Do you remember a few weeks ago for the Lens-Artists Challenge we were challenged with doing some AI photography? It wasn’t something that I ever thought I would play around with, but I did it for the challenge and since then I’ve thought about it a lot more.
I had seen that you could do it with Adobe Photoshop and it got me thinking. Of course, I had to investigate. I had no idea how to do it. I found an online course that showed me some stuff and I have been playing around. Now, I’m not quite ready to share other stuff, still learning, but I did think it would be fun to try creating images like I did for that other post and see what PS does with the instructions.
You can find that other post here, Lens-Artist Challenge #302 – Artificial Intelligence and Photography.
I thought it might be a good idea to go back and see how the others were and then see how PS did with the same and similar requests.
The first one in the other post was, create an image of Melbourne with no people and foggy morning with sunlight.
Shall we see what PS came up with.
I know Melbourne and while a couple of the buildings are similar it isn’t Melbourne. However, there is the feel of it, and I could imagine it being very similar to how the city would look on a morning like that.
The next one was to create an image of a black church sitting at the base of a mountain with dried grass everywhere that looks like a photo.
PS’s response was this.
I thought out of all of them these were the most realistic. I wonder if people didn’t know AI created them would they know they weren’t photos?
I also did the one my friend asked me to, which was to create an image of the Melbourne skyline and the Yarra River but add some famous New York buildings as well and also replace the Princes Bridge with the Tower Bridge.
I got this.
So there does not seem to be any New York buildings and no Tower Bridge. Though, it is very much like the feeling you get of Melbourne on a lovely spring day. Again some of the buildings are right, but this isn’t how Melbourne looks.
One of the groups that I enjoyed in the first post, besides the church ones, were the railway tracks with the abandoned buildings beside them. So I wanted to try that one again.
I asked PS to create an image of an old abandoned building that is beside train tracks in the middle of nowhere with storm clouds, photographic.
I decided to change it a bit and instead asked for an abandoned mansion. Let’s look at what PS gave me.
They aren’t bad, and not sure how photographic they are. Still, I will have to keep trying many of these to see how we go in the future.
While I was doing the above one I wondered what I would get if I substituted a ghost town for an abandoned mansion.
Here are the results.
Maybe not exactly the results I was expecting, but it is a good start.
One thing I do really like is that I can decide the size or orientation. It is good that they aren’t all square. They are still low res, but I haven’t played to see if you can change that. More playing needs to be done.
I still think AI has a long way to go, but I think if I were going to create images I would be more tempted to use Photoshop. I have only just started playing with it and I do want to learn a lot more about it. See where I can go with it.
I also want to use it with my own images, but I still have a lot to learn about that. I will keep posting here and let you know how I get on with it all. I’m hoping to take some photos this weekend in the city that I can use to play with. We will see.
These seem more successful than some others I’ve seen, although cities seem to pose a general problem for most programs. I think the black church shots might fool people who have never seen that church at Búðir in Iceland in ‘real life’, in its actual setting (I have a photo here if you’re curious: https://www.toonsarah-travels.blog/stunning-landscapes-of-iceland-the-day-the-sun-shone/) And I rather like some of the railway tracks ones.
Thank you Sarah, I was thinking the same. I do wonder if the problem with the cities is more to do with Australia and the image base that Adobe uses doesn’t have that much stuff from here. I should try doing some quintessential Aussie images and see what it comes up with. I was surprised when I saw your photo that it did use the actual church, or a very good likeness. I asked for the other stuff, so that is on me. Thank you for showing me that image, it was good to see. Thanks for that, means a lot.
The ones with the church and the railway lines look very ‘cut and paste’ to me. I can see that you may find A1 helpful if it improved because you do like to do a lot of processing.
I don’t know how much AI I would use, probably some because I can see that some things I do would be easy with using it, but still, it isn’t that great so far. Though I still think it is amazing that you can put in a few commands and get images. I agree, not quite amazing yet, but who knows what the future will bring. Still, I won’t be getting rid of any of my photography gear. Thanks RJ.
I’ve seen some obviously AI generated images online with owls and other wildlife. On many I’m finding you can still tell the difference. A sharp close up photo with no feather detail is a big clue. People not into photography think it’s the real thing. Some of the images are not of any owl species that exist. I don’t mind people taking their own real photos and making them artsy etc. with AI. People have been changing out skies and lighting for a long time when creating images. Some other areas I’m not keen on.
No doubt that AI is photographic or photo real at this stage, but I’m sure it will happen. I’m having fun creating these images, more just to see what is possible and what it comes up with. I do agree that many of us will use it on our own images. I do so much to my images now, so I can’t see why I would take the easier way. I guess we will see where it all takes us. Thank you Sharon.
When a photo generated isn’t a thing that actually exists I am not comfortable with it. Enhancing lighting , eliminating noise, or being artistic with an existing photo I have no issue with. People have photo stacked for a while or combined different photos.
I am getting confused with everyone assuming that I think these are photos, they aren’t.
They are created by a computer, not a camera. I know the difference. We see things all the time of things that don’t actually exist, look at movies, look at computer games. None of my images, well very few of them, show the world as it is either. I love manipulating them and seeing how I can change them.
Good to introduce it as ‘Playing Around’, Leanne. Because that’s what it is; it has nothing to do with photography. The results have ‘no soul’ and somewhere in the half unconsciousness I feel that it’s fake. I have been wondering when I would consider to use it in photography. First things that popped up were red-eye and spot-removal. But those are already available in photo-edit-software; but sometimes the ‘spots’ are not so easy to remove; but with AI it could be done in a split second. A step further would be removing annoying objects that are unavoidably included in the frame. But I think I’d rather take a few steps aside; or not take the shot at all. Is it useless then? No, it is great for advertisement photography; but maybe it’s better to say that it wil replace advertisement photography totally, very soon.
I wonder sometimes, if I did some and put them up and didn’t say what they were would you pick that they were fake? I might have to try it. As my husband was saying today all new phone cameras are using AI so you may be using it and don’t even realise. I think if there are aspects of AI that will help you with your image you will use it. Imagine taking a beautiful photo but there is something in it that you hate and you can’t get rid of it, but you can ask AI to do it for you, wouldn’t you use it? I should have read your comment right through before answering instead of reading it sentence by sentence as I respond. Sorry. Who knows what will happen with it. I have a post coming out tomorrow discussing more about this. Thank you.
Interesting experiments!
Thank you Geoff.
Interesting outcomes
I agree Nora, thank you.
Very intersting and not bad either. 🙂
Thank you Sherry, I agree, not bad.
Impressive creations!
Thank you Egidio.
Looks like you had fun and PS didn’t do that bad. I wonder if you described one of your own pictures, how close would PS come to duplicating it.
I did have fun Anne, and I agree, PS did a pretty good job. I find if you put something really specific like a building it doesn’t do well, but who knows how long it will be before it can.
Interesting, thanks for the follow-up post. You are doing a lot better with PS than I am; I understand from others that the main differential is the quality of the instructions, or “Prompts” that you provide the AI engine. If so, you have some good prompts. Looking forward to future posts on this subject.
Regards,
Robin
Thank you Robin. I find it quite easy, so far. I don’t know why. I have also watched a few videos on how to do it as well. I was watching a course too.
Oh, I think the AI images are awful! They all shriek ‘ersatz’, and lack personality and soul. ‘Real’ photographers rule!
I saw your comment Margaret and thought oh no, but then I thought back to what people were saying about digital photography over 20 years ago and I think your comment would be along the same lines, now most of us are using digital cameras. While I don’t think AI will take over “real” photography, but there is probably a place for it in some way. I think it could get to the point of it being hard to tell the difference.
I know it’s early days, Leanne, so of course AI will improve. However, I think to my eyes it will always look ‘overprocessed’. I limit my processing to cropping, straightening horizon lines and possibly altering the light a little and I really don’t want to do anything else. And on the whole, I prefer the photography of others where the processing is understated. So AI may struggle to please me!
I get what you are talking about Margaret, but your response also surprises me, you know I manipulate my photos a lot and I mean a lot. I do everything I can think of to make them look like something else. I love to see what I can do with them. I turn day time images into night images.
And that last is something I can’t imagine ever doing. The only reason I do any photo-manipulation at all is to return the scene to what I believed I saw.
I love doing all that stuff, sometimes I think I enjoy the processing more than the taking photos, then I go out and have a great time. I don’t think it matters, you have to do what you love doing, and if it isn’t all that manipulation then that fine, nothing wrong with it.
I don’t know whether to marvel or to weep at requests of clever(er) software in their creation of images. 😬
I don’t know either. I can’t help thinking there is no point ignoring it, but maybe work out how to use it for yourself. I certainly prefer taking photos, but I’m looking forward to seeing how I can use some of it on my photos. Time will tell. Thank you.
Dear Leanne
we like these AI images much better than the ones before. They are a promising beginning.
The last picture we like best.
We think that’s a good way to make photography more arty.
Thanks for showing your tries – we are very interested what comes out of it
The Fab Four of Cley
🙂 🙂 🙂 🙂
That is good to hear, I guess time will tell how it goes. Thank you.
All great views and emotions.
Thank you.