Unlock the Mystery of Taking a Reflection Photo in Your Own Backyard

[DaveC426913]

This is a 'How did I do this?' teaser.

A couple of notes first:
- The picture I include is simulated (Googled for a generic shot). Sorry, best I could do without the original. Ignore the fact that a] it's not me, and b] that the pic is indoors.
- Even though this is a simulation, I actually did do this. This isn't just a thought experiment or a trick.

So:
While wandering in my own backyard with my camera, I saw a reflection of myself and took a picture.

Note the orientation of the image. Note that the camera is tilted in relation to its own reflection.

- I used a normal camera
- it would have looked the same using a film or a digital camera, it would look the same on screen or as a photoprint
- I did not alter the image a la PhotoShop (except cosmetically)
- i.e. there are no "tricks" here
- I did not alter the environment, do any setup of any sort. It was indeed, a spontaneous shot.
- it is not unique. With a bit of searching, I could have done this at somneone else's house.

Q: How could I get a picture like this?


Two-part answer:
The big one is: how is it possible?
The little one is: how did it happen in my backyard?

 

[ceptimus]

Possible spoiler:

::With two mirrors at 45 degrees to each other. One vertical like a normal mirror, the other sloping like a house roof.
Maybe a greenhouse in your garden? The roof and wall glass can act as mirrors, particularly when it's dark. Or one of those windows that opens with a horizontal hinge.
::

 

[Gokul43201]

I don't see what is special about the picture ? How do we know that the picture is rotated with respect to the camera ?

If there was a mirror on the floor, and you bent down to take a picture of yourself taking a picture, how do you decide which is the right orientation for display ?

If you are telling us that the image is, in fact rotated with respect to the object, that's a teaser worth thinking about.

PS : I don't understand how Ceptimus' solution works !

 

[Galileo]

Since the guy is holding the camera horizontally (with the shoot button on top), the photo should be such that the width is greater than its height.
(If you wanted to shoot something tall, you'd have to hold the camera vertically).
So the dimensions (width and length) of the photo are reversed.

And I have no idea how this is done. I can figure flipping or mirroring the picture. But rotating it 45 degrees.. nope.

 

[DaveC426913]

"If you are telling us that the image is, in fact rotated with respect to the object, that's a teaser worth thinking about."
Yup. That is evident in the picture. (The camera is horizontal, how can the image it took be vertical?*)

Edit: Oops. Didn't quite answer this. Horizontal and vertical are with respect to the camera itself, not to any exterior object like the ground. Even if I were pointing at the ground, a camera could not take a picture of itself with its horiz vs. vert format rotated.


"PS : I don't understand how Ceptimus' solution works !"
Me neither.


"I can figure flipping or mirroring the picture. But rotating it 45 degrees.. nope."
Uh, 90 degrees. (You're probably thinking of the mirror(s) being at 45 degrees.)



*Trivia: NOT THE REAL ANSWER - (nor any kind of hint): They used to make a 35mm "half-frame" camera back in the 80s. It wasn't all that popular, and went out of fashion quickly. The idea was to take almost as good shots but get twice as many on a roll (or, conversely, pay less per number of pictures). It used normal 35mm film, but rather than having an exposure area that was 36mmx24mm, it exposed only one half of that - 18mmx24mm - which incidentally made the image taller than it was widewith respect to the camera. If the camera in the pic were a half frame camera, you'd get a vertically oriented image like the sample.

This possible answer is disqualified for failing three conditions I imposed in the teaser:
- the "normal camera" condition
- the "same as digital camera" condition
- the fact that I actually did this

 

[ceptimus]

"PS : I don't understand how Ceptimus' solution works !"
Me neither.

Oh well, it wasn't the answer you were looking for then.

If you get a pair of ordinary mirrors and play with them, you'll soon see how to rotate an image through 90 degrees. A good way is one of those bathroom cabinets that has two sliding mirrors as doors. Leave one of the mirrors in place, lift the other one out, and while holding it at right angles to the 'in place' mirror and 45 degrees to the vertical, look at your reflection in it.

 

[DaveC426913]

Your answer is partially viable.

With your setup I will indeed get an image of myself rotated 90 degrees, true. But I will no longer be looking at myself in the image - I'll get a picture of my semi-profile (turned 45 degrees to my right/left).

I guess I'll have to remember to add that to the initial description if I ever use this again.

 

[DaveC426913]

OK, hint #1:

It was a reflective surface. The surface was not flat. What shape will produce the effect?

 

[vikasj007]

i think the required arrangement can be made by using three mirrors, to get a rotated image of myself.
i'll try my best to explain the arrangement.

-- the first mirror--

if i am standing on the ground, the mirror is in front of me, at 45 degrees from being normal to me(the way all looking glasses are), towards, say, left.

-- the second mirror--

the mirror is in such a way that the light rays from the first mirror fall on it again at an angle of 45 degrees,(here the mirror is at 45 degrees to the flat ground).

-- the third mirror--

light again falls on this mirror at 45 degrees, this time the mirror is at 45 degrees to the ground, and 45 degrees to me(the same as the first one, but the first one was standing vertically, and this one is at 45 degrees to the ground also.)

this way we can get a rotated image of ourselves.

i hope i was able to make myself clear, though it could have explained better by making a real 3D model of it. it is difficult this way, but just try and visualise this and you'll get my point.



though the only thing is that i cannot think of any such thing having such arrangement, in my or anyone else's backyard.

maybe you can think of that.

 

[DaveC426913]

I don't know if it will work, I'll have to trust you on that one. Unfortunately, it will still show a reflection of you from a partial profile.

Here's a hint to make it way more difficult, and way easier at the same time:

There is only 1 (one, uno, un) mirror!

 

[NateTG]

Here's a hint to make it way more difficult, and way easier at the same time:

There is only 1 (one, uno, un) mirror!

It would have to have a pretty funky shape to show something rotated 90 degrees without distortion. Radially symetric mirrors of any kind will not cause an effect like that unless the mirror is large enough to be acting as multiple reflective surfaces.

 

[ceptimus]

Maybe a flat mirror half immersed in a tub of water, intersecting the surface of the water at some funky angle would do the trick.

 

[vikasj007]

I don't know if it will work, I'll have to trust you on that one.

well, i have to admit that i did not actually try the combination with some mirrors. i came up with all that in my head.

but it still does not seem to me that this can happen with just one mirror, if it is a plane mirror...is it?

 

[DaveC426913]

It would have to have a pretty funky shape to show something rotated 90 degrees without distortion. Radially symetric mirrors of any kind will not cause an effect like that unless the mirror is large enough to be acting as multiple reflective surfaces.

Well, there is some distortion, now that you ask. Although, the distortion is limited to the fringes, and it could be considered "a normal type of photographic distortion".

But the mirror shape is really quite (OK, somewhat) simple.

 

[K.J.Healey]

no, its not rotated. What he's saying is that its basically stretched, because the camera is set up to take a picture at the normal aspect ration of W > H , but its showing as H > W. Thats the problem.
So what shape would make something 2-3 times as tall as the initial image? Probable a curved surface of some sort. It would be convex with respect to the viewer, such as the outside of a fender on a motorcycle, or the side of an aluminum propane tank(large).

Is that more along the lines?

 

[NateTG]

...

That would be a pretty lame answer. He's already, more or less, eliminated photo cropping by describing the half-frame camera above. Using a mirror that provides linear stretching is not going to change the shape of the print.

 

[DaveC426913]

Now that I ponder the questions and theories proposed, I realize this teaser can be simplified enormously:

Throw away the camera. (It provides a reproducible demonstration, but has no effect on the phenomenon.)

So:

I am standing in my backyard. I see my own reflection rotated 90 degrees. Period.

Hints so far: It is not a series of angled, flat, mirrored surfaces. Indeed, it is a single, mirrored surface.

 

[Gokul43201]

I am standing in my backyard. I see my own reflection rotated 90 degrees...

...about the axis perpendicular to the "plane" of my body.



If your body type is cylindrical, conical, spheroidal, or something that is not easily indicative of a plane, refer to say, the picture of the Vitruvian Man, as a guide.

 

[DaveC426913]

How about:

'rotated 90 degrees about an axis drawn between myself and the reflector'

?

 

[RandallB]

I’m envisioning some kind of a large concave reflector, possibly ornamental but more likely something used to direct a beacon of light. The function of the curve is deeper than inside a sphere but probable not hyperbolic I’m not sure.
If viewed straight on and into the point where a filament might be it would both reverse the image left to right and flip it up to down. But as view is moved off the center line the reflection begins to rotate.
When taking a photo any flaw in the center of the mirror may need to be hidden, say in the dark part of the camera lens?
I’ve never seen one but my imagination says maybe it could be so.

 

[K.J.Healey]

I can't wait to hear the answer. I can't imagine anything that could be found in a backyard that could do this besides a camcorder lying on its side on top of a tv screen. But he's saying its a single reflective surface. I can't think of a single surface being able to rotate an image without possible having some refraction gradient glass over it, or absolutely crazy physical properties.

If anyone has an idea, sketch it up, id like to see what you think.

 

[DaveC426913]

So, I'll hold off a bit longer then?

You'll kick yourself.

 

[RandallB]

So, I'll hold off a bit longer then?
You'll kick yourself.

We'll put on soft shoes don't hold off, give it up.

 

[gerben]

has it anything to do with water/ice?

 

[DaveC426913]

Answer

The mirror is helical or corkscrew-shaped.

It was witnessed while looking very closely at a cheap plastic chrome-plated icicle ornament that is sold in every department and craft store.

 

[RandallB]

Like the one's on the Wind Chimes - now I wish I'd have gotten one - have to find them again to take a closer look.

 

[vikasj007]

The mirror is helical or corkscrew-shaped.

It was witnessed while looking very closely at a cheap plastic chrome-plated icicle ornament that is sold in every department and craft store.

but will it give a clear image as you have shown in the picture.

as far as i can imagine it(i don't really have something of that sort with me) the image might be rotated but it would be a blurred image. i mean it is very difficult to get a sharp image with a that shaped object.

 

[DaveC426913]

It will give as clear an image as a chrome-plated plastic surface can. Blurriness is not a factor, though some distortion is.

Some real-life mitigating factors:
- the thing is only one inch wide, so I'd have to get very close. My camera won't focus that close. That's why I couldn't show a real pic.
- the reflection I see has some distortion that is proprtional to the distnace from centre. It is like a wide angle image - reflections near the edge recede rapidly. But that's because I can see the whole device. Zooming closer will eliminate these distortions.