I have seen plenty of photographs that have been so overbaked in post processing that that they have an otherworldly or fantasy quality to them. Personally, I don’t like such photographs, but they seem to have some popularity with the general public.
Recently, I’ve been playing around with AI photograph generators. It’s my impression that many AI photographs have such qualities.
Another common AI photograph quality is lack of definition. There seem to be too many pieces of the photograph out of focus. Then too, many AI photographs are just botched up (literally). Finally, many AI photographs just don’t look right.
Nevertheless, in some cases the AI generated photographs look real. It’s difficult, if not impossible, to tell that they are not traditional photographs.
The bottom line is that you need to play around with words and make endless experimental runs using different words with the generator to craft the photograph you want. But maybe it’s just easier and more productive to go out and take the picture.
That said, my experience is just a one-off. You need to experiment yourself with AI to see what it can do for you. It’s not going to go away. And in 5 years it will work much better than it does now.
First, you need to distinguish between AI assisted post processing and AI generated photographs. AI assistance is present in many photograph editors now, and it makes the normal thing you do easier and faster.
An AI photograph (image) generator, in contrast, creates a photograph from words you enter into the generator. You process the resulting photograph by changing the words and re-running the AI generator to see what you’ve got until you’ve got what you want. There are numerous AI photograph generators. Try one.
AI: white horse green pasture snow mountains [DALL-E 2 photography]
AI: white horse green pasture snow mountains [Adobe photography]
AI: white horse green pasture snow mountains [Microsoft photography]
See what you think. AI generated photographs may be adequate for many situations. But in my opinion, AI generation is currently not useable for aesthetic, realistic, color photographs; that is, the type of photographs that most of us strive for.
Nonetheless, AI generation can be useful for turning photographs into impressionist art or abstract art and for comparable processes.
Picasso style AI image of the white horse above (Micrpsoft)
AI: a image of fermenting wine grapes painted by salvador dali acrylic paint [by Dean Busquaert]
AI is the new parlor game
I found myself with three other retired people after dinner. They all had their noses in their phones. They were making AI images, AI poetry, AI songs, and AI videos and then presenting the results to each other. The results were not particularly appealing, in fact boooring. Rudely, I dozed off until my spouse kicked me awake, and I had to pay polite attention. Of course, you know if AI is already a parlor game that it’s also well entrenched in internet platforms. YouTube, I’m guessing, has more AI video than the rest of the digital world combined.
Finally, you need to understand the combination of AI assisted post processing and AI generated images. Software such as Adobe’s Firefly incorporated into Photoshop (beta – moderate learning curve) enables you to easily apply complex processes to your photographs but to add AI imaging as well. In any case, this is a path to endless fiddling around.
Not only is this powerful digital technology but apparently Adobe is doubling the price of Photoshop (or maybe there will be two versions of Photoshop, one without AI generation).
AI is a technique for which you need to know the capabilities and limitations to determine how you might use it. Maybe a parlor game after dinner is the avenue to knowledge after all?
Five Photography Club members did a photo walk/drive in the Williams Creek Reservoir area on June 22. It was a gorgeous day, with a few fast moving clouds in an otherwise blue sky. Our goal was to photograph mountain scenery, reflections, wild flowers and whatever else we might see. There were still a few massive patches of iris, as well as other wildflowers in bloom. We managed to find two of the elusive forest gnomes; carvings of bearded gentlemen in stumps near the lake. Below are a few photos of the day, from Dave Anderson, Andy Butler, and Joe Sinclair. Click on the photos to see larger versions.
The July meeting of the Pagosa Springs Photography Club will be held on Wednesday, July 12, 6:00 p.m., at the Community United Methodist Church, 434 Lewis Street in Pagosa Springs. Our speaker this month will be Greg Holden. Greg’s topic will be Learning to See Creatively.
Greg Holden is a photographer from Longmont, Co who uses his photography to share with others how he sees the world. He enjoys creative photos of little scenes often overlooked by others rather than postcard landscapes. He often explores abandoned places and captures the details of the layers of paint and rust on a discarded automobile. Greg enjoys teaching others and sharing his passion for photography and speaks around his home in the greater Denver area, but also continues to speak and judge virtually, at clubs throughout his former home in the MD and VA areas. You can see Greg’s work on his website imagesunderfoot.com .
In his presentation, Greg will discuss his approach for finding different viewpoints in common scenes and how he creates interesting abstract images by isolating a part of a scene. Whether it is iPhone photos taken in his kitchen or DSLR images from local parks, Greg’s photos will demonstrate that you do not have to travel to exotic locations or spend hours using software manipulation to make creative images.
Our meeting will begin at 6 p.m. with socializing, and the program will begin at 6:30 p.m. This will be a hybrid meeting, also available on Zoom. The Zoom link will be emailed to members; others who wish to attend may request the link by email to abutler@mac.com. Photography Club meetings are open to the public.
The Pagosa Springs Photography Club promotes educational, social and fun interactions between all who enjoy making and viewing great photography. The club sponsors educational programs and outings to help photographers hone their skills. We welcome photographers of all skill levels. Dues for 2023 are just $25 ($35 family). For more information about the club, and to download a membership application, visit our website at https://pagosaspringsphotoclub.org/about/ .
Digital art, regardless of how you make it, comes out flat and two-dimensional. I suppose one exception is a 3D movie and another is a 3D photo sterocope (antique). But this article concerns only two-dimensional photographs. The significance of two dimensions is that a work can be printed on a printer, shown on a screen (e.g., computer monitor), painted on a canvas, displayed on other flat surfaces, or any and all of the preceding. That’s something to keep in mind when trying to decide what is a photograph. You need words to make distinctions (labels, categorical names). There is no third dimension to help make distinctions.
To begin this analysis, there are some people who argue adamantly that only a photograph untouched by manipulation is a photograph. This is an absurd dogma at a time when every photographer has access to a powerful digital darkroom (e.g., Photoshop). Even in the predigital era, for example, a noted photographer Ansel Adams made his fame by clever manipulation in a chemical darkroom.
Thus, manipulating photographs is nothing new. Digital technology has just made it easier and more powerful. And almost every photographer uses such technology.
The problem is that when one manipulates a photograph too much, is it still a photograph? Too much is a judgmental question. And no one owns the definition of the word photograph. Consequently, any dogma that declares that some photographs are not photographs or are bogus photographs is obviously deficient.
So, let’s take some examples. Suppose one accentuates the color saturation in a photograph (e.g., in Photoshop). Is it still a photograph? Remember Fiji Velvia film? A very colorful film with accentuated color saturation. It’s difficult to claim a Velvia photograph is not a photograph.
This photograph is +100 saturated in Photoshop (looks like Velvia)
For another example, it turns out the moon is tiny in most landscape photographs. If one pastes in a larger moon, is it still a photograph?
Of course, it is. And what if you subtract something from a photograph (a common practice)? Is it still a photograph? Or is it a bogus photograph?
Unfortunately, to some people these techniques are cheating. But since no one owns the word photograph, the cheating analysis can’t stand.
HDR
If you overuse HDR (high dynamic range) processing in Photoshop too much, your photograph starts to look like an artist’s color pencil drawing. And with more even HDR processing, it eventually becomes a color pencil drawing, in effect. Is a heavily HDR-processed image still a photograph?
If you think there is an immutable definition of photograph, where do you draw the line? Obviously, some manipulation may be acceptable to you, but other manipulation is not. The fact is, where you draw the line is your dogma (perhaps your evangelistic pursuit) and probably doesn’t exactly match anyone one else’s.
The latest leap into this quagmire is a photograph created by AI (artificial intelligence). Is an AI photograph really a photograph? Of course, it is. Who is to say it’s not?
Recently Boris Eldagsen won a prestigious Sony world photography award. He then revealed that the photograph was AI-generated. The photograph is very compelling. But some people felt betrayed. The photograph was created by a careful selection of words fed into an AI generator. Does that make it any less a photograph than an Ansel Adams’ mountain severely crafted in a chemical darkroom? Certainly not.
Perhaps someday there will be accepted labels that distinguish photographs created in different ways. Until then, it’s all very fuzzy and controversial.
Eldagsen’s AI generated prize-winning photograph
Where’s the creativity?
For AI, there’s the proper selection of words to run the AI generator. This is a creative act and can be a very complex chore. In addition, when one gets a desirable photograph from AI, one might be likely to do some digital manipulation to make it better. It could be said that this is more work than running out and shooting a normal photograph.
AI unrelated to photography
AI has implications for societal and technological changes unrelated to photography. Many even seek legislation as an attempt to control the evolution (revolution) of AI. That’s a matter that will not immediately relate to a definition of photograph.
AI: bald eagle selfie
There are, however, places for unmanipulated photographs taken with a camera. One is journalism. Manipulated photographs used in journalism are appropriately unethical. Journalism is expected to tell (show) the absolute truth. Likewise, in the justice system, unmanipulated photographs are required for accurate evidence. Even in the insurance business (and other normal businesses), unmanipulated photographs are important. Think about the photograph of damaged vehicle used for a claim adjustment.
There is also implied accuracy. An example is taking photographs for a travel brochure. It is implied that the photographs are not misleading (or at least not too misleading).
Then, of course, if one represents his or her photographs to be unmanipulated—but they are manipulated—that’s fraud. And fraud is socially unacceptable and even illegal in some cases.
Nonetheless, when it comes to creativity, anything goes. All artistic tools are at the disposal of creative photographers. David Hockney (leading artist) uses iPhone photographs as sketches to start his paintings. Painters, going back many centuries, used camera devices (e.g., the camera obscura) to create sketches to start their paintings; when this was first revealed a few years back, many curators and collectors thought of it as cheating. What do you do about cheaters from the 1400s? Fine them? How many guilders? And how do you collect the fines?
Photographers have been adding and subtracting images to and from photographs for decades (even in chemical darkrooms). If a photographer wants to call their work a photograph, who is to say it’s wrong (or bogus) regardless of how it was made?
Below is my photograph created by super, super saturation in Photoshop. Is it a photograph?
If it’s not a photograph, what do you call it? This gets back to the first paragraph. One needs to distinguish between photographic creations by labels, because photographs are all two-dimensional (flat). You can’t distinguish between them using a third dimension (e.g., statue, base relief). And with exploding modern technology, we don’t have widely-accepted labels to make distinctions, specifically among different kinds of photographs. This situation is likely to be temporary, perhaps for the next 25 years. But until we have the new labels (widely used), we’re stuck with ambiguous words like photograph.
For instance
Perhaps aig could become the word to label photographs created with AI; that is, AI generated. Perhaps addphoto and subphoto could become the words to label photographs created by adding or subtracting part of the image. Etc.
In the meanwhile, curators, art show judges, collectors, and the like are overrun by the new technology and are prone to dogma that they can’t easily define. This causes a great deal of friction in the photographic community.
The lesson to be learned is that one must evaluate each image (work of art) for its aesthetic value: how it tells a story, or other subjective criteria. How it was made is of no concern. It’s either evocative or informative, or it’s not. And if the artist calls it a photograph, it is. Only when photographs are used for special purposes (e.g., travel brochures, justice, journalism) is reality de rigueur.
This article just scratches the surface of complications for the word photograph. For instance, expectation is a factor. If one submits a photograph to a New York City gallery, the definition of a photograph is likely to be wide. If one submits the same photograph to a rural-county-fair-photography contest, the definition of a photograph is likely to be narrow.
This post processing is a little over baked (coupling on a train), but it’s still a photograph.
So, don’t let anyone tell you your photograph is bogus or not a photograph.
“When you show me a moving, heartfelt image that causes me to laugh, cry, think, or otherwise react, I couldn’t care less how you did it. I don’t care if you used a pinhole camera or film or an iPhone. What I care about is the image.”
Although we are used to plenty of bluebird days here in the shadow of the San Juan Mountains, that isn’t always what we get. And for our June 2 wildflower outing, we had clouds and cool weather. Luckily, cloudy skies can still be great for photographing wildflowers! Several members of the Club spent the morning wandering through meadows and forests near Pagosa Springs to enjoy and photograph spring wildflowers. Thanks to Gary Webb, Rob Hagberg, Mark Guenin, Kristine Rubish, Andy Butler, Chris Plemons, and Dave Anderson for participating. Below is a sampling of the images. Click on an image to view a larger version, and use the arrows to. move from one image to the next.
What is cropping a photograph for the sake of discussion in this article? It’s changing the shape of the image into something other than a rectangle. Thus, when it displays on a digital screen, it has a new shape.
Create a new image in Photoshop the same dimensions of the photograph and make sure it’s all one color (white is the default).
2. Make the opacity of the image 0. This is the background layer, and the background hash will show as you work.
3. Create a second layer.
4. Import and place the photograph onto the second layer covering the entire layer. Then rasterize the layer.
5. Use the rectangularmarquee tool (or another selection tool) to create a shape that you want to cut away.
6. Erase that shape; be careful to erase everything within the shape.
7. Create more shapes to cut away and erase them until you get the crop (new overall shape) you want.
8. Merge the second layer to the background layer using the merge layers tool.
9. Save as a .png (PNG – portable network graphic).
How will this display? The cutaway portions of the photograph should not show (should be invisible) on most digital displays. (On some rare digital displays, the background hash will show leaving the photograph a rectangle with the background not invisible.)
Note: this will work only if saved as a .png photo files.
If you’re going to cut away portions for a metal print, make the background layer a florescent color with full opacity so that the photo service will know where to cut. After the cuts, the metal print will have a creative shape.
Most photo services do not offer cutting. One that does is MagnaChrome:
Pagosa Springs Photography Club 