Cognitive Daily

Take a look at the following two circles. At the center, they’re both the identical bright white. But which one seems brighter?

Let’s make this a poll:

I’m not sure if this illusion will work when respondents know the objects are the same brightness, but naive viewers will reliably rate the circle on the left as brighter — this is called the “glare effect,” and it occurs whenever there’s a gradual gradient around a circle or other shapes (the gradient must approach the color of the shape as it gets closer to the shape itself). I found the illusion so powerful that I had to close the picture window I was using to create the image while I wrote this post.

To me, the glare effect is so strong that the image feels like it is burning into my retina — even though I know the white background of my word processor is just as bright as the white circle. In fact, when Daniele Zavagno and Giovanni Caputo were first researching the phenomenon, subjects complained that it was very uncomfortable to look at images with the glare effect — they were seeing afterimages, as if a bright light had been flashed in their eyes. But are viewers really seeing stronger afterimages when they experience the glare effect?

Hongjing Lu, Daniele Zavagno, and Zili Liu developed two experiments to test this question. In the first experiment, they showed viewers plain circles and squares, without the glare-effect-causing gradient, for varying periods of time, at two different levels of brightness. When the objects disappeared, viewers were instructed to press a button when they could no longer see an afterimage. While results varied for each individual, they followed the same pattern. Here’s a typical result:

As you can see, the longer the image was presented, the longer the afterimage persisted. And brighter images led to longer afterimages. These results — both statistically significant — were as expected; the experimenters just wanted to confirm that the relationship between afterimages and image viewing time was measurable at the brightness levels their equipment could display.

The procedure for the second experiment was the same, but a different set of images was used:

The image on the left is a control while the image on the right is the glare effect image. The reverse gradient was used for the control to ensure that the overall luminance for control images was similar to the glare effect images. Here’s the surprising result:

The afterimage actually persisted significantly longer when there was no glare effect! Even though the glare effect makes images seem brighter, the apparent increase in brightness had no impact on the afterimage, and in fact actually seemed to diminish the afterimage’s strength.

The authors suggest that the glare effect isn’t actually occurring in the retina or some other low-level visual processing area. Instead, it’s probably a reflex to protect us from looking at truly bright light sources like the sun. The discomfort I feel when looking at the pattern that causes the glare effect is simply the result of an over-ambitious protection mechanism.

Lu, H., Zavagno, D., & Liu, Z. (2006). The glare effect does not give rise to a longer-lasting afterimage. Perception, 35, 701-707.

Comments

#1 Mark Chu-Carroll
August 17, 2007

Fascinating – I see the image *without* the halo as absolutely, unambiguously brighter.

I’ve got a perceptive impairment that fouls up a lot of spatial perception related things… I wonder if that’s somehow related? It doesn’t seem likely, but since it is a sight-related impairment, and I’m clearly outside of the “normal” perception, it’s worth wondering..
#2 Dave Munger
August 17, 2007

Mark:

Do you still not see the effect after you stare at the image a long time? What about in your peripheral vision? That’s where I find the effect most dramatic. And again, the fact that you know both images are equally bright might be affecting you (although in this study the researchers themselves participated and still noticed the effect).
#3 Brian
August 17, 2007

For me, the glare effect image was not as bright. In fact, I saw what appeared to be a bright ring around the outside of the circle boundary, but the interior of the circle appeared grayish.
#4 Zachary Tong
August 17, 2007

I’m not sure I would rate the left as “brighter”, but it is certainly more disrupting. The image on the left affected my peripheral vision much more than the image on the right, and left significant afterimages despite being much less high contrast (I would have assumed the other way around, with high contrast leaving a sharper, more defined afterimage).
#5 arby
August 17, 2007

Same as Brian, #3. I noticed it before I read what I was supposed to see. rb
#6 Scotty B
August 17, 2007

2 cents: At first glance the right appeared brighter (and I voted accordingly) then I went back and stared at it for a bit and the left became brighter as I stared.
#7 Alunfoto
August 17, 2007

There are all sorts of issues with looking at images on computer screens, since rendering will vary from monitor to monitor. On my colour-calibrated monitor, the haloed circle on the left looks distinctly less bright in the middle than do the sharp-edged one. Still, my image-editing software tells me that both circles have the same hue; completely white. So there should have been a third alternative in the questionnaire, “equally bright”.

The suggestion that haloed images is an evolved mechanism to avoid looking at the sun seems a bit odd to me. The only thing that cause halos around the sun is particles in the atmosphere that diffuse the light, so a sharp-edged disc of light in the sky should be more dangerous to look at than one with a halo.

Best,
Jostein
#8 Mats M
August 17, 2007

I think I also see it as Brian (#3).

It could be that our vision takes information from neighboring neurons, note that the gradient suggests that a brighter registration is more plausible (gradients usually continue) and makes us ‘see’ a brighter value than the screen can produce.

It would be interesting to see if by changing the slope/size of the gradient it might be possible to make this perceived circle even brighter.
#9 P.Tane
August 17, 2007

What I see is simalar to Brian, but the “glare effect” makes me feel like the image is shining in my eyes, and it does appear brighter as a whole, even though the inside looks dimmer than the rim of the circle. That sensation intensifies whith my peripheral vision to the point of being uncomfortable. Its afterimage also lasts longer for me.
#10 Susannah
August 17, 2007

For me, it’s the one on the right. I saw it as noticeably brighter at first glance, before reading the post, and nothing has changed on repeated tests (going away and coming back later). In fact, the left image seems grey in the center, with a white rim and then the blur on the outside.
#11 christopher
August 17, 2007

ditto susannah and brian.
#12 Freiddie
August 18, 2007

I agreed about the glare effect. But the gradient, apart from the glare effect, also causes a somewhat opposing effect, because gradients “soften” the color by blending into the background, while the other one (right) seems to have an abrupt edge that contrasts between the white and black. This is probably why there are still quite a number of people who believe the contrary. I also noticed that the one with the glare effect seemed to be “whiter” while the other one seemed “yellowish”. The funny part is that the glowing one “varies” in brightness when I move my head up and down since I’m using an LCD screen!
#13 Nigel
August 18, 2007

My first impression was that the left hand figure was brighter – certainly it is more glaring, and I can understand why some people have said it is hard to look at – but on looking more carefully, and looking back and forth between them (but before reading any comments), I found I saw the same thing reported by Brian, P.Tane, Susannah and others. The left figure seems to have a narrow bright ring around a dimmer center disc. The ring looks as bright, or perhaps even brighter, than the right hand circle, but the inner disc looks dimmer. I had read where you say that they are actually identical before I saw this, but surely if that had biased me it would have been toward saying that they look identical, not that the left figure displays this more complex pattern.

Of course, the left hand figure is objectively brighter, in the sense that there is more white there, albeit spread over a bigger area. That might have something to do why it initially appears brighter. The thing as a whole is brighter, but when you pay attention to the inner disc (equivalent to the right hand circle, with which you are supposed to be making comparison) that no longer seems so bright.

Frankly, this ring and disc illusion that so many of us are seeing seems to me to be more interesting than the rather unsurprising result about afterimages. I believe the conventional wisdom is still that afterimages are caused by the depletion of the light sensitive pigments in retinal receptor cells, and you would surely expect the rate at which they are depleted to depend on objective, not subjective, brightness, on the actual amount of light hitting the cells. My guess is the ring and disc effect has something to do with on-center/off-surround (or vice-versa) retinal ganglion cells, but I am by no means sure quite how that would work (or if it is really the explanation).
#14 Zachary Tong
August 18, 2007

Jostein (#7): Regarding the evolved aversion to a halo: I would bet it isn’t an evolved trait so much as a perceptual adaptation by each individual. I’m sure every human, at some point in their life, has stared directly into a bright light like the sun and seen the resulting halo. And also felt the pain of looking into such a bright light. I would bet this is more of a learned cause-effect than an evolved one.
#15 Xanthir, FCD
August 20, 2007

I definitely perceive the halo as brighter. It does bother me, like looking directly at a lit bulb from across the room might. On the other hand, the plain circle looks perfectly normal.

It’s hard to tell which is brighter in my peripheral, though.
#16 Steve
August 23, 2007

Totally with Brian: “For me, the glare effect image was not as bright. In fact, I saw what appeared to be a bright ring around the outside of the circle boundary, but the interior of the circle appeared grayish.”
#17 mbs
August 28, 2007

At first I found the center of the “glare” circle to scintillate (and it did seem brighter), but after a bit of staring I found that the center did darken to a gray leaving a brighter halo. I would guess that this is due to the visual cortex having a suppression response from saturated/glare-affected area.
#18 Patrik
December 19, 2007

Peculiar!
Intitially it seemed as if the left circle was much brighter due to the glare effect that Dave describes. However as I continued looking at the left circle it became grayish with a really glowing outer rim (described by many here). It’s a little bit similar to watching the sun’s corona during a lunar eclipse but then the center is of course much darker and not just slightly grayish as in this case. I perceive this corona effect to be stronger than the glare effect that I first experienced.