'One in a Million' iPhone Photo Shows Three Versions of the Same Woman [View all]
Last edited Sat Dec 2, 2023, 12:36 AM - Edit history (1)
https://petapixel.com/2023/11/16/one-in-a-million-iphone-photo-shows-two-versions-of-the-same-woman/
Standing in front of two large mirrors, Tessa Coates reflection does not return the same pose that she is making, and not only that, but both reflections are different from each other and different from the pose Coates was actually holding.
While Coates was holding one arm up and another down, the reflection on the left is seen holding both arms down, and the reflection to her right is holding both arms up to her waist.
An Apple technician named Roger explained to her that an iPhone is not a camera, its a computer. It takes a series of burst images very quickly even though its not a panoramic or a burst, explains Coates.
Roger informed Coates that she moved her arms as the camera took a series of images from left to right and it made a different image on each side of the photo.
When you click the shutter on your smartphone it takes multiple images and chooses the best elements in a very short space of time. AI makes the decision.
Much more at PetaPixel link above.
Primer on Computational Photography:
https://petapixel.com/computational-photography/
So much for phone evidence photos. I get very good jpegs from Nikon's internal CPU, and they can be tweaked in the menu options. I do minimal processing of the jpegs with Apple's Preview app or GIMP, because the JPEGs are so nice.
Years of film photography teach you to get the shot "right" every time, or as close as you can get.
When I do process the raw images, it will be with RawTherapee or DarkTable, both free and open source. Maybe when I take a blah picture, which may be never.
But it will be fun. 📷
Any misspelled or goofy words here are the fault of autocorrect.
Edit: John Gruber at Daring Fireball is a doubter.
https://daringfireball.net/linked/2023/12/01/fake-iphone-computational-photography-glitch
If you're tired and can't sleep, there's a discussion at Hacker News with 377 comments.
https://news.ycombinator.com/item?id=38482085