Gizmodo 
Enter your username and password.
-
posts about #andrewhicks more → 
Mathematician Creates Impossible, Rule-Bending Mirrors
Tip your editors:
tips@gizmodo.com
Editorial Director:
Brian Lam | Email | Twitter
Editor:
Jason Chen
Email | AIM | Twitter
Features Editor:
Wilson Rothman
Email | Twitter
Senior Contributing Editors:
Jesus Diaz
Email | AIM | Twitter
Mark Wilson, Reviews
Email | AIM | Twitter
Contributing Editors:
Matt Buchanan
Email | AIM | Twitter
Adam Frucci
Email | Twitter
Sean Fallon
Email | Twitter
Jack Loftus
Email | Twitter
John Herrman
Email | Twitter
Dan Nosowitz
Email
Chris Mascari
Email
Danny Allen
Email | Twitter
Rosa Golijan
Email | Twitter
Chris Jacob
Email
Columnist:
Brendan I. Koerner
Interns:
Don Nguyen
Email
Kyle VanHemert
Email
Comment Intern:
Nick Ellenoff | Email
Comment Account Questions:
Comments@gizmodo.com
| posts about #andrewhicks more → |
Mathematician Creates Impossible, Rule-Bending Mirrors |
Original material is licensed under a Creative Commons License permitting non-commercial sharing with attribution.
02/25/09
It's actually performing the flip between you and the mirror. So if you're in the wrong spot it will just look blurry.
02/25/09
all true, except the lens vs mirrors thing. a concave mirror exhibits some of the basic effects of these mirrors, and yet is not lense-like.
02/25/09
So why would a stair climbing robot need such a mirror? It would just create a part which could go out of alignment, and it's not like doing a transformation like this on an image is particularly cpu intensive. In fact, I was doing stuff like this years ago on a 486 in realtime.
02/25/09
"If you can produce such an image using a mirror, then there's no reason you couldn't recreate it agorithmically as well."
there are plenty of good reasons why you can't recreate it algorithmically in real time
02/25/09
oh man shoulda kept reading the whole comment. putting emphasis on 'real time' makes me look like a jackass now.
anyways, if the robotics maker could go without one more moving part, i'm sure they would. i won't pretend to know why they're going the mirror route, but i don't doubt that there are several sufficient reasons.
02/25/09
02/25/09
02/25/09
02/25/09
02/25/09
it's like the prosecutors trying to talk about bittorrent and failing... talking optics without understanding is a little off.
02/25/09
Optical distortion: An optical aberration caused because the transverse magnification may be a function of the off-axis image distance. May be positive (pincushion distortion), or negative (barrel distortion).
There is no pincushion or barrel distortion in the mirror's reflection - all lines are straight as stated in the article.
02/26/09
Actually...they're not. I didn't notice this right away, but if you look at all the lines that radiate outwards from the center, yes, they are all as straight as a rail. However, if you look at the vertical line just to the right of the man's right elbow, it does bow outwards a bit.
Really, there is no way to make a wide-angle lens or mirror that won't distort straight lines in some manner. All you have to do is point that thing at a sheet of graph paper and you'll have proof positive.