I have a problem that I thought I'd throw out there. Joy and I are giving a talk on holography which we're going to split into two parts. I give the history, science and techniques of holography and Joy gives 3D perception and it's exploitation in the arts. The problem is this:
The major causes of 3D perception are vergence (the eye twists together to focus on a spot in 3D space), accomodation (the lens contracts to focus) and binocular disparity (the difference in position of the image on each retina from the fovea). Stereoscopy works because primarily of vergence and disparity. However, holography is seen as 3D because it captures phase information. So how do these correlate? How do phase differences affect either accomodation and/or vergence and/or disparity. In other words, what's the connection between the classic explanation of 3D in holography (phase information) and 3D in stereoscopy?
the way I look at it is this:
the phase information in a hologram allows the accurate replication of the wavefront emanating from the object.Hence the causes for the 3D perception are exactly the same as the causes for the 3D perception of the real object- probably a mixture of all three.
One way to explain this in lay terms is that the phase part of a hologram determines the DIRECTIONS of the rays
One way to explain this in lay terms is that the phase part of a hologram determines the DIRECTIONS of the rays
I agree with this entirely. I never use the term "phase" when explaining holography, partly because I don't quite understand it myself, and I have to think through it every time. Here's how I look at it:
Let's illuminate the object with laser light. Now put a glass plate where the plate would go, and look through that glass plate. We see the object in perfect 3D. Now, all we are seeing is the light scattered by the object, and no more. If we could record all that light information as it hits the plate, and then 'replay' the light later, then by definition as long as we look through that window, we will still see the image of the object.
The easiest way I think I can explain it is consider the light falling on the plate as a huge number of rays of light, each with a different intensity and direction. Holography records this intensity and direction for each and every ray, and on reconstruction, replays the 'continuation' of these through the plate. It's as if the rays are frozen as they hit the plate, and released on reconstruction. Hence we are completely fooled into thinking the object is still there.
Regarding depth cues, all depth cues are maintained in holography (i.e. classic full aperture laser illuminated holography), and so there is really no reason to mention them at all. Other 3D methods use one or more depth cues, so we have to enumerate them to evaluate the quality of each method.
In holography, the original wavefronts of the light coming from the object are accurately recreated on reconstruction (viewing) due to the phase information. The radius of curvature of these wavefronts is proportional to the distance from the object's position and the viewer's eyes. Holography recreates the original vergence, accomodation, and disparity that was produced by the original wavefronts. For vergence, the two eyes, when focused on the same spot, must follow lines of sight which intersect at the point of focus. These lines of sight are perpendicular to the recreated wavefront at each eye. For accomodation, the lens of the eye must adapt to properly focus the light from the point of focus - and this adaptation exactly corrects the curved wavefront to focus the light to a point at the retina - therefore the accomodation is proportional to the radius of curvature, which provides depth cues to the brain.
I think it's the curvature too. I think the eyes orientate themselves to a direction perpendicular to the wavefront and so cause binocular disparity. The direction vector, mentioned by dcgman and Kaveh, would be perpendicular to the wavefront and so the eyes would orientate themselves along the direction vector of the rays origin, but I think the actual wavefront is the more important factor. Usually, this distinction is small, however it could be relevant for holograms in crystals. In a birefringent crystal, the wavefront and the ray vector are not necessarily perpendicular.
I also wonder how this would bear on aberrations. If an ideal spherical wavefront were distorted, the radius of the wavefront would be different for each eye and so the direction vector to each eye would be different. Would this cause a "false" disparity? This might lead the viewer to place the object in a different depth and/or orientation than it should be.
Another factor that I think operates at a subliminal level is eye movement - even when attempting to gaze steadily at an object, humans tend to unconsciously scan a scene, flitting in a chaotic way between areas of interest. For close objects, the entrance pupils move about laterally in the image wavefront, viewing the scene from slightly different angles. I believe these slight rapid perspective changes contribute to the perception of a real, solid object in holograms as opposed to the static, somehow unreal impression a stereoscopic image evokes.
I haven't seen the possible connection to depth perception mentioned elsewhere, but I expect the idea is obvious enough to have been mentioned somewhere.
This is something I had not thought about. If a "point of interest" in an object were at a different depth to the "main" object, would you enhance false perspective? For example, consider a hologram of an airplane facing you with very bright props. Would the brightness of the props, relative to the rest of the plane, cause this to be a point of interest and so artificially elongate the plane?
By the way, you can experience this effect by closing your eyes, lightly touching your eyelids and turning your head. There's a very odd pulsating feeling under your eyelids.
"In other words, what's the connection between the classic explanation of 3D
in holography (phase information) and 3D in stereoscopy?"
Beyond pure physics, both might have in common to use (artificial) techniques that actually PRODUCE a 3D image. Both require a viewer to reconstruct the image. In some ways both methods en-code images that we
viewers subsequently de-code.
So in the case of holography, instead of merely stressing on the accuracy of an object's wavefront reconstruction, I would also mention its inherent
artificiality: by setting up a scene of objects, illuminated by highly monochromatic, polarized light and through interference, we are getting a hi-fi replica of the objects.
I've always found it interesting that a hologram, which is always talked about as an exact replica of reality, is always seen as a thing in itself. The real world is experienced by different people in different ways, even to the point of color perception, but the hologram is never related to human visual perception from a physiological point of view. There's always this explanation of the hologram as "exactly" reproducing the light from an object but no two people, who get the same light from the "real" object, ever see the real object in the same way. There are always small perceptual differences.
We've all seen the optical illusions a la Escher, but it would be interesting to create such an "Escher" hologram.
Very nice! Yours?
What I meant was a contradictory or apparently nonsensical image as in the standard illusions (http://www.optillusions.com/) that work because of our sense of perception in both 2D and 3D. Because a hologram has perspective and parallax, many of these apparent contradictions are resolved by simply looking at the image from a different perspective (the famous triangle made of three rods comes to mind). I was wondering if there was an apparently contradictory or nonsensical holographic image that worked only becuse it was a hologram, ie used the propeperties of holographic reconstruction (maybe with deliberate aberration) to "work" the illusion. I used Escher because that was a famous, apparently nonsensical image.
As I said to Martin, I think too much is said of how true-to-life a hologram is giving people the impression that it is life. I thought such a hologram may be a way of saying something about visual perception in a medium that represents life purely through diffraction.
It is my hologram, but I have not made it, and I am not sure who has.
I understand what you mean about a 3d illusion and I have thought about that many times. Unfortunately I do not think it is possible to make such an illusion. At least not in the sense as the Escher illusions where you have a 2d picture seeming to illustrate something 3d but getting contradictory when lines do not behave as they would in the 3d world. I have problems in seeing that you could get a similar but 3d-ish effect with holography, since in most ways it behaves as objects in the real world. But perhaps you can distort things so you get some weird illusions.
In a way a hologram is a very good illusion since there seem to be an object where there really is just reflection or diffraction of light.
I really like when holographic artists, I mean real artists using holography, make holograms with the aid of HOEs. It is beautiful when they make 3d pictures of twisting and bending light. In a way that is to make illusions and to "create" a new world you can look into that you otherwise would not be able to see. At least that is my assumption. But is it like that? My question is if it is possible to make an identical picture not using HOEs but instead with other things like mirrors and lenses? Then it would be just a "plain" 3d copy anyway and not a 3d illusion. I hope you understand what I mean.
I think I see what you mean. I think Rudy Berkhout's work probably falls into this category. He creates flowing lines and ribbons of multi-colored light patterns by basically using twisted metal sheets. I tried to find a photo of his stuff but couldn't. Not very net-savy!
I've seen photos of wooden models based on Escher drawings and other optical illusion sources and they all depend on having the viewer look at the models from a particular direction (of course).
It occurs to me that with some 3D software you could create a set of views that once put together in a stereogram creates a hologram of an apparent real-world scene that could never exist in our 3D world.
Now, where did I put that list of lcd sources? The programming for this would be fun.
---------------------------
If someone says it can't be done but they haven't tried it, don't believe them. http://www.dragonseye.com/Holography
"I think too much is said of how true-to-life a hologram is giving people the impression that it is life."
That's also my impression. Creating mere copies (whatever their perfection) of real objects might turn out to become kind of a trap under certain circumstances.
If I am not misled, most work (?) in art holography has been done as transmission holograms, and quite often as installations making use of "abstract" colored light patterns formed by a grating.
This may not answer you question but have you thought about this difference? This may be the phase explanation.
When viewing a stereo image each eyeball sees a single flat image with no perspective and thus the single eyeball focus remains constant, even when scanning the image. Now, when viewing a hologram with one eye, the eye sees different parts of the image with perspective. As the eye scans the holographic image the lens of the eye must deform to focus that part of the object to the retina. I would imagine this to be more noticeable in bright light as the iris is larger.
If you hold your finger up at arms length and view it with one eye you will see that the distant background is out of focus. When you focus on the distant background the finger is out of focus.
So, although the differential distance in the different areas of a hologram or real 3-D object is low, the eye may send readjusting clues to the brain for altering the shape of the eyeball/lens ever so slightly, probably not noticeable, for focusing the different parts and these clues may be detected or subliminally used for realism. This single eye focus probably plays a small role in the overall 3-D interpretation of an object or hologram as the differences between the two eyes perspective is greater and thus more dominant. But I bet it makes the difference between a cool 3-D stereogram and a WOW that looks like a real object hologram.
Basically, this is accomodation. I wonder what the resolution in 3D space is? I mean, if you made a hologram of a bunch of dots in 3D space arranged to look like a hammer (I believe the CGI people call this 'particles'), how close would those dots have to be to have the brain say, "That's a hammer" as opposed to, "That's a bunch of dots that looks like a hammer". Remembering that in real life accomodation, the lens changes continuously, while in this kind of system, the eye would have to accomodate "digitally". What would happen if you made a bunch of dots to look like a hammer, used a diverging ref, then viewed the hologram through a layer of glass and twisted it. The dots would fly apart at a greater rate at the edges of the twist and seem to explode - special effects explosion holography?.
Basically, this is accomodation. I wonder what the resolution in 3D space is? I mean, if you made a hologram of a bunch of dots in 3D space arranged to look like a hammer (I believe the CGI people call this 'particles')
Or just points.
how close would those dots have to be to have the brain say, "That's a hammer" as opposed to, "That's a bunch of dots that looks like a hammer".
They wouldn't have to be very close at all (constellations are a good example). Depending on the angle you could get away with very few points and people would see the hammer.
Most people also don't see (in the mental sense) any difference between "looks like" vs "is". I don't remember where the example comes from but show someone a painting of a boat and ask them what it is. Some large proportion of people will say "it's a boat" and the rest will correctly answer "it's a picture of a boat".
Actually this probably came from a joke about technical people being too literal
Remembering that in real life accomodation, the lens changes continuously, while in this kind of system, the eye would have to accomodate "digitally". What would happen if you made a bunch of dots to look like a hammer, used a diverging ref, then viewed the hologram through a layer of glass and twisted it. The dots would fly apart at a greater rate at the edges of the twist and seem to explode - special effects explosion holography?.
---------------------------
If someone says it can't be done but they haven't tried it, don't believe them. http://www.dragonseye.com/Holography
damn, this bit scrolled off my screen and I forgot to reply to it...
Remembering that in real life accomodation, the lens changes continuously, while in this kind of system, the eye would have to accomodate "digitally". What would happen if you made a bunch of dots to look like a hammer, used a diverging ref, then viewed the hologram through a layer of glass and twisted it. The dots would fly apart at a greater rate at the edges of the twist and seem to explode - special effects explosion holography?
Hmmm. Is the layer of glass flat? If, using CGI conventions, the axes are X - left/right, Y - up/down and Z - in/out of the screen or hologram, what axis is the glass rotating along? Y? Or is it just rotating around Z?
At any rate this sounds like a fun experiment.
---------------------------
If someone says it can't be done but they haven't tried it, don't believe them. http://www.dragonseye.com/Holography
"Hmmm. Is the layer of glass flat? If, using CGI conventions, the axes are X - left/right, Y - up/down and Z - in/out of the screen or hologram, what axis is the glass rotating along? Y? Or is it just rotating around Z?"
Assuming the hologram is in the x-y plane, it's rotating around y. I was trying to say that as you rotate the hologram with a flat glass above it, the thickness of glass at the edges would be larger than the thickness in the centre. This extra thickness would cause greater refraction of the image light at the edges, in particular in a particle system, the seperation between particles would be greater. The particles should therefore seem to fly apart faster as you turned the plate at a constant rate. I had not considered a non-flat piece of glass, but if the glass were curved upward at the sides this effect should be enhanced.
