Ceci n'est pas une photo by jlmather

Ceci n'est pas une photo

P.S. nor is it a book.
That’s pretty cool. No idea how you do that but it’s a great effect.
January 18th, 2023  
@keeptrying Thanks Linda! Glad you like it … but is it still a “photograph”?
January 18th, 2023  
It’s definitely a photograph and it is art. Mmmmm
January 19th, 2023  
I've seen the audiologist, now it's time for the optician. Love what you've done and the colours
January 19th, 2023  
@clifford @keeptrying Thanks, both: this is just me “thinking aloud” about how I feel about departing from a faithful representation of physical reality, when the resulting image is no longer a “photograph” by any conventional definition (whether “creative” or not) and becomes “digital art”. The image here is based on an essay by landscape photographer Guy Tal, posing the question “what is real”? I see the result as more of an exploration of that question, than a photo of the page of a book (black type on a white page). Sorry if I ruined your eyesight, Cliff: I’ve probably given you a headache now!
January 19th, 2023  
Ha ha! Very clever John and effective! Is this AI imagery?
January 21st, 2023  
@marshwader No, it’s not AI. I just photographed a page of Tal’s book “More Than Just A Rock” and played around with it.
January 21st, 2023  
@marshwader BTW, most of what is marketed as AI is nothing more sophisticated than looking through a rippled piece of glass. They are just nonlinear filters, as opposed to linear filters which are much easier to design. Many noise-reduction filters, for example, are nonlinear (but relatively simple to design, and they’ve been around since before the marketing men caught on to the term “AI”). Move your eye in front of the rippled glass, and the view changes in a way which may appear difficult to predict. Give an AI filter exactly the same input, and you’ll get exactly the same output, unless the design of the filter incorporates a randomised number. The term “AI” tends to be used when the design of the filter is so complex that the only way of doing it is by (essentially) trial and error: “show” the filter an image, make an assessment of the “quality” of the output, and use that to drive an adjustment of the “weights” in the filter (the relative proportions of the input, which are mixed and combined in the filter); try, try, and try again, with multiple images to check that the output is “good” for a range of inputs. Once you’re done, you have your piece of rippled glass.
January 21st, 2023  
Leave a Comment
Sign up for a free account or Sign in to post a comment.