A lot of questions about resolution. I put a notice on the video, but seems to have been missed by many. This is still limited to the 1024 x 1024. I will be sure to let you all know, when they release a higher resolution option.Remember, it's still beta. For now, use this work around ruclips.net/video/a7SDzpEGdCY/видео.html
I was about to leave a comment regarding that. This feature is cool for out of focus "Bokeh" backgrounds but at 1024 its not useable for much else, at least in one large go. One side at a time makes more sense to ensure max resolution. By the way, for those not familiar with AI Terminology, This new feature is actually really called "OUTPAINTING" Adobe likes to rename things to give the impression that they invent everything from scratch.
We’re getting spoilt by all these AI tools. My most often use is to remove objects and distractions. It used to be a slow and tedious task, but now, it’s like magic
In every case , the result is low res, it will not fill hight resolution photo with same resolution content, the best way is to use fill with content aware tool, not generative fill.
The Generative Expand in v. 25.0 works fine until I tried to edit from LRC. The file would not open. Adobe support had me revert to 24.7, which resolved the problem. The person said that the problem will be resolved in version 25.1 due to be released soon.
No offence to the content creator, who is only demonstrating a new tool in PS. However, as a passionate photographer, I don't understand the undue emphasis on automation of editing software. Photography is about telling stories articulately, but it has lately become telling stories 'artificially.' Only an artistically disabled would need AI gimmicks to overcome lacklustre photographical skills. AI enhanced images are a work of fiction and shouldn't be confused with photography, just like movie stunts shouldn't be confused with reality!
It's a fantastic tool and help. But now we are going into a fantasy world, so we have to be careful and add a comment about what we have done to the photo after my opinion 😅😮
The commentary boils with fake people who maybe even calls themselves photographers. Well, going from a photographer with own art, to a client that orders others to make large parts of their pictures is fake and a lie. Soon fake pictures will flow everywhere, and lying fake people will say this photography is mine. How fake and stupid the humanity has gotten..
there is an issue/error with photoshop beta 25.0 - when sending an image from lightroom to photoshop beta and back. It has a glitch. I had to revert to the previous version and all went back to normal. Anyone else experience this?
Is it only me or is using AI crushing the excitement (good or bad) for the journey through which we had to travel to get our results. For saving time and work yes its brilliant but the feeling you had editing your first photograph years back is gone isnt it?
Is there a way to replicate the fill result but in different angle of the same object? lets say you want to expand the same photo taken from different angle, Challenge!!!!
This is awesome! I struggle when it comes to cropping a picture to a size of my desires. This new feature makes it less difficult and also opens many possibilities. Thank you for the tutorial once again Colin.
Generative Fill, unfortunately fills the selections with low res nonsense. Maybe soon it will be hi res nonsense. Take a close look, if it has any objects, buildings or people its just garbage. Its for use as a last resort when you have messed up.
thats interesteing I thought i had the latest BETA but it doesnt have this feature and there is no update available in cloud - can you advise which BETA version number to confirm?
@@chasebros4948 So I did a bit of research. It is still using the same mechanics as generative ai. The only difference is that there are less steps. So at the moment, sadly, we're still stuck to using 1024px, hopefully that evolves soon.
AI is changing how we do photography and I am not sure it's for the better. At what point is a photo no longer a photo but rather a compellation or composite and not really the work of the photographer?
@@photoshopcafe I know, but I felt it's worth expressing a point of view when we're bombarded with AI developments on everything but especially photography. Not everyone loves it. Still liked the video, thanks ♥️
All those tools are about serving, not learning. And in The 90s we said “the future is bright, we all need sunglasses”. My guess is that it more so goes in the direction of sticks and stones.
Can you explain how this is different than generative fill? Only difference I see is using the crop tool to generate rather than cropping first and then using generative fill.
Great video as always! Low res fills pretty much limit usefulness of the feature. I made the mistake of furnishing an image to a client without noticing it and they were not happy! When do you expect the GF/GE to be ready for prime time?
wow that's crazy! Question, could you supplement an image so that it seamlessly bl3nds the expanded portion? Let's say that I've some beautiful foreground and background images which I'd like to integrate into an image. Is it capable of doing that? Secondly, is this reaching out to their server or is it happening within the computer?
Just saw this yesterday in a video from Brian Matiash. I have the new version installed, but haven't played with this function yet. Definitely adds some fun to the editing process.
@@photoshopcafe it's not a large language model at all. it's a diffusion model, and so I guess you just don't even know what you're talking about, or you're paid by adobe. either way, toodles.
Wow, that sounds very exciting! But how does that relate to the 1024 x 1024 pixels that could previously be inserted as the maximum size in Generative Fill. Does that no longer apply?
Hi Colin, Do you have any advance info on when these features will no longer be in beta but added to regular PS like was done with the remove tool. My win 11 doesn't play nice with beta.
how do we get the regular generative ai box now? this tool suffers from the 'violated blah blah blah' and the putting a . trick doesn't work on it. I just want to use the regular generative AI function but now when i crop anything i'm automatically forced to use this new generative expand tool
Hi Colin. How come when you only expand the photo you end up with a high resolution (300x300 px), but when you fill with a prompt and expand at the same time the new area is low resolution. Am I missing something? Thank you for your tutorials!
Nope doesn't bother me at all its just for a few seconds and its done with I found it OK . Not like its going to kill me or make me pregnant so why worry.
Thank you for this latest very informative feature demonstration. I love the fact that this has been built into the crop tool, with no need for selections to be made. I am looking forward to all of this AI progression to come eventually into the full version of PS.
Unbelievable. Just incredible. What has emerged in the past year is nothing short of miraculous, can you imagine what the next 5-20 years will bring? I can't - it's beyond me, but I know I'm excited about it!
I wonder if - among all the praise for PhotoshopBeta - you also experience the extreme amount of censorship when trying to prompt generative fills on images that either include people or one of the words on their blocklist. Any way to work around this? Your opinion? I've got prompts like "concrete slab" refused...
So let's put our cards on the table: this is obviously "generative fill" combined with "expand canvas" so you don't have to use the marching ants. Convenient. I get it. Here's the problem. If you do the same process manually, you marching ants the top part and genfill, THEN you marching ants and genfill the bottom. As two separate steps, the AI does two separate text-to-image method calls. But with "generative expand", it's ONE method call. If you do it all as one single method call, you will lose a ton of sharpness and granular detail in the new image. This is especially true for any image bigger than 1080x1080 pixels, as the program stretches and distorts it's own genfill content. It's much better to do this manually in chunks with marching ants rather than all in one go.
The canvas expansion using generative fill is great whether using the former long hand method or the new one. One caveat that I found myself having to deal with almost always is having to add a brightness adjustment layer linked to the generative layer to better match the brightness. At times I also had to use the AI erase tool to get rid of the hair line(s). Hopefully that will get better with newer versions.
Nice one Colin. That was a lot more information than the videos I have seen about this. I am still wondering what on earth they have lined up for the major version update in October? It will have to something like using your webcam to read your mind, so you no longer need keyboard or tablet/mouse.
@@photoshopcafe I know that. It puzzles, me, though, that photographers who fuss over minor distractions in their photographs add major distractions to their videos. You don't actually need to remind people every time, especially those already subscribed. If you have a series of videos, you can give a preview of the next, and promise that subscribers will get early notifications. Then, it's a benefit to me, but not a favour to you.
@@MONTY-YTNOM You wouldn't have subscribed otherwise? A reminder at the end perhaps, promising that you will find out immediately a new video becomes available?
A lot of questions about resolution. I put a notice on the video, but seems to have been missed by many. This is still limited to the 1024 x 1024. I will be sure to let you all know, when they release a higher resolution option.Remember, it's still beta. For now, use this work around ruclips.net/video/a7SDzpEGdCY/видео.html
I was about to leave a comment regarding that. This feature is cool for out of focus "Bokeh" backgrounds but at 1024 its not useable for much else, at least in one large go. One side at a time makes more sense to ensure max resolution. By the way, for those not familiar with AI Terminology, This new feature is actually really called "OUTPAINTING" Adobe likes to rename things to give the impression that they invent everything from scratch.
👍🏽
We’re getting spoilt by all these AI tools. My most often use is to remove objects and distractions. It used to be a slow and tedious task, but now, it’s like magic
More time to be creative rather than solve time consuming annoying tasks!
World is moving on, keep up or get left behind, stop bickering boomer.
Or just do it right from the camera in the first place, this isn't photography anymore if half the photo is AI generated.
In every case , the result is low res, it will not fill hight resolution photo with same resolution content, the best way is to use fill with content aware tool, not generative fill.
The Generative Expand in v. 25.0 works fine until I tried to edit from LRC. The file would not open. Adobe support had me revert to 24.7, which resolved the problem. The person said that the problem will be resolved in version 25.1 due to be released soon.
Thanks for sharing
Thank you Colin! It's 3AM here, but I will definitely be trying this tomorrow morning!
No offence to the content creator, who is only demonstrating a new tool in PS.
However, as a passionate photographer, I don't understand the undue emphasis on automation of editing software. Photography is about telling stories articulately, but it has lately become telling stories 'artificially.' Only an artistically disabled would need AI gimmicks to overcome lacklustre photographical skills. AI enhanced images are a work of fiction and shouldn't be confused with photography, just like movie stunts shouldn't be confused with reality!
It's a fantastic tool and help. But now we are going into a fantasy world, so we have to be careful and add a comment about what we have done to the photo after my opinion 😅😮
The commentary boils with fake people who maybe even calls themselves photographers. Well, going from a photographer with own art, to a client that orders others to make large parts of their pictures is fake and a lie. Soon fake pictures will flow everywhere, and lying fake people will say this photography is mine. How fake and stupid the humanity has gotten..
there is an issue/error with photoshop beta 25.0 - when sending an image from lightroom to photoshop beta and back. It has a glitch. I had to revert to the previous version and all went back to normal. Anyone else experience this?
,,,waiting to see how well developed this feature will be upon release,,,anticipate higher monthly fees
Is it only me or is using AI crushing the excitement (good or bad) for the journey through which we had to travel to get our results. For saving time and work yes its brilliant but the feeling you had editing your first photograph years back is gone isnt it?
Not in favor of adding mountains and lakes. Otherwise this is a time-saver compared to creating blank areas, circling them, and generating.
Use it how it best serves your needs. Personally I’ll use it more like the drone image, to give more flexibility
Is there a way to replicate the fill result but in different angle of the same object? lets say you want to expand the same photo taken from different angle, Challenge!!!!
This is awesome! I struggle when it comes to cropping a picture to a size of my desires. This new feature makes it less difficult and also opens many possibilities. Thank you for the tutorial once again Colin.
Anytime
Generative Fill, unfortunately fills the selections with low res nonsense. Maybe soon it will be hi res nonsense. Take a close look, if it has any objects, buildings or people its just garbage. Its for use as a last resort when you have messed up.
thats interesteing I thought i had the latest BETA but it doesnt have this feature and there is no update available in cloud - can you advise which BETA version number to confirm?
Is the generated resolution still 1024px or have they increased it in this update?
I was going to ask the same thing. Wondering if the 1024 resolution constraint applies to generative expand like it does to generative fill.
@@chasebros4948 So I did a bit of research.
It is still using the same mechanics as generative ai.
The only difference is that there are less steps.
So at the moment, sadly, we're still stuck to using 1024px, hopefully that evolves soon.
Still 1024x
I am on Photoshop beta current version and the Generative Expand is grayed out. How do i correct that?
AI is changing how we do photography and I am not sure it's for the better. At what point is a photo no longer a photo but rather a compellation or composite and not really the work of the photographer?
Thank you so much for sharing and going over this feature.
Is there an ethical obligation on the photographer to state that he has used AI tech to make his image look better?
They're blurring the line between fantasy and reality.
It’s up to you how you use it. You can keep it as real, or crazy as you want.
Great tip. Cheers from Nova Scotia.
Excellent! Great tutorial as always!
That option would be great on video 😊i
thank you, great video!
Very cool but I would eventually like to see resolution matching. Large areas are currently a problem if you have very high resolution images
01:00, should be "speed up" not "sped up"🤫
Truth in photography still counts, so I don't really want AI generated content, clever as it is.
It's a tool, like Photoshop always has been. Its up to you how you want to use it.
@@photoshopcafe I know, but I felt it's worth expressing a point of view when we're bombarded with AI developments on everything but especially photography. Not everyone loves it. Still liked the video, thanks ♥️
Too good sir. Thanks 🙏🙏
I will wear that baby out!
Good skill, but I am used to the filters to take a shoot, like Kase or Nisi. It's significantly reduce my PS job.
You can do this already by using the square selection tool and then invert it. Then do generative fill
We know. The point of this is to save some steps
@@photoshopcafe Oh okay my bad
All those tools are about serving, not learning. And in The 90s we said “the future is bright, we all need sunglasses”. My guess is that it more so goes in the direction of sticks and stones.
Can you explain how this is different than generative fill? Only difference I see is using the crop tool to generate rather than cropping first and then using generative fill.
? at 2:28 you are changing the resolution correct? And can't a plug in do this for ya?
Why would you want to add a plug-in for something photoshop can easily do?
Great video as always! Low res fills pretty much limit usefulness of the feature. I made the mistake of furnishing an image to a client without noticing it and they were not happy! When do you expect the GF/GE to be ready for prime time?
Problem Lightroom “edit with “ Photoshop beta 25 . Have you encountered that Colin ? Thank you
I did. there is a problem transitioning from Lightroom to photoshop beta and back. I had to revert to the previous version
It’s a known issue
Scary Colin .
wow that's crazy! Question, could you supplement an image so that it seamlessly bl3nds the expanded portion? Let's say that I've some beautiful foreground and background images which I'd like to integrate into an image. Is it capable of doing that?
Secondly, is this reaching out to their server or is it happening within the computer?
Gen fill and expand are cloud based
I have another tut on combining images with ai (I think that was your first question)
Is resolution of expanded areas still limited to 1024 pixels on longest edge?
Just saw this yesterday in a video from Brian Matiash. I have the new version installed, but haven't played with this function yet. Definitely adds some fun to the editing process.
Nice beta update informations thanx a lot
You’re welcome
Great time saving information. Thank you.
all just repurposing tech made freely available on local machines using SD months ago.
There are many LL models out there, SD is just one of them
@@photoshopcafe it's not a large language model at all. it's a diffusion model, and so I guess you just don't even know what you're talking about, or you're paid by adobe. either way, toodles.
Imagine what if Illustrator got ai tools just prompt and get Illustration as done
It's still extremely low res though, right?
yup
looks like still limited resolution
Yup
This is crazy, we really are in another generational leap
Yup
Wow, that sounds very exciting!
But how does that relate to the 1024 x 1024 pixels that could previously be inserted as the maximum size in Generative Fill. Does that no longer apply?
it said near the end that that still stands
great
This is unexpected but very useful
Merci
Hi Colin, Do you have any advance info on when these features will no longer be in beta but added to regular PS like was done with the remove tool. My win 11 doesn't play nice with beta.
Im not allowed to say, even if I do. But I don't know
So that is why Photoshop was acting different today 😂 I really pay better attention
Lol!
We want matching resolution!!!
💯💯💯💯💯
I think we will get it eventually
how do we get the regular generative ai box now? this tool suffers from the 'violated blah blah blah' and the putting a . trick doesn't work on it. I just want to use the regular generative AI function but now when i crop anything i'm automatically forced to use this new generative expand tool
Selection tool evokes regular gen fill like before
my friend, when you are still in that situation try putting in the prompt just a dot and try to generate. Let me know if it worked for you :-)
great video
Thanks
Hi Colin. How come when you only expand the photo you end up with a high resolution (300x300 px), but when you fill with a prompt and expand at the same time the new area is low resolution. Am I missing something? Thank you for your tutorials!
Nope doesn't bother me at all its just for a few seconds and its done with I found it OK . Not like its going to kill me or make me pregnant so why worry.
@@MONTY-YTNOMwtf kind of AI bot response was that? 😂
It’s low res both ways
Pl explain image sharpness thro AI
It does it’s best to match what’s in the image. It’s resolution limited right now
Thank you for this latest very informative feature demonstration. I love the fact that this has been built into the crop tool, with no need for selections to be made. I am looking forward to all of this AI progression to come eventually into the full version of PS.
what version number is this?
Beta
Holy crap!
Is this still using a low-res generated image, which gets blurry when scaled to match image size?
Great feature and great video. Always in the ball as usual Colin. 👍🏻
Unbelievable. Just incredible.
What has emerged in the past year is nothing short of miraculous, can you imagine what the next 5-20 years will bring?
I can't - it's beyond me, but I know I'm excited about it!
A great update that will indeed save some clicks. Thanks Adobe.
Yes it does!
I wonder if - among all the praise for PhotoshopBeta - you also experience the extreme amount of censorship when trying to prompt generative fills on images that either include people or one of the words on their blocklist. Any way to work around this? Your opinion? I've got prompts like "concrete slab" refused...
My guess is the corporate lawyers are exercising an abundance of caution. Hopefully we will come to a happy compromise
my friend, when you are still in that situation try putting in the prompt just a dot and try to generate. Let me know if it worked for you :-)
Is it resolution limited like generative fill?
good question
Yes the same algorithm
yes it seems so still :-/
So let's put our cards on the table: this is obviously "generative fill" combined with "expand canvas" so you don't have to use the marching ants. Convenient. I get it. Here's the problem. If you do the same process manually, you marching ants the top part and genfill, THEN you marching ants and genfill the bottom. As two separate steps, the AI does two separate text-to-image method calls. But with "generative expand", it's ONE method call. If you do it all as one single method call, you will lose a ton of sharpness and granular detail in the new image. This is especially true for any image bigger than 1080x1080 pixels, as the program stretches and distorts it's own genfill content. It's much better to do this manually in chunks with marching ants rather than all in one go.
Another feature to expand the photographic artist's toolbox. Thanks for the video demonstrating the feature.
I also think this is very useful
Wow.... awesome! The things I learn from you! Thank you
Glad to hear it!
Uhmmmm, hoy crap...great video.
Thx
The max 1024X1024 size remains an issue. Filling a 24mp photo, piece by piece in not the proper solution.
The canvas expansion using generative fill is great whether using the former long hand method or the new one. One caveat that I found myself having to deal with almost always is having to add a brightness adjustment layer linked to the generative layer to better match the brightness. At times I also had to use the AI erase tool to get rid of the hair line(s). Hopefully that will get better with newer versions.
Interesting, I haven’t noticed that.I’ll be on the lookout in future
I wish the fills were at the same resolution as the original.
Very clever but pretty useless for professional reults with the constraint on resolution. Look forward to that limitation being removed.
Nice one Colin. That was a lot more information than the videos I have seen about this. I am still wondering what on earth they have lined up for the major version update in October? It will have to something like using your webcam to read your mind, so you no longer need keyboard or tablet/mouse.
Now if only adobe would use some intelligence and learn how to save files accross a network.......
This feature is incredibly beneficial for me.
Me too
This is a step further. The previous version of GF could do the same result, but with more work with selections. Thanks for the nice tutorial.
thats what i thought..
UH MAZE ING🎉🎉🎉
Its getting better and better.
And it will continue to do so
I have hit the subscribe button. 😊
Thanks
thank you for sharing.
My pleasure
This new feature is going to save me a lot of time!
I very nearly unsubscribed when you interrupted the story. I don't understand why RUclipsrs think that's a good idea.
It’s not easy building a RUclips channel
@@photoshopcafe I know that. It puzzles, me, though, that photographers who fuss over minor distractions in their photographs add major distractions to their videos.
You don't actually need to remind people every time, especially those already subscribed. If you have a series of videos, you can give a preview of the next, and promise that subscribers will get early notifications. Then, it's a benefit to me, but not a favour to you.
@@oneeyedphotographer it got me to subscribe
@@MONTY-YTNOM You wouldn't have subscribed otherwise? A reminder at the end perhaps, promising that you will find out immediately a new video becomes available?
Crazy