Have you tried it? It will compile, but it seems like the native structure wrapped by the Movie dies as soon as you create it.
Here's the critical section of some test code I wrote:
grabber = new SequenceGrabber();
SGVideoChannel vChannel = new SGVideoChannel(grabber);
System.out.println ("got video channel, sourceVideoBounds = " +
", videoRect = " + vChannel.getVideoRect());
/* this doesn't work - QT native object is non-existent
Movie movie = Movie.fromSequenceGrabber(grabber);
QTComponent qtc = QTFactory.makeQTComponent (movie);
Component c = qtc.asComponent();
Run this with a camera attached (I'm using an iSight) and here's the output (line breaks inserted for clarity):
got video channel,
sourceVideoBounds = quicktime.qd.QDRect[x=0.0,y=0.0,width=1600.0,height=1200.0],
videoRect = quicktime.qd.QDRect[x=0.0,y=0.0,width=1600.0,height=1200.0]
quicktime.QTNullPointerException: The QT native object
represented by this QTJava object is no longer valid:
So, we have some low-level access to the camera pixel data via the video channel, but trying to use high-level concepts like the Movie bomb out: we can't even println it! And taking out the println doesn't help - you'll just get more QTNullPointerExceptions later on when you try to idle the SG.
I hope to be back soon with a piece on what you can and can't do with QTJ capture. In the meantime, file a bug on QTJ capture being broken. Apple says volume of bug reports is a key determinant of where they apply development effort.