The Chinaman's Hat Illusion

[elkcub]

Whenever I acquire new binoculars, which is frequent nowadays, I take them out back and view familiar things. This is not only convenient, but also provides a ready means of comparison. Several views are of well-manicured bushes, like oleander or pittosporum, and can be viewed purpendicular (i.e., normal) to their surfaces at distances of 20-60 ft. Well, here's the thing. When I stare at such a dense, full-field, leafy surface for a few moments, it invariably becomes dish shaped, rather than flat, and resembles the inside of "Chinaman's Hat," the outcropping of lava off the coast of O'ahu, Hawaii. It's an amazing and strong effect. http://www.hawaiiweb.com/oahu/sites_to_see/chinamans_hat.htm

Years ago I thought something might be wrong with my binoculars, or maybe me (still possible), but every binocular I have produces the effect — although not to the same extent. Some hats slope up from the center almost to the edge. Others are bowl shaped for about 1/3 the field diameter and then flatten out to a disk, which forms the hat's brim. The illusion takes somewhat longer to develop with one eye, and seems to be facilitated by two.

Although I'm anxious to hear any explanation that makes sense, it seems to me that the perception may be the result of field curvature, or possibly astigmatism, where the brain tries to make sense of differential focus radiating from the center of the view. So, the Chinaman's Hat percept is created.

Other than the desire for an explanation, one thing that interests me is the possibility that the illusion might be used to measure some meaningful optical property of the binoculars. It wouldn't be difficult to standardize scores for the hat's crown depth, crown diameter, rim, and so forth.

Well, anyway, that's my New Year's contribution for those with discretionary time — and a manicured pittosporum bush.

Ed

[Henry B]

Link doesn,t work.

[Andrew Rowlands]

Link doesn,t work.
Oh yes it does ...









Link direct to photo (http://www.hawaiiweb.com/images/chinamanshat_3.jpg)

[zurtfox]

Oh yes it does ...
Link direct to photo (http://www.hawaiiweb.com/images/chinamanshat_3.jpg)

Oh no it doesn't!




(for me)

[FrankD]

I could not get it to work either.

[lucznik]

Neither link works for me.

[Katy Penland]

I think something's wrong with the hawaii site itself. Maybe exceeded bandwidth? Keep trying, maybe it'll be available later. :t:

[Curtis Croulet]

The link does not work for me.

[elkcub]

Gads, I get it every time I try. However, photos are at this site: http://www.webshots.com/g/d2004/1-nw/44005.html

And, just to get you started, one is attached — and another is now on Post #1.

[Henry B]

Page still not available.

[marcus]

It worked for me too. The first time I clicked on the picture it did.

[Henry B]

Gads, I get it every time I try. However, photos are at this site: http://www.webshots.com/g/d2004/1-nw/44005.html

And, just to get you started, one is attached — and another is now on Post #1.
OK,thanks. I can get on here.

[AlanFrench]

Although I can't say exactly why you see this illusion, similar illusions are quite common. Although your brain generally knows which way the relief on lunar craters goes, they sometimes get turned "inside out," either in photos or through a telescope. Testing telescope mirrors can reveal various bumps and holes, and sometimes they get reversed. Once in a while I notice this can happen with the unaided eye on features of the landscape.

Clear skies, Alan

[elkcub]

Although I can't say exactly why you see this illusion, similar illusions are quite common. Although your brain generally knows which way the relief on lunar craters goes, they sometimes get turned "inside out," either in photos or through a telescope. Testing telescope mirrors can reveal various bumps and holes, and sometimes they get reversed. Once in a while I notice this can happen with the unaided eye on features of the landscape.

Clear skies, Alan

Alan,

I'm rather familiar with reversible or ambiguous figures as they're called, and the explanation is fairly well undersoood. However, from your response it doesn't seem that you've experienced this particular illusion yet. Let's put it this way: there are craters on the moon, but my bushes (probably) don't have shapes like the inside of a hat. Of course, until there is confirmation from others who have the same experience, it's unique to me. :king:

Generally speaking, illusions occur when the brain has to make sense of data that don't otherwise fit together. Sometimes the perceptions are reversible, but in this case it doesn't appear that way to me. Maybe it would to others. The apparent image size increase with roof binoculars is another non-ambiguous illusion.

Ed

[AlanFrench]

Alan,

I'm rather familiar with reversible or ambiguous figures as they're called, and the explanation is fairly well undersoood. However, from your response it doesn't seem that you've experienced this particular illusion yet. Let's put it this way: there are craters on the moon, but my bushes (probably) don't have shapes like the inside of a hat. Of course, until there is confirmation from others who have the same experience, it's unique to me. :king:

Generally speaking, illusions occur when the brain has to make sense of data that don't otherwise fit together. Sometimes the perceptions are reversible, but in this case it doesn't appear that way to me. Maybe it would to others. The apparent image size increase with roof binoculars is another non-ambiguous illusion.

Ed

Ed,

No, I guess I have not, nor did I quite understand what you were describing. I hope things don't start appearing like the inside of a hat! |;|

[zuiko]

Hi,

I haven't experienced the illusion you mention.

I suspect however that it might be similar to an illusion I sometimes see in a different situation. When I look through a grille door - the ones with the diamond patterns - because there are similar parts of the structure involved - my brain fuses the image in such a way that non-corresponding parts of the metal grille are fused giving the illusion of the object being more near or further away than what it is.

It has to do with the amount of vergence one is exerting with ones eyes for that particular correspondence.

So two areas of bushy leaves may be being fused as corresponding in your brain leading to the illusion that I sometimes see with the grille.

Random dot stereograms have a similar effect.

Here is a good example of a "Chinaman's hat" random picture stereogram which might give you an insight into the phenomenon...

[Henry B]

What doe,s the second image represent?

[elkcub]

Hi,

I haven't experienced the illusion you mention.

I suspect however that it might be similar to an illusion I sometimes see in a different situation. When I look through a grille door - the ones with the diamond patterns - because there are similar parts of the structure involved - my brain fuses the image in such a way that non-corresponding parts of the metal grille are fused giving the illusion of the object being more near or further away than what it is.

It has to do with the amount of vergence one is exerting with ones eyes for that particular correspondence.

So two areas of bushy leaves may be being fused as corresponding in your brain leading to the illusion that I sometimes see with the grille.

Random dot stereograms have a similar effect.

Here is a good example of a "Chinaman's hat" random picture stereogram which might give you an insight into the phenomenon...

Zuiko,

The donut stereogram on the right is really nice, althogh I've seen several others, and a computer algorithm to make them. The donut does bear certain similarity to the 'Chinaman's Hat' percept I described, but not quite. (I am an experimental psychologist with longstanding interests in 3-D photography— and have read most of the classic books on the subject.)

Obviously, the dot stereogram was engineered to allow the donut to float above the flat background when viewed "wall eyed" for 3-D. (Forget cross-eyed viewing for me. My Mom said my eyes might get stuck ;)) What I'm talking about is observing a flat, unorganized, i.e., "random," leaf pattern where the leaves are all different sizes and orientated in different spatial directions. Throw in some branches here and there as well. This is not a repeating pattern as with a grille, and doesn't lend itself to vergence based (con)fusion. In addition, once visualized, the percept doesn't disappear readily even when the head is moved to a different section of leaves. In other words, it's indifferent to the particular patterning, and can be induced monocularly as well.

If you would view the situation as described with binoculars, I believe you'll experience the correct Chinaman's Hat percept (I made up the name, so it can't be found anywhere else). One question, of course, is why does the percept take this particular symmetric form, since it could actually be anything or nothing. I think it has to do with the optical properties of the binoculars one is looking through, since the leaves become progressively more and more out of focus with distance from the center—but not exactly the same for each binocular.

Incidentally, to make sure this is a function of binocular use, I have viewed the scene with my handy-dandy 1x binoculars — two 4.5" tubes with a 1" hole in each. This has a FOV of ~12.6 deg. and zero magnification.

Thanks,
Ed

[elkcub]

What doe,s the second image represent?

Henry,

I've been unable to visualize the left attachment myself. It's called "Eye on Gorilla" at http://www.ied.edu.hk/has/vrdemo/rds/. It may require cross-eye viewing. The one on the right will produce a donut floating over a flat surface with wall-eyed viewing, — i.e., the eyes are made to look straight ahead without converging. It takes some practice, and is similar to staring. Once seen, it can be made to reappear more quickly.

It would really be better to observe foliage through binoculars like I mentioned than go off on a tangent with dot stereograms. However, the strength of the CH percept is indeed facilitated by binocular, rather than monocular, viewing. Still, it can be seen monocularly.

Ed

[zuiko]

Obviously, the dot stereogram was engineered to allow the donut to float above the flat background when viewed "wall eyed" for 3-D. (Forget cross-eyed viewing for me. My Mom said my eyes might get stuck ;))

Hi Ed,

In addition to the wall eyed image, in the second donut stereogram, I can actually see three different stereograms with various amounts of convergence (cross eyed viewing). So four images in total.

In the first level there is a donut extending into the page; the second level creates a double curved patform in the centre, with two levels of scooping into the page; and a tertiary image where there is three platforms in the centre with four borders, and three levelled scooped out structure extending into the page.

I'll try what you describe taking particular care to observe.

[zuiko]

The only other thing I can think of especially if it occurs with one eyed viewing is that your brain has particularly strong ability to judge from monocular cues to depth.

http://ahsmail.uwaterloo.ca/kin356/cues/mcues.htm

I'll try it out on our front hedge. I seem to be able to see all the weird things - the rainbow effect with DLP screens etc.

[elkcub]

The only other thing I can think of especially if it occurs with one eyed viewing is that your brain has particularly strong ability to judge from monocular cues to depth.

http://ahsmail.uwaterloo.ca/kin356/cues/mcues.htm

I'll try it out on our front hedge. I seem to be able to see all the weird things - the rainbow effect with DLP screens etc.

I anxiously await a report of your experience. The foliage must fill the entire field of the binoculars.

[zuiko]

Well I tried on two occasions in overcast and then bright sunshine conditions at my front hedge as well several more distant ones.

Standing from about ten metres away the foliage filled the frame, and despite several minutes of looking I wasn't convinced I could definitely see anything like what you describe.

The binoculars were hand held as steadily as possible.

I did notice that when I looked slightly above and below eye level that a sort of keystoning distortion was visible - ie. the field appeared tilted. When I panned the binoculars slowly there was also a very mild pincushion type distortion visible.

Perhaps it's binocular dependent but if I ever see it I'll post back on this thread.

Best wishes.

[elkcub]

Well I tried on two occasions in overcast and then bright sunshine conditions at my front hedge as well several more distant ones.

Standing from about ten metres away the foliage filled the frame, and despite several minutes of looking I wasn't convinced I could definitely see anything like what you describe.

The binoculars were hand held as steadily as possible.

I did notice that when I looked slightly above and below eye level that a sort of keystoning distortion was visible - ie. the field appeared tilted. When I panned the binoculars slowly there was also a very mild pincushion type distortion visible.

Perhaps it's binocular dependent but if I ever see it I'll post back on this thread.

Best wishes.

Zuiko,

Maybe it's idiosyncratic to me. No one else has yet indicated that they get the same perception. If so, I'm just a special person. :king:

My trimmed bushes are about 12 ft. tall, and I view them from 20-60 ft. away at eye level. Try focusing on a single leaf in the center. The percept should appear within 30 sec.

-ed
PS. The foliage needs to be pretty dense. I get it with each of my 13 binoculars, roof or porro. The hat shape is different though, sometimes simply a prominent dimple (depression) in the center.

[elkcub]

Ed,

No, I guess I have not, nor did I quite understand what you were describing. I hope things don't start appearing like the inside of a hat! |;|

Alan,

Based on our recent side discussion about flat fields for binoculars used in a 3-D world, I've found an interesting and perfectly satisfactory explanation — even though you haven't yet seen the phenomenon. To appreciate the effect look directly at a flat cut hedge with lots of leaves so that it fills the visual field. Carefully focus on leaves in the center of the field and hold steady. Just relax. Don't try to force anything. The illusion should appear in several seconds. It looks like the inside of a hat, which flares out to form a fuzzy flattened rim. Like CA, once you see it it will appear with little provocation (the first time I saw it years ago, it startled me).

My explanation is that this results from field curvature and distortion working together, coupled with the fact that the hedge has some depth and the leaves are slightly different distances away from the observer, let's say within a range of 8-10".

Due to field curvature, the leaves adjacent to the ones focused on in the center, and somewhat closer to the observer, will remain in focus from the center to some angular extent limited by the depth range of the foliage. Due to lens distortion, leaves more and more off axis, but still in focus due to field curvature (and DOF), will progressively become larger in the projected image. Since the off-axis leaves are both in focus and increasingly larger, the brain makes sense of this curious situation by creating a bowl shaped field — the inside of the hat. The larger off-axis leaves must appearer nearer to the observer to maintain size constancy.* To account for the rim of the hat, at some off-axis point the field goes out of focus and looks blurry. This is the dreaded "edge softness" showing up in 3-D. Finally, the 3-D stereo percept is enhanced by the fact that that slightly different field patterns are presented to each eye — stereopsis.

The nice part about experiencing unique visual phenomena, of course, is that nobody can really argue with the explanation. :D

Over and out,
Ed

* Because it generally appears like a hyperbolic surface of revolution, the shape is consistent with either the system's Petzval surface, or its three dimensional distortion surface. Both are symmetrical around the viewing axis and convex.

[elkcub]

In post #25 above to Alan French I have explained the Chinaman's Hat stereo illusion to my own satisfaction. For those who are at all curious I would encourage trying to visualize it under the conditions described. For those who succeed, I would be pleased to discuss your perceptions or alternate explanations. Absent that, this thread is now at an end.

Ed