Really great compilation you found, Jay.
Thanks, one of the guys at Sundance pulled it up, so I bookmarked it.
I good have done without the incessant looping of the track from Vertigo.
I'm okay with it since it helps you focus on the visuals. Also, probably copyright blah blah handwaving.
I'd say the most effective shots were from Vertigo, Apollo 13, Road to Perdition, Poltergeist, and E.T.
Indeed.
Most of the others seemed, for lack of another word, pretentious [edit: wrong word came to mind earlier] on the part of the director or cinematographer.
Indeed -- more of a "look what I can do!" That's the same complaint I have with the current trend of handheld cameras and snap-zooms.
Questions, since what I know about cinematography could fit on the head of pin: is the lens somehow linked mechanically to the dolly, causing the zoom (and focus tracking) to follow along, or does the camera operator do what he needs to do in the view finder?
For most of those shots, not even that. The camera operator had the viewfinder and the pan and tilt wheels. The focus puller sat to the side and operated all the lens controls, which were typically pinioned into knobs (see below) precisely so that they could extend through holes and slots in the blimping. But the focus puller couldn't see through the viewfinder. Even with the advent of cable linkages, the job of focus puller was retained simply because geared heads still required both hands to operate.
The dolly was pushed by two or three stout lads, none of whom could see anything the camera saw.
How did they do this with older, blimped cameras that didn't have a reflex view finder? Did the focus puller have to know what rate to change the focal length and the focus?
Yes. Older cinematographers relied on taped distances for focus, and standalone viewfinders for focal length. The focus puller would measure the distance with a tape attached to the lens linkages, with the zero point set at the focal point. The knobs for focus and zoom had a white index mark. Behind the knobs is a white glass (or later, plastic) disc on which he can mark temporary settings with a grease pencil. So he measures the beginning and ending of each shot for focus and zoom, and marks the corresponding index positions in grease pencil. During the shot, he is responsible for manipulating both knobs at the proper rate between his index marks.
Similarly the dolly position at the beginning and ending of each shot is marked on the floor (or dolly track) with spike tape. It's the job of the dolly grips to push the dolly at a precise, repeatable speed through the shot. That's the case even when you're not attempting a dolly zoom.
The pinnacle of today's state of the art is fully programmable motion-control rigs, built in some cases from industrial robots. These can be programmed to accept and imitate even handheld camera movements, also at scaled speeds to account for different frame rates. The cinematographer can literally move the camera physically by any means and the rig will record its motion precisely. The lens adjustments can also be programmed in advance, also by capturing the "artful" manipulation of the controls by an expert focus puller.
Less expensive are servo-controlled remote heads that can be placed on dollies, cranes, and other traditional camera rigs. The camera operator and focus puller sit side-by-side at a remote station, often with the director, and operate remote versions of their respective controls. The operator, for example, has traditional pan and tilt wheels that behave as in a normal head, but simply encode the motion for repetition (in real time or recorded) by the servos in the remote head.