I think it’s the way the chequered diamonds alternate in orientation. In this image, the top two lines of diamonds have had been given a standard unchanging orientation, and the top blue line looks straight.
Oh, good one. That does work for me.
What’s the right place? That was one of the first things I tried when I saw it, but no matter where I put the ruler, and no matter how much I knew the bars on each side started the same distance from the top or bottom on each side, I couldn’t see anything but slanted tiers of blue.
Aha! Nice job.
The circles are concentric, not a spiral.
I couldn’t find the animation on my iPad, either, but on my desktop, I saw it fine. It’s an animated GIF, the first top on the page (the one right under the heading that says “Scroll Down for Video”). I guess if it doesn’t load in all the frames for whatever reason, it’s just a static image of the illusion.
It looks like the little black & white diamonds make the big black and white squares look slightly like this:
<<<<<<<<
>>>>>>>>
<<<<<<<<
which makes you think that they’re slightly smaller on the left, and slightly bigger on the right
which only makes sense if the blue lines are diagonals.
Cool illusion.
One thing I find interesting is that, if the image is big enough that I can’t focus on the whole thing, and move my head along the lines, it starts to look like multiple diagonal lines beside each other for a split second before looking like one long diagonal again. I think it’s because I’m trying to follow it at a diagonal, which means I eventually get too low.
If you stand back about 20 ft and look at it ,it does appear almost straight but still not perfect. It might be because my brain remembers the first image I saw. I think it is all based on the little diamonds being off set.
Powerful illusion! One way to see it is an illusion is that when you look at the black squares you will see they are all the same size so the apparently converging/diverging lines must be parallel.
Except I see them as different sizes as part of the illusion!
Ohhhh.
holy crap, you’re right.
I was completely taken in - I was thinking that the effect in question was the slight bowing that I can’t top seeing of the **vertical **lines, which makes them look curved to the left.
This one is way more impressive than I thought
I had the same idea, and stood 35 feet away and looked at the image, fully expecting to see the lines as parallel. Nope. Just as slanted as up close. I’ll ask my brother, who has not yet seen the image, to look at it from 35 feet and report back.
Related to this, I was surprised that it still looked slanted even in the little thumbnails of photos on my phone. There was no way I was seeing the little diamonds at that scale. At least not consciously: when I moved away from the phone, they finally straightened out! So weird.
They looked slanted to him also. So at least for him, not seeing it previously isn’t relevant.
Wait, he saw them as parallel, or do you mean he saw them as slanted? It sounds like you mean slanted there, since you’re saying “also.”
Correct, see edit. Thanks.
I was about to ask the same!
Another very cool illusion, this time with a “three dimensional” mojo: https://twitter.com/slackerinc/status/1050255308906524672?s=21
With this one, I’m finding that unless I take quite a bit of “time off” from it, when I return to it it only looks one way (with everything facing up).
I think:
[SPOILER]Consider four types of shadow:
(a) Light at left, plate concave.
(b) Light at left, plate inverted.
© Light at right, plate concave.
(d) Light at right, plate inverted.
Note that the difference between a and b shadows — shadow is on opposite side of plate — is much greater than the difference between b and c shadows — shadows are on same side but have different shapes. (b is concave, 1st-quarter moon; c is convex 2nd-quarter moon.)
In a real scene you’d see only a & b … OR c & d. But here you lots of b and a few c.
You start by seeing the b’s since they outnumber the c’s. Inverted; light at left – no problem. The c’s catch your eye. Whoa! Something’s wrong. The simplest way for brain to become satisfied again is to treat the b shadows as c shadows. As noted above, the error in the difference is smaller than the alternatives.[/SPOILER]