Someone please explain how this is possible...

[Evin Grant]

Apparently Sony has built a prototype camera that can do single lens 3D.
I really don't understand how this works so please enlighten me, I though you needed interocular distance to get 3D how do you get a stereo image with one lens, even if you do split the image circle in half?

Full press release...
http://photorumors.com/2009/10/01/3d-is-in-your-future/#more-1858

 

[Mark Pugh]

take a look down an optical viewfinder while moving your head around a bit. (Focussed on something in the foreground)
It's three dimensional - you can see around the foreground object...

 

[Noah Kadner]

Yeah I didn't get this either. I'm guessing the 3D effect is less convincing than a two camera setup. Though if it worked it would be genius.

Noah

 

[Jacek Zakowicz]

Apparently Sony has built a prototype camera that can do single lens 3D.
I really don't understand how this works so please enlighten me, I though you needed interocular distance to get 3D how do you get a stereo image with one lens, even if you do split the image circle in half?

Full press release...
http://photorumors.com/2009/10/01/3d-is-in-your-future/#more-1858

Pretty ingenious...
The vertical line where the mirror arrows are pointing is the interocular.
Now just add dual anamorphic adapter on the back, square ccds and voila cinematic 3d

 

[Mark Pugh]

Though if it worked it would be genius.

Noah

... and a hell of a lot easier to shoot.

 

[M Most]

Apparently Sony has built a prototype camera that can do single lens 3D.
I really don't understand how this works so please enlighten me, I though you needed interocular distance to get 3D how do you get a stereo image with one lens, even if you do split the image circle in half?

Think of it this way. When you look at the world through two eyes, you're not looking at two different images, you're looking at one image from two slightly different perspectives. When you look through a lens, you're still looking at the same image. If you align your eye with the exact center of that lens - as a typical imaging plane of a sensor or a film gate does - you get one perspective. But if you move your head a bit, you get a slightly different perspective, based on the direction of your head move. You can see this by looking down the barrel of an optical viewfinder. Move your head to the left and right and the perspective shifts. What Sony seems to be doing is taking advantage of this by moving the two "cameras" behind one main lens rather than have two discrete taking lenses. The interocular distance is controlled by the mirrors that reflect these two slightly different perspectives into two separate imagers. My guess, based on the diagram, is that the two center mirrors can be physically moved (pivoting around the common axis), thus shifting the viewing axis and changing the apparent interocular distance accordingly. The limitation, obviously, would be in the maximum interocular that can be achieved, limiting its effectiveness somewhat for very long, wide shots that require large interocular to present a true 3D efffect. Another limitation would likely be that neither "off axis" view would be at the optimal focus point for the main lens, and there would likely be distortion issues to overcome, but I could be wrong about that, and Sony's engineers are pretty clever people.

 

[D Fuller]

Would this be why large format cameras seem to have a qualitative difference in the way they draw images than smaller-format lenses?

 

[Joseph Ward]

I would guess it would have more to do with the angles of mirrors and software algorithms. It has two sensors which is good, some cameras use one sensor with whatever 3D Lens attached.

http://gizmodo.com/308659/adobe-tin...sing-camera-and-software-[updated-with-video]

 

[Kwan Khan]

Very Smart...

 

[Pawel Achtel]

Isn't the FOV halved and the interocular distance limited to half of the front lens diameter?

 

[Jeff Kilgroe]

Pretty much. But that's where the anamorphics come into play, as Jacek mentions.

IMO, I would much rather shoot dual-lens 3D and the more I think about it, the EPIC 617 looks like the perfect camera body for the job. I'd elaborate if I wasn't on my way out the door...

 

[Dylan Macleod CSC]

By having a single taking lens don't you diminish the 3D effect? You are not getting the same slightly different perspective from the offset of each eye? Isn't this why 3D is shot with two cameras (or two lenses)?

Seems like the 3D effect is not going to be as effective with this method.

 

[Pedro Guimaraes]

It's actually a pretty old concept....thing about 3D is that it's over 100 years old and people have tried hundreds of ways of achieving 3D. Some that now are more viable in the digital age.

This kind of stereoscopic relay system has been used alot in stereoscopic microscopes before.

The Relay lens does have light loss. The front surface mirrors using the the mirror box section do not loose any light since they are only reflecting the light hitting them after the relay lens. Lens selection and focal length are also limited by the size of the mirror box and corresponding mirrors.

John Rupkalvis and I developed a similar system using a single Phantom65 camera last year. Unfortunately I can't release pictures or give more details. We also have single camera RED One prototype system in the works. No light loss, single camera 3D system. To me it's the holy grail.......dual camera systems all have issues.
I'm Tired of steadicaming beam splitters and heavy parallel rigs. ; )

This sony camera should work very well......The inter-ocular is limited by the size of the mirrors and mirror box. Should be great for interior shots etc....

we will all follow these developments closely. Shooting with 2 cameras and beam splitters can be very limiting and very problematic.

Purpose built 3D cameras will be the ultimate answer for 3D production. I'm very happy sony and panasonic are both exploring stereo cameras seriously....

A few more links...same release I think
http://www.televisionbroadcast.com/article/88078

http://hdtv.biz-news.com/news/en_US/2009/10/01/0006/sony-develops-240fps-single-lens-3d-camera

 

[Pedro Guimaraes]

Yeah I didn't get this either. I'm guessing the 3D effect is less convincing than a two camera setup. Though if it worked it would be genius.

Noah

It does work, and the 3D effect is the same as a 2 camera system

 

[Erik Widding]

The inter-ocular is limited by the size of the mirrors and mirror box.

May be so. But you are still limited to an inter-ocular distance less than the diameter of the front element.

 

[Dylan Macleod CSC]

It does work, and the 3D effect is the same as a 2 camera system

I'd have to see it to believe that it could be the same as a 2 camera rig.

Hold a CD box centered in front of your eyes vertically with the end facing you. Close one eye and you see one face of the side of the box. Close the other and you the opposite side of the box.

A single taking lens is only going to see the skinny "end" of the CD box.

There is less information to create the illusion of depth, no?

 

[Sven Seynaeve]

In some way then , and IMHO , it should be better if they come up with system supporting 2 lenses, where you could change the ID , but using all this with a single camera system having 2 sensors and recording a stereo stream of data.

 

[Michael Lindsay]

John Rupkalvis and I developed a similar system using a single Phantom65 camera last year. Unfortunately I can't release pictures or give more details. We also have single camera RED One prototype system in the works. No light loss, single camera 3D system. To me it's the holy grail.......dual camera systems all have issues.

I heard about this at IBC and got excited.. When can we hear more?

regards

Michael L

 

[Pedro Guimaraes]

I'd have to see it to believe that it could be the same as a 2 camera rig.

I have Footage to prove it :smile: both RED and phantom65. We shot a few tests for some big projects.....Unfortunately, my lips are sealed with NDA's. In fact we even went as far as a IMAX film out. In the film out were other footage of various other 3D systems including film systems.

I have to run now, but maybe I can explain more later. I'm running mad doing some last min. testing as I'm leaving to location this weekend. I'm DIT and stereo assist on a 3D feature shooting on REDs (beamsplitters).

 

[Mikael Lubtchansky]

Usually when shooting 3D you'd like to adjust the distance between the 2 images to increase of decrease the 3D effect, depending on subject, lens used...

Indoor with some foreground elements can have a quite close setup (approaching the distance between our 2 eyes).

Landscape shot with no or further foreground will need a greater distance between the 2 cameras to get a faked but more effective 3D.

Any special FX, macro, or just a close-up may require adjusting distance too.

I imagine this kind of camera will produce a pretty good 3D effects for any interior scene /news etc but would lack flexibility for subtle 3D adjustment in a $$ feature film shoot. Could very well help some people create more 3D content in a quick and easy way while other will experiment with modular DSMC RAW stereo rigs ;-)