Hacker News new | past | comments | ask | show | jobs | submit login
The McCollough Effect (cheswick.com)
56 points by geographomics on Sept 1, 2014 | hide | past | favorite | 16 comments



After gazing for a few minutes at both grids, I do see the effect but the green is much more prominent for me than the magenta (which I see, but much less pronounced) despite gazing back and forth at both for about the same time.

The page would be a bit better if it incorporated some basic image controls, IMO, to allow you to rotate the image, zoom it, etc.

I downloaded the image and did some of that locally and zooming in and out is pretty interesting because the effect becomes less pronounced (for me, anyway) as I zoom out away from 100%. It seems like the size of the "haze" around the bars remains constant as you zoom even as the size of the black bars changes, so as you zoom in the effect becomes less and less pronounced until eventually the black bars get so big that it seems to disappear altogether.


When I first looked at the black and white lines just now, I already saw a faint green and magenta glow in them. I hadn't even looked at the colored images yet.

Then I remembered that several months ago I tried this. Could it have possibly lasted that long?


I also experienced this and saw the image a few months ago. That the effect lasts that long is sort of unsettling.


Yep, wikipedia cites a 1975 study "Extremely long-term persistence of the McCollough effect" in saying that the effect can last up to three months in some cases, after prolonged exposure.


I tried this several years ago and still see green and pink when I look at the black and white lines. The green in particular is quite strong.


Same here, though the effect is not as strong as it used to be. At first glance I see the colors, but then they are gone and all I see is the black and white stripes.


Citation needed on this: but from what I understand, the way your brain processes vision is that some of what you "see" is in fact the past. The visual cortex keeps recently seen things around and the incoming visual information is mixed with the past images to create the present (almost like compression in a video). Thus it would make sense that after staring at the green lines for awhile your brain is using past information when looking at new lines and hence the illusion.


This (what you described) is basically how most optical illusions of the type of "look at this thing then look at this other thing" work, but this one is a lot stranger than most in that the effects of it can reportedly last for up to 3 months.

I can't vouch for that directly, but I'm still seeing it an hour later and some other responses here have mentioned that they actually are still seeing it months later.

Other optical illusions that rely solely on the mechanism you're talking about fade in under a minute.


For anyone who's interested in full-screen colored grids, I threw them together in CSS.

Green (horizontal) – http://codepen.io/ChristianBundy/full/iaAcq

Purple (vertical) – http://codepen.io/ChristianBundy/full/moshj/


Thanks, I much preferred your full screen version to the small ones on the post. However, you've swapped the pink and green round compared to the original.


I already have it without switching from the grids to the B&W images, just by concentrating on one and then the other I see fading aftereffects from the image I looked at before, takes a long time to fade to nothing.

Uncanny how reproducible that is. I need to look much longer at the coloured bars before that same effect starts to be present on the B&W image. It's almost as if the screen has burned in.


"I know what you are thinking: this is a simple afterimage effect. If you think so, walk away from your terminal until you think the after image should be gone. Go home and try it in the morning. Then take a look. Or better, simply rotate the image. Well, maybe that isn't so simple with a CRT, but you could rotate your head."

What you describe is a simple afterimage effect (http://en.m.wikipedia.org/wiki/Afterimage#Negative_afterimag...).

The MacCullough effect is different because it lasts much longer (Wikipedia claims 3 _months_)

The McCullough effect is thought to be due to adaptation of edge detectors. 'Normal' after images are thought to be caused by adaptation of cones.


things like this always make me think about how things like great cooking, visual art and music (im sure other stuff too) have managed to implicitly understand complicated and nuanced behaviors of our senses.


I trained one eye as instructed, while closing the other, then rotated my phone 90° and trained the other eye.

The effect has lasted for hours: one eye sees green-horizontal-purple-vertical, and the other, the opposite.


Last time I tried this, even my tiled floor and the blinds on the windows suddenly had colors. I think it will be triggered by almost every kind of striped pattern.





Guidelines | FAQ | Lists | API | Security | Legal | Apply to YC | Contact

Search: