Two questions. I'm sure there are several more examples than the Secret Window movie, but since I just watched it the material is fresh in my head.

1. How do they do those so long camera moves, ie: at the beginning of the movie when the camera starts down along the river, comes up the beach, goes up the side of the house, in through the second story window, over the banister down to the first floor couch, all in one (semmingly) shot?

2. How do they get such clean composites of the same actor interacting with themselves? Ie: when Johnny Depp's character starts talking to himself towards the end of the movie...

Thanks,

Jason

Cool, thanks for the reply. So for question 1., how is a digital blend accomplished, as far as the actual shoot is concerned? I mean, do the shoot the first part of the sequence, move the camera, shoot the next part, move the camera, etc., and if so, how do the get the camera back in the right place?

Thanks!

Jason

navarro_parker

1. I don't remember that scene specifically. But long tracking shots aren't new. They could be digital blends of several different shots. Brian DePalma used some clever non-digital methods for the beginning of "Snake Eyes"

2. I didn't think the composites were *that* clean. It obviously looked like an effect. Ideally, you film both versions on a chroma key background. You matte them out and composite them on a background plate. That's how they can overlap. And you switch their layering order when they walk around each other. It's usually very well choreographed, with the first version playing back for reference when the second take is done.

See the Two Towers Extended Edition DVD for some terrific effect tests putting hobbits and humans in the same shot (both walking around each other) with motion control cameras.

ari_valen_levinson

Jason, generally some morphing software (such as the plugin by ReelVision for AE) is used to combine two takes. They do as much of the pan as they can in one shot, then back up about two seconds and begin the second part. The morphing software is then used to blend the overlapping two seconds.

As for the seemless integration of two performances by one actor, see Adaptation. It is easiest when the camera is locked off, but otherwise a motion control rig is advisable. I think nararro underestimates the use of manual masking in such a scene. I know that the one time I did a short in which the actor had to interact with himself, I used manual masking the entire time. A collegue of mine did a similar concept but used greenscreen for both of his actors (using only a background plate of the scene) and we both agreed that the masked version is better. Also, I found that I often had to edit the performances to make the timing fit perfectly.

Ari, do you have a link for ReelVision? I Googled and couldn't find it...

Thanks!

Thanks for the info! Helps very much!!!

navarro_parker

Yeah, I didn't mean to overcomplicate things. Ari is right about the magic you can work with a locked off camera. (I don't recall the camera being static in Secret Window)

....just for S & G's, for those of us that didn't see the movie, a link to the movie (trailer, effect, whatever) would help.

- Joe

here it is:

<http://www.apple.com/trailers/sony_pictures/secret_window/>

seth_hancock

Hi all,

I have found this thread to be extremely interesting. I am a newbie to Post Production and a lot of jargon/lingo is still somewhat foreign. Can you please tell me what "Locked Off Camera" means in shooting? Thanks.

Seth