Anaglyphs, 2d to 3d

[tonyk]

I would welcome any advice on conversion of 2d images into 3d anaglyphs.
I am trying to speed up the process of cloning the 'holes'.
Anybody with any expertise or advice in this area would be appreciated
Tony

------------------

Moderator note:

Also see this thread: http://www.retouchpro.com/forums/pho...ereograms.html

[DannyRaphael]

Hi Tony:

Gotta confess I'm not familiar with the term anaglyph. Would you mind posting a link or two to a 2d that's been converted to a 3d. That might stir some braincells.

Thx/Danny

[tonyk]

Danny
Thanks your response
Anaglyphs are images formed from two shots taken of the same scene where the camera position mimics the view from each eye position (left eye, right eye). The red channel from one image is combined from the blue/green channel of the other. This combined image is viewed with red/cyan glasses such that the brain is fooled into seeing the image in 3d.
By taking any single 2d image it is possible to mimic this effect by duplicating the 2d image and manipulating this copy to createa left eye and right eye pair. The two images combined as above etc
Sorry but I know ive not explained it well.
A good explanation of the 2d to 3d method can be found on http://www.dddesign.com/temp/nsa02/2dto3d.htm

Regards
Tony

[Fluffbutt]

Are you working from two images, left and right, or are you trying to get a 3d from a single image?

If left and right (I use two single use camera glued together - right spaceing, press shutter button at same time).

All you need to do is open both images, and copy the red channel from the right image and paste it overwriting the red channel of the left image. (oh ahh or is it red from left into right image? oh well you'll see the very first time you do it...)

Edit> "any single 2d image" opps, me guilty of not reading properly

I went to that url and as far as I can see there's no real way of automating this process - it requires a creative brain to decide just what should be forward and what should be backward of the main image plane.

For eg, the train requires more work than the horse (ass?) would....

[DannyRaphael]

Tony:

After reading the additional info and checking out the link I knew what you meant, having donned the red glasses in the past to see 3-D horror film epics of the 60s.

As Fluff said, the selections, masking, etc. would have to be done manually.

Depending on the steps it looks like at least parts of the process could be automated, e.g., opening a file, saving as L and R versions; if you got into the channel swapping part, those steps could be automated too.

Anyway, are you familiar with Photoshop actions?

- - - - -

Fluff:

Great idea on the assembling the disposable cameras and doing the simultaneous snap!

~Danny~

[gina]

Hi,

The anaglyph is only one of the methods of getting to see 3d image information.
For real in depth information you should start of with two images. For example the trick with two inexpensive camera’s glued together.
At http://www.starosta.com/3dshowcase/ihelp.html
you’ll find the method to see such stereopictures without any glasses or other parafarnalia. The bonus is it works for color pictures too.
If you only want the anaglyph look I have some action that just uses the images in the red and green channel and shifts them a little. I tried it but it didn’t make much difference to me. Mail me if you want it. I’ve tried this approach at work as some time ago my boss asked me to try to make the anaglyphs for visualising the surface of an etched aluminum surface. Once we bought an image program facet stereo that did just that but we couldn’t find it at that time. So someone gave me the tip to use corel pp. I used corel photopaint8 and PSP 5.03 and this works: (mind you I had made the two stereo images before, so I started with two slightly different photo’s).
- One image is colorized red.
- the second one is colorized cyan, blue or green, it depends on the color of your 3d glasses.
- I make a copy of the red image, make a new layer and paste the blue, cyan or green image on top. Change the blending mode from the green layer to overlay. You should see an anaglyph image then. The details for the right eye should be shifted to the right and no more than 0.5 cm or our eyes get to confused.
You could try this same method with one image if you like and shift the top layer yourself. I don’t know if you get the really depthty information you get with real stereophoto’s because the image is not only shifted but there is an angle difference too.
Hope it was a little helpful and not too confusing (I’m really bad at explaining)


bye

ps I try posting the action.I think i got it from the adobe exchange

[gina]

Hi,


Retry with the action after a tip from Danny

gina

[Mig]

Read pages 344-345 of Deke McClelland's Photoshop5 Bible for the complicated answer. You can probably find it in later versions of this book, which you'll find anywhere. No need to buy the book, just read how it's done.
For now, eaiser way, experiment by dragging or nudging the red channel over to the left. Put your funny glasses on as you do it. Have fun.

M

[tonyk]

Thanks for all the info on anaglyphs. Just got back to my computer after a week away...withdrawal symptoms terrible..
My problem with creating 2d to 3d is trying to avoid cloning of the 'holes' created with shifting areas of of the duplicated image. I dont think it can be avoided so means slow and tedious work
and Im too impatient to do this.
Actions speed things up considerably and I thank everyone for their help
Tony