Source: Engadget
Article note: This picture is pretty much a complete description of "Computational Photography" features. It's mostly cheap bullshit that makes wrong images, and I technically work in the area.
A woman was photographed standing in front of two mirrors with an iPhone camera, but the actual photo shows three completely different arm positions. The arms are in different locations in mirror number one, mirror number two and in actual real life. Is it Photoshop? Is it a glitch in the Matrix? Did the woman take a 25-year trip inside of Twin Peak’s black lodge? No, it’s just a computational photography error, but it still makes for one heck of an image.
It all comes down to how modern smartphone cameras deal with photography. When you click that camera button, billions of computational operations occur in an instant, resulting in a photo you can post online in hopes of getting a few thumbs up. In this case, Apple’s software didn’t realize there was a mirror in the shot, so it treated each version of the subject as three different people. She was moving at the instant the photo was taken, so the algorithm stitched the photo together from multiple images. The end result? Well, you can see it above.
Smartphone camera software always pulls from many images at once, combining at will and adjusting for contrast, saturation, detail and lack of blur. In the vast majority of cases, this doesn’t present an issue. Once in a while, however, the software gets a tad bit confused. If it was three different people, instead of one with a mirror, each subject would have been properly represented.
This is something that can actually be recreated by just about anyone with an iPhone and some mirrors. As a matter of fact, there’s a TikTok trend in which folks do just that, making all kinds of silly photos and videos by leveraging the algorithm's difficulties when separating mirror images from actual people.
This article originally appeared on Engadget at https://www.engadget.com/what-did-an-iphone-camera-do-to-this-poor-womans-arms-201507227.html?src=rss
