How-To Tuesday: Make 3-D photos

(best viewed with 3-D glasses)

This week's how-to is a fun one, we're going to show you how to make 3-D photos with any digital camera and some free software. We'll also explain how 3-D photography works and as a special treat, we've got a gallery of 3-D gadget photos to view along with how to make 3-D photos from NASA images.

What are 3-D photos?

Before we get started, we figured it would be good to explain why 3-D photos work.

Humans over time have evolved with many capabilities which offer an advantage over non-humans (at least for now) opposable thumbs for gripping stuff, big brains for figuring out stuff, and Binocular vision for seeing two points of view at the same time. This allows us to perceive depth (i.e. see things in three dimensions). The ability to tell if a tree branch is 3 feet or 3 inches away is pretty important, as is being able to tell if that lion is 30 yards or 30 feet away. Both our eyes are on the front of our faces and each see a slightly different view of the same thing. Close one eye, then close the other, you'll notice there's a slightly different view of the same thing. Our brains put these together so get a 3-D image of our surroundings.

Anaglyphs, the type of 3-D photos we're going to make, do the same thing by tricking our eyes into doing the same thing they normally do, just with a flat picture. The anaglyph is a single image (usually black and white) that has red/blue "outlines" on it which, when viewed with 3-D glasses, appears to jump out at you. The image gets processed so each eye sees a slightly different view and our brain combines them to give the effect of depth perception.

If you have 3-D glasses handy, look at a color wheel from any image editing program, you'll notice that the red areas are bright and the blue areas are dim when viewing through the red lens, and the opposite is true when viewed through the blue lens. This effect gives the images depth, lighting, and the overall 3-D effect in our brains.

The reason a lot of 3-D photos are in black and white is that with a color photo red and blue objects aren't quite as visible since we're blocking those colors out once we put on the glasses. So in our example we're going to make the images black and white. Color images work, you'll just need to avoid those colors in your photos.

Other types of 3-D...
There are other types of 3-D photos. Interlaced, which requires LCD glasses, is usually expensive, Wall-Eyed, those "Magic Eye" posters you need to stare at for a minute to see the images, and Cross-Eyed, two images side by side that look 3-D when you cross your eyes. If you're interested in making those types, Google around for resources, but anaglyphs are the easiest ones to make and only require cheap red and blue filtered glasses, so we went with those.

And speaking of the glasses, it's time to make some.

Getting started
For this how-to, you likely have everything you need as far as hardware goes, the only item you might not have are the 3-D glasses, but those aren't too hard to find. In general, there are a lot of ways to make 3-D photos, but here's the way we made ours and what we used.

Ingredients

• Digital camera
• PC (With Windows XP, SP1 or above)
• Free Software, Callipygian 3D Photo Editing Software.
• 3-D Glasses. These red and blue 3-D glasses usually come with a comic book, or some special video game/TV promotion. We got ours at Borders books in the kids section with a 3-D Dinosaur book. If you can't find any at a comic or book store, you can also order some from many places online, Ward's Natural Science has 15 pair for $15.75.

Installing the free app, Callipygian 3D

It's a good idea to download and install "Callipygian 3D" now, depending on your system set up, you may or may not need the .NET Framework (the app will tell you) if you do, you can get it here.

If you don't have a PC with XP or use a Mac or other system, we'll go over how to make these manually later in this how-to.

Taking the photos
We're using a Sony T1 digital camera, taking the photos at VGA quality (640 x 480), though you can take them at any resolution you'd like.

Now this is the most important part of the how-to: you're going to need to take two pictures of your subject/scene. The first and second picture should be at the same exact level (height) and only one to two inches apart. We're basically going to reproduce what your eye would normally see. If the images are taken too close or too far, the final image won't look 3-D.

It takes a few times to get the best images, but with digital photography you're not wasting film, so we usually take a few dozen. It's usually best to be 10 to 15 feet from the subject and to have a few objects like people or robots to give the photo depth. Outside photos tend to look the best. If you have a tripod to keep the camera level that'll help out, also be careful not to turn the camera in on the second photo, each one should be parallel like your eyes. If you're taking pictures of people, make sure they stay perfectly still for each photo (for this reason, we prefer robots).

When we make these 3-D, we convert some of them to grayscale images, while you don't need to do this, they tend to look better. If you want to shoot images and plan to keep them color, you'll be able to tweak the images in the 3-D application.

Just to recap...

1. Find a subject that will be still, and has foreground and background objects to give the photo depth.
2. Take the photo 10 to 15 feet away from the main subject.
3. Move the camera (not you) one to two inches to the right or left, at the same level. If you camera has a viewfinder that helps too.
4. We usually take a photo, move to left on the second one to keep it simple.
5. Take lots of photos, it's only electrons.

Getting the images on your PC

After taking your pictures, import them on to your PC the way you usually do, we just pop in our memory card and drag the folder off the card.

We usually keep our images in the pairs we took them, we'll even rename them and add a R or L to the image(s) to keep track of them as well.

Using Callipygian 3D
Now that we have a pair of images (one taken a couple inches away from the other, of the same scene) and converted them to grayscale, we can now import them in to the 3-D image maker.

You should have installed Callipygian 3D by now, click Start > Programs > Callipygian3D.

The main menu has two panels, a right and left one.

At the bottom of the menu is a pulldown list of Selection Styles, choose "Arbitrary (anaglyph)".

Click File > Open and choose an image for the left and right images, or you can drag and drop the images directly on the right and left areas on the menu.

If you took the photos in the wrong order, you can flip them by clicking Edit > Swap Left and Right.

Now that the photos are in, select the area on the left photo you'd like to make 3-D, click and hold the mouse button down while dragging the mouse.

Once you select the area click View3D > Anagplyph. You'll now see a preview window where you can make some adjustments, the sliders moves the image right or left, up or down, it's a good idea to put your 3-D glasses on now as you adjust them and get the best 3-D image.

There are other options at the bottom of the menu, where you can tweak the images colors depending on how you shot it, we usually choose Red/Cyan (black/white) but you can also choose other the other settings depending on the photos you shot. A little side note, Cyan is a blue color for the non-designery folks out there.

Red Cyan Full Color-Default mode that takes the Red component from the Left image, and  the Cyan component from the Right Image.

Red Cyan Tweaked Color-Similar to above, but with primary colors (R, G, B) desaturated to  reduce retinal rilvalry.

Red Cyan Grayscale Red-The left channel is converted to Grayscale using the NTSC weights,  and mapped to the red channel. This makes an easier-to-see 3D image,  at the expense of color fidelity.

 

Red Cyan Black White-Both Left and Right are converted to Grayscale. This makes an easyto-see 3D image, but with no color information.

Green Magenta Full Color-Uses Green/Magenta instead of Red/Cyan. For LCD projectors viewed  with polarized glasses (not the red and blue ones).

Magenta Green Full Color-Uses Magenta/Green instead of Red/Cyan.

Play around with each one with your 3-D glasses on to see which one works best.

When you've got your 3-D image the way you want it, click File > Save 3D View. This single image will be saved to your local system at the same resolution it was shot (for our example, VGA).

We usually save ours at the best quality and crop / edit in other image applications if needed.

And that's it! We usually keep a few of these images on our laptop and PDAs just for fun and even have a flat panel display at home that plays these files so when folks come over they can see some neat 3-D photos. You can also print them out

Here's the final image.


Making 3-D photos manually...
If you don't have PC or Windows, you can use PhotoShop or Gimp and make the photos manually, it takes a bit longer, but here's how to do that.

Just about any photo application will do, PhotoShop, PaintShop Pro, GiMP—as long as the application supports the red, blue and green  color channels to be changed.

For this example, we're going to use PhotoShop.

Open you files, we had named ours BOTL.JPG and BOTR.JPG.

Convert both files to grayscale (this usually makes the 3-D effect "pop" more) Image > Mode > Grayscale.

Go back to the left image and put it in RGB mode (Image > Mode > RGB).

Go to Window > Channels and select the Green and Blue Channels (Shift + Click). The Image should turn blue-ish.

Go to the right image, and go to Edit > Select All (Command + A) then Edit > Copy (Command + C).

Go back to the left image, and click Edit > Paste (Command + V). Then Click the RGB Color Channel in the Channel Window. The image should now have Red and Blue colors.

Click the Red Channel in the Channels Window, then Click the "Move Tool" in the Tool palette, move the image to align the 2 images over each other. You can use the arrow on click and drag it with the mouse, now is a good time to put on the 3-D glasses, once you line it up properly it will appear 3-D, if there is extra image information crop the image, then save it.

That's it!

Bonus Tip...
NASA's Mars Rovers (Spirt and Opportunity) take 2 photos with their navigation and hazard-avoidance cameras, just like we did for our examples. If you go through the gallery of RAW images, you can find pairs of images and make your own Mars photos.

Opportunity RAW Images can be found here.

Spirt RAW Images can be found here.

They way to tell which ones are pairs are to match up the URL (web address), the last part of the address has a R (for right) and L (for left).

For example, here are two images you could make a 3-D image of...

Link 1.
Link 2.

Note the only difference in the links are the R and L, and the minor shift in the photos.

Here's the final image, all 3-D'd up.


We spotted this on a JPL article (also about making 3-D images) and there are literally thousands of images to use.

3-D Gadget Gallery
Last up, here's a few quick 3-D photos we made of some of the gadgets and bots around here, enjoy!

---

Phillip Torrone can be reached via his personal site: http://www.flashenabled.com or  [e-mail]

[Pieter @ Dec 19th 2005 2:26AM]

You do realize that to make a grayscale image in Photoshot, you can do Image -> Adjustments -> Desaturate, right?

[Michael Tseng @ Dec 19th 2005 2:26AM]

man you are sooooooo spoiled. Who's your daddy?

[Trachalio @ Dec 19th 2005 2:26AM]

http://www.ishkur.com/captions/index.php?num=4

Sorry, just couldn't resist.

[decompyler @ Dec 19th 2005 2:26AM]

Pieter: There is nothing wrong with the way the author suggested how to do desaturate an image. Everyone has their personal method and that doesnt make it wrong. You do realize that your method isnt the easiest, RIGHT? (ctrl + shift + U)

[the Other michael @ Dec 19th 2005 2:26AM]

any way to get a briefer excerpt in to the RSS feed? All those images are nice--in the full post....

[twobits @ Dec 19th 2005 2:26AM]

Interesting anaglyphs. Most of them show greater disparity of the image at greater distances (Eg. the cell phone thing with buildings in the background is very strong) This shouldn't happen if the image pairs were shot by parallel optical systems (as you suggest the orientation of the camera on the second shot should be). However, you will notice that human eyes will not remain parallel when looking at close objects - they tend to converge on the object in focus - resulting in a more complicated disparity relationship which clearly the brain is able to deal with and may prefer. These non-parallel or converged pairs that you select as optimum may even be needed to match what the brain-eye system expects when looking at the images within arms length. Cool.

[Amy @ Dec 19th 2005 2:26AM]

Thanks so much for sharing all this info. Can't wait to try it myself.

[Modesty B Catt @ Dec 19th 2005 2:26AM]

...I'd do this with two of those boxy disposable camera stuck together side by side and reinforced with lollypop sticks (yay, for low-tech superglue).

Then try really hard to press the buttons at the same time for those action shots.

Scanner and Photoshop 2.5, sorted.

[Dan Gapinski @ Dec 19th 2005 2:26AM]

http://www.sciplus.com/singleItem.cfm?terms=10096&cartLogFrom=Search

[Jesse @ Dec 19th 2005 2:26AM]

twobits, think of it like your screen was a window--if your eyes are focusing on the screen, then if there's no displacement between the two halves of the 3D image of an object, it will look like it's right up against the "window" (for the same reason that 2D images look like they're the same distance from you as the monitor they're displayed on), while things that are further apart will either look further in the distance or like they're popping out of the window, depending on whether the red half of the image is to the left or the right of the cyan half.

BTW, I have a bunch of 3D images I've made on my webpage at http://www.jessemazer.com/3Dphotos.html , including some scenes from DVDs I converted to 3D (at http://www.jessemazer.com/3Dphotos.html#movies ) by taking nearby frames in scenes where the camera was panning sideways and using them as the left eye/right eye parts of a single 3D image.

[Jesse @ Dec 19th 2005 2:26AM]

Oh yeah, and I should also mention you can get a free pair of 3D glasses by sending a self-addressed stamped envelope to this company:

http://www.rainbowsymphony.com/freestuff

Just make sure you include a note asking for red/cyan glasses, since they have other kinds.

[Adam Villani @ Dec 19th 2005 2:26AM]

It's also possible to just train your eyes to focus in such a way that two adjacent images are brought together. This works with ordinary photographs, film or digital, that you don't have to manipulate in any way except placing them next to each other. Of course, they have to be photographed correctly first. But that's fairly easy to do... just remember that you're simulating two eyes with the camera. Take the two photos a few inches apart to simulate normal human vision, or farther apart to imagine how a giant might see things, which is useful if you want a 3-D shot of a landscape far away.

I took some photos from the south rim of the Grand Canyon where I took the one shot, moved maybe 10 meters to the side, and took another shot. Stick the pictures next to each other, and have your eyes bring a common object together, and you've got a 3-D image. A stereoscope can help if your eyes aren't focusing correctly.

[Barry Berman @ Dec 19th 2005 2:26AM]

For 3d within AE, we use an adjustment layer with the following effects. 3D Glasses, set to Red/Blue LR, and choose your left/right layered photos appropriately. Then also add to your adjustment layer, Shift Channels, shift, Green to Blue. This makes your image grey scale, except for the red channel but minimizes ghosting.

Alternatively, learn to cross your eyes for cross eye 3D.

Also, I prefer to not use regular cameras and use a Stereo 3D camera from the 50s, called a Realist. It has a unique look and is more fun. Here's a link to a recent roll I shot.

http://www.3dmation.com/images/stereo_photos/stereo_082104/cross/index.htm

Barry

[Barry Berman @ Dec 19th 2005 2:26AM]

Oh, forgot to add, you need to also rectify your photos. Often times there will be a vertical offset between your left and right exposure, and that's confusing for your brain to resolve. You should align the features to make it easier for your brain. Many pictures above need to be rectified.

Barry

[Morely @ Dec 19th 2005 2:26AM]

If you have a machinist friend, you can get him or her to make a simple bracket that fits on top of your tripod and under your camera; flip it to the left to take the first shot, then to the right for the second, and you'll always get the exact same parallax distance.

All it takes is two metal plates, four "arms", eight screws to hold the arms to the edegs of the plates, and a suitable screw to hold the camera to the top plate. The bottom plate should be drilled and tapped for the tripod screw. The whole thing when assembled becomes a parallelogram when viewed from the front or back.

[Jim Lewis @ Dec 19th 2005 2:26AM]

Using Photoshop, I just put elements in their own layer, then shifted the blue and red channels. The mountain picture is a much than the somewhat botched capitol building picture.

[Grfnbrg @ Dec 19th 2005 2:26AM]

http://dictionary.reference.com/search?q=callipygian

[davidmin @ Dec 19th 2005 2:26AM]

Wouldn't it be easier just to use a stereo film camera and scan it in? My wife has one from http://www.3dstereo.com/ . They're cheap and fun if a bit bulky. I might try this on some of her photos.

David

[gummih @ Dec 19th 2005 2:26AM]

have you tried printing your own glasses with clear A4 slides for an incjet printer? What exact colors would be best (exact red and cyan?)

[Ben @ Dec 19th 2005 2:26AM]

Is it really necessary to take two photos? Cant you just take one picture and crop off a little of the left side for one eye and crop off a little of the right side for the other eye?

[Jesse @ Dec 19th 2005 2:26AM]

Cropping the same photo differently wouldn't work--the 3D effect comes from the fact that you have two images of the same scene from different perspectives, so if there's a tree in the background next to a person's head in the foreground in one image, in the other image the person's head may be partly blocking that same tree because the camera's line-of-sight was a little different. These sorts of differences in what your two eyes see are how your brain interprets depth.

[Morps @ Dec 19th 2005 2:26AM]

Interesting that a thread started on this subject. A couple of weeks ago, I started pondering the same idea. I too had found the Mars Rover images and decided to create a full screen 360 degree panoramic of them. Of course, that immediately got me thinking about whether the same thing could be accomplished with one image. [link to QTVR pano: http://homepage.mac.com/morps/blogwavestudio/LH20040318210020/LHA20040811064545/index.html]

I started Googling around and really couldn't find anything. So I started mucking around in Photoshop. I created a Photoshop action that would create an anaglyph from a single image. It's not perfect, but it does work (see link at bottom of paragraph). After running the action, you do need to nudge the Channels a bit. But, for a poorman's anaglyph creator, I think it does a decent job. If any of you try it, lemme know what ya think. [Link to anaglyph sample and action: http://homepage.mac.com/morps/blogwavestudio/LH20040318210020/LHA20040818071633/index.html]

[William @ Dec 19th 2005 2:26AM]

"Humans over time have evolved with many capabilities which offer an advantage over non-humans". True, but binocular vision is not one of them. This capability is shared among humans and non-humans alike.

But the article itself is very interesting, nevertheless.

[dinoj @ Dec 19th 2005 2:26AM]

Interesting thread this, full of goodies I shall have to link to... Anyway, it is possible to put the raw stereo pairs from the Mars Rovers together to create near-true-color stereo pictures. A friend of mine who works at the Adler Planetarium in Chicago worked out a scheme for doing this, and we put them up at http://astro.uchicago.edu/cosmus/projects/marsstereo
You'll find the pictures there in both anaglyph (i.e. use red-cyan or red-blue glasses) and side-by-side format (a.k.a. twinview). The latter is used for displaying them using two projectors and polarized lenses (see http://www.geowall.org) and for printing on Holmes cards.

You'll also find stereo pictures from various observatories around the country at the same site, but they're in twinview format
http://astro.uchicago.edu/cosmus/projects.html#stereopix
These were taken with a single digital camera and a bar, or two digital cameras synchronized with a single click, and then aligned with cheap ($100) software called Pokescope (http://www.pokescope.com). You can align, crop, and rotate stereo pairs with Pokescope, as well as effortlessly switch between anaglyph and twinview. I'd be curious to know if it actually does more than Callipygian.

[Forrest O. @ Dec 19th 2005 2:26AM]

Want to display your stereo pairs without anaglyph glasses? Check out the Wiggle Stereoscopic Viewer: http://wiggle.sourceforge.net/ . Be sure to crawl around the samples directory to see more: http://wiggle.sourceforge.net/samples/ . The download includes an html page that aligns your images and outputs the required html to embed the flash applet.

[Chris Lowery @ Dec 19th 2005 2:26AM]

The time for space trick was demonstrated by Jim Gasperini in animated GIF...
http://www.well.com/user/jimg/stereo/stereo_list2.html

Using Flash to transition between them was my idea, see Old Stone Gate on that page.

Wiggle is very cool but doesn't follow the "duh" convention of using mouseleft-mouseright to move the 'camera' and blend the images. The image should oscillate unless the user mouses into it.

When using the Flash technique, it's best to take 3 pictures (left, right, center) and blend between those. Unless you also want to do up-down.

Come on Wiggle!

-chris

[Chris Lowery @ Dec 19th 2005 2:26AM]

I take it back, it does support mouse look, in a 'mode' you can enter by clicking on the image. the 'focal length' idea is very useful. I'd like to see this UI for Wiggle:

up-down = focal length
left-right = mouse look
click = pause/unpause
double-click = modes/menu

Optimised color anaglyph is terrific, but most people don't have red-blue cardboard 3D glasses. It's almost an icon of geekness. For intuitive 3D impact, a stronger Wiggle can change the web.

[Allan Silliphant @ Dec 19th 2005 2:26AM]

In recent years the majority of 3D glasses have
shifted from RED/BLUE to RED/CYAN, which facilitates seeing the full color spectrum better. This greatly improves skin tones. Paper
glasses have several serious shortcomings, the
obvious is that they get mangled in a few days!
They can't be cleaned easily, but less well
known is the fact that the red filter tends to
shift focus perception, due to it's different place in the spectrum. (Wave length in nm's)
With plastic glasses, it is possible to slightly correct this stressful mismatch with
a minimal diopter correction. When this is done the eyes of the average person can relax and
view for extended periods in comfort. A popular
line of "professional grade" plastic glasses,
called Anachrome provides glasses, so enhanced.
They also have a set of processing and shooting
protocols which help to make better color
anaglyphs. A key feature of these is the
advice to allways use less than human eye
seperation, (stereo base) in shooting or rendering CGI for stereo. The norm is about 1/2
the typical. This allows the overlaid images to partially "self-mask". In other words only a
very few pixels are showing the "carrier color"
which makes the pesky red outlines on typical anaglyphs. There are hundreds of image examples
on the website for these glasses and technique.
which is www.anachrome.com. Anachrome images
are so closely overlaid that most can be posted small on line as a regular color picture. When
a viewer wants to see the 3D detail better they
simply open the image to full size. There is a
directory of other sights that use the method in varying degrees with links. The site also
provides some tips and instructions toward
making broader spectrum, more saturadated
color anaglyphs with a series of corrective steps. To some extent, these more colorful
type of anaglyphs are being refered to as
"3D Anachromes", or A3D images.

[Robert Swirsky @ Dec 19th 2005 2:26AM]

Thanks for featuring my "Callipygian 3D" software.

I'm working on a MUCH IMPROVED version, and I'd welcome any suggestions you may have.

[Alan Eliasen @ Dec 19th 2005 2:26AM]

It's been my hobby to take 3-D photos for years and years. Anaglyph is okay, and one of the few techniques that's easy to post on the web, (although I'm the only person I know with the required glasses,) but I vastly prefer using plain old photographs and a viewer like the ViewMagic, which can be obtained from a few places, like: http://www.berezin.com/3d/ViewMagic.htm

These have excellent optics (front-surface mirrors,) and are simple, robust, and easy to use. I've had mine since you bought them direct from the original inventor.

All you do to make 3-D pictures is to take two photos from different angles with any camera, and then place one photograph 4 inches above the other (I tape them to a piece of paper, or just set 'em on a table.) You get stunning, photographic-quality images *in full color.*

Nothing adds interest and depth to your pictures like 3-D does. Details you'd never notice pop out at you. And you can produce effects you'd never be able to see with the naked eye. I have pictures of Mt. St. Helens taken from about 5 miles away. I moved at least 100 yards between taking pictures, and you can see the whole 3-D shape of the mountain and the crater. Really amazing stuff, and recommended for anyone who is a photo nut.

[amogh @ Dec 19th 2005 2:26AM]

good post.

[amogh @ Dec 19th 2005 2:26AM]

nice post.

[Douglas Hoyt @ Dec 19th 2005 2:26AM]

Another possibilty to consider is Picture Window Pro 3.5 (http://www.dl-c.com/Temp/) where they describe their capabilities as:
"Picture Window's stereo transformation lets you prepare pairs of stereo photographs for viewing, either with red/blue colored glasses or with a standard side-by-side stereo viewer. The horizontal and vertical offsets between the two images can be adjusted for the best viewing while previewing the results on the screen."

P.S. I have taken stereo images, mostly with a Stereo Realist camera, since around 1970. You get spectacular images if you compose the picture correctly.

[nate lucas @ Dec 19th 2005 2:26AM]

I'm attempting this process with Photoshop and I am having no success. I think I lose it when I try to paste the right image onto the left/blue image. My image just goes back to being a grayscale image. No blue and/or red. What am I doing wrong?

[elena @ Dec 19th 2005 2:26AM]

wallawallawalla

[Allan Silliphant @ Dec 19th 2005 2:26AM]

Here is a simple method to do color anaglyphs in Photo Shop 6.5:
Open the two oictures side by side. Look at them
with eyes in CROSS-Viewing mode. If they don't form a comfortable stereo image, reverse their positions. Once you can see stereo, change each with AUTO LEVEL FUNCTION. 95% of the time the image will then look better. Next check the
RELATIVE MAGNIFICATION of the two images by over-laying them in LAYERS.(this is in the Windows column) In the pull-down menu "normal"
will be the default, use p.down menu to change to SCREEN. (This only works when two images are actually laying one on the other. If only one layer SCREEN cannot be selected.)
Move the right hand image over the left and try to find any difference in SIZE or ROTATION.
Correct the difference by altering the upper
image back at its source. You can both rotate
a few tens of a degree or more, and change size.
By trial and error get them to overlay as good as possible. This will greatly reduce ghosting later. Now go to history and remove the overlay.
Now, use the CURVES function in the Image/adjustment column. Remove all RED from the image on your left (which is actually the RIGHT camera view) then using CURVES remove GREEN, then BLUE from the image on your RIGHT
(which is ACTUALLY the LEFT camera image).
Slide the CYAN image onto the RED image. Move it
around until the MAIN SUBJECT is in register.
right to the pixel ! (checked by zooming in)
at the contrastiest line of demarkation you see.
The go over to the LAYER column and go down to
FLATTEN IMAGE. DO IT. then Crop! Then add
a little extra chroma to countact the color loses filters induce. That is a basic Anachrome,
done to level one standards.
Allan Silliphant
Anachrome 3D Group

[Allan Silliphant @ Dec 19th 2005 2:26AM]

Here is a simple method to do color anaglyphs in Photo Shop 6.5 or 7.0 :
Open the two pictures side by side. Look at them
with eyes in CROSS-Viewing mode. If they don't form a comfortable stereo image, reverse their positions. Once you can see stereo, change each with AUTO LEVEL FUNCTION. (95% of the time the image will then look better). Next, check the
RELATIVE MAGNIFICATION of the two images by over-laying them in LAYERS.(this is in the Windows column) In the pull-down menu "normal"
will be the default, use p.down menu to change to SCREEN. (This only works when two images are actually laying one on the other. If only one layer, SCREEN cannot be selected.)
Move the right hand image over the left and try to find any difference in SIZE or ROTATION.
Correct the difference by altering the upper
image back at its source. You can both rotate
a few tens of a degree or more, and change size.
By trial and error get them to overlay as good as possible. This will greatly reduce ghosting later. Now, go to history and remove the overlay.
Now, use the CURVES function in the Image/adjustment column. Remove all RED from the image on your left (which is actually the RIGHT camera view) then using CURVES remove GREEN, then BLUE from the image on your RIGHT
(which is ACTUALLY the LEFT camera image).
Slide the CYAN image onto the RED image. Move it
around until the MAIN SUBJECT is in register,
right to the pixel ! (checked by zooming in)
at the contrastiest line of demarkation you see.
Then, go over to the LAYER column and go down to
FLATTEN IMAGE. DO IT. then Crop! Then add
a little extra chroma to countact the color losses, filters induce. That is a basic ANACHROME, digital anaglyph,
done to LEVEL ONE standards.
Allan Silliphant
Anachrome 3D Group

[Maddy @ Dec 19th 2005 2:26AM]

I believe that those "wiggling" gif or flash images are a good way to simply PREVIEW the stereo effect. They allow viewers to see the relation between near and far objects without having to use 3D glasses or perform eyeball tricks. I would not go so far as to call them actual stereo images. They are only previews, presented as flickering monoscopic images. For true stereoscopic effect, both left and right images need to be presented simultaneously to both eyes; or at least fast enough that the brain can merge the two images. Let's not confuse the public any more by trying to convince them that "wiggles" or "t-f-s" are stereo images - they are only previews.

[romulus @ Dec 19th 2005 2:26AM]

cross-eyed or parallel (i prefer the latter after years of "magic eye" stereograms, others prefer cross-eyed) images are arguably the easiest to make. take two pictures, then copy them into either side of the same image. or skip that step and put them next to each other on a web page.

i've done parallel 3-D photography of a still subject (a messy room) with a *film* camera, simply stand as still as possible and take a picture with either eye, trying to keep the camera as level as possible each time.

though this method is perhaps not as solid, it does address the "human focusing" problem that twobits mentioned earlier.

dunno why you would need special software to create anaglylphs. why can't you use Photoshop's channels feature? basically, activate only the red channel and paste in the left image, then activate only the blue and green channels and paste in the right channel. with all channels active, the result should be an anaglyph. (an inverse of this method is also useful for convering anaglyphs back into two images for use as parallel/cross-eyed pairs.)

[Ana @ Dec 19th 2005 2:26AM]

I had my comic book glasses ready and must say these look really good. The robot in the leaves is kinda cute ;-). Ana

[Benny @ Dec 19th 2005 2:26AM]

Dude you have like every cool gadget and gizmo that exists! 2-3 Aibo's ... that 80s robot, a segway, AND a roomba!!! WHOOSH!!!!!

You live in the future! All I have is this lousy G5 :)

Worldpeace,
Ben

[Zvi Oren @ Dec 19th 2005 2:26AM]

Do you know where can we purchase 3D posters?

Zvi

[JORGE LUIZ @ Dec 19th 2005 2:26AM]

Photos 3D

[Vin @ Dec 19th 2005 2:26AM]

does anyone know how to make 3D illustrations?

[Sarah @ Dec 19th 2005 2:26AM]

I found this whole gallery of 3D images that were made from a train window. http://www.hardtoremember.org/train/index.php?mode=anaglyph

[Lippo lippi @ Dec 19th 2005 2:26AM]

RE: Do you know where can we purchase 3D posters?
here:
http://www.gengotti.it