Generative fill course 20%off YTLC20: photoshopcafe.com/shop/generative-fill-in-photoshop-introduction-course/ Written steps: photoshopcafe.com/photoshop-2024-generative-fill-tips-16-fast-tips-and-tricks/
Thanks for another great video and keeping us informed of all the changes and new PS features. It's nice you don't waste time on flashy introductions and telling us how great you are, because we already know that. You stick to the point, give a lot of info without the fluff that we don't need. Or want actually.
7:49 - What 3 nuclear plants in immediate surroundings didn't achieve in 3 decades, Photoshop mastered in just a few seconds... Thanks for the precious tips, and the smile :)
I was doing some portrait editing today and I had shot a bit tight on one of the subjects ( of course the best shot). I used the generative expand on the right side of the portrait which include the subjects hair and it is an amazing job! I was able to then do a proper crop. Many thanks for your concise useful info!!
Hi very interesting thanks. A question your thumbnail of the lady guitarist - first examples, goes from the original photo to your generative fills, the thumbnail however ends with he in tshirt, shorts etc how to you replace the trousers with the shorts?
Hi Colin, very good tutorial and very helpful. In the early days of the beta version generated fill always worked very well but since it is in the official Photoshop version I often have problems. Especially with changing the background. I select the subject, invert and write what should be generated. The background is changed but so is my motif. Faces are distorted or half of a hand is missing. I never had these problems in the first beta version. Do you know what could be the reason for this? Thank you very much
Thank you for the great video. There are some great new tools. But I am a bit confused about the addition of weapons, armor, or other things in the scene. I wonder where these elements (images) come from and if I can use them for commercial purposes in my work if I use this option.
@@photoshopcafe I don't think so because the generative Layer of the truck contains just the space of the selection, as you said. The whole background is on another layer
Good day Colin. Love all your tutorial videos. Seen lots of them & learned lots. I have a question that doesn't have to do with this video and wonder if you can give me a reply. I have Lightroom & Photoshop. I bought a new HP all in one computer. Approx. 6 months ago and since Photoshop Beta, I've been having computer issues when both programs are open. It is very very slow and then freezes. I did contact Adobe and issue not resolve. What type of computer are you using Mac & Windows?
Great information! I have issues when I try and eliminate people. I make my selection of the person and leave the space blank and it sometimes generates another person and won't stop doing that.
I purchased your video tutorial on Generative Fill Photoshop. In the segment with you on the bridge and subject selection, masking...you mention select radius for seeing edge, and in my version of Photoshop 2024, I cannot find the "radius" when I go to see edge? I'm not using the beta version but 2024. How do I manage that edge? I love the tutorial so far, but a little disappointed that the tutorial seems to be made with beta version and not latest version. Will you have a tutorial available with the 2024 version? Thanks
Hi there. There are no changes with the final release and beta, there's nothing to change they are the same. Radius is in Select and Mask - under Edge Detection (a little disclosure arrow is there, easy to miss). Also make sure the top tool on the left is selected (it is by default). If you have any other issues, please contact me at photoshopcafe.zendesk.com/home where it's easier for me to provide more timely and detailed support.
wait... they now add in the whole image for backgrounds and mask out the subject? Game changer.... I needed a generated background for a 2nd picture and had to get the subject out first *lol*
Then you need to use Photoshop to remove her 6th gloved finger haha. Hilarious quip about Gattaca. This is incredible, and thank you for this straightforward simple and thorough tutorial.
Colin …what are your thoughts please . To generative expand first downsize the entire image to 1024 …generative expand …then upscale entire image to a higher resolution . Perhaps a viable alternative to several 1024 ‘boxes ‘ Personally 99.9pct of the time I use generative for getting rid of things . Works well in the vast majority of instances
Looks good at first except the quality is not there for real commercial use. If you work on low res images or soft focus/blurry images it works well. I am waiting until it can make high res images, quality still not there.
I watch all your videos Colin, because you offer such great advice and demo interesting things, BUT, and this BUT isn't really aimed at you but everybody putting up videos on Generative Fill, or Expand,. Whats the point because its all so low resolution to be useless if you were thinking of printing anything, this new technology is only really useful for playing around for a laugh or for putting things up on social media where quality isn't an issue. If Adobe ever manage to get the resolution right I`d be interested, but then you'd probably only get one credit a month! LOL
Great tips. The one issue I have consistently found with AI generated content is that it seems incapable of accurately depicting human hands. If you look at the tip where you ask for it generate a glove, you'll note that she has 6 digits in every version. My assumption is that when programming AI, it has confused concepts like "humans have 5 fingers" and "humans have opposable thumbs" and come up with "humans have 5 fingers and 1 thumb." I've yet to see an AI engine create an accurate model.
That was the point of my Gattica joke. Its part of the process and will change soon enough. Mid journey had the same "journey" now it can count digits and do perfect faces. Gen fill will get there too and do better faces as well
@photoshopcafe lol, sorry, wasn't familiar with the Gattica reference. I just signed up for Midjourney as well. Last week I just entered "human hands." Every result had 6 digits in each hand. Evidence that AI is clearly giving us the finger. Lol
Not at all existed about the direction Adobe is going. Sure, innovation is inevitable, but they’re essentially optimizing their product to the average Joe. Cutting out the professionals that have supported their product for decades is no way to run a business. They fully realize our talent out performs their AI 10/10 but as long as their subscription is cheaper than us, they know the masses with fold.
I knew I would found such a comment among those "yes sayers" :) Including AI will cheap down all the work. Then who will buy Adobes overpriced products? Yes maybe the average Joes who dont mind spending 20 bucks in the month (without noticing it is 240 bucks in a year 1200 in 5 years, etc...for a single product) for the AI, what will make some uncreative, ugly and weird graphics with few clicks. Good job Adobe and co. for killing the graphics design industry.
Lets be honest, it's just a toy isn't it ? Any professional trying to produce work to a tight schedule is going to be disappointed. There are 5 problems with generative fill. The first is the algorithm. For example you ask for 'an old fashioned car' - and you get old car wrecks'. So you have to waste time trying to second guess the bot's understanding of what you type. Secondly it's slow - so multiple regenerate to get a better image is very time consuming. Thirdly the generated bit has a fringe of the background around it so it blends in. This means you can't resize or reposition it because the fringe won't match. Fourthly the censorship (aka Community Standards) are child-like. I tried 'swimwear' the other day and got refused. It's just not reliable enough to use every time - so trying something to see if it works - only to get the orange box after it has thought about it for a bit - is just another waste of time. Finally the floating panel for the generative fill is a gimmick. The last thing we need is more clutter. Much better to right click on the selection and choose from the menu.
I agree. It seems to totally ignore the concept of artistic composition where you move things around and change size and perspective. And, as you say, the Community Standards are frustrating. I tried putting some cars outside a house today. Apparently you can't have "Rolls Royce" or "BMW"; but you can have a "Jeep" . Thinking it might be something to do with brand names I tried "Dyson vacuum cleaner" and got something that look like R2D2 with a handle on the top. ???????
Dude, it JUST came out of beta less than a month ago. 2-3 years ago, being told we’d soon be able to do something like this would have been mind blowing. And ai is constantly improving at an insane rate. Just… relax a bit. 😂 You’re certainly not obligated to use it.
I find it to be useful time-saving tool and I use it on professional work all the time. It's integrated in my workflow with the other tools. I use it for what its best at, like the other tools.
Nothing like creating some half baked AI photography. Editing is cool until you let AI do your job. Personal opinion this is for lazy photographers and photo editors.
My tip: Do not even try to expand images when women are present on the image, 99% chance to get blocked, even if you own full copyright and there is zero nudity involved. I'm not sure how you were even able to add the shirt to your model without getting blocked.
Could it be that it's calculating the amount of bare skin that it sees ? So non-nude woman wearing a bikini ( for example ) triggers Prude Mode based on mathematical percentages. You might get it to work by selecting a rectangle at the edge of your image and copying it onto a new layer. Turn off other layers beneath then expand the copy.
Instead of leaving the "What would you like to generate?" box empty, if you get blocked, try putting a single . (period) in the box instead. It kept telling me I was violating their terms because I was trying to expand a red awning on a dark brown house, but flew right past it when I added the . to the "What would you like to generate?" box.
It seems to hate tight clothes like crossfit models wearing leggings, equestriennes in breeches etc. At this point i have basically given up. The dot . trick does not work for me in those cases, either.
Generative fill course 20%off YTLC20: photoshopcafe.com/shop/generative-fill-in-photoshop-introduction-course/
Written steps: photoshopcafe.com/photoshop-2024-generative-fill-tips-16-fast-tips-and-tricks/
generative fill is legit my favorite thing ever right now, its so cool
I know. How did we ever manage without the Generative Fill feature? 😄
It's fun and a time-saving tool
@@Stoy-boywe worked. Hard.
Thanks!
Thanks so much. I really appreciate it! 🎉
Thanks for another great video and keeping us informed of all the changes and new PS features. It's nice you don't waste time on flashy introductions and telling us how great you are, because we already know that. You stick to the point, give a lot of info without the fluff that we don't need. Or want actually.
7:49 - What 3 nuclear plants in immediate surroundings didn't achieve in 3 decades, Photoshop mastered in just a few seconds... Thanks for the precious tips, and the smile :)
Lol yeah
Regenerating part of a previous generated item. Can't believe I didn't think to try it! Thank you!
Glad to help :)
very interesting and informative. Will have to practice these to try and get up to speed. Thanks.
I approve this straight forward demo, looking forward to more updates as the software keeps updating.
Excellent info, as always!!!
thanks
The quick mask mode tip is awesome! Can’t wait to try!
Excellent lesson Colin. Thank you
the generative expand tool has so many possibilities....love that tip
Thank you so much, Colin. You have the most helpful videos of anyone else I have seen.
Happy to help!
great tips!
Fast and informative - what a guy!
I was doing some portrait editing today and I had shot a bit tight on one of the subjects ( of course the best shot). I used the generative expand on the right side of the portrait which include the subjects hair and it is an amazing job! I was able to then do a proper crop. Many thanks for your concise useful info!!
Thank you very much Colin, always learning so much with such good information you give us.
Thanks
Great reference...great tips!! Thank you.
thanks
Well done...good to know
Well explained!
Thanks
Thanks man!!! I lost the Contextual Task bar as I closed and could not find.
Great tips!
Glad it was helpful!
Hi very interesting thanks. A question your thumbnail of the lady guitarist - first examples, goes from the original photo to your generative fills, the thumbnail however ends with he in tshirt, shorts etc how to you replace the trousers with the shorts?
Hi Colin, very good tutorial and very helpful. In the early days of the beta version generated fill always worked very well but since it is in the official Photoshop version I often have problems. Especially with changing the background. I select the subject, invert and write what should be generated. The background is changed but so is my motif. Faces are distorted or half of a hand is missing. I never had these problems in the first beta version. Do you know what could be the reason for this? Thank you very much
Thanks.
The tip about trimming file size by deleting unused variations is a game-changer for me. It'll definitely help with managing large projects.
Thank you for the great video. There are some great new tools. But I am a bit confused about the addition of weapons, armor, or other things in the scene. I wonder where these elements (images) come from and if I can use them for commercial purposes in my work if I use this option.
The part at around 5:13, just select the truck and invert the selection on the truck layer
That doesn't work. If you do it the way you suggested, it replaces the background on the entire image
@@photoshopcafe I don't think so because the generative Layer of the truck contains just the space of the selection, as you said. The whole background is on another layer
What you responded, Colin, was my experience as well.@@photoshopcafe
When might you use generative fill to remove something rather than the remove tool? Any advantages/diadvantages of one method over the other?
I made a video on that ruclips.net/video/ODSAToMS-is/видео.htmlsi=NXVX2-Bj6iNOZNw4
Good day Colin. Love all your tutorial videos. Seen lots of them & learned lots. I have a question that doesn't have to do with this video and wonder if you can give me a reply. I have Lightroom & Photoshop. I bought a new HP all in one computer. Approx. 6 months ago and since Photoshop Beta, I've been having computer issues when both programs are open. It is very very slow and then freezes. I did contact Adobe and issue not resolve. What type of computer are you using Mac & Windows?
It sounds like you need more memory on your computer, or a dedicated GPU
Thanks for the info and quick reply.
Great information! I have issues when I try and eliminate people. I make my selection of the person and leave the space blank and it sometimes generates another person and won't stop doing that.
I purchased your video tutorial on Generative Fill Photoshop. In the segment with you on the bridge and subject selection, masking...you mention select radius for seeing edge, and in my version of Photoshop 2024, I cannot find the "radius" when I go to see edge? I'm not using the beta version but 2024. How do I manage that edge? I love the tutorial so far, but a little disappointed that the tutorial seems to be made with beta version and not latest version. Will you have a tutorial available with the 2024 version? Thanks
Hi there. There are no changes with the final release and beta, there's nothing to change they are the same. Radius is in Select and Mask - under Edge Detection (a little disclosure arrow is there, easy to miss). Also make sure the top tool on the left is selected (it is by default). If you have any other issues, please contact me at photoshopcafe.zendesk.com/home where it's easier for me to provide more timely and detailed support.
And thanks so much for supporting the CAFE by getting the course :)
well said
wait... they now add in the whole image for backgrounds and mask out the subject? Game changer.... I needed a generated background for a 2nd picture and had to get the subject out first *lol*
Then you need to use Photoshop to remove her 6th gloved finger haha. Hilarious quip about Gattaca. This is incredible, and thank you for this straightforward simple and thorough tutorial.
Hi mine's not working after using any kind of tool the generative fill box won't appear
Colin …what are your thoughts please . To generative expand first downsize the entire image to 1024 …generative expand …then upscale entire image to a higher resolution . Perhaps a viable alternative to several 1024 ‘boxes ‘
Personally 99.9pct of the time I use generative for getting rid of things . Works well in the vast majority of instances
You will lose a lot of image quality doing that, but it will match albeit a softer image
@@photoshopcafe if upscaling with PS but not so with Topaz Photo Ai or their Gigapixel Ai 🤔
Looks good at first except the quality is not there for real commercial use. If you work on low res images or soft focus/blurry images it works well. I am waiting until it can make high res images, quality still not there.
It's coming, untill then, use it on smaller segments and use other tools for larger segments like I do
I've had a problem where all I wanted to do is crop an image and it starts the generative fill routine and wasting my time. What did I do wrong?
When you select the crop tool the bar at the top has a box that says “generative fill”. Click this and select background
the crazy part with expand is if you went below the model and the dog, it would add body parts
You’re the PS Man Colin!
Thanks
how to buy Firefly generative credit
I watch all your videos Colin, because you offer such great advice and demo interesting things, BUT, and this BUT isn't really aimed at you but everybody putting up videos on Generative Fill, or Expand,. Whats the point because its all so low resolution to be useless if you were thinking of printing anything, this new technology is only really useful for playing around for a laugh or for putting things up on social media where quality isn't an issue. If Adobe ever manage to get the resolution right I`d be interested, but then you'd probably only get one credit a month! LOL
03:30 Asking for a Whale and getting a dolphin. Wow 😂
Great video, thanks. It’s cool stuff, unfortunately digital has made the real art of photography obsolete.
When just removing stuff ..to avoid block type . and nothing else in text box .
I do that if I get a warning. Making more informed selections reduces those warnings.
Can you show how to restore old damaged photos (paper) with AI?
We need a remove hair strands asap
remove tool or spot healing
Great tips. The one issue I have consistently found with AI generated content is that it seems incapable of accurately depicting human hands. If you look at the tip where you ask for it generate a glove, you'll note that she has 6 digits in every version.
My assumption is that when programming AI, it has confused concepts like "humans have 5 fingers" and "humans have opposable thumbs" and come up with "humans have 5 fingers and 1 thumb."
I've yet to see an AI engine create an accurate model.
That was the point of my Gattica joke. Its part of the process and will change soon enough. Mid journey had the same "journey" now it can count digits and do perfect faces. Gen fill will get there too and do better faces as well
@photoshopcafe lol, sorry, wasn't familiar with the Gattica reference.
I just signed up for Midjourney as well. Last week I just entered "human hands." Every result had 6 digits in each hand.
Evidence that AI is clearly giving us the finger. Lol
@@barbarianatgate2000 LOL!
Not at all existed about the direction Adobe is going. Sure, innovation is inevitable, but they’re essentially optimizing their product to the average Joe. Cutting out the professionals that have supported their product for decades is no way to run a business. They fully realize our talent out performs their AI 10/10 but as long as their subscription is cheaper than us, they know the masses with fold.
I knew I would found such a comment among those "yes sayers" :)
Including AI will cheap down all the work. Then who will buy Adobes overpriced products? Yes maybe the average Joes who dont mind spending 20 bucks in the month (without noticing it is 240 bucks in a year 1200 in 5 years, etc...for a single product) for the AI, what will make some uncreative, ugly and weird graphics with few clicks.
Good job Adobe and co. for killing the graphics design industry.
I tried generating white shirt and it couldn’t even get the color right .. right now this tool sucks for me
This of course means no more believing what you see.
hello can you do PS tutorial for beginners guide
I have done a few beginners tutorials already, but I'll do more in the future
Gone are the Photoshop tutorials on the thousands of tools in Photoshop. All Photoshop channels are only talking about generative fill.
my last several tuts weren't on gen fill
Great tips, thanks. Jesus has gone quiet on his RUclips channel and his website isn't working. I hope he's OK.
He just got married. He'll be back soon
Georges Bush of Photoshop
Lets be honest, it's just a toy isn't it ?
Any professional trying to produce work to a tight schedule is going to be disappointed.
There are 5 problems with generative fill.
The first is the algorithm. For example you ask for 'an old fashioned car' - and you get old car wrecks'. So you have to waste time trying to second guess the bot's understanding of what you type.
Secondly it's slow - so multiple regenerate to get a better image is very time consuming.
Thirdly the generated bit has a fringe of the background around it so it blends in. This means you can't resize or reposition it because the fringe won't match.
Fourthly the censorship (aka Community Standards) are child-like. I tried 'swimwear' the other day and got refused. It's just not reliable enough to use every time - so trying something to see if it works - only to get the orange box after it has thought about it for a bit - is just another waste of time.
Finally the floating panel for the generative fill is a gimmick. The last thing we need is more clutter. Much better to right click on the selection and choose from the menu.
I agree. It seems to totally ignore the concept of artistic composition where you move things around and change size and perspective.
And, as you say, the Community Standards are frustrating.
I tried putting some cars outside a house today. Apparently you can't have "Rolls Royce" or "BMW"; but you can have a "Jeep" .
Thinking it might be something to do with brand names I tried "Dyson vacuum cleaner" and got something that look like R2D2 with a handle on the top.
???????
Dude, it JUST came out of beta less than a month ago. 2-3 years ago, being told we’d soon be able to do something like this would have been mind blowing. And ai is constantly improving at an insane rate. Just… relax a bit. 😂 You’re certainly not obligated to use it.
I find it to be useful time-saving tool and I use it on professional work all the time. It's integrated in my workflow with the other tools. I use it for what its best at, like the other tools.
Just make sure there's no skin in your image or the Adobe morality police will stop you with an error.
There is skin on mine. But I know what you mean. It will get better in time
Nothing like creating some half baked AI photography. Editing is cool until you let AI do your job. Personal opinion this is for lazy photographers and photo editors.
Много глупав клип
My tip: Do not even try to expand images when women are present on the image, 99% chance to get blocked, even if you own full copyright and there is zero nudity involved. I'm not sure how you were even able to add the shirt to your model without getting blocked.
Could it be that it's calculating the amount of bare skin that it sees ?
So non-nude woman wearing a bikini ( for example ) triggers Prude Mode based on mathematical percentages.
You might get it to work by selecting a rectangle at the edge of your image and copying it onto a new layer. Turn off other layers beneath then expand the copy.
Instead of leaving the "What would you like to generate?" box empty, if you get blocked, try putting a single . (period) in the box instead. It kept telling me I was violating their terms because I was trying to expand a red awning on a dark brown house, but flew right past it when I added the . to the "What would you like to generate?" box.
It seems to hate tight clothes like crossfit models wearing leggings, equestriennes in breeches etc. At this point i have basically given up. The dot . trick does not work for me in those cases, either.
I've found I get fewer warnings when I select larger areas, or use the well-known period (full.stop) trick
Thank you!
Thank you!