Author Options:

Could a prism and lenses be used to make a stereoscopic video with only one camera? Answered

I want to make my own 3D video and I only have one camera. could I make a camera attachment in a "T" shape and use a prism to split the camera's focal point to the left and right then have each direction hit a 45 degree mirror to turn the focus back to the front of the camera and set a filter on each side of attachment that will correspond to filters on lenses of 3D glasses, and have this make a stereoscopic video with one camera? Could anybody who has experience with prisms and lenses verify if this would work before I start building a prototype...


This is rather old but I was looking for something like this. I would like to make something like this for a Raspberry PI 2 with a camera. Here are some more interesting links to different setups using mirrors:



here is a video of just useing a prism

This is a phenominal topic and I really like the berezin website that CaptGreg mentioned. This is something I have done with multiple cameras.

Absolutely.  Lens replacement systems which generate 3D stereo exactly as your diagram indicates, are available for more DSLR cameras.  See http://www.berezin.com/3d/3d_lens.htm for standard lens or http://www.berezin.com/3d/macro3d.htm for the macro version.  They replace the lens and have a a built in lens focusing optical system.

For 3D stereo movies, you could see if you can find a camera which will take the Berezin 3D lens system or make up your own from parts from Edmund Scientific or other optical suppliers. 

Try asking Berezin if they have 3D adaptors  for any video cameras,

I would start by fooling around with the prisms from a pair of dollar store binocular.  Crazy glue the prisms back to back.  See what you can register on the camera lens or remove the lens and see what registers on the CCD imaging chip inside your camera.

Now you will need a 3D stereo display system that will process the video footage and separates the video streams into whatever format your stereo display system requires.  Most modern GPU's have a 3D stereo jack on the card which will drive the IR/Ultrasonic transmitter used with active shutter glasses.

Have you considered using a laser on a Holographic table? the single beam is bounced around mirrors after being split with a prism. One beam illuminates the object by going through a lens and spreading the coherent light on to the object to be reflected on the front of the film while the other beam is sent to the back of the film through a lens to illuminate the back of the film and set-up a defraction pattern. Find a way to do this with ordinary light.

Yes, it will work. But not very well. Commercial units are available that do just that. It will take some tinkering with the focal length of the lens to prevent a dark circle around the outer corners of the frame. The big disappointment in most of the units I've seen is the fuzzy band in the middle between the two frames. There isn't a clean separation like two images places side-by-side in photoshop. But if you've got the parts, go for it! It will be enjoyable and very educational for you, even if you later move to another process. As for color filtering, I'd leave that to post-production.

Sterescopic vision requires that slightly different images go to the two eyes (a retinal disparity). Since these images could both be viewed by the same eye, it won't work--you'll just get a double image. It would be like viewing an anaglyph (red/green image) without the glasses. The red lens makes the red image disappear; the green makes the green disappear, thus producing a disparity. It might work if you could film with oppositely polarized lenses and view with similar lenses--not sure.

light is broken upwards when it enters the prism. then it wont be reflected down but with an angle. i dont think it will reflect the way you want think of asymmetrical solutions (where one color goes straight thru and one is to the side) or use half transparent mirror instead of the prism

I like that Idea of using one-way mirrors my only thought of that would be that the mirrors might cause a ghostly haze if it starts reflecting an image of the camera's lens itself...

it may happen with any optics. i think you can minimize this if you enclose the entire system in tunnel with rough black walls

I think CameronSS's comment "The problem with this setup is that both images are recorded on the same film/CMOS/CCD" would begin to apply to your device if you switched to polarized filters.

Ordinary cameras are designed to record red, blue, and green on separate channels (and sometimes a few other colors). So, as you described it originally, red and blue images falling on the same part of the sensor can be distinguished.

But ordinary cameras are not designed to record different polarizations on separate channels, so the two images differing only by their polarization falling on the same part of the sensor would just be overlapped/overexposed with no way to separate them again.

As he said, you would need to devote an area of sensor for each image, or take the images at two different times. The IMAX 3D camera is basically two cameras operating simultaneously, side-by-side.

I had read this in the past but forgot about it. Thanks for bringing it to my attention again. This gives me a great starting point.

The problem with this setup is that both images are recorded on the same film/CMOS/CCD. A 3D image is created when two images are recorded of the same scene, each slightly different. The only way that this could work is if it split the image from each lens and put it on each half of the sensor, or perhaps taking two pictures in rapid succession, one with each lens covered.

SWV1787 said he's using anaglyph filters on each lens, so only the green sensors are getting light from both lenses. (Or from neither lens, if the filters are carefully selected.)