Generative fill: Convert any image into art | Photoshop Beta + BEST prompts
HTML-код
- Опубликовано: 1 июл 2024
- In today’s tutorial, I’ll be showing you how turn any photo into an oil painting, sketch, crayon drawing, comic art and more in just a few clicks + stay tuned as I’ll be revealing a list of my favourite prompts for you to use.
Link to the image I've used so you can follow along:
www.pexels.com/photo/blue-box...
To be the first to hear about new and trending design tips, be sure to SUBSCRIBE and leave me a comment of any tutorials you want to see next!
UPDATE: For best results (especially with some of the drawing prompts) try with images similar to mine with a white background and inanimate objects. 'Oil painting' and 'Water colour painting' have produced the best results with all types of images including portraits!
This feature does NOT work in the current Photoshop Beta or current release of Photoshop. I have tried these steps across several videos and the only result is a new layer of generative paintings. I am useing the PS Beta 25.10.0 release. I've followed each step exactly as described. Is anyone else having this issue in this release? What's missing? I've tried this in both 8-bit and 16-bit modes and it simply will not work.
same here.. it did work in the beginning :(
@@vandereclame2898 Same here, I think it has to do with changes in photoshop?
Unfortunately in the latest psd beta this feature is no longer available. However it is available in the current Psd version
AWESOME! Thanks for share this great tips!
Outstanding! Thank you very much. You made my Sunday!
No problem! Glad I could help 😊
Thanks for the tutorial. After writing an Action, the process is effortless.
Nice one!
Thanks for your sharing! I found it tricky with some kind of original photos, such as 3D render.
When I apply this method on 3d render, it didn't work and just show some blue paint on that.
thanks, needed this for a project on my art college.
Awesome, glad I could help. Hope it goes well!
Great tutorial. Thanks
Thank you, glad you found it helpful 😊
This wicked. Thanks
No worries, glad it helped😊
I tried your method works pretty good except how do I get more detail when I'm trying to do the images
Great video. One trick that I have tried is doing everything you have done but when finished with an image that I like I put that image over the original and then play with the fill or opacity of the generated image. I have been able to get some great effects and sometimes I have increased the percentage of gray and then when the final image in generated (especially in oil paint) I can manipulate it to where it is more pleasant.
Awesome Tip! Will have to give it a try.
Great technique! Hope Adobe will give us a less convolutedd way of making partial selections soon...
I hope so too!
Helpful. Thank you.
No worries!
Thank you, but the generativr=e window keep on forcing the remove background or select subject on me. Any idea on how to by pass these quick promps? Thanks a bunch
Thank you so much. I've been lost.
No problem 😊
Thank you so much :)
You’re welcome 😊
Wow! That was amazing! Thank you for demonstrating this effect! 👍 my question is: Are there any other ai tools that can convert our images into artistic style effects, instead of photoshop? Thank you. You have a new subscriber to your channel.
@eurossocial the great thing about Photoshop is that you can apply this effect so that it distorts your original image instead of replacing it. There are lots of Ai tools out there but results might differ from your original image or requires complex prompts to achieve the same effect. My favourite Ai image generator at the moment is leonardo.ai :)
Wow, mind blowing, congratulations, i have subscribed to your channel. I have one question, can i do something similar on a face? I want to add same shadows, pencil drawings etc
Thank you! You can give it a go, however I haven't had great results with faces yet as things become distorted. Hopefully we should see this improve as Adobe develops this tool.
Secondly, when i create a alpha channel now, it now will only make a white alpha channel instead of the black i see in your example. Previously, i could get it to create the black and do the fill but now it will only create a white alpha which seems to break the process.
6videos and finally I get itthanks somuch
Nice one! Happy to help 😊
I am doing everything according to the instructions and it is not working in my case. i have the latest photoshop beta.
this effect is insane, there are so many possibilites with this...
Very exciting!
Amazing! I had no idea you could do this... suddenly Photoshop AI gets about 10x more useful
I know! Such an amazing tool!
I would really like to figure out how to find a prompt that mimics the style of the old 1980s hand painted movie posters.
Try asking chat gpt for a list of prompts for your theme! Something like “write me 10 image prompts for ‘1980’s hand painted movie posters’.” Might be too specific but worth a try. Something simpler like ‘vintage painting’ might work better 😊
Awesome, thanks so much.
Help me out with my thinking here;
30% Brightness for the new channels layer you create would be about 1 stop (depending on the colour space your using) below middle grey (128 (ColourMatchRGB) or 145 (Adobe RGB) 147 (sRGB)), meaning the darker you go, the more pronounced the effect would have in impacting your image? Is that correct?
Again awesome, thank you!
Thanks for watching, glad you found it helpful!
Think of channels as a mask, so the whiter your channel (100% brightness) the more it is going to apply the generative fill layer and distort your image. The darker it is the less it will affect your image (0%B). So that’s why 30%B seems to be the sweet spot in the middle. Same concept applies for any colour space you are using.
Hope this helps!
@@Studio.Shephrd Oh yep of course! (and awesome!) thanks heaps 🤩🙌🧡
Hi! nice video. How did you get your marker pens image? Marker pens just generates images of actual markers. Thanks
For best results with the drawing prompts, try with images similar to mine with a white background and inanimate objects. You can also be more descriptive with the prompt to get better outcomes such as 'marker pen drawing' 'marker pen sketch' 'scribble sketch with marker pens' etc. Hope this helps :)
@@Studio.Shephrd alright sounds good 👍
I’ve tried this every way possible still can’t get it to work. It keeps generating a big coloured in canvas. I’ve even used the same photo as yours. Please help me. Thank you.
Update: it works!! I was trying it on the new beta version and for some reason it doesn’t work there. But the regular photoshop it works. Results are amazing!!!!! Thank you so much!
@@xsharkzones2840 How did you get this to work? The regular Photoshop doesn't have generative fill I thought
Nice tip. I tried it with a portrait, there it didn't really work. I tried different gray scales, but they all look bad…
Hey, sorry to hear you’re having troubles.
Different images will get various results. I found images with white or light backgrounds worked better. But overall following these steps should produce similar results. Maybe try with a similar image to start and see how you go.
Hopefully as Adobe develops this tool it will become even better in the future! Hope this helps 😊
how to make a generative fill button immediately without selecting objects?@@Studio.Shephrd
@@Studio.ShephrdCould those problems possibly caused by the current resolution limits of the beta?
TNX Shephrd, It was a very useful article
Useful and concise without boring talk
Is there a source where different styles can be found?
To insert and write the correct prompt like the style you wrote
@samikuza thank you 🙂 Not that I know of, at this stage it's just being descriptive about what you want and a bit of trial and error.
Thanks. I tried it several times but it doesn't work at all for me. What can be the problem. I have done exactly the same steps as in the video.
Hey, sorry to hear you’re having problems. At what step are you having trouble?
I couldn't make this, i don't know why appear red after control click on the grey chanel
Can you please update this or make another video showing how to do this in the current version. 25.10.0 because I followed your exact instructions and what I get is for an example, I have a picture of a woman holding a baby sitting on the end of a couch. It left the background the same but the area that the woman is sitting actually turned into a gray fluffy triangle shaped pillow looking thing on one, on another when the prompt is "Crayon drawing", it tried to turn the woman into a blue crayon but only part of the crayon. Nothing I try works in the current version of Photoshop Beta.
Try it on a regular photoshop, i've tried just now with the (regular) photoshop 25.9 and it worked just fine! although you may think it does not have an AI generative fill, it indeed does have it.
That I’ll try. How did you figure out the trick with the additional alpha channel? Is that somewhere hidden in a manual or do you have good connections to Adobe? 😉. Please add a download link in the video description to that car to follow the tutorial, thanks.
Thanks @rolandrickphotography, using channels is one of the tricks I have seen circulating the design community. Basically all you're wanting to do is create a transparent selection so generative fill blends with your image rather than replacing it. I have added the link to the image in the video description now :)
@@Studio.Shephrd Thank you so much 😊🙏
@@Studio.Shephrd I did that now with an image of mine, doesn't work at all. Not even close. Posting links is unfortunately no longer possible on YT, the comments got deleted. Can I use an e-mail address to send you a WeTransfer link?
@@rolandrickphotography Link to the image is in the video description, still available there
@@Studio.Shephrd Yes, thanks. I meant I can’t post a link here to my image which doesn’t work with that method. Sorry if I wrote that unclear.
Is there a way to get the same effect but on another photo? Say I want the oil painting effect on one file, and also want the same oil painting style on another photo (seperate psd), is there a way to do this?
Hey, not that I know of at this stage unfortunately. Each time you generate it creates a new unique effect so your best chance would be to keep generating until you get one similar.
This is not working for me. Nothing like the results you are showing. I followed the instructions- any ideas of what I am doing wrong? Thanks.
Hi, hard for me to tell without seeing or knowing more. I would say most important step is the transparent masking stage which affects the final outcome. Check this video out that uses a different approach to do the same thing - hopefully you have better results ruclips.net/video/M2o7q8xyRQo/видео.htmlsi=lXAoVCJAnq6ZTg9A
Thank you for your constructive reply- I will check it out. I appreciate you taking the time!@@Studio.Shephrd
Cool! Is it work for bulk processing!
Hey @sergeyafanasiev3896, not at this stage as each variation is unique.
What's the reasoning behind adding the extra channel? I'm not being a smart-ass, just curious to understand how/why this makes it work.
Basically all you're doing is creating a transparent selection so generative fill blends with your image rather than replacing it. This can also be done with quick masks using the same 30%B colour fill in your mask.
I use that technique in this next tutorial for reference ruclips.net/video/FPzcaIgBhDo/видео.htmlsi=XpV3ZGxBu02sC4se&t=59
Thanks for watching!
@@Studio.Shephrd thanks for the explanation
Hi, same issue as everyone else...the generative fill did not look like the video.I'm not sure if there is a step missing but it doesn't seem to work.
Try starting with the prompt oil painting, this seems to be the most consistent across all images. Some prompts don’t seem to work as well on certain images so it requires a bit or trial and error. Hope this helps.
how to make a generative fill button immediately without selecting objects? 1:34
So basically when you create a channel with 30% brightness you could think of it being 30% transparent, so even though it doesn’t show anything as selected you actually have your 30% grey channel selected. This helps the generative fill tool blend the two layers better.
Hope that answers your question 😊
The problem on my photohop beta is; even though i apply %30 percent brightness, it still fill with white cover.
Hey, make sure you’re applying to the correct channel and have the foreground colour selected before you fill. Watch the video again and check you haven’t missed any steps.
I have latest photoshop beta, and I followed all steps but it's generating a blue image. Have tried so many times
Try it on a regular photoshop, i've tried just now with the (regular) photoshop 25.9 and it worked just fine! although you may think it does not have an AI generative fill, it indeed does have it.
This process is driving me mad. I managed to create the same variations you used on one document, but when i tried again, it now won't create the same effects. It now just creates the generative fill over the image, and when i go back and follow the steps exactly it still now does the same thing. I cannot get it to convert the image I want, into the variations of generative fill but rather, it just creates its own image over the top of mine.
@drfunk1986 sorry to hear you’re having issues. If you haven’t already check out this other video of mine that uses quick masks instead of channels to create the same effect.
Turn any image into water color art:
ruclips.net/video/M2o7q8xyRQo/видео.htmlsi=9It6k0jy-obeTYeS
Hope this helps
No idea why when i hit RGB after foreground colour fill everythings all red hues.
Hey, potentially you have a quick mask open or in the wrong channel. If you're in a quick mask the red shows what area is selected, you can toggle out of quick masks by clicking 'Q'. If you're having trouble with this method I'd recommend watching this video which uses quick masks instead of channels to create a transparent selection. ruclips.net/video/M2o7q8xyRQo/видео.htmlsi=yNkehE_A51wfNfIT
Hope this helps :)
Instead of turning my photo into a crayon drawing, it just messed up the proportions and made my hand look weird when erasing my face. Do you know what might have happened to my generative tool on Adobe?
The effect will react different with different images. Make sure to follow all the steps and play around with variations until you get the right look :)
all Skeetch Skeetch Skeetch
1:11 now go where?
Can this be done e on photopea? I don’t have photoshop.
Hi, not familiar with Photopea but I don’t believe so. These are unique tools to Adobe Photoshop
If you don’t want to create a new channel, you can just make a 30% mask. Seems a simpler way to do it. But either works fine.
Thanks for the tip! You’re right, I’ve seen people use that method too, both work just as well 😊
¿Cómo sería eso?
how would that be?
@ Press Q (to enter quick mask mode). Fill with 30% grey. Press Q again to exit.
I have a new video dropping tomorrow using that alternative technique so be sure to check that one out!
hello
is this still available feature in adobe photoshop?
Hey, yes generative fill is still available in Photoshop :)
@@Studio.Shephrd thanks
Yes i found it later on today :)
I don't think this works with faces. Or does it?
Yes should still work with faces, but might not be as effective as it would cause a bit of distortion so can get some odd looking results.
Check out my latest video to see one way to use generative fill on portraits.
Thanks for watching!
КРУТО!
@Artlantida Thanks!
Adobe made this function wildly inconvenient (( This alchemy infuriates(It doesn't work, did it according to your instructions, 10 times and nothing worked. Photoshop 25.9
so Niw Ziiland doisn't usi thi littir E at all?
😂
what about reverse?
I haven’t tried! But by all means give it a go and report back. I would assume it wouldn’t be as effective in producing realistic images as it is at producing textured ones.
Use it while it's in Beta... Credit system coming this fall...
Hopefully not, but imagine there will be some extra subscription costs to cover the expense of running AI..
I'm not satisfied with the pixel art result. It would be better if smaller pixels
If you’re looking to create pixel art maybe check out my other video ‘convert image to pixel art’ in illustrator😊
Where's the list
Listed at the end of the video :)
doesnt work what so ever
Hi - great utility. But does anyone actually understand what is going on here. It's all a bit difficult to comprehend for me.
It drew me a container full of crayons. Boo
Try generate more variations if there aren’t any to your liking. Click generate again and it will create 3 more variations to choose from. And if you still aren’t successful then try with some alternative images and they might work better 😊 Hope this helps
hmmm, basically just scribbled in the photo of my dog like a kid with a fistfull of crayons angrily going over a photo. Zero idea there is or was ever a dog there...
Hey @k-Sal, make sure not to skip this step here 00:50, this will reduce the intensity of the generative fill so it will show more of your image! If you have followed all the steps then maybe try it on a different image and see if it makes a difference.
I did this with a portrait, the results were AWFUL....and I mean, DREADFUL....
are you australien
New Zealand 😊
ok it's close@@Studio.Shephrd
increible... aunque el pixel art se ve 💩
Photoshop AI is pure dumb in painting. None of the results came out as shown in here above.
Hey, sorry to hear it’s not working well for you. All images react differently to the prompts, as this is still in beta we should hopefully see this improve in later updates but for now it’s very much experimental. Have a play with other prompts and you could try being even more descriptive than I have for various results. Hope this helps:)
resuts were horrible and even as an amateur i would have been embarassed to present them. i guess there is a a reason you didnt use people. much like firefly/ps the results were terrible. many of the faces presented looked nothing like those in the original. its his or miss, and mostly miss, at best.
Hey, at this stage it is very much experimental but there are a lot of great uses for it already. I’ve had some good results on portraits but just with certain prompts as you’re right it definitely struggles with faces. But as this is only Beta it’s a pretty good start and exciting to see how this will develop.
Best thing: Just hire a real artist!! Get passion, real skill and love for the image ❤
What if you have no budget tho?
Thanks @UC-Music, I think this will never replace artists as originality and human expression can never be replicated artificially. However this is a fantastic new tool that offers new opportunities for designers to be creative.
This is bs doesn’t even work
Hey there, what’s not working?
Bullshit
As an artist I would just like to take the time to say, No stop this garbage.
How does it feel to be contributing to the erasure of artists and illustrators? Painting and drawing is skill that takes decades to learn. Don’t participate in technologies that would discourage someone from learning how to paint or draw, or technologies that replace careers. There is nothing impressive or helpful that you are doing here. Creating art is fulfilling and brings to others joy. Replicating art with AI is pathetic. 😢
@stephenschildbach940 I understand with the rapid rise of Ai it can be daunting for artists who worry about being replaced. But I think Ai is merely a tool to enhance artistic processes, and you will see new artists emerge who use various Ai tools to create unique work. However the essence of art, being personal expression and exploration of different media will always hold value and will not be replaced.
@@Studio.Shephrd I won't deny that AI could be occasionally used to "enhance the artistic process", but unfortunately it will be predominantly used by people who are "takers" not wanting to work for anything or not not wanting to pay anyone for for a skill that not many people can do without years of training. It creates a society that thinks they can get any creative result they want without putting any effort or supporting a creative person. Using AI merely produces a dopamine hit for the button pusher, missing the whole point of making art in the first place-it is a human creative process. Not to mention art is fun! And, why would you have a computer do something for you that was so enjoyable and fulfilling? There is no doubt in my mind that AI applied to "making art" is a huge negative sum loss that harms society and artists. AI is anti-human.