Can you see the edges between the diamond and the background? Typically, the motion occurs when you can see the edges, and disappears when you are far enough away to make the edges too thin to be seen.
I don't think so, for me the illusion is not just uniformly upwards throughout the cycle, but also includes a sudden downward shift, so its more like sawtooth motion illusion., so if it were to move down I predict it would look more like a somewhat static diamond jumping down in steps... it would be interesting to try in order to determing the exact illusory motion profile, if it doesnt seem static between the jumps but seems to wiggle up and down before jumping...
Edit: if someone would try this, I would recommend using a larger diamond with thick enough edges and viewed from a distance such that pixelization effects are ruled out...
I've seen the reverse phi illusion effect crop up a few times in various places. [0] But as the paper says, it typically involved changing the luminance of the foreground object.
It's cool to get a breakdown of exactly what makes this effect tick.
Are there any other good non-static optical illusions such as this? As in illusions which require some element that changes over time cyclically in order to produce the effect?
> Are there any other good non-static optical illusions such as this? As in illusions which require some element that changes over time cyclically in order to produce the effect?
That is excellent. I developed a trick when playing piano where I can roughly count two things at once for a short period of time (Any Murakami fans? Hard Boiled Egg and the End of the World made me try it!).
When I tried to do that to count the flashes the one I was looking at felt like it sped up to match the other one but I couldn't count both at once. My brain was well and truly tricked.
There is a large collection of gifs based on the reverse phi illusion in a tiny obscure subreddit that I unfortunately can't find again. I think most of the submissions were made by a handful of users and they never referred to the effect by name. It was mostly artistic applications such as an angel infinitely rising upwards just by flashing two images. If someone finds it again, please let me know!
I too tried to find a few collections I've come across, but no luck... not too sure of what search terms to use. Was hoping someone else would come in clutch.
It's even worse with the second of these gifs: if I look at it frame by frame, with ample pauses, the scene still keeps rotating in the same direction. I know that two of the four frames must return to the other position, but I can't see it.
It might be fair to say that it is moving but also staying in the same position, as mentioned at the end of the article. To say it's not moving is like saying an animated sprite in a video game isn't moving either - sprite movement is just an "illusion" of continuous motion rather than the discrete appearance and disappearance of separate images that it really is.
I wonder if when film or flipbooks were first invented, people considered them to be an optical illusion too?
I wonder if there's any application for them in VR to counter motion sickness? Instead of a stationary object looking like it's moving, you have a moving field of view looking like it's stationary...
* Dwitter has a built-in `R` for RGB color, `S` for Math.sin, and `x` for the 2d context of the canvas.