3D Steadicam Tests With Two RED Cameras
Published : 22nd May 2010
I tested out two RED cameras on my Steadicam rig in the P+S Technik freestyle flyer :
With the varying and unpredictable movement of Steadicam; who should I recommend to pull interocular or convergence (depending on your 3D strategy). A stereo-optician? or a focus puller on a stereo-opticians advice. I feel I need two qualified focus pullers, one for focus and one for interocular/convergence. Am I wrong?
With one camera pointing up rather than pointing down. Do I need to put one camera upside down so that the scewed image from rolling shutter goes in the same direction. I have had conflicting advice and am now very confused. I am easily confused as I am a Steadicam op not a stereo-optician.
Thank you for your time
On most of my 3D shoots we've used an AC to pull Convergence ... since they are use to working with lens pulls. I was that person on Captain Eo on the "A" camera rig, although we had the unusual situation of usually leaving convergence on the back of the set. Convergence was then re-set in post. I still was needed to reload the upper camera of our 65 mm 3D rig. On some shoots the convergence puller is told to keep convergence a bit behind focus/subject matter. On Avatar I believe Convergence was electromechanically linked to focus. On some shoots a Sterographer will also talk to the Convergence puller via headset.
On some huge productions we've had the luxury of each (8) rig having its own focus puller, convergence puller and interocular puller, besides operator. All remote. That's a LOT OF MONITORS!!! The interocular puller was always listening to one of the (3) Sterographer/Dp's.
It's nice for the convergence puller to have his/her own monitor if they have to work independently of a 3D expert ... following guide lines ... on free form shoots/concerts/sporting events. They can thus see exactly where convergence is.
A(C)vailable in South Pasadena, CA
>> On most rigs with motorized remote controls for Interaxial and convergence, the two controls are >>on one handset, (focus and zoom on a Preston or C-Motion type handset)
The crew-member with the most likely skill-set for setting Interaxial (often referred to as Interocular) and pulling convergence during a shot. is a first assistant (focus puller) as the job requires
accurately judging distance and remembering what the settings should be. Where it differs fundamentally from focus pulling is that many situations call for the convergence setting to be different from focus. - You do not always want the actor to appear right at the screen and have the room going closer and further... sometimes you want the actor to walk into or out of the screen plane.
There are both stylistic and technical concerns with regard to both IA and convergence settings. It is much less common on that IA needs to be adjusted within a given shot, but it might need to be adjusted from shot to shot, especially if changing focal lengths, but bouncing these settings all over the place can be counter-productive too.
Just as with colour, data-handling and etc, discussions should happen before shooting begins between post-production and the DP and camera crew to decide on some conventions regarding how depth will be handled.
Mark H. Weingartner
LA-based VFX DP/Supervisor
As a 3D Tech/DIT I will work with the First to operate the I/O per the DP/stereographers decision as the stereographer usually has eyes on the monitor. I agree an AC should pull I/O but some productions prefer not to bring anymore crew then they need to. It’s a different scenario if your working Broadcast as the folks in the truck prefer to set I/O.
Btw not all Red Sensors are created equal and can be up to 10pixels offset with-in the sensor mounting.
3D High-speed Data
>> As a 3D Tech/DIT I will work with the First to operate the I/O per the DP/stereographers decision >>as the stereographer usually has eyes on the
I've had a few shoots where I have been asked to do so as well. I've done it begrudgingly and under protest as I really don't want to set a precedent...
Los Angeles DIT/VC/Tech Guru
>>With the varying and unpredictable movement of Steadicam; who should I recommend to pull >>interocular or convergence
Hi Thomas. In my opinion, the person who changes the 3D settings has to understand them, even if they follow orders. So I will choose a 3D technician. But all 3D technicians I worked with are former AC's or DoP's so...
To answer your other question : on the freestyle rig the horizontal camera is normally positioned, the vertical one is its reflexion in the mirror, so the lower side towards steadicam and upper side towards what you are shooting. The two images are flipped, horizontally. Rolling shutter is a vertical problem so it would be an issue if the two images were flopped (vertically).
And this is to launch the debate : according to me, you should not change the convergence during a shot, as convergence sets the stereoscopic base by setting the infinity point. Changing convergence during a shot can be very disturbing if it's too violent or two strong. NB : when I say convergence, I'm talking about the angle between the cameras, and not the point where they converge.
Also, changing IO during a shot, as strange as it can be, is virtually invisible. I know it's not the "natural way" our eyes are working, but two cameras are not two eyes.
What you should be concerned of is : changing IO means one camera is moving.
So you can lose your balance...
Hi Thomas, et al!
A couple of very good points here.
1) rolling shutters and mirror rigs.
Yes - you need to make sure that after reflections etc the shutters are rolling the same way otherwise on horizontal movement the images will "tear" in opposite directions, pretty much ruining the 3D
effect. I've seen this - not pretty!
So - Physically Flip the camera that is looking at the reflection - then in post you will need to apply a horizontal flip to this image (to match the one looking straight through the mirror) rather than a
vertical flip. Works a treat. You will still get rolling shutter distortion but at least the two eyes distort the same way and you preserve the 3D effect.
2) Who should pull IOD and convergence?
Well it needs to be a coordinated approach - i.e following the plan as mapped out by director and stereographer. Other than that I am not precious about it but AC's tend to be better at judging distances and remote moves than I am - I know that much!
You can adjust convergence in post even if the move is not perfect. In fact those that advocate shooting parallel and converging in post have really just taken that process to the extreme.
There is much debate about whether convergence should follow focus.
This is what your eyes tend to do in the real world so it is a good starting point but these parameters are independent and you can converge on a different plane to where you focus - it produces a
different effect. To me it is another tool to help tell the story -discuss the effect you are trying to achieve with your friendly neighbourhood stereographer !
IOD is the other (independent) parameter we have at our disposal.
Should you change it during a shot?
Sure you can - if you want that effect.
IOD basically controls the perception of depth in the scene.
More IOD = more perceived depth.
Use this subjectively.
All I would say is please stay within the mathematical divergence limits so you don't bend our brains when we try to watch it! This may be a reason for changing IOD during a shot in itself - say
you start converged on an object very close to camera and then move off to a scene with more distant subjects visible - you might break divergence limits unless you reduce the IOD. In that case you could adjust the IOD dynamically during the shot - hiding the IOD move in the camera move - just like we hide convergence pulls in focus pulls.
Otherwise, if you don't change IOD during the shot you have to start at the smallest IOD that you will need during the take - and be happy with the perceived depth this produces at all points
through the shot. Again a subject discussion to be had with a stereographer.
There are 400 new conversations we could start right here.
Go gadget forum!
. physicist . stereographer .
. maker of 3D rigs and IOD calc . www.speedwedge.com
Personally, I do not see the DIT position and the Convergence puller's positions as an obvious combination or skillset melding...
Mark H. Weingartner
LA-based VFX DP/Supervisor
> What you should be concerned of is : changing IO means one camera is moving.
> So you can lose your balance...
Not something you need to be concerned about with this rig, since both cameras move in opposite directions when adjusting IA (in order to maintain balance)...
Nice video Thomas - good to see the setup. How about lens control, sync generators and 3d monitoring though? Any thoughts on how manageable you'd find it with all those extras? Interesting to know the total weight with all that gear attached too. And, the way the cameras mount on the Freestyle rig is correct in regards to rolling shutters - they are both rolling in the same direction, so you needn't worry about mounting cameras upside down. That only applies to rigs with a different design (such as the Standard P+S Mirror Rig).
>> Re: 3D Steadicam tests with two RED cameras
How we are getting ready now is have the 1st pulls focus & 2nd pulls convergence and IO from the monitor. From an overlap feed either on the cart or truck.
E Gunnar Mortensen
1st AC Local 600
Sharp Wit & Sharper Focus
Good information, Leonard, except for this:
>> So - Physically Flip the camera that is looking at the reflection -
This physical flip is not needed if the reflection camera is mounted below the mirror as in the Swiss rig and the P+S Freestyle (the rig that this thread is about). These "under slung" style rigs are also
easier to flag for flares, glare, and internal reflections in most lighting situations (light from above).
It should also be noted that on rigs that aren't built correctly for rolling shutters, you can flip either camera, you don't need to flip the reflection camera. Just make sure you only flip one.
VFX & Animation Design