AI & Photography

https://creator.nightcafe.studio/

Photography has to do with the art to create pictures. At first by painters it was not seen as art. You pressed a button and the picture was there. For a painting you had to work for hours and you needed the skill to be able to paint what you saw. At least when we talk about landscapes, seascapes and portraits; pictures that had a large reality.

But later in history this changed and nowadays photography is seen as art. As well in color as in black and white. Photographers are seen as artists too. Maybe because after ‘just pressing the button’ the picture is manipulated a lot to make it look totally different. With modern technology like Photoshop this is easy to do, but in former days black and white photographers did the same already in a different way.: they shaded parts of the picture of gave other parts extra light (dodge and burn, the same terms that are still used in Photoshop and Gimp). So the picture did not get equally lit over the whole image. That alone was already an art. He remarked about that: The negative is the score, and the print the performance. When he had done his working prints and made the final one, there was only one of it. It was unique. And in that way a true work of art. Here is an example of a mapped print. It was done by the master printer Pablo Inirio. It is thought that Adams did something similar. He printed most of his pictures himself and had a tone mapping system for that.

You can read more about it here: https://www.timagesgallery.com/blog-2/did-ansel-adams-photoshop-his-images

Adams said: There are always two people in every picture: the photographer and the viewer. That was the way he saw photography.

With AI, created by a computer we appear to differ a lot more in opinion if it is art or not. In a way it is, cause there is only one, original picture. You can tell AI exactly the same ‘promt’ (the description of the picture it has to make) and still it makes another picture. Here we are not talking about skills of the computer tho, but about a random outcome. The writing of the description takes skill. It has more to do with writing and in that way ‘ painting the picture’ . Still art or not, it is a great way to compose new material.

A nice test would be to take one of my pictures and then describe it to the AI, see if it will look a bit the same or not.

I will use NightCafe for that. It is free to use and there is a pro version, with many more possibilities of course.

Here are the results.

Here is the picture I used.

Here is the description:

An undergroundstation with the platform in the middle and two tracks on each side. On the right side the train to Heiligenstadt arrives. There is a clock hanging from the roof, signs with the times trains arrive. There are people coming from the train with bags and suitcases walking towards the viewer. The ceiling is made of metal and the floors are shiny with markings for blind people to find their way to the exit

The two above look really nice, but they are no underground stations. The third down neither.

And here are a few of the results

The program makes mistakes too, like these ones:

Well it was fun playing with the program. I hope you had fun reading it. 

And it might be fun to try it out yourself in for instance NightCafe. There are more websites around, but I am just familiar with this one.