Digital Photography and post-processing in 2024
A.I. In Photography, 2024 Edition
It is a paradigm shift, a game-changer. Maybe even a miracle.
“Adobe Announces All New AI-Powered Creative Cloud Release Wednesday, September 13, 2023 09:02 AM”
That is not even four full months past as of this writing. I was all over it like a cheap suit when it appeared in beta. I was gobsmacked and remain so. I still can’t believe what it can do for my photography. It is a paradigm shift, a game-changer. I’ve been tootling AI’s horn since day one, specifically as it applies to digital photography and post-processing thereof. I’ve acknowledged its dangers, especially its misuse, even criminal misuse, and I have no illusions that it will not get worse, but I am crazy about it. Nuts. Over the moon.
I saw a video this morning about creating faux fog in the background in Lightroom. The creations were beautiful, seamless, painterly. The author stated that he had intended to use the natural fog already there, but it lifted before he was fully set up, so he shot the woodland scenes anyway, deciding to see if he could fake fog in Lightroom. A commenter rather rudely attacked using the principle that doing anything to a photograph makes it no longer a photograph, et cetera, yada yada. I disagree (as did the author). I see little difference between what we, as editors, do and what painters do. The painter can put things in that were not there and simply not paint things that were. When I started noodling with Photoshop years ago, the first tool I used was the clone stamp.
I subscribe to Adobe’s Creative Cloud (Photography), which comes with Photoshop 2024 and Lightroom Classic on my desktop tower. For 37₵ per day (you can’t buy a small McDonald’s coffee for that), I am always on the cutting edge, as Adobe updates both powerful apps at least once a week, sometimes more, I suspect because the technology is changing so fast. Now in my eighty-second year, I have trouble keeping up, but I’m in there swinging. I love what I can do with it that — thirty years ago, when I pegged my guns and thought I’d never work as a photographer again — photographers of my generation never even dreamed about. The thought never occurred to us, so it was never a fantasy, not even a delusion.
- I can change the apparent depth of field of a picture after the fact. Ah, go on wit’ ya!
- I can seamlessly hide distractions with a few clicks of a button on a machine. Get outta here!
- I can sharpen a picture after the fact, so you’d never know I missed focus.
- I can even double its size with no loss — even increase resolution. You been kissin’ the Blarney Stone!
I sincerely try always to disclose in the caption what I’ve done because I personally know many photographers who bitterly resent feeling fooled.
That list is just a few things I can do, and I sincerely try always to disclose in the caption that I’ve done it because I know many photographers who resent feeling fooled. If you work for a newspicture agency, you will be fired and banned for life for doing it at all.
A few of my favorite things
- Adobe Photoshop Lightroom Classic, latest version
- Adobe Photoshop 2024, latest version, plus Beta
- Topaz Labs Photo AI combines all three earlier tools of Sharpen AI, DeNoise AI, and Gigapixel AI. I have the full suite but now use only Photo AI.
I have always had a folder of pictures I titled Desktop Wides. Initially, they were all supposed to be cropped to a 16:9 aspect ratio to fill the screen with pictures that were meaningful to me and my family when run as a desktop slideshow.
That old folder aged out. The photos were no longer up to my lofty standards (ah-hem); more than half were not cropped to 16:9, so I decided it was time for a complete update with newer material. In the process, I pulled out all the AI stops.
If this doesn’t interest you, that's perfectly fine. I love doing it and yapping about it. I never wish to bore folks.
For Example…
In the Bleak Midwinter
We’ve all done it: gone out looking for but not quite finding pictures. On a clear but bitter day at Lincoln’s Shoemaker Marsh, my buddy and I were mostly cold and disappointed. I shot this footbridge, but aside from Oscar and me, no soul was around. It's a pretty pointless picture, but I hope I learned long ago that if it’s there and you’re there, shoot it.
Above is what I made from it, sow’s ear to rawhide purse (I hope). The clear blue sky was, for me, too clear, so I subbed in a canned sky that I thought improved it. The crop is 16:9. The foreground was burned down and softened with a graduated filter.
The improvements, such as they were, made a better photo, but a picture of a footbridge with no feet on it is beyond meh. Photoshop’s latest generative fill allowed for making a quick lassoed selection of the space where I wanted yumans. The prompt was “three hikers.” It took several tries to get the proportions right and considerable tedium moving them to inboard the fence. If you enlarge them, you’ll see they are far from perfect. Photoshop has the same problem with people as most do at this state of the art. From a distance, though, would you know?
The Green Canoe at Sunup
I like to get out during the Golden Hours. At dawn, there is nobody around. I shot this landscape but felt something was missing.
Cropping the original down to 16x9 was easy, and I lost nothing important, but I thought it needed a boat on the water. I made an elliptical selection with the lasso tool and prompted Photoshop to add a green canoe with a wake in that space. It isn’t perfect, but at this “distance,” it convinces me. I see two people in the boat. The wake shows it is gliding gently to the right. Generated fill even got the light of the barely-there sun right.
When Your Subject is Too Big for Your Set
My DIL received a vase of roses from a student. I brought it to my tiny tabletop set and discovered it barely fit. I could have tried to find another place to shoot it, and I did make several tight closeups of the flowers.
But what I wanted was the whole vase in a horizontal (landscape) frame with the 16:9 aspect, like this, only I’d shot it vertically (portrait):
Mexican Dancers
I love this picture at its regular 3:2 ratio. It was cropped in the camera. I like that the young man has leapt clear of the floor. He is a classical Mexican dancer, stoic and severe. I wanted to add this to my Desktop Wides folder, but the fundamental problem is that it can’t just be cropped. Either I lose the boot or the hat. This was a job for (tootle tootle tooo…) Photoshop Generative Expand. It’s part of the same suite, allowing you to increase the canvas anywhere you need, and the gen. fill will fill in the blanks.
I “cropped” from left to the 16x9 ratio, opening a white space that I prompted the AI to fill with the formerly half-dancer. I wasn’t displeased except for the hat. In Mexican dance, usually, only the men are covered.
I deleted the hat brim on the left, then used the Object Selection Lasso to outline the dancer’s head, making a tight selection. I expanded the selection by five pixels all around, then wrote the following prompt:
“The back of a Latina woman’s head with black hair in a single braid garnished with Mexican colors.”
I was chuffed! I think the illusion is nearly perfect. The flowers are not in Mexican colors, but she is not the main subject. I got what I wished for.
Huffalbos in Custer State Park, South Dakota.
The challenge with this picture was that simply cropping to a 16:9 ratio would crop out the two bison happily munching the moose moss in the background. The picture needed them for compositional balance. I was shooting with a long zoom out of the slowly moving car window, so precision cropping in-camera wasn’t very precise.
This turned out to be trickier than it looked. I first tried using the Content-Aware Move Tool, but it was unsuccessful, leaving parts of the animals’ bodies behind.
Instead, I object-selected each bison and then experimented to find that the expansion needed was a whopping two dozen pixels each to capture the entire animal together with the needed amount of grass. Once done, they could be dragged and dropped much closer to mom and kid (calf), cropped to 16x9, et voilà! I get giddy.
If you’re still here, can you stand one more? It’s an early one from when Generative Fill was first introduced in Beta. I have made more adjustments since.
Mauling Mantids
A thing about photography that has always annoyed me is being forced to play it as it lies. I wanted the picture as you see it: two mantids, probably male and female, getting ready to duke it out. (Spoiler: the male loses). Obviously, as-is, the picture is a cluttered mess.
I made others, close-ups, ground level, the usual protocol, but dangit, I wanted this angle, but crikey, just look at it: concrete walks, signs, backgrounds both too dark and too sharp, a fairly dumpy memory snapshot.
But I had been a full-time working pro, f’cryin out loud! 😫 Maybe with the new toys, I can fix this.
I did my usual editing in Lightroom Classic, then took it into Photoshop 2024. All the concrete has been seamlessly replaced with grass using Photoshop’s generative fill. The sign has been “disappeared” with the new AI Remove Tool, part of the growing suite of essential retouching brushes.
But I want to direct your attention to the background blur, or bokeh. Photoshop and Lightroom, at least the desktop versions, have attained Depth Blur, which allows you to control faux bokeh. In Photoshop 2024, it is in the Neural Filters, so you may not have it. Lightroom Classic, also 2024, is currently in beta while Adobe engineers fine-tune it. They encourage feedback from users. Depending on which versions of Photoshop or Lightroom you have, some of these tools may not be available.
All of these effects could be done by a skilled operator using Photoshop 3, released for Windows with layers in November 1994. The difference over thirty years is automation. Most of these new tools are nearly perfectly automatic. They select what needs selection and magically do their thing. If the app misses, you go in with a small brush and correct the mistake. Jobs that once took hours of eye-straining drudgery can now be done in minutes. For professional retouchers, it’s heaven.
So is it still photography?
I contend that these are still fundamentally photographs. They are based on original photographs. The AI has been trained on other photographs to infill places where the base pictures were lacking. Looked at one way, it’s high-tech airbrushing. Being an honest, ethical person, I disclose except for cloning out something minor, like a fireplug. I am neither a journalist nor an evidence photographer; it is not illegal for me, but if I were forced to disclose every “fireplug” I deleted, I’d be doing it on nearly everything.
There you have it. I’m gonna siddown and shut up now.
📸As always, gratitude for looking in. I sincerely appreciate it! Questions in the comments will be answered promptly, so please ask.😊👍
