h a l f b a k e r yPoint of hors d'oevre
add, search, annotate, link, view, overview, recent, by name, random
news, help, about, links, report a problem
browse anonymously,
or get an account
and write.
register,
|
|
|
Since the eyes see two pictures and interpret them into one image, it is possible that if we show an image created from both pictures, you would be able to "look thru" the picture and see interpret it correctly in two dimensions (unless you closed one eye).
I suppose with motion this would be particularly
true.
So all that's needed is this: two digital cameras at eye distance watching a scene (same focus for both cameras).
Then digitally mix and impose both pictures one on the other. No polarizing no glasses, just show the blurred picture movie and voilla! (or ?) .... you look "thru" the picture (or "passed" the picture) and see a 3d scene, with no extra machinary.
Waiting for results of your experiments.
Hmmm. I dunno.
http://bz.pair.com/fun/stereuhoh.html
29 Oct 02 | [35Kb image] I shot this pair as converging on the model object with ~4 inch spacing between camera film planes (CCD planes?). Doesn't exactly jump out at you. Am I missing something? [bristolz, Oct 04 2004]
Stereogram
http://www.eyetrick...ograms/thealien.jpg
A 3-D Alien, to those who can see these things. [Amos Kito, Oct 04 2004]
(?) Animated Stereogram Video
http://www.worldlan...m/Products/8688.htm
For those who think printed ones aren't nauseating enough. [Amos Kito, Oct 04 2004]
Stressed?
http://www.noise-to...chives/stressed.pdf
When you continue searching for the focal point... [pashute, Mar 21 2006]
Hybrid Images
http://cvcl.mit.edu...onroe_einstein.html
See second anno below from [gtoal] [gtoal, Aug 02 2007]
Stereogram
http://www.stereosc.../gallery/index.html
For cross-eyed viewing (pretty much what you descibed--except the images are side by side) [cowtamer, Aug 03 2007]
[link]
|
| |
This is exactly how the red/green or red/blue stereo glasses work. Without the colour tinting increasing the contrast between the images many people (I think) will not see the 3d effect. I'd be interested to see how well this works if you get an image up on the web. It would be interesting to compare its effectiveness to the same image as a red/green stereogram. |
|
| |
This is your baby, though. If it's going to be researched it will probably have to be done by your hand. |
|
| |
I could *maybe* imagine this working for very simple images but, how could the eye/brain know which parts of even a mildly complex picture represent left and right eye images? I have serious doubts about this one. |
|
| |
Would the camera views be set to converge at some point in the distance or remain parallel? |
|
| |
Yup [bristolz], kinda what I expected. That's enough experimenting to satisfy my curiosity. Gracias. |
|
| |
[bris] The stereuhoh needs some work. Maybe the spacing? With a Stereogram [link] you need to do serious eyeball aerobics. There's something different about yours -- I think you need to compensate for image size. You're gonna need guide marks, two "dots" to converge, to give your eyes something to focus on. I dunno where the dots should be positioned, maybe have the original "dot" someplace just below & in front of the subject, so it's part of the photo. Even then, the resulting image will still, of course, be "blurry". Many people can't see Stereograms and will certainly never see 3-D in a superimposed image. |
|
| |
Sorry, won't work. The cameras need to route the information to individual eyes, somehow. Otherwise, as waugs says, you're just seeing two images blurred together. |
|
| |
[waugsqueke], I have an animated Stereogram video. It's pretty amazing, but gives me a splitting headache in 5 minutes. That's bad, since it took many hours for me to figure out how to "see" the thing. I can see printed ones fine.
The big advantage of SIRD Stereograms is there's just "random" noise in view until the very moment you focus on the "3-D image". Any "blur" is hidden as noise again.
I'd bet some people could view a superimposed photo in 3-D, but it would take practice. And it's such a simple idea, it must have been tried already. We see double images all the time. The ones that converge are perceived in 3-D. |
|
| |
Amos, the two dots idea sounds great! |
|
| |
I'm not talking about SIRD, and it's true "many people can't see stereograms" but many more (at least 85% of the population) can. |
|
| |
Others - go ahead fishpeople, "won't work" is not very scientific is it? Now you ask why the brain would do this: simple: If there is something recognizeable and it's MOVING there's all the reason the brain can do the job. |
|
| |
Consider any movie or TV. You are seeing long scenes of nothing in between the pictures. The eye and brain completely ignore that. |
|
| |
Bristolz thanks! But your picture wasn't "moving". The idea is to see (and hear) something moving (a head talking) and then your brain will "get it" (possibly). |
|
| |
And with Amos' idea of two dots it would be even easier! |
|
| |
Wiggle your monitor around a little. |
|
| |
Bris wow! I looked at your picture for a while, focused in back (like the SIRD Amos gave in his link) and found that I do have a 3d feeling of the laptop. |
|
| |
The problem is that for a single picture, it seems not to make sense, and I keep on coming back to the original. Also, it seems your picture doesn't have both pictures intertwined, just in a certain area one superimposes the other. Am I right? |
|
| |
Intertwined? Not at all sure how I could do such a thing. It's two images, layered, with one at 50% opacity. Simplistic, I know. |
|
| |
The idea of intertwined is that there should not be any "mix of colors". Instead each pixel is from a diferent picture, and no color compensating is done in the new picture for the mix. |
|
| |
Also, possibly, shading each picture with a different color (e.g. green/red) it may help the eyes. |
|
| |
Bris, I looked again at your picture and it DOES jump out at me. The problem is that I still see the two parts (in each eye) and don't ignore the double info, as I do in SIRDs. |
|
| |
I believe a movie would compensate for that. |
|
| |
//Also, possibly, shading each picture with a different color (e.g. green/red)//
You've just rediscovered the red/green stereogram. |
|
| |
Amos: Is that video of a random dot stereogram? If so, I'd love to see it. |
|
| |
Try looking at your elf quite close up in the mirror. Let your eyes relax, and allow your left eye field-of-vision to superimpose over your right (or vice versa) magic eye style, the reflection of your eyes should be similar enough in appearance to 'lock' together in the middle, the same way as rods banananas do. you will now appear (to your elf at least) to have three eyes. This can be quite unnearving whilst pissed. |
|
| |
waugs, rayford: you are ignoring anything that doesn't agree with your pre-conception. This is something to be experimented not philosofised. How would the brain compensate? How does it compensate for movies (still pictures shown at an interval). How does it compensate for head movements. How does it compensate for two disparate images received from the two different eyes? Some serious answers were given here. Please read them again. And waugs, if you want to stop just washing an idea away without any thought, take a few minutes to stare "in back" of the picture. Perhaps adding the two dots would do the trick for you. I tried it on any other twosome parts and waited till the picture "converged" into one. Then the picture jumps out at you. (Unless you never see any 3d images in which case I rest my case). |
|
| |
Bristolz, could you put two dots (which should converge) on image? |
|
| |
//Is that video of a random dot stereogram?//
[st3f], yes it is. If I can find my tape, you're welcome to borrow it for as long as you like. I've found a [link], too. |
|
| |
Damn. I was assuming an MPEG rather than a physical tape. You meing USian, and me being a brit seems to render this impractcal. Thanks for the offer, though. |
|
| |
Thank you waugs. I thought you are simply dismissing the idea. I will take the challenge and put some more theory into this posting to prove my point, within a few days. In any case it seems to me so easy to achieve (for someone with two simple cameras and some graphic software) that it's worth the try. |
|
| |
I have never been able to decipher a Stereogram |
|
| |
An easier way to make 1D, 2D and 3D images appear to be blurry is to use copious amounts of a canned form of water mixed with alcohol. <Homer>Beeeeeeeer</Homer> |
|
| |
Sorry, this honestly doesn't work. |
|
| |
I know, there's a load of you just screaming 'but why, you're just dismissing it' but I'm not. I work in the displays dept of a large national lab, just in case you question my credentials. |
|
| |
The brain works out 3D by comparison of the two *different* images it receives from the eyes. For the most part, although you don't notice it, you can still see the two superimposed images outside the centre of your field of vision. For example, point your finger straight towards you and look at the end of it from quite close up - you can still clearly see the two images of your finger. |
|
| |
The suggestion of just letting the brain sort it out doesn't give the brain 2 images to use, it just gives it one. The brain can check the focus by making the pictures match, and then work out from the focus distance how far away it is. There's nothing to push it to make it 3D. That blurred laptop does look a little more interesting than I expected it to, but only when I looked at some other part of the image, so that the double image was blurred. |
|
| |
This is why stereograms work. When you look at them straight, nothing. But when you cross your eyes, suddenly the brain has two different images, and if you cross them far enough it can 'lock' at a different focal distance than before. It then uses the parallax difference to work out the distances. |
|
| |
All 3D technologies just use different methods of giving 2 different images to the eyes - red/blue specs, or cross-polarisation, or LCD shutters, or of course 'true' 3D volumetrics. Sharp's LCD just uses two 'sweet spots' taking advantage of LCD technology's inherent limited viewable angle. |
|
| |
Someone keeps mentioning the 'long distances' between successive film images as evidence the brain can perform magic as regards image interpretation. The truth about moving image is just that the eye is not the most responsive detector ever built and it integrates over long times as it measures the rgb values. That's why things are smoothed out, however jerky they are, be it video, a refreshing CRT, or a 'moving' caret. |
|
| |
Oh and before you say I just can't do stereograms, I've never met one I couldn't do. |
|
| |
But there's a larger load of us who were skeptical to begin with. |
|
| |
//Try looking at your elf quite close up in the mirror. Let your eyes relax, and allow your left eye field-of-vision to superimpose over your right (or vice versa) magic eye style, the reflection of your eyes should be similar enough in appearance to 'lock' together in the middle, the same way as rods banananas do. you will now appear (to your elf at least) to have three eyes. This can be quite unnearving whilst pissed.// |
|
| |
My wife and I used to give each other the "cyclops stare". |
|
| |
This idea might be salvagable in a *small* number of cases -- when the viewer is colour-blind in one eye and the good eye is also the dominant eye. |
|
| |
I have observed in proper stereograms that if you have a dominant eye, you get the full 3D effect even if the image that the weak eye is seeing is badly out of focus - your mind has already adjusted to using a poorer image for depth cues and the better image for detail. |
|
| |
So let's say the left eye sees red as green: you have a photo of a greenish object as the main image, with a red-tinted ghost of the object superimposed for the left eye. The right eye will concentrate on the real object, but if the colour balance is right, the left eye ought to get a stronger green image than the right, which it will use to provide depth cues (whereas the right eye subconsciously ignores the red image which is weaker and unnatural) |
|
| |
No cross-eyed squinting necessary, but it should only work for a small percentage of the population, if it works at all. Personally I prefer the entirely workable idea filed under "3D Contact Lenses". It would be too cool to have wall posters that could only be read by people wearing polarized contacts - a real life version of "They Live" :-) |
|
| |
The point I was making (4 years ago) was that when watching a movie, you are "following" an object in 3d. This can help the mind and give it an extra anchor, which COULD cause the eyes to dismiss the extra info. |
|
| |
Recently I have seen (at Datasafe but I dont think its their product) a 3d "wide-screen" TV, without glasses. Everyone who sees it stops in their tracks, but when you try to read something, you get a terrible headache and dont want to continue looking... |
|
| |
OK, here's another way it could work. Again, this relies on poor vision in one eye - but not colour-blindness this time... |
|
| |
Let's assume its your left eye that has a somewhat fuzzy focus compared to your right eye. |
|
| |
Create a left-eye image using the low spacial frequency components of the left half of a stereo pair, and a right-eye image using the high spacial frequency components of the right half of the stereo pair. Superimpose the two images. (See the MIT site in the links if you have no clue what I'm talking about) Now walk backwards from the image until you get to the point where you see the low frequency image more strongly. |
|
| |
Somewhere just before that point both images should appear about equal in perception. Remember, your good right eye should *not* yet see the low-frequency component predominating because it can still see the high frequency image that your left eye has now lost. |
|
| |
Hopefully, the two images should combine to give a stereo effect. |
|
| |
Definitely worth an experiment, probably not too hard generating the hybrid image - maybe harder finding someone with the right level of poor eyesight in one eye to test it however! |
|
| |
A link to the Hybrid Image site at MIT should now be in the section above. |
|
| |
[gtoal]OMG! He/She was down the pub the other night!
(May have involved Beer goggles) |
|
| |
Very much baked. They are called stereo pairs--you just cross your eyes and see the 3D image. See link. |
|
| |