Research on the visual rabbit illusion takes a leap forward | Hacker News Reader

Research on the visual rabbit illusion takes a leap forward

kyushu-u.ac.jp

48 pointsgeox40 comments

The gif frames are "left, off, right, right". I.e. there are only two flashes, but the second lasts twice as long. I assume somebody either didn't understand how gifs work, or they accidentally dropped frames in export.

Note that there is also a version of this that has two flashes and the illusion is a third flash. The gif, as described, is intended to be the version where there are three real flashes.

Thanks for working out the cause of the bug, I thought I was going mad.

You mentioned another version of the gif. Is this one we can find somewhere?

I don't know if there's a gif specifically, but here's a video: https://www.youtube.com/watch?v=yCpsQ8LZOco

Interesting..

I immediately "didn't get it" because I only saw two flashes.

Wow I have such a big brain - I only perceive reality.

That's a different one, that's one with auditory beeps and only two flashes. It's not the same as the illusion in the paper.

Some commenters conflated the two, which is why I acknowledged it as a different "version".

This has been driving me mad. I thought I was broken.

> In this illusion created by researchers, the first flash appears to the left, and the second and third flashes are at the same spot to the right

What third flash?

Oh. Good. It’s not just me.

After covering the first part of the sceen, I still fail to see the 3rd flash

Not only does the gif only have 2 flashes, but the caption of the gif and the article text differ. The caption says there are 2 flashes on the right, the article immediately under the caption says 2 on the left.

It does qualify the two-on-the-left version as "in traditional experiments", but yes, even then it's still not 100% clear if they are describing two different versions there on purpose or by accident.

These kinds of illusions are something machines doing tracking & estimation already experience. The estimator is trying to fit a path that makes kinematic sense onto noisy measurements of position. When I see these all I can think is that your brain clearly feeds you "filtered" and "estimated" corrections for you to decide against.

This type of perception-before-decision is ubiquitous in robotics.

There's 2 hard problems in robotics: Perception and funding.

I wonder to what extent our brains also back-fill in other, more complicated cause-and-effect information. If a certain cause-and-effect is more logical to our brains, does it simply recreate the memory of our perception to the more logical one?

Our eyes have a very narrow area of full-fidelity vision. The brain creates an illusion of us seeing a large field of view by having the eyes do a lot of quick movements very fast[0], while blanking our vision during the movement itself so we don't notice it[1], then stitching the result together into one perceived image, while eliminating, interpolating or faking anything that was happening during transitions or outside of sampled areas[2], and slapping a common timestamp on the whole composite.

And that's just the tip of the iceberg.

[0] - https://en.wikipedia.org/wiki/Saccade

[1] - It's why you can never see your eyes moving when looking at your reflection in the mirror - but you can see it when you look at the image of a selfie camera on your phone, thanks to the processing lag.

[2] - Hence the analog clock illusion, when you look at a regular analog clock with a second hand, and the first second seems to last longer than it should - the brain is assuming the hand is stationary and it takes until the clock ticks again for it to realize its mistake. This illusion notably doesn't happen on continuous-motion clocks.

Can confirm that the gif only has two flashes, although the second one appears to be two frames long.

That is the whole point of the illusion, that there are two filled circles that flash and your brain perceives it as 3 flashes because they come so fast that they are processed as a temporal chunk before you become aware of it.

Seems like that's not the same illusion - there are multiple closely related illusions here.

> That is the whole point of the illusion, that there are two filled circles that flash and your brain perceives it as 3 flashes

Where is this from?

I could be missing something, but everything I can see in the article and the abstract of the paper indicates that the point of the illusion is that showing three flashes causes us to misplace the second flash as being located at the midpoint of the first and second.

Yeah it's supposed to be two flash positions only, but I missed that the last one is supposed to flash twice which it apparently doesn't in the gif so that was the confusion.

No, that isn't it, re-read the article. The illusion is supposed to be that three flashes that are not evenly spaced, or are even spaced out of order (1--3-2) get perceived by the brain as evenly-spaced an in order.

Look at the explanatory diagram with the labels "What our eyes see" and "What our brain "sees"" (the third image on the page).

The gif is just broken. It's supposed to have a space between the second and third flashes.

Yes I see the last one was supposed to flash twice, but at 2 positions. The explanation is the same though as far as I understood the article, that the brain processes all flashes as a single temporal chunk and that is what causes the temporal re-alignment. It kind of resembles the famous letter-jumbling where you can still read the text if you re-arrange the letters inside the words if you read fast enough.

I’m seeing the illusory flash.

Here is a version with beeps too.

https://m.youtube.com/watch?v=yCpsQ8LZOco

This makes a little more sense than the article's demo, but I still only perceive two flashes, even with the audio. Maybe it's also environmental...

For what it’s worth I’m viewing it fullscreen on a phone.

The version with three beeps and two flashes is a different illusion. This one is supposed to have an actual third flash and no beeps.

> This one is supposed to have an actual third flash and no beeps.

Nope, it has only two flashes in reality :)

If, like me, you are perceiving three when focusing on the + then the illusion is working. When I focus on the line where the flashes appear then I can clearly see there are actually only two flashes.

No, the article is a different illusion. Three flashes.

https://streamable.com/an7hso

I recorded it and slowed it for you

And just in case I was going mad I checked the frames of the gif. There are 50 frames. Frame 15 has the left dot, 17 and 18 have the right dot.

Either the gif's not loaded right or I'm more broken than I thought, but I'm only seeing two distinct flashes - one left and one right...

You have to focus on the cross, and the demo on the page is slowed down. I don't know why they don't have the original speed there.

What the article describes at the end sounds a bit like how many folks' memories may work. At least mine often has that issue unless I consider those memories.

A year afterward, there wasn't any inconvenient gap there, it was all like a quick training montage, and 1 -> 2 -> 3 nicely spaced. "I didn't spend months writing that proposal, that was like a week tops wasn't it?"

Reassured that others also don't see the third flash.

If you're going to work on subtle detail oriented things such as illusions, it seems like this ought to be noticed and resolved really quickly, which doesn't fill me with massive confidence in the rest of the research!

Since the illusion example in the article is poorly described and implemented, here's a better example in video form: https://www.youtube.com/watch?v=yCpsQ8LZOco

Thanks but again this seems to be somewhat related but not identical to the illusion described in the article.

(Sure, the illustration / GIF / captioning on the OP aricticle is really poor and confusing)

But the illusion described in the article involves three actual flashes (without beeps) and the illusion is aboutthe LOCATION of the second flash, not that the second flash itself is a phantom / non-existent / imagined one.

Is this like the nerd version of “when you see it”? I can’t seem to see a third flash.

I'm only seeing 2 flashes no matter what I do... am I doing something wrong? Even if I cover parts of the screen I still only see 2 flashes.

There are two flashes in the gif specifically, because some people will apparently see a third one "in the middle" between the real two. Hence the "illusory rabbit". I only see two though, and it seems I'm not alone.

That is a different twist on the illusion. The gif is indeed supposed to have three flashes.

Mhm, and the twist then is that the second flash seems to be in the center, even though it is to the right?

Reply on news.ycombinator.com