I came across a video on YouTube by, FStoppers, about a new feature offered in Photoshop Beta called, Generative fill. I was wondering if anyone has tried it, and if you have what do you think? Or are you going to pass on it?
I also found a video from a photographer named Chris Parker, who is against it and he shares his opinions as to why to give you a different perspective you might be considering it.
I've been using it for a few months now, and is really impressive. Just for removing dust spots or random things out of photos, where I used to use content aware fill, I find generative AI much quicker and much better results.
Here's one example of where I've used it. I have a 10.5mm fisheye lens that filled the frame for a cropped sensor (non-circular fisheye). I took a hack saw and shaved off the built in lens hood off the lens, then used it on a full frame sensor to get more a circular effect, but a full 35mm sensor still isn't big enough to capture the full circle.
This is the inside of my wash machine, and you can see the sensor isn't big enough to capture the full circle.
I did the shot again this year, but this time I pulled it into photoshop beta and used generative AI to add to the top and bottom of the frame to complete the circle.
And if I didn't point out that I used AI to complete the circle, most people wouldn't have noticed.
Yes, I've had a go with it. I think it could be useful for some things. I had a shot of a gannet in flight which I cut the tip of its wing off. Using the generative fill, I was able to include the whole wing with sky above it. Sorry I can't work out how to include the examples. I tried it after seeing someone else example of an owl that had been too tightly cropped. He increased the size of his document and used the generative fill. It made a very good job of adding branches from the tree around it, and giving the owl space to breathe. I was very impressed.
That Parker Photographic video pretty much summed up my thoughts. There are folks who do really impressive photoshopping, but I’m not one of them. I prefer to stick (mostly) to typical darkroom functions…. Crop, exposure, burn&dodge…. Levels, etc. I do like the control I get doing b&w conversion on the computer, but also use in-camera b&w function some.
That said, “you do you”, and if this kind of tool spurs your creativity, then more power to you.
@lsquared - Thank you for replying with your thoughts Larry.
I've learned a lot of my Photoshop and Lightroom skills watching YouTube videos and many times when I'm looking for help like removing a busy background for instance, when I find one, the person who uploaded a video is using a plain solid background. Of course that's no help at all. Fortunately, with the new object recognition feature it is usually a lot easier to do.
I don't know who I am any more. I thought of myself as a wildlife, flowers, or landscape photographer. But when I see a photo like the one in the first video where they had the picture of the bull and changed the background to have it standing on a night time street, I wonder if I should go the way of composting? But then, like Parker indicated, is it really your work anymore?
@skipt07
I fall into the ‘old school camp’ and give an enthusiastic thumbs-up to Chris Parker’s review. I really hope Adobe offers a PS subscription without Firefly and generative fill.
I have used generative fill and I find it quite useful. I use it in much the same way as @sudweeks and @yorkshirekiwi. It is very good at removing unwanted elements from a shot and also for expanding the canvas for those times when, for whatever reason, I just didn’t allow enough space around the frame. And before someone says “get it right in camera” we all know there are times when that just isn’t possible, especially when it comes to unwanted elements like signs or garbage cans. That said, I find it more than a little worrisome that significant elements of an image could be created out of thin air and passed off as a the work of the photographer. That doesn’t sit well with me. What boils down to, I guess, is the integrity of the photographer.
@cocokinetic - Thank you, Karen for your insight. As you might tell from my question, I'm conflicted. While I have been using Photoshop for years I only started to subscribe two years ago and have greatly enjoyed the features that are offered with it. But a part of me would like to dabble in this newest AI, just to see what I can create. I wish you could have the ability to have both the regular Photoshop 2023 and the Beta installed at the same time. But I don't think that's possible. My concern is that if I install beta then decide to go back, will I be able to, or how much of a hassle would it be?
Oh yes, the Photoshop 2023 doesn't have the Firefly feature generative fill included.
I'm guessing that Chris Parker is primarily a portrait style photographer where he can have more control of his shoots.
We've all had situations where we take a number of shots of a subject and the best overall had a part of the subject missing or a photo that the subject is closer to the edge than we would like. But having beta you can correct this. I see nothing wrong with that at all. It's where we start adding elements to a photo that wasn't in the original and calling it your creation.
But then too, and I've done it myself, I've seen exquisite landscape photos that are mind blowing and having used Photoshop and Lightroom, I wonder, is that what they really saw or how much of it was created in Photoshop?
Years ago when I got into photography, I came across Ansel Adams' work and I thought that I wanted to mimic his Zone System in my work. I learned about a man named Fred Picker who had a business called Zone VI that taught the Zone System. I thought that if I could find out what settings I needed to use to get the medium gray, when I took a picture of something it would look like an Adams' print right out of the camera. But I was wrong. I later learned that Adams spent large amounts of time burning and dodging his prints to achieve the final product.
So, I'm only guessing, but I think if
Adams was around today he'd be using Photoshop and/or Lightroom to achieve his signature prints. But he only manipulated what was there. He added nothing.
@skipt07 HI Skip, you can have both versions of Photoshop on your computer at any one time. If you just have cc2023 open your files automatically open in that version. If you have the beta version it will ask you if you want to open the file up in the beta version.
@cdcook48 Yes I agree with you about adding generated images, even when doing composites. I just think Beta will make things easier for selections etc.
I regard Photoshop Generative Fill as another tool to use. Since it manifested itself I have used it for editing my own photographs. As a one-off, I have used it to create the starting image for the next WWYD challenge.
From the updated Beta version of Photoshop, I have probably used the remove tool far more than generative fill. There are similarities in the two tools when used to remove someone or something from an image and in my view represent a significant improvement on content-aware fill.
I think there is also a skill set to be learned, particularly with selecting words in the generation prompts and the size and shape of the selected area to be replaced on an image.
I will not be using it to create a photograph of myself driving a sports car at the top of Mount Everest.
I like it, but it is another tool for me to use.
Write a Reply
Sign up for a free account or Sign in to post a comment.
Tweets ahoy!
Join the Group
Pin With Me
There's also a way to have photoshop AI use a photo as a reference, then generate an image based on that. I've played with it a little, but didn't get as good of results as these: https://twitter.com/actionmoviekid/status/1673540357449973760?s=46&t=XsS1uW8sb6s6s8lKVH3plg
This is the inside of my wash machine, and you can see the sensor isn't big enough to capture the full circle.
I did the shot again this year, but this time I pulled it into photoshop beta and used generative AI to add to the top and bottom of the frame to complete the circle.
And if I didn't point out that I used AI to complete the circle, most people wouldn't have noticed.
That said, “you do you”, and if this kind of tool spurs your creativity, then more power to you.
If you uploaded the image to your Project pages, you can post it like you do for any challenge or Flashback Friday.
Thank you for your reply and in put Carole!
Sorry about the bad joke.
I've learned a lot of my Photoshop and Lightroom skills watching YouTube videos and many times when I'm looking for help like removing a busy background for instance, when I find one, the person who uploaded a video is using a plain solid background. Of course that's no help at all. Fortunately, with the new object recognition feature it is usually a lot easier to do.
I don't know who I am any more. I thought of myself as a wildlife, flowers, or landscape photographer. But when I see a photo like the one in the first video where they had the picture of the bull and changed the background to have it standing on a night time street, I wonder if I should go the way of composting? But then, like Parker indicated, is it really your work anymore?
I fall into the ‘old school camp’ and give an enthusiastic thumbs-up to Chris Parker’s review. I really hope Adobe offers a PS subscription without Firefly and generative fill.
Oh yes, the Photoshop 2023 doesn't have the Firefly feature generative fill included.
I'm guessing that Chris Parker is primarily a portrait style photographer where he can have more control of his shoots.
We've all had situations where we take a number of shots of a subject and the best overall had a part of the subject missing or a photo that the subject is closer to the edge than we would like. But having beta you can correct this. I see nothing wrong with that at all. It's where we start adding elements to a photo that wasn't in the original and calling it your creation.
But then too, and I've done it myself, I've seen exquisite landscape photos that are mind blowing and having used Photoshop and Lightroom, I wonder, is that what they really saw or how much of it was created in Photoshop?
Years ago when I got into photography, I came across Ansel Adams' work and I thought that I wanted to mimic his Zone System in my work. I learned about a man named Fred Picker who had a business called Zone VI that taught the Zone System. I thought that if I could find out what settings I needed to use to get the medium gray, when I took a picture of something it would look like an Adams' print right out of the camera. But I was wrong. I later learned that Adams spent large amounts of time burning and dodging his prints to achieve the final product.
So, I'm only guessing, but I think if
Adams was around today he'd be using Photoshop and/or Lightroom to achieve his signature prints. But he only manipulated what was there. He added nothing.
From the updated Beta version of Photoshop, I have probably used the remove tool far more than generative fill. There are similarities in the two tools when used to remove someone or something from an image and in my view represent a significant improvement on content-aware fill.
I think there is also a skill set to be learned, particularly with selecting words in the generation prompts and the size and shape of the selected area to be replaced on an image.
I will not be using it to create a photograph of myself driving a sports car at the top of Mount Everest.
I like it, but it is another tool for me to use.