Welcome to our community

Be a part of something great, join today!

Register Log in
  • Hey all, just changed over the backend after 15 years I figured time to give it a bit of an update, its probably gonna be a bit weird for most of you and i am sure there is a few bugs to work out but it should kinda work the same as before... hopefully :)

3D Rig

I

IanK

Well-known member
Joined
Nov 22, 2007
Messages
212
Reaction score
0
Points
16
Hello,


I am about to shoot something in 3D with two Reds or possibly 2 Canon 5D Mark IIs. I did several tests with two HVXs recording parallel to each other and came up with unsuccessful results. The two lenses were fairly close but alas, it did not work... I noticed many pictures of people shooting Reds side by side? How is this possible? Even if the two bodies were touching would the distances of the lenses still be too far away? I guess what I am asking is how close to the two lenses need to be in order to get a proper stereoscopic image (if the subject were about 20 ft away) on an 18mm or so? I have seen the PS Technik rig but I would love to get away with not renting a beamsplitter rig. Thanks.


Ian
 
Side by side shooting on large format cameras is only really appropriate for landscapes and wides. Close up shooting requires greater convergence, and as such smaller cameras or beam splitter rigs.
 
Hey Ian,

Just to offer a slightly different perspective. There are a lot of stereoscopic rules of thumb that are bandied about, but in reality there are some situations where fairly wide “base separations” can be used. It really depends on the total depth of the scene that you want to capture as well as the basic near and far distances of that depth range. For example in architectural stereo photogrammetry using very precise set ups and precision wide angle lenses it is possible to shoot a façade of an ancient ruin with the cameras spaced nearly a meter apart (800mm) and only eight to ten meters (25ft) from the façade. The range of depth variation of the subject from a 25ft shooting distance may only be about four feet. In such a case it is possible to record and capture extremely fine three dimensional detail of the surfaces and produce stereograms that are easy on the eye. This is because the range of depth of the subject is fairly shallow. However if you stood a group of actors in front of this façade there is no way you could use a 800 mm (inter-ocular) base separation. In such a case, for special technical applications (i.e you want to extract the 3d geometry from the stereogram) it would be conceivable that you could use a base separation of 200-300 mm, because the façade is opaque i.e. the farthest visible distance is about 25ft. However if there was a large arch through the facade where you could see trees and grass in the distance, then there would be no way you could use 200 mm base separation. What would happen when you set up your horizontal image separation/translation HIT (on screen in “post” or viewed on site) the figures and the façade would be well resolved but the trees and grass would not be stereo resolvable and would appear as a very distracting double images. Conversely if you adjusted the HIT so that the trees were stereoscopically resolvable (by the eye and brain) the figures and the foreground would be a painful double image. Therefore for cinematic and (non technical VFX) a much smaller base separation is required to capture a large range of depth. The problem here is that subjects appear as flat cardboard cut-outs disposed at various depths. I think the key to good stereo is to have the ability to control the total range of depth, so for example if you are shooting human figures against a green screen or solid background you could certainly get away with side by side shooting although the figure might have a slight hyper stereo effect. However you would not be able to do close up on a face using side by side RED ONEs for example, a different camera or mirror/splitter rig would be essential. I would personally like see more use of side by side rigs and splitter rigs only be used for more close up applications, that way the size of the splitter rigs could be kept to a minimum. Side by side can be very good if you want to capture very large sites at a distance, base separations of the order of several meters can be used to good effect as well. We have also developed special (software) techniques that enable one to shoot at wider base separations that are easy on the eye but don’t produce the cardboard cut-out effect, but instead preserve very fine 3d surface detail (but I won’t bore you with that); but I think side by side shooting will become more common place in the future.

Ian, how were you planning on synching up your two canon 5D mark IIs? Just interested if this is possible?

Cheers
Eric
 
The HVX tests probably didn't work because they weren't genlocked. (This is very important.) It is possible to genlock HVX 200 but you have to open up the camera and do some "hacking".

Also you do realize that most cameras (including the RED) have a slight misalignment of the sensor within the body. Which means when you put 2 cameras on a flat plate you will have vertical disparity. Of course you can correct this in post but you start losing pixels/resolution.

But It sounds like you are an enthusiastic neophite to 3D and that's great, but you should really should do some research about stereoscopy and learn more about the craft. Getting proper 3D as you found is not as simple as just putting 2 cameras side by side. There are many considerations. Focal length, stereo base, convergence, stereo window violations, ocluded material, miniturization, giantisim etc...etc.....There are many great books about stereoscopic theory/photography/filmaking at your local library. 3D has been around for over 100 years and there is alot of knowledge out there to learn. Obviously there is alot on the internet as well.

3D is very subjective and many experts will have theories about the "right" way to do things. In the end you have to learn what those are and then experience them for yourself.

-Learn the Rules
-Break the rules
-Develop the rules for yourself and your camera system.


One of the biggest things that has hindered 3D success in history was badly shot material and viewers then having a bad experience with it and swearing off 3D as a form of enjoyable entertainment.

So the way I see it it is my responsability as a filmaker/cameraman/Stereographer to shoot proper 3D that will result in a positive expirience for the viewers.

So I hope you will continue to seek more knowledge about stereoscopy and it's creation.

If it was easy everyone would of been doing it.......

A lot of new production tools/technology is now enabling educated professionals to create better and better 3D content more than ever before....

just a tip, I do alot of my learning with my stereo Nikon D200 dslr camera.

http://www.reduser.net/forum/showthread.php?t=29035
 
Good words have been written here answering your question, still I like to give you few:

Positioning two Cameras in Parallel axis is by far my favorite way of shooting 3D (stereoscopic), instead of using the cameras in a "L" position such as my two RED #851 and #852 were used on "My Bloody Valentine", but they are more then one element to consider when preparing a rig for S3D (stereoscopic 3D), I will try and point a few, obviously perfection in S3D is a B%#$!@ to achieve and takes tone of money and man power, but look at what they have done with Fusion 3D and Pace, on Cameron's Avatar, just unreal..

1) The cameras if fixed must be placed at a distance of 2.5" and so have to be the lenses.

2) Number one, is just a standard starting point, but if you really want to have a perfect S3D rig, the you will have to mount the cameras in such a way that the cameras can be moved away and closer to each other.

3) Further more to point 2, you will also have to be able to point the cameras towards the inside/front, allowing for the lens to act as if they were your own yes, so that when focusing closer you move the camera back further away allowing for the lenses to come closer together like a triangle, and when focusing further away then you will position the cameras and lenses almost parallel to each other, and move them accordingly to the distance of were your Plane of focus must be, just remember that its pointless to film something g too far in 3D as it looses the 3Dimentionality there for becoming 2D.

4) As said before cameras must be in perfect sink, in fact I'll go further, the cameras must be an exact match of camera, and lenses, and not only be form the same manufacturer but they must be the same model and of the same year.

There is so much that goes in to play in the world of S3D that it requires tone of study and practice, but if you love it, then you'll get there.

ciao
 
KETCH ROSSI said:
Good words have been written here answering your question, still I like to give you few:

Positioning two Cameras in Parallel axis is by far my favorite way of shooting 3D (stereoscopic), instead of using the cameras in a "L" position such as my two RED #851 and #852 were used on "My Bloody Valentine", but they are more then one element to consider when preparing a rig for S3D (stereoscopic 3D), I will try and point a few, obviously perfection in S3D is a B%#$!@ to achieve and takes tone of money and man power, but look at what they have done with Fusion 3D and Pace, on Cameron's Avatar, just unreal..

1) The cameras if fixed must be placed at a distance of 2.5" and so have to be the lenses.

2) Number one, is just a standard starting point, but if you really want to have a perfect S3D rig, the you will have to mount the cameras in such a way that the cameras can be moved away and closer to each other.

3) Further more to point 2, you will also have to be able to point the cameras towards the inside/front, allowing for the lens to act as if they were your own yes, so that when focusing closer you move the camera back further away allowing for the lenses to come closer together like a triangle, and when focusing further away then you will position the cameras and lenses almost parallel to each other, and move them accordingly to the distance of were your Plane of focus must be, just remember that its pointless to film something g too far in 3D as it looses the 3Dimentionality there for becoming 2D.
ciao
Click to expand...

We have a local company that works a lot in 3D and they tell me that in fact converging the cameras can create a lot more problems if the subjects are reasonably close ie say 10 feet or less because you then have offset the deep background which is actually more problematic and difficult to correct for in post.
 
KETCH ROSSI said:
NOt me.. the cameras, but unfortunately I never was down South on the set :(.

ciao
Click to expand...



I think most of the Movie was shot in Pittsburgh PA...

My Bloody Valentine is a great "Case study" on how to make money with new technology and a not so original Horror plot!

No disrespect intended here at all...This group made a lot out of a little!

Dave
 
Michael Hastings said:
We have a local company that works a lot in 3D and they tell me that in fact converging the cameras can create a lot more problems if the subjects are reasonably close ie say 10 feet or less because you then have offset the deep background which is actually more problematic and difficult to correct for in post.
Click to expand...

Well yes, but if and when done properly by a good Stereographer, then the converging of the lens towards the subject in close proximity will be a natural movement and focusing just like we do with our own yes when putting something closer to our eyes we converge to that object in order to focus, and so it has to be done wit S3D in order to fully benefit form the Depth and Plane of focus.

This is in fact not always applied by Stereo rigs as it is more complicated then just leaving two cameras parallel to each other, but movement of the camera bodies apart form each other, as well as the angle change of the lens them self creating a Triangle, is imperative to get the best images and not give vision issues to the viewers as it will create much less issues to deal in post when shot this way, Pace uses this very system configuration with the Fusion 3D rig, used in AVATAR and soon to be used in Resident Evil4 "Afterlife", personally I don't see other way, I have not learned to o many old ways as I much prefer learn the new and more sophisticated ways used by the Stereographers that I come to admire, no doubt they are many ways, they always are.

ciao
 
You must log in or register to reply here.
Back
Top