03 February 2008

Spinning lady illusion explained

This post solves an illusion/riddle named Spinning Lady, doing rounds in the Internet and demonstrates how to get rid of the illusion.
First publication of Spinning Lady illusion appears to be in the Australian "Herald Sun" October 07. (Thanks to Aaron who showed it to me then) {Some have claimed that Nobuyuki Kayahara 2003 is its author} Apart from disagreeing to take it as a test to see if I am Right Brained or Left Brained (I am Wrong Brained by the way.), I argue that this illusion is created accidentally and I identify two central points that make this illusion possible. By reversing these two points, I prove that the illusion can be disambiguated. There are also arguments that all silhouettes are ambiguous which is a bit like saying "all ambiguous things are ambiguous"! This post is about why at all they are and in particular this one is ambiguous.

If you closely look the at the animation (click on the image), some will see the silhouette rotating in clockwise direction, others will see it rotating in anti-clockwise direction. Most, I have heard will see clockwise rotation. We should be able to switch between these rotations by focusing on the shadow of the legs and imaging the rotation in a direction opposite to what we see.

Anurag Sinha identifies a feature that is also responsible for the illusion. I am here to hammer the last nail on the illusion's coffin. One thing you will notice when the silhouette appears to have changed the direction is, the change of not only the grounded foot but also the floating hand. Left hand appears to float when the rotation is anti-clockwise and right hand appears to float when the rotation is clockwise.

How can this happen? I mean when is our left hand our right hand? The notion of left and right and hence the notion of clockwise and anti clockwise are based on our ability to distinguish front and back! If we watch a clock without a dial, from behind we will see it moving in anti-clockwise direction. Thus if we have no way of distinguishing the front and the back, we have no way of deciding whether the motion is in clockwise or anti-clockwise direction. If the cues are confusing, we end up in an illusion; some thinking in one way and other in another way.

There appear to be two points of confusion here: One is in the foot that is grounded, other is in the floating hand! Look at these snap shots:
In the snapshot above left foot is on the ground right hand is floating
In the snapshot above right foot is on the ground left hand is floating
I drew eyes on the face of the silhouette (also boobs). Now we can distinguish between the front and back of the silhouette! Can`t we?

Thus the illusion is created not only by the ambiguity in resting feet judgement but by also the total lack of cues to distinguish between the front and back of the silhouette. We should note that we can see front and back for certain range of angels but these are not the distinguishing part of the rotations. Thus If we can introduce cues to tell the difference we can stop the illusion.

Further, I argue that this illusion is an accident; an accidental creation when an avatar rotating in "no one knows what direction" was converted to a silhouette (Only the author can tell and most importantly it does not matter!).

I did an experiment with this illusion. I added cues to distinguish front and back of the silhouette by drawing boobs and eyes to make it rotate in anti-clockwise direction. I also had to add depth cues to the feet to disambiguate anchoring.

Now click and watch the edited animation. (It is not polished yet). Can you see it rotating in both directions? NO? You can only see it rotate in anti-clockwise direction. Even if you can switch, my edits force you back to think anti-clockwise rotation. We can edit the animation to make it rotate in the clockwise direction too.
Thus I have shown that the perception of rotation of this silhouette depends both on the depth cues and the cues to distinguish the front and back of rotating objects.

And yes, this is not a right or left brain test. Real left and right brain tests are elsewhere. (click)
Postscript:
This was created by Japanese designer Nobuyuki Kayahara. This is a Bistable illusion caused by lack of visual cues for depth.

8 comments:

  1. Anonymous9:41 pm

    I could make it switch by covering the top of her body with my hand for a few seconds, then when I uncovered it, it was still going in the same direction but the editing made no sense and looked like nonsense on the image not like a face or boobs. In other words, it didnt force me to mentally attach it to the dancer to identify it as "front"

    ReplyDelete
  2. Dear anonymous I think you need little more patience to appreciate what is going on. Your inability is not a proof for something being nonsense. If you have serious arguments to make counter my explanation.

    ReplyDelete
  3. I think we should rotate other totally black silhouettes to see if the same phenomenon is true for rotating trucks, trees, chickens, etc. I think we will find that the "illusion" is not an illusion at all. It is simply a rotating silhouette, which devoid of other clues, can be imagined to move in either direction. The true fascination with this figure is the pleasing form. No where near this much attention would have been given to a spinning chicken, lol!

    ReplyDelete
  4. I think we should rotate other totally black silhouettes to see if the same phenomenon is true for rotating trucks, trees, chickens, etc. I think we will find that the "illusion" is not an illusion at all. It is simply a rotating silhouette, which devoid of other clues, can be imagined to move in either direction. The true fascination with this figure is the pleasing form. No where near this much attention would have been given to a spinning chicken, lol!

    ReplyDelete
  5. @jvankirk I think we agree! "I argue that this illusion is created accidentally and I identify two central points that make this illusion possible. By reversing these two points, I prove that the illusion can be disambiguated. There are also arguments that all silhouettes are ambiguous which is a bit like saying "all ambiguous things are ambiguous"! This post is about why at all they are and in particular this one is ambiguous."

    ReplyDelete
  6. Anonymous10:05 pm

    The signals connect with concentrated nerve locations, which in turn spread the signal
    to the whole abdominal area.

    Here is my web page - http://iweboo.com

    ReplyDelete
  7. Anonymous5:16 am

    You make some good points but you're missing a key factor; the figure isn't spinning in either direction, it is a 2D image swinging from side to side, not a 3D image rotating.. the "fact" it appears to be spinning, that is the greatest illusion of all.

    ReplyDelete
  8. Anonymous 5.16 am, Thanks for stopping by and commenting. This is not a 2D image swinging. I have reviewed all the frames of GIF image while creating my edited version. This is a 3D silhouette - a 2D projection obviously of a rotating 3D silhouette. I saw 2D projection because our displays obviously can not display in 3D. (I am sure you are aware all 3D things we see on screen are a projection/shadow on a 2D plane). Thanks for stopping by and commenting.

    ReplyDelete

Please leave a note about what you think about this write up. Thanks.