Picture-in-Picture: Switching Between the Two?

Question

Hey gang -
First a bit of background: my video editing is purely as a hobbyist and focused primarily on my race track activities.  I do a lot of in-car video when I'm at the track.  Each video is about 20 minutes long or so, depending on how long the track session goes.
I'm considering buying a second camera and pointing it backwards to get the rear-view-mirror sort of effect.  Part of the idea is for amusement to watch how quickly the cars I pass shrink back behind me... :-)  So the idea was to make a PiP video with the rear-facing camera view perhaps 45% its normal size and in the lower right hand side of the main video.  I watched the online tutorials on that and it seems fairly easy to do (if not computationally expensive).  I even played around with it using two of my older track videos just to see how it works.  Easy stuff.
OK, the question: is there any way to easily switch views?  In other words make the rear-view camera the primary and the front camera the PiP?  Is there a way to do that without cutting the two video streams and swapping them between V1 and V2 in the timeline?  Again, I'll have 2 20-minute video streams (and a separate audio stream).  Is there a way to do that transition while leaving the two streams intact?  Or do I have to cut?
Sorry if I'm using the wrong terminology here.  Still new to this stuff. :-)
Thanks!
jas

Answer 1

Well, in general, it's actually more of disk thing, because you're playing multiple video streams simultaneously. Functionally, it's equivalent to creating a PIP with two or more video tracks visible. Of course, depending on the nature of the footage, e.g. AVCHD, it can become a very CPU-intensive process, as well.
I'm not sure if hardware MPE/CUDA really has much influence on multicamera editing or not, to be perfectly honest. Theoretically, since a CUDA GPU handles things like scaling--and that's more-or-less what you're doing in the multicamera monitor--its conceivable that hardware MPE works there. It really all depends on whether that part of the application was written to take advantage of CUDA or not. One of the Adobe engineers who seems to work most closely on GPU acceleration sometimes pops into threads like this to clarify such questions--I don't know if anything definitive has ever been stated before.