07.09.2023  •  Photographers NEWS

0
 

CLOUD! - WE! Shoot it, aka Michael Compensis & Thomas von Salomon, present one of the world’s very first hybrid works of photography and AI with commercial character

In their personal project ‘Cloud’, the two car photographers from WE! Shoot it, aka Michael Compensis & Thomas von Salomon, get to the bottom of what’s possible at the moment with AI as an imaging medium. The result is the world’s first spread of commercial character for which an AI and photography hybrid was used.

“When Michael began, quite early, at the beginning of 2023 to experiment with text to image and image to image AI, I was still very skeptical. It all looked like a nice thing to play around with, but I didn’t really see any sustainable use of AI in our photographic work,” Thomas von Salomon tells GoSee.

A half a year of working with it intensively, doing research and a bit more training later, it was an entirely different story.

“It is actually quite similar to photography. Of course, you can create a few nice pictures pretty fast in the beginning, but if you really want to be able to prompt properly and get reproducible, target-driven results out of AI, it does take a little more. A deeper technical understanding and experience are crucial,” says Michael Compensis.

“It became clear relatively quickly that certain content could be generated well through prompts, and some content, on the other hand, not at all. AI is in part good at merging elements in an unusual manner. This makes images possible that could hardly be realized at all or only under extreme difficulty by means of photography alone, and it also enables the development of ideas in the design and creation process that only a year ago were unimaginable. But when it comes to the real depiction of objects, you reach the limits of what you can do with AI pretty fast.

The prompt ‘Porsche 911’ may then contain important cross-references from the training material in the generated image so that it’s possible to recognize a Porsche, but it has little to do with the real car upon taking a more detailed look. Which is why, we knew fairly quickly that serial commercial images could only be generated with a hybrid method.” Michael Compensis concludes.

“Our intention was to use the capabilities of AI as an imaging technology to create a spread with a surrealistic look. As a basis, we decided to shoot the photographic elements in the daylight studio of Inmotion in Allgäu because that would already result in a very graphical appearance with daylight. We really had a lot of luck with the weather because the light and sky themselves already had an otherworldly appearance to them.” says Thomas von Salomon.

What the Inmotion Studio did above all provide the duo with was an empty stage with enough depth to give the AI free space. The car in front of the lens was the Mercedes A-Class. All other elements were generated by the creative duo with various AI solutions.

“Thanks to the extensive experience we gained through months of testing how to prompt, we were able to generate the corresponding elements reproducibly and in a uniform visual language. In doing so, it was above all also important to us to create a certain tension between the individual motifs with corresponding size ratios, which tends to be rather difficult to achieve with generative AI. But I think we did a very good job of it.

The challenge then, in the end, was once again merging it coherently with the photographic elements in post production. Also, we had to first overcome the obstacle that most AIs are limited to a relatively low resolution up until now because we wanted to use only excerpts of some of the generative elements. In some cases, this made it necessary to use a resolution ten times what the AI is actually capable of delivering.” says Michael Compensis, who took care of post production in-house this time. “A hybrid approach was then taken here again with different AIs for generating and scaling.”

The result in the end is an impressive spread with serial character that consciously explores the question of reality and surrealism in artificially generated images. By the way, the motifs are all at least 8600px wide, which means they can be presented in large format without a problem.

The creative duo WE! SHOOT IT is represented by Kelly Kellerhoff, Berlin.

GoSee : we-shoot-it.com

 
Featuring: Kelly Kellerhoff
WE! SHOOT IT

WE! SHOOT IT

 
WE! SHOOT IT

WE! SHOOT IT

 
WE! SHOOT IT

WE! SHOOT IT

 
WE! SHOOT IT

WE! SHOOT IT

 
WE! SHOOT IT

WE! SHOOT IT

 
WE! SHOOT IT

WE! SHOOT IT

 
WE! SHOOT IT

WE! SHOOT IT

 
WE! SHOOT IT

WE! SHOOT IT

 
0
0 Comments