canadians are gentle and kind enough to call Floh-duh "a great place"! 😀
very good
Her breakfast combinations are wild… either ice cream with orange juice or fruit with coffee 😅😅
I don't know but from my experiences the removing background still doesn't give me the result which I would expected.
I have to leave and get some work done. Where can I find this video later in the day?
download all? See favourites?
Hi. what if i want to keep same subjects/desings to make a story. for example i like an elephant and a kid standing in front of the elephant with a kind of vectors/ ilustrartion look. now i want to generate the same kid riding the same elephant. i give FF the first image as style reference adn it did not worked. did you ever try somrthing like that.
Has the Generative fill in photoshop changed since this video was made? I tried to remove a person from a photo, so added no text prompts, just clicked 'Generate' and it kept replacing my person with some weirdly freaky people.
Brother it was suppose to do the Adobe firefly master class and more than half class was about photoshop😧
🎉
How can I get firefly Adobe is it iOS Mac book or pro ipad
13:06 Making me hungry 4:30 am!
BJJ isn't tiddlywinks. You are going to get injured. Lots of back pain, lots of fatigue. But lots of highs too and a way more interesting life. BJJ is great stress relief. My advice to anyone is, if there's something wrong with your body or your mind, take time out and fix those problems. Then return to train. Time out is great.... you can come back rejuvenated with new focus and in 5 or 6 weeks, it'll be like you've never left.
I appreciate the effort but pre plan. It felt completely unorganized. Once again....thank you for your efforts!
Burger looked like a vegan burger...lol
Hi from Costa Rica, you know?
🎯 Key Takeaways for quick navigation:
00:00 📜 *Introduction to Image to Line Art Workflow*
- Introduction to the challenge of turning any image into line art.
- Mention of the workflow using Comfy UI and generative AI.
- Emphasis on the complexity of the task.
01:01 🧩 *Essential Requirements for the Workflow*
- Listing the custom nodes and models required for the workflow.
- Instructions on installing necessary custom nodes and models.
- Highlighting the importance of having these components.
02:09 🧰 *Setting Up the Workflow Nodes*
- Explaining the initial setup of nodes for the workflow.
- Grouping nodes into categories such as models, image and resolution, prompts, and more.
- Describing the purpose of each node in the setup.
16:29 🔀 *Connecting Image and Control Net Processing*
- Detailing the connections between the loaded image, control net, and other nodes.
- Explaining the importance of maintaining image resolution.
- Mentioning the image processing flow in the workflow.
26:50 🎨 *Adding Post-Processing Effects*
- Introduction to post-processing effects, including upscaling.
- Connecting nodes for upscaling and enhancing the generated line art.
- Explaining how to fine-tune the results for different image types.
36:59 🖼️ *Examples and Practical Usage*
- Demonstrating practical usage with various image categories.
- Highlighting the flexibility of the workflow for different image styles.
- Mentioning additional nodes for background removal and customizing results.
Made with HARPA AI
Looked for the first thing in your list: "Introduction to the challenge of turning any image into line art" ... it's not at 00:00. Clicking on the time stamp link takes me to where he's restoring an old photo. At least that first time stamp doesn't match, so don't think the others will either. Not sure where these time stamps came from.
Firefly is free for three days. After that 398 rupees for one month.
The only way I can see the problem is
👍🏼👍🏼👍🏼👍🏼👍🏼👌🏼😎
😎
Working on all these 'things' but never quite finishing them... too buggy, slow, bloated or gives unruly 'warnings' on benign image content. Why can't you just FIX WHAT YOU STARTED?
McDonald from Zambia Africa
As long as the image quality of the generated expansion areas is as poor as it currently is, this functionality is unfortunately hardly usable.
Szkoda ze nie rozumiem,znam tylko włoski!Pozdrawiam p. Posła!!!
супер
Adobe are making a serious attempt to kill the artist in humans. Some are going to say that it will inspire people to make art. Really?
Typing "gorgeous blond woman, laying in the grass, an impressive castle in the background, a sunset" is not art.
What's next? Typing a few words, and you got yourself a movie? Replace human actors by AI actors?
Where is all of this going to lead to? Who the heck wants to live in such a world in which AI is even replacing musicians, singers, authors, actors, newsreaders or even teachers?
Humans are supposed to be pleased by humans, and not by AI. What a fucked up world it's going to be. Oh wait, it already is.
It's art, perhaps not to your taste, but it is unmistakably art. Art is a vast realm of expression, encompassing a multitude of styles and forms. This dynamic field witnesses an ongoing interplay between classical traditions and the emergence of new, contemporary expressions.
For example, think back to when Dadaism first appeared; many people didn't get it. They thought it was strange and didn't fit the traditional mold. However, it paved the way for innovative movements like abstract expressionism and pop art. Even graffiti art was met with resistance at first, and now we have digital art, which is becoming more and more popular.
Some people engage with art casually, while others discover hidden talents within themselves. Art is an inclusive world, allowing participation even for those without traditional artistic skills.
In essence, art, while it may not be everyone's favorite, remains a space for diverse and profound expression. History shows us that resistance to new art often turns into appreciation over time, making art an ever-evolving tapestry of expression.
@@kamu747in the loosest sense of the word, it might be considered “art” but it isn’t your art. How can you be an artist without artistic skill? Artistic skill is what defines an artist. Typing “Teddy bear sitting in the snow wearing cowboy boots” doesn’t take artistic skill and it isn’t art. It’s akin to taking your car to a mechanic and asking them to rebuild the engine. A week later you’re going around telling everyone that you’re a “master mechanic” that just rebuilt their engine? You’re neither a master mechanic nor someone with the skill set required to properly rebuild an engine but it’s fun to pretend… I guess. ai is not a tool - it is a replacement. The more ai is used and relied upon, the more creativity will become stagnant. Why wouldn’t it? You’re asking a non-human algorithm to predict and define something that is inherently human.
@rh9967 @rh9967 I "hear" you or rather I understand your concerns. Artistic skill is subjective. At 1st glance and to some in general, a Rothko painting might lack artistic skill. Some may view it as scribbles and paint splash.
Anyway, I've had this discussion countless times & as much as I'd like to engage you in this wonderful discussion, at this precise moment, I find myself a little tired. So please accept the following text, which I'm copying and pasting here from one of my previous back and forth discussions on youtube comments. I'm disclosing this in case you ever encounter it elsewhere. But it addresses precisely what you're talking about, I say what I say from experience, having extensively made an enormous collection of AI generated art and from interacting and appreciating other artists' works, its out of respect of your opinion (for I once held the same thought) that I'd like you to read the following so that maybe you might understand the implications of AI generated Art in the context of Artistic merit, its long, but I hope you'll read it:
In the realm of art, there exists a spectrum of sophistication. On one end, there are those who provide generic instructions such as "Cat with a hat at the Beach," resulting in creations that fall short of established artists and animators. However, at the other end of the spectrum, we encounter highly skilled digital artists who have elevated prompt engineering to an art form worthy of awe. These creators infuse their AI-assisted designs with the essence of their own minds and the passion of their souls. These outputs truly stand out amidst the sea of uninspiring AI-generated works. They deserve our admiration, and their efforts epitomize the essence of true artists.
Just as anyone can wield a pencil and paper without skill, they may be categorised as amateurs in the world of art. Similarly, in the intersection of art and AI, there are amateurs and there proficient practitioners, commanding deep knowledge of prompt guidance and complex instructions crafting genuinely original creations that can't be duplicated It's essential to remember that when discussing AI-generated art, the ultimate genesis lies within the human mind, as these creations wouldn't exist without human initiative.
Uniqueness is another facet of this fascinating landscape. No two adept digital artists envision the same artwork. Some individuals, not well-versed in traditional painting or drawing, harness the human imagination to conceive initial drafts, which they subsequently reverse engineer into a prompt or modify using AI. Others blend various AI-generated elements to craft entirely new artistic pieces, akin to the methods employed by notable pop artists like Andy Warhol and Roy Fox Lichtenstein. Are we to assert that their works lack artistic merit? I firmly believe that their creations merit all the respect one can give to a formidable artist.
These common prompts everyone is now familiar with, such as the one you've described, "teddy bear sitting in the..." are the entry level. Think of them like colouring books for children. It's a starting point accessible to everyone, then for those who truly are interested, there's a deeper and expansive world to explore, just as with any interest we discover deeper layers the further we progress in our development. There's more to this than just one-liner prompts. There's a world that intermingles prose, descriptive writing, and technical digital illustrative skills. These are domains that recruit immense mental creativity. Creativity isn't dead. It's found a new outlet.
This territory is novel, and given sufficient time for the field to mature, I hold a confident outlook that artworks collaboratively generated by humans in conjunction with AI image generators, along with the emergence of more advanced AI tools, will undoubtedly earn their place within the annals of art. We stand on the cusp of witnessing the emergence of pioneering legends in this burgeoning field.
@@kamu747 writing is certainly an art form. Did Stephen King write The Shining? Yes. Did Stephen King bring The Shining to life as visual representation of his own story via film? No. This is an excellent example because there exists an alternate version of The Shining that follows the book more closely and, where, King had a more direct involvement in its direction. Visually, it is completely separate from Kubrick’s version. Writing is an equal but separate skill form. Like a painter painting an image, uploading the image for ai to convert into a novel (written entirely by ai using an image for reference) and then calling themselves a writer. They are no more a writer than the person prompting ai to create an image is a visual artist. Now they even have prompting for prompts. So, now, ai is writing prompts for itself. All you have to do is give it a vague set of visual instructions like “Teddy bear sitting in the snow wearing cowboy boots.” Prompt generator creates the rest. I guess we can agree to disagree.
i wonder how these presenters always get so brilliant results. If i type "a blond hovawart wearing red cape" i get pictures of people instead of dogs . . .
Interesting re sty ing ad.
I hate them too 12:37
Gawat.
Question: does Adobe require its customers use North American English when activating image generation?
Comment: Adobe and other US software vendor’s use of “Evangelist” to describe a technology smitten employee is silly. This has everything to do with the legacy of American Founding Fathers’ Puritanism, not contentment in one’s occupation.
I started watching this, but paused to look at how old this video is, and two months makes it completely irrelevant. So sorry Howard. If you can hit me with some Firefly 2 information, I am all ears... and eyes
takes forever to do anything
The beta period has been ongoing for six months and the forced watermarks make the images created by Firefly useless. Please stop adding watermarks to images.
Did you pay for an account?
@@quantumtimelines2846😂😂😂😂
On the contrary. They should add more watermarks.
seems like basic looking previous iterations cannot be done. Even Pinsky cannot find his!
we white want. lol when he scrambled his words like he was a robot breaking down.....
It's a shame they cancelled that show
"you may have saw?" :-) cmonnnnnnnnnnnnnnnnnnnn man
everybody likes a southern dialect but without the actual accents. I done seen this before.
@@PHlophe So you done seen this before. Really?? 😂I also want to have saw this.
Firefly give me ugly results, pay for that, no thanks
Hey Adobe. Wake up. I’m trying to follow along to use your product but you have a teaching moment full of commercials. Goofballs.
Bad quality and very biased results, just usable for eavy workload of bad images in pseudo batch.
Absolutely useless tool for real creators
Everything looks so generic. The only company I can see using this is some low budget company needing generic artwork for daily turnarounds with clients who have no eye for quality. And that's coming from a guy who has done work for billion dollar companies who design by committee. They would absolutely trash this generic look.
Ok billion dollar guy, your experience applies to each and every case and we care a lot about your opinion. 👍
Just remember. Look at this as a starting point, not your destination.... Use these tools together. I'm pretty sure you're going to find yourself in a successful place! That's my opinion!
Ya, but is it always going to stay that way? I have worked with Photoshop for 25 years and it didn’t just stay the same since the 90s. Change took decades, then years, then months, now were seemingly on the verge of weeks. Also 95% of the population could not care less about whether a logo was created by a human or not, only graphic artists care. The good news is traditional non commercial might see a renaissance because human beings will care about craftsmanship, time, and love put into what another human created. Ads, logo, commercial art. I really have my doubts.
@@fredthomson2384 : I can't see any big brands paying an ad agency millions of $$ for something done in a few minutes using AI. Especially when most of brands have a certain look they like. Maybe as prelim ideas you bring to a committee and say..."What about something like this?" but no way for anything final to press. Right now you can search AI generated designs and everything looks like it came out of a CGI animated cartoon: Skin smoothing, unnatural lights, mist, soft focus, etc. Very odd plastic look. And then it's all the same concept and placements of things. Also caught some details on different designs pulled from the same sources.
@@prettywikked11This is only the beginning, it seems. It’s all moving very fast. In the near future no one will tell if this guy in the video will be real or AI.
@wazid khan india