(image copyright: the movie, Haunted 3D. used for educative purposes only. click each for larger)

Stereoscopic 3D Depth Continuity:

In traditional 2D movie making the task of maintaining continuity from scene to scene is that of the Script Supervisor (and Director). This includes “depth” as well. For instance, if a scene features a person walking down a road into the distance, the next camera angle can’t show the person to have traveled half way down the road in a matter of a couple of seconds, unless this is done intentionally as part of an effect of speeding time.

But again in 2D movies, it’s easy to cheat depth. That’s one of the ways those stunts involving a person crossing a railroad track at seemingly the very last second while a train approaches. The use of long lenses compresses depth and makes it look like the train is a lot closer than it actually is.

In Stereoscopic 3D, this is very near impossible to do. The one thing that stereoscopic 3D movies give us, is the ability to record and present “Spatial information” or the depth dimension of a location.

Take a look at the two images above. They are from (in our opinion) a well made 3D Bollywood movie, Haunted 3D. The stereographer, Brent Robinson has done an excellent job throughout the film in maintaining depth continuity along with other S3D best practices. While it can be said that he has been conservative with the depth budget throughout the movie, in hind sight, the movie works in both the cinema and on a typical 3DTV without causing any harm to the audience.

S3D Depth Continuity is more than just Cut matching:

In the scene above, both the shots “match” in depth continuity in a number of ways:

1) The actual “deepness” of the corridor match in both shots shown at two different times in the movie (one way to ensure this, is of course to shoot all scenes involving a particular set, at the same time)

2) This means that the lens settings are the same

3) The Interaxial and /or toe-in (convergence) if any, is noted and re-applied when shooting at the same location.

3) That the time taken to travel the distance is approx the same depending if the character or camera is previously shown as running or walking.

That example is by no means the only scenario to show Depth continuity. There are many other depth continuity scenarios that need attention. Now take a look at the images below:

(image copyright: the movie, Dark Country. used for educative purposes only. click each for larger)

The image above is a severe breach of Depth Continuity in a single take. It is not a cut but an actual tracking shot around the main characters, but the depth abruptly bounces due to an interaxial change. Thus the entire depth of the scene gets disturbingly changed.

Another example of a depth mismatch is in cuts. In the pictures above, the first shows the same scene with one 3D depth setting, this then cuts to a solo shot of one of the characters speaking, and then cuts to the second image above, showing a completely different depth. This is another example of no Depth Continuity.

Depth Continuity – The Script Supervisor’s Job? The DPs or the Stereographer?

That question is open to debate at this stage of 3D movie making, as there is still an unsettled question of how much importance should be allotted to a Stereographer. As time goes by and more experience is gained, it’s natural that the DP would want to take charge of the entire composition of the 3D “frame”. We use the word frame in quotes because today in 3D filmmaking, a lot of established terms are open to new interpretation. At Real Vision we prefer to call the “frame” the window, because that’s what the frame becomes… a window into 3 dimensional space.

Getting back…should the script supervisor be in-charge of the depth script as well as the movie script? or should the depth-script then be in the hands of the DP, and can he/she also pay attention to maintaining depth continuity in addition to his/her other tasks.

Simple depth continuity as in the scene above can be accomplished to a certain extent in post by H.I.T. but again, HIT can only solve for scenes such as the one above, which does not contain too many foreground objects that may otherwise cause disturbing stereo window anomalies.

Depth Script Design and Implementation:

As 3D movie making evolves, one debate that will soon be over is whether there is a need for a stereographer or not. We would argue yes, there is a need. For the very same reason that there is a need for a 1st and 2nd AC, as well as the need for a Best Boy and Gaffer… for the smooth running of 3D film shoot, a stereographer is certainly recommended for this one reason alone: The designing and implementation of a Depth Script (comprising depth budget of scenes, taking care of depth continuity during shoot and in post with the editor) and many other 3D related tasks.

Can these jobs be offloaded to the script supervisor or DP?  Yes, in indie productions one person fills many shoes, but in a well planned (and financed) 3D feature, the depth script should be with the Stereographer who sits with the Director and DP during a shoot, and with the Editor in post.

(image copyright: the movie, Drive Angry 3D. used for educative purposes only. click each for larger)

The Edit, Editor and Depth Continuity:

In the scene shots above taken from the movie Drive Angry, the image on the left shows …

Read the rest in the Book, “THINK in 3D” via Amazon: Paperback & e-book (also on iPad via free kindle app)

  • Chuck Comisky

    Good article, here are 3 of 10 guidelines for good stereo as developed on AVATAR:

    INTER-OCULAR distance varies in direct proportion to subject distance from
    the lens:  The closer the subject, the
    smaller the inter-ocular.  The farther
    the larger.  A shot of the Grand Canyon
    from half a mile away may have a 5′ inter-ocular.  A shot of a bug from a few inches away may
    have a 1/4″ inter-ocular. 
    Inter-ocular tolerance is subjective, but there is a constant value of
    background split which cannot be exceeded.

    CONVERGENCE is almost
    always set on the subject of greatest interest, and follows the operating
    paradigm for focus — the eyes of the actor talking.  If focus is racked during the shot to another
    subject, then convergence should rack. 
    An exception to the rule of following focus exactly is a shot with a
    strongly spread foreground object which is NOT the center of interest (such as
    in an OTS), in which case a convergence-split may be used (easing the
    convergence forward slightly, to soften the effect).  This should be combined with control of
    inter-ocular to yield a pleasing result. 
    Convergence splits are limited by high contrast edges at the plane of
    interest, which may cause ghosting in passive viewing systems.
    moving shots.

    • Anonymous

      Thank you for weighing in. It’s hard to add to or counter the opinion of someone who’s worked on a 3d opus.

      So in being respectful and fully acknowledging the insights you have gained and kindly sharing, from being first hand involved in the 3D spectacle that was Avatar, I would like to offer my view-point on the converge-on-subject-of-interest.

      You do say “almost always” and this is key to my agreement with you! convergence would/should almost always be on subject of interest, but i’d clarify my stance with

      a) this is the ideal/best way to do OTS and reverse angle shots when subject of interest fills most of the frame (it also makes for easy depth continuity)

      b) However, in scenes in Avatar, you have two characters in a room that are not in closeup or medium closeup and yet focus/convergence racks between them. This has the disconcerting effect of squashing the dimensions of the room. If we use the analogy of real life third party observer (the camera) looking in on such a conversation between two characters… in real life we would not jump a few feet forward to listen to the person at the back speak (the effect of converging on that person) and then jump back to original observer position as the front person speaks.

      Some more observations: in S3D I believe that a rack focus needs to be “motivated”. This would ideally be done with a camera move while racking focus so that *all* ambiguity of where the Director wants the audience to view is removed.

      If we use the standard 2D version of rack focus with a fixed frame and no camera movement to re-frame the new subject so that the subject becomes the most obvious center of attention, then we stand the chance of leaving residual visual artifacts that the Viewers brains need to resolve before it moves on (rapidly) to re-focus according to the directors wishes.

      This does not mean that the motivated move while rack focusing has to be boring and completely remove the previous subject from the frame…it just needs to remove any possible ambiguity of where the audience should look, so that they dont end up watching a blurry part of the image. (Over a 1 or 2 hour period of such brain activity, a headache is sure to ensue)

      one more observation. The art and craft of representing a nice 3D image or roundness of 3D on something as far as the Grandcanyon or a city block or indeed a high up shot of people on a street needs careful consideration.. again motivation is key.

      Example: is it ok to have people looking like miniatures, or a toy city (due to a 5 foot interaxial?) This is a very possible end scenario with such interaxials and needs constant attention from stereographers/moviemakers.

      Yet I fully agree that you need to do hyperstereo interaxials to represent some depth in these far off scenes. My preferred method is to establish the shot but quickly move the camera to reframe something of more interest closer…before the brain starts questioning the hyperstereo scene in front of it

      the art of framing for 3D is where it’s at.

      Best Regards.