PDA

View Full Version : OT: Accidental 3D pictures



Evan
12-06-2010, 09:28 AM
I was reviewing some of my many pictures recently and suddenly realized that I may have a considerably larger collection of 3D images than I thought. When I photograph something I commonly take several images rather than just one. As it happens, if hand holding the camera it is very likely that a small horizontal shift may occur which inadvertantly produces a 3D image pair.

Then I noticed that there is another way for a certain type of 3D image to form. Rather than moving the camera the subject can move instead. In this instance the subject is clouds which are a favorite subject of mine. My interest in astronomy extends to anythng related to the sky, both night and day. Searching through my collection I have found some interesting example of 3D images that were formed by the movement of clouds instead of the usual method od displacing the camera.

This was the image that first alerted me to the possibility. It was taken from the top of Mt. Kobau at a star party. The images are encoded with an x for crossed eye or an = for parallel viewing. In some cases an interesting effect is produced both ways. The type of 3D effect produced here is a semi synthetic effect and does not necessarily represent the actual spatial relationships that existed.

http://ixian.ca/pics8/3dclouds.jpg

This next set are clouds back lighted by an intense aurora with a trace of light pollution on the right from town.

http://ixian.ca/pics8/aurora_x.jpg

http://ixian.ca/pics8/aurora_s.jpg

This one is different in that the 3D effect is real as I took two images out the window while driving. This gave a very long baseline that reveals the 3 dimensional shape of the isolated cloud.

http://ixian.ca/pics8/3dvirga.jpg

continued

Evan
12-06-2010, 09:32 AM
http://ixian.ca/pics8/3dvirga_s.jpg

http://ixian.ca/pics8/anticrep.jpg


This is best in parallel viewing mode. These are a rare phenomenon called "anti crepuscular rays".
http://ixian.ca/pics8/anticrep2.jpg

aostling
12-06-2010, 09:55 AM
Evan,

Your "heavens in hyperstereo" collection is a real treat. That's the first time I've looked into an aurora, in 3D. I look forward to more of this, as you continue to take such pairs during the season.

vpt
12-06-2010, 10:05 AM
I'm not seeing the 3d. Nice pictures though!

Evan
12-06-2010, 10:18 AM
Not everybody can see 3D this way for reasons that remain unknown. My wife cannot see it at all no matter how hard she tries. I suspect it may have something to do with eye dominance. I am largely ambidextrous and that may also extend to vision. Someone with a strongly dominant eye probably has strongly dominant visual processing in the brain as well.

Mcgyver
12-06-2010, 10:25 AM
I'm not seeing the 3d. Nice pictures though!

centre two images and go cross eye...there'll be in the centre of your view an image of that looks like two overlays....try with eyes to get these two images to line up and all of a sudden when eye/mind registed it you will lock onto it and be looking a three d image. Just don't stay that way :)

aostling
12-06-2010, 11:11 AM
You may need to resize the stereo pair to match it for your particular monitor. Download the jpg to your desktop, then insert it into a cell in an Excel spreadsheet. The image can then be easily moved around the screen, and resized with the mouse. This is particularly helpful for parallel pairs, which need to have "homologous points" separated by a distance close to your interocular spacing (68mm for me).

dp
12-06-2010, 11:56 AM
That is pretty amazing. I have some similar photo series I've taken and I'll have to examine.

This phenomenon is something I discovered as a kid. I found that by staring at the springs of the upper bunk bed I could merge adjacent rectangles at which time the surface became 3D. Same with bathroom floor tiles - the smaller the tile the easier, but with practice I could merge repeating objects wider than my eyes are apart (not looking cross-eyed - I must have looked like Marty Feldman, though).

http://www.lifeaftercoffee.com/wp-content/uploads/2006/04/marty_feldman.jpg

vpt
12-06-2010, 12:06 PM
That is pretty amazing. I have some similar photo series I've taken and I'll have to examine.

This phenomenon is something I discovered as a kid. I found that by staring at the springs of the upper bunk bed I could merge adjacent rectangles at which time the surface became 3D. Same with bathroom floor tiles - the smaller the tile the easier, but with practice I could merge repeating objects wider than my eyes are apart (not looking cross-eyed - I must have looked like Marty Feldman, though).

http://www.lifeaftercoffee.com/wp-content/uploads/2006/04/marty_feldman.jpg




I have been able to do something similar since a kid as well. If I stare long enough at pretty much anything it somehow my eyes focus differently and it will seem as though everything I look at is far away from me even though the objects are only feet away.

OP if I hold the monitor up to my face and go cross eyed and whatnot I can kind of see what you guys are talking about. It is like binoculars.

Evan
12-06-2010, 01:56 PM
I think you have it then.

Here are a few more.

This is from the Banff-Jasper National Park in the Rockies, both types.

http://ixian.ca/pics8/massif_x.jpg

http://ixian.ca/pics8/massif_s.jpg

This one just doesn't bear being reduced enough to view in parallel.


http://ixian.ca/pics8/3dwinter.jpg


If that makes you feel cold then this one will warm you up a bit. It seems that even an amorphous background will produce a 3D effect for the foreground. Also crossed eye.


http://ixian.ca/pics8/3dfire.jpg

Optics Curmudgeon
12-06-2010, 09:11 PM
Wow, nice effect. One thing for people that have a hard time "free viewing" stereo pairs to try is using two lenses, the objectives from that pair of 7X50 binoculars that got left on the roof of the car and fell off work. Hold one in front of each eye and vary the distance between them to get the images coincident.
Another 3D effect can be seen when imaging solar features in two slightly different wavelengths. This pair was 6563.3 Angstroms for the left image and 6562.3 Angstroms for the right one. These shifts correspond to +/- 22 kilometers per second in velocity, so material approaching the instrument appears to be out front.
http://i106.photobucket.com/albums/m278/ocudge/Spot11080922stereo2.jpg
Poor daytime seeing usually produces spurious effects but this one came out pretty well.

Tony Ennis
12-06-2010, 09:26 PM
Nice pictures, Evan. I got the narrower pictures to look 3D a lot easier than the wider ones. In fact, none of the wider ones worked for me.

The 2nd 'rock face' picture and the fire picture worked the best.

Evan
12-06-2010, 09:29 PM
Very cool. I wish I had some filters that narrow.

aostling
12-06-2010, 09:30 PM
Even ordinary objects can mesmerize, in 3D. Here is a stereo pair of pond scum I took with my Stereo Wollensak about twenty years ago.



http://i168.photobucket.com/albums/u183/aostling/Screenshot2010-12-06at62316PM.png

This is arranged for parallel viewing. I look at my mounted stereo slides with Kodaslide II viewer, which has a pair of outstanding eyepieces. That is 3D viewing at its best.

Evan
12-06-2010, 09:35 PM
Tony,

Just capture the screen and stick in your favorite image program and reduce them.

ulav8r
12-06-2010, 11:03 PM
The crossed ones were easier. The parallel worked with my eyes crossed a little less and took a little longer to get right. The 3D effect was greatest on the trees and snow, it was great. Once I got the first one right, I could scroll down to the next one with no extra effort.

rohart
12-06-2010, 11:51 PM
A couple of years before digital SLRs came out I made up a fixture to hold two identical film cameras and fire them off electronically to make stereo pairs of scenes with a decent amount of movement. I wasn't aiming quite at stereo images of Evel Knievel in mid flight, but you get my drift.

Then came the digital revolution. Now, my original digital SLR can be bought quite cheaply, so when I get a round tuit I'll ressurrect my plans. I suspect I'll have colour rendition problems, but I'll no doubt find a few ways around that.

I found your evening silhouettes very good, Evan. But now you have made your serendipitious discovery, how about capitalising on it.

How about taking two shots in close succession from a moving vehicle ? Or taking two quick shots of traffic on a motorway. The eye can make up its own mind what was moving and what was still.

aostling
12-13-2010, 12:55 AM
Another 3D effect can be seen when imaging solar features in two slightly different wavelengths. This pair was 6563.3 Angstroms for the left image and 6562.3 Angstroms for the right one. These shifts correspond to +/- 22 kilometers per second in velocity, so material approaching the instrument appears to be out front.
http://i106.photobucket.com/albums/m278/ocudge/Spot11080922stereo2.jpg


I can see the stereo effect at the flares. But I don't understand how the two photos being at slightly different wavelengths produces the necessary parallax. Could the stereo separation be caused by the sidereal rotation of the sun during the period between the two exposures?

dp
12-13-2010, 02:23 AM
I can see the stereo effect at the flares. But I don't understand how the two photos being at slightly different wavelengths produces the necessary parallax. Could the stereo separation be caused by the sidereal rotation of the sun during the period between the two exposures?

Parallax can only be achieved if the pictures were taken at two different times which is probably likely. Differential refraction can happen at different wavelengths but that variation may not arrive in the brain as 3D. Or it might, depending on what you've been doing with your brain in the moments prior :)

Evan
12-13-2010, 06:53 AM
But I don't understand how the two photos being at slightly different wavelengths produces the necessary parallax.

There is none if the material is approaching directly toward the camera. However, because the material approaching with any obliquity is traveling at different velocities as imaged by wavelength, the faster will have moved further from the original site of eruption than the slower. At any given angle that means the faster will offset by an amount proportionate to the difference in velocity and the viewing angle. It isn't a parallax view of the same material, it is a view of different material at different velocity.

I have found a few more examples in my photo collection. In particular this first one may be rather unique. It shows a rainbow only a few feet distant formed by rain falling from a blue sky on ground that isn't wet in bright sunlight. On top of that not only can you see the rainbow you can also see the raindrops in 3D. It may take a minute to pick them out but once you do they are easy to see.

These are all in crossed eye format. They don't lend themselves to being reduced as much as necessary for parallel format or they don't look right for some odd reason.

http://ixian.ca/pics8/3drainbow.jpg

http://ixian.ca/pics8/3dsunrise.jpg

http://ixian.ca/pics8/3dclouds2.jpg

This last image truly benefits to an amazing degree from being in 3D.

http://ixian.ca/pics8/3dflower.jpg

A.K. Boomer
12-13-2010, 06:54 AM
I cannot get any of the pics to work doing the crosseyed thing but the pics in the OP that have the road (second from last esp.) I don't need to do a thing and it screams 3d at me - the road narrowing sets the stage for distance in one direction --- the fanned out rays set the stage of distance for the other - it effectively tricks my mind into seeing everything else in stages in-between (just looking at one pic), very cool.

We see with our brain not our eyes.

aostling
12-13-2010, 10:17 AM
I have found a few more examples in my photo collection. In particular this first one may be rather unique. It shows a rainbow only a few feet distant formed by rain falling from a blue sky on ground that isn't wet in bright sunlight.

These came out very well, even the raindrops and the rainbow. I like the hypostereo view of the flower. Also, thanks for not showing a close-up of an insect's face (those can give some people the creeps).

Evan
12-13-2010, 10:33 AM
Think about the rain drops for a moment. The images were taken at different times. What's going on?


BTW, here is an intentional 3d of my 6" telescope and fake 3D of the Pleiades.

http://ixian.ca/pics8/3d6scope2.jpg

http://ixian.ca/pics8/3dstars.jpg

aostling
12-13-2010, 11:38 AM
Think about the rain drops for a moment. The images were taken at different times. What's going on?

BTW, here is an intentional 3d of my 6" telescope and fake 3D of the Pleiades.


I ascribed the 3D raindrops to the reflections of the sun from them. The raindrops would move, but the reflections would stay (generally) in the same spot. I'm not convinced by my explanation, though.

How did you shift the stars in the Pleiades to give that wonderful stereo view? Was this random, or based on known distances?

Good grief, your new telescope is complete, and mounted! That sets some sort of record for project completion haste.

dp
12-13-2010, 12:08 PM
Interesting how the fake 3D Pleides images reveal the involuntary motion of the eyes as you move your attention from one star to the next. That motion causes the light points to instantly merge to a single point while adjacent points in the immediate area split to two images, giving the 3D affect.

Evan
12-13-2010, 01:39 PM
The raindrops are chance alignments of the raindrops in the two images that occur purely by accident. Because there were so many drops the chances are high that some will be in approximately the correct spatial alignment to produce a stereo pair even at different times.

The stars are arranged by leaving the left image as original and with my eyes crossed then shifting the stars in the right image using a select/move brush in PSP. The right image star is moved to the left or right by an amount that produces the desired result. This is constrained by not wanting to overlap an already existing star. A general rule is that stars of similar brightness that are very close together are probably gravitationally bound and should stay reasonably close. Another is that very bright stars are probably closer than stars of similar spectral content that appear dim.

The distance to the individual stars isn't known with sufficient precision to generate a "real" 3d image of the cluster. Hubble was able to determine the average distance using parallax measurement to be about 440 light years but the Pleiades represent the current absolute limit of the technique using the orbit of Earth as a baseline. Further improvement will await the deployment of a telescope in an orbit at the L2 Lagrangian point of Jupiter perhaps.

aostling
12-13-2010, 02:15 PM
The raindrops are chance alignments of the raindrops in the two images that occur purely by accident.

An interesting phenomenon. Now that I think about it, yours is the first stereo view of rain that I've seen.

Sorry, I mistook your old 6" telescope for the new project. The brain failed to click -- I know that the new design is dramatically different.

Optics Curmudgeon
12-13-2010, 02:30 PM
I can see the stereo effect at the flares. But I don't understand how the two photos being at slightly different wavelengths produces the necessary parallax. Could the stereo separation be caused by the sidereal rotation of the sun during the period between the two exposures?
In a normal 3D image pair the effect is due to the difference in relative positions of features in the images. Here the difference is in intensity. In the left (redshifted) image features moving away appear relatively brighter than stationary ones, the opposite is true for the blueshifted ones in the right image. The images were taken seconds apart, and the only significant spatial shifts were due to atmospheric turbulence. The 3D effect here is a visualization of velocity instead of depth. The effect was first used by Robert Leighton in 1960 to visualize velocity fields on the sun, leading to the discovery of the 5 minute solar oscillation. I do it just because it looks cool.

Joe

Evan
12-13-2010, 03:56 PM
What instrument are you using?

Optics Curmudgeon
12-13-2010, 05:26 PM
What instrument are you using?

It's a digital spectroheliograph. A coelostat feeds a 75mm F/15 objective, through a relay lens system and a scanning prism to a 1.25 meter Littrow spectrograph. The exit slit is replaced with a relay lens and the detector out of a business card scanner (1275 pixel linear CMOS detector). Some hardware fools the thing into thinking it's still scanning cards. It has an image scale of 1 arcsecond per pixel, and a passband of 0.4 Angstrom. A tilting plate lineshifter allows arbitrary wavelength shifts of +/- 1 Angstrom, these shifts can be made quickly because the plate is mounted on a galvanometer scanner. Like any astro project, it's a work in progress.

Joe

Peter.
12-13-2010, 05:43 PM
Can 'seeing' these images cause any lasting damage to your eyesight?

I can 'see' Evan's raindrops and the flower in good 3-D, and the telescope easily 'jumps' out of the monitor and gives the best image for me by far. I find that very little effort is needed which is uncomfortable for a few sceconds unti I get the focus lock then I can relax and hold the picture.

Trying to 'see' OC's flares however is painful for me. I can bring the two pictures together with some small effort by they won't 'lock' and my eyes quickly become strained. After trying that image for 30-45 seconds I find even my normal focus is blurred for a little while afterwards.

rohart
12-13-2010, 06:11 PM
All the 3D images are excellent, Evan. My only comment is that the one of your telescope is made too great a separation for the object distance. The 3D is overdone - the scope is fine, but the background doesn't connect.

The shot of the Pliades reminds me. When I was younger, and had a tendency to stare into space, inner space, or at least indoor space, I developed the following trick. When listening to HiFi on a tape player with two level vu meters, I would cross my eyes to superimpose the two meters. The needle would become one, but would now describe a 3D movement, into the meter and out, according to which channel was putting out more level at the time.

Just fun - no real meaning other than level, but a good way to monitor both levels simultaneously.

Evan
12-13-2010, 06:59 PM
Can 'seeing' these images cause any lasting damage to your eyesight?


No, none at all, lasting or otherwise. Crossing your eyes is no different than looking at something up close. In fact, for those having trouble seeing the crossed eye images, hold your finger about halfway between the screen and your nose and focus on your finger. You should also then be able to see the 3D image beyond your finger.