[net.video] converting film to video

lauren@vortex.UUCP (Lauren Weinstein) (04/14/84)

This is still usually done in much the same manner that has been used
since the early days of television.  The "film chains" which do the 
job are essentially just a movie projector with a lower than
normal intensity bulb pointed into a video camera.  There are usually
some mirrors and lenses in between to help with size adjustments and
such, but the basic idea is quite simple.  To avoid flicker between
the 30 frames/sec (NTSC) frame rate and films's 24 frames/sec rate,
the projector (which would usually have a synchronous motor) has a 
special shutter and gearing installed that displays some frames "twice"
rather than the "even" display rate that conventional projector shutters
provide.  This difference cannot be noticed by the viewer but allows
a match between 24 and 30 frames/sec.  

If you try this with an ordinary projector (i.e. by aiming a video
camera at a movie screen) it will work, but there will be flicker
caused by the frame rate mismatch.  For some simple applications, this
might be acceptable.  Be sure NEVER to aim a video camera directly
at a normal projector -- the high intensity lamp might burn out the
vidicon in the camera!

--Lauren--

dmmartindale@watcgl.UUCP (Dave Martindale) (04/15/84)

The film chains which transfer film to NTSC video do not usually show
some frames once and some twice.  NTSC uses 30 frames per second, but
60 fields per second.  The film chain will hold one frame of the film
in position for 2 fields (=1 TV frame) and then hold the next film
frame in position for 3 fields (=1.5 TV frame).  The net result is
24 film frames/sec matches 60 TV fields/sec, while the motion isn't
much disrupted.

However, any home videotape that I've seen which started out as a film
seems to have every 4th film frame recorded on two successive TV frames,
with one TV frame for each of the other 3 film frames.  This gives
a frame rate match again, but I would expect the motion to be less
smooth since some frames are held for twice as long as others, compared
with the 3:2 ratio of the system described above.  I couldn't see any
problem at full speed, though.  Anyone know why this system is used?
The only thing I can think of is that in stop-frame operation, each
TV frame comes from a single film frame and so would be a stable image.
Using the 2:3 method causes 2 of every 5 TV frames to be made up of
two fields which came from different film frames - these would have
very pronounced flicker or jitter if displayed in stop-frame.

The Europeans have none of these problems, since they use 25 frames/sec
for both film and video.

ntt@dciem.UUCP (Mark Brader) (04/25/84)

[Just in case.]
	The film chains which transfer film to NTSC video do not usually show
	some frames once and some twice.  NTSC uses 30 frames per second, but
	60 fields per second.  The film chain will hold one frame of the film
	in position for 2 fields (=1 TV frame) and then hold the next film
	frame in position for 3 fields (=1.5 TV frame).  The net result is
	24 film frames/sec matches 60 TV fields/sec, while the motion isn't
	much disrupted.

Equipment is now available which will substitute various alternative frame
patterns for the normal 1-1-2-2-2-repeat described above.  The result is
that the movie still shows smoothly on TV, but occupies a different length
of time from what it did before.  For instance, a frame pattern of
1-1-2-2-2-3-3-4-4-4-5-5-6-6-repeat would result in a playback 1/15 faster
than the original.  This has been used to fit movies into the available
time slots without the need for a lot of editing or for excess minutes
slopping over the end.  The only particular show I have heard of this
being done to was "Chinatown", but I expect it is becoming standard practice.

Mark Brader