How to shoot a seamless take at different zooms
 

Hey gang,

I am looking for some guidance. I saw this video by Scott Kelby and really like how it's edited.. specifically how it's seamless dialogue, but the camera cuts in and out at different zoom levels while he's talking:

http://www.scottkelby.com/blog/2008/archives/1351

Does anyone know how this was done? Do you think they had multiple cameras rolling during the take? Or do you think they used 1 camera and did different takes at different zooms? Or do you think it was shot with just one camera at a wide zoom, and just zoomed in closer in post?

I am really interested in learning how to do something like this myself, so any helpful hints or advice on this would greatly be appreciated. Thanks!

Hi Deke -

I think it's likely it was shot in one take, pretty seamless, although it's hard to tell without being able to go frame by frame - it did appear that the hand and head motions were seamless, not "cut together".

With the low res of YouToob, you could probably zoom in post, but it looked to me like there were slightly different angles, so I'm guessing at least two if not three cams were rolling.

I've been working on doing some similar stuff (really liked the lighting in that video, that's been my struggle), and multicam is a in my estimation the way to go. Lighted right, you can have several smaller cameras and get a great end result for about the same $ as one "big" cam. Shooting live events, I've come to the conclusion that 3-4 smaller cams will get a better end result than one or two big ones in terms of covering all the possible angles, unless you've got the luxury of a high budget shoot with manned cams!

If you are shooting in HD and delivering in SD or lower you could do this with one camera. There is enough resolution with HD to zoom in about 200% and still be at full SD resolution, so no softening of the image.

How to shoot a seamless take at different zooms
 

Looks like it was shot in at least two takes. Cuts come at the beginning of a sentence and it doesn't look like the shots had to match up much. I believe it was shot all the way through in medium, then all the way through in closeup.

I'll agree with Richard - it is at least two takes, more likely 4, although some of the reframing could have been done in post. The edits take place at the natural rhythm of the speech, the mic was placed in a consistent position close to the subject in all cases so there's very little difference in timbre between the shots (though if you listen carefully there is a very slight difference in the wide shots versus the CUs), sounds like they were careful to use room tone to bridge any gaps in the edited speech, and the music also serves as a unifying bridging element to smooth the flow of the edited dialog. It also sounds like he is an accomplished presenter, used to producing a consistent delivery over repeated takes.

Both techniques (multiple cameras on a single take vs one camera and multiple takes) can be successful. Results will vary based on things beyond equipment and production technique. The skill of the performer has an enormous impact on how good a single camera approach will look as does the complexity of the staging and the performance itself.

This was a pretty simple stage and setup. In my experience, differences in head, shoulder, arms and hands between takes will be noticable when you sit down to edit (when it's too late) and will dictate when you switch. In a more complicated staging and setup with multiple actors, handheld microphones etc... it gets harder to make multiple takes look good. Then there's the audio as mentioned previously with similar issues of technique and performance.

So, overall, if you look at it in terms of a range, multi-cam and multi-operator might be the best and then everything else is a comprimise off of that with varying effort/skill required in the perverbial "ummmm I'll fix that in post".