HOW WOULD YOU IMPLEMENT THIS? Stereoscopic Rendering (3D)

#1

Steroscopic Rendering

From wikipedia:

Stereoscopy, stereoscopic imaging or 3-D (three-dimensional) imaging is any technique capable of recording three-dimensional visual information or creating the illusion of depth in an image. The illusion of depth in a photograph, movie, or other two-dimensional image is created by presenting a slightly different image to each eye.

After the success of the 3D release of Beowulf there seems to be a renewed interest in producing 3d content in 3D.

Since this a new area for most 3D users, I seems a ripe opportunity for Lightwave.

[ul]
[li]How would develop tool sin Lightwave to facilitate this?[/li][li]How would you select the proper focus point? How would you implement the Depth of field in 3d?[/li][/ul]
Looking forward to your comments.

-R

#2

There already in LW and work quite well…we’ve been using them at the school for movies for quite some time. If your ever in Florida come by and check it out!

#3

Well, I dont see the problem of producing the frames but getting them to run properly on a standard PC monitor. Can Real-D movies be played back on a PC?

#4

I recall seeing a very short 3D stereoscopic video created many years back in Lightwave showing off the rare “eagle in a barrel”.

Anyone know a good source for decent quality stereoscopic glasses that aren’t the typical
paper crafted construction?

#5

I’ll see if I can find out where we get ours from

#6

There are multiple techniques for stereoscopic display.
Some involve two projectors using polarization filters and glasses with opposite filters.
Some involve shutter glasses and the one which is worst quality wise uses red and green glasses.
Most of them are already working rather nicely in LW, as William said.
In my experience though the eye separation allone can sometimes be problematic. It can be good to have a focus for the two eyes to center on. This is something that LW seems to be missing at the moment.
One can work arround that though by simply parenting two cameras to a NULL with a NULL as a target and rendering them separately, then performing the rest in post (depending on wat the technique needs).
CU
Elmar

#7

The focusing of the eyes on a single point in space is called convergence, and you want to make sure that it occurs as far away as possible. If you converge on a subject that’s reasonably close everything behind it will have cross-convergence. Viewing a 3D scene like this will lead you to the quickest headache you’ll ever get. Don’t get me wrong, you do want convergence and the closer you can get it the better. That’s why interior scenes work well for stereoscopy. For outdoor scenes you just need to set your convergence extremely far away. You can cheat by using a shallow depth of field to make everything in the background (behind the point of convergence) out-of-focus.

#8

Have not worked with it, but having a focus is a must as far as I can see.
Setting up the targeting and parenting like you explained is easy though.

Wasn’t there a solution that doesn’t need glasses too?
can’t find that link… :confused:

#9

Lightwave has had stereoscopic rendering and tools for convergence since 1997. I had requested from Allen Hastings to develop tools for a 3D animated film that we produced at Vinton Studios. Stereo works best with ortho or parallel cameras. Converging on a point is the biggest problem with eyestrain. However somtimes you want to create a 3D scene that has a convergence point set to push the 3D more. In this case, all you have to to is target Lightwave’s camera to a null and set that null to where you want the convergence to occur.

When you enable the stereoscopic rendering tab under the camera panel, Lightwave will offset each L/R pair of images 1/2 the eye sepeartion value from center. If you have the camera orient costrained to a target (convergence) each camera will toe-in during render. However you have to wathc for this. Don’t oover due this toe-in because it can cause severe distorions in background and cause the viewer’s eyes to go wall-eyed which is very uncomfortable.

#10

to develop tools for a 3D animated film that we produced at Vinton Studios

You mean “Will” Vinton studios? I’m a bit late as I only found out about the studio at Christmas when I watched Return to Oz on TV. The animation was simply brilliant and I’m surprised it didnt win the VFX Oscar for 1985. :shrug:

Oh well, back to the topic at hand I suppose…

#11

Ah, :slight_smile: Found a link at Danimator’s page.
http://3dfool.blogspot.com/2008/01/3d-television.html

#12

Yes thats the same Will Vinton. They were bought out by Phil knight of Nike, and renamed LAIKA. They are doing a stop motion 3D feature right now called “Coraline”

#13

Yes, I know that you will use it rarely, but it is sometimes needed. I remember us doing this a lot here at a 3ds- MAX production house, where we did not have any built in tools for this back in the days (its been about 8 years now).
Personally I havent done any stereoscopic stuff since then, but I have checked the tools in LW and they work sufficiently well. As I and Danimator said, you can do this in LW with several techniques as well. I was not aware of a convergence setting. I only know of the workarround with targetting the Null. Danimator, can you point me to the location of that setting?
CU
Elmar

#14

There is no setting for convergence, its either on or off depending if you have the camera targeted to a null. You will also need to disable the Pitch and Bank on the camera as convergance should only affect the heading. Its also a good idea to put you camera into a rig that will allow the movement to be speratly controled on other nulls. I like to build a vitural dolly and crane with my setups. This will get a more realistic camera in the end.

#15

Thanks Daniel, that explains why I have never seen it :wink:
CU
Elmar

#16

Wondering, anyone knows if glasses-free 3D televisions work the same way, so the LW stereoscopic-method can be used?

http://www.youtube.com/watch?v=M-TPxytihwI

#17

A stereo cam setup is not that difficult to set up.
1 Null with as childs a left eye cam and a right eye cam, with 8cm inbetween.

Having 2 cameras means you have to render 2 different scenes. If you use the stereo
function from LW it will render 2 eyes in 1 go.

  • P -
#18

The thing about 3D is you have to seperate MOST viewing methods from the generation method. Humans have two eyes, we see the world 3 Dimensionally from birth. two renders that replicate that view (in scale) is all you need. Once you have those two images, you can set them up to view in almost any format including glasses free 3D. Many of those systems will work better with more than one stereo pair, but can do amazing things with just one pair.

Remember, LW stereo rendering is workable for any system be it Print, anaglyph, porlarized, real-d, dolby 3d, IMAX 3D, Sensio, Feild Sequntial, Pulfrich, etc. There is no new stereo technology. Stereo hasn’t changed since the civil war when the photgraphed battelfileds in 3D. We still have two eyes and thats all you need to do is recreate that. The only thing that has improved is the method of delivery (projection) and the knowledge of how to set it up.

#19

Thanks,
that basically answered my question :slight_smile:

It’d be cool to test this some day

Home-made 3D
http://www.scec.org/geowall/

#20

http://www.stereoscopy.com/reel3d/polarizing-glasses.html