But it would be tough to calcuate the location of a splice based on the diameter. In addition to the reel diameter, it would also need to check the film diameter. The film diameter would expand or contract, depending on the humidity and winding tension, confusing the computer. Even a shaft encoder might not be accurate enough. Perhaps using a frame counter (an optical type to avoid wearing the film during high-speed) to record the frame count between reels when making up... then using that to find the splices at breakdown.
I'm waiting for the day when platters are "wired" into the projection system, in the same way the projector and sound are all controlled by the automation. Today, the platter's only "connection" is usually a wire for a platter-wrap detector.
I was thinking that someday (if e-projectors don't make film obsolete) someone might come up with a common interface and language, like how manufacturers are trying to standarize infra-red remote-controls for audio/video gear. Wouldn't that be nice; no control wiring between automation/projector, sound processor and platter.
They already have this: its called a tension sensor- or if you have Component Engineering failsafes/autos, motion sensors as well.
Aaron
First, calculate the number of "layers" of film on the reel:
Layers = OD of film on reel - Hub dia. of reel / Thickness of film.
Next find the circumference of the layer of film at the exact center of the reel:
Circ. = OD / 2 * <pi>
Then multiply the circumference times the number of layers and you have it!
The idea here is that exactly half of the layers will be smaller than your average and half will be larger. For every one that is smaller there will be one that is larger. The two will cancel each other out and you will be left with little or no error at all.
I wrote a program for my HP calculator that does this. All I have to do is punch in the measurements of the reel. I have found that there are MANY places where inaccuracy can creep in:
Polyester film is 120 µm thick. Acetete is 140µm. I have heard that Fuji film is slightly thinner than Kodak. I don't know for sure. Still, my point is that there is approximately a 15% difference there. When you try to make calculations based on the wrong figures that can really throw your numbers off.
Your measurements have to be super-duper accurate! If the film is only 120 µm thick then theoretically you have to be accurate to within LESS THAN 120 µm! An error of just a few microns will really add up over the course of a couple of thousand revolutions!
You don't know if the film has been wound onto the reel with perfectly even tension. If it isn't then you are introducing errors into the calculation. Again, even a 1% error will really add up on you!
Splices and wrinkles in the film will throw off the calculation.
When you are calculating the length of a reel, those errors can be cancelled out. For instance, if you make 3 separate sets of measurements and average them out you can minimize measurement error. If you use larger reels of film and you meaure very accurately you can get less errors because your inaccuracies have a tendancy to cancel each other out. (The theory of very large numbers?)
I made a super large "caliper" out of a meterstick and was able to measure the film on a platter. If I worked really, really carefully I was able to estimate the time it would take to play that movie and be accurate to within +/- 5 minutes. It was fun but I soon figured that if you count the number of reels and multiply by 20 you could come up with a number that was almost as accurate AND you could estimate how much film is LEFT in the movie. You can't measure and calculate THAT while the platter is running!
The point I'm trying to make is that computer control of any film transport system by measuring the reels or counting the revolutions would be rife with errors. If you wanted to be super accurate it would be better to run the film through a film counter and measure the actual length of the reel. Then you make a similar arrangement on the MUT and program it to stop at a comfortable distance BEFORE the end of the reel. The problem is that you have to run the film over a sprocket to do that. Even if you were able to make a "machine vision" system to count the number of sprocket holes that go by you STILL have to run the film through another set of rollers. The more rollers or sprockets you run the film through the more chance you have for damage. I just don't think this is feasable.
I DO like Brad's idea of somehow sensing the splices. The only problem I see with that is that by the time the machinery has stopped the splice will have gone past. You would have to somehow back the film up, either by hand or by telling the system to back up slowly until it finds the splice and finds it.
Frankly, I'm not a fan of computer controlled doo-dads. There are just too many variables to account for. Humans naturally account for things that computers can't. For instance a human can predict where the splices will be and start slowing the platters down in time. As I pointed out above it would be really tricky for a computer to do the same thing.
I think a well trained operator can outdo a computer controlled projector any day. Computerizing things just to have computers is a waste of time and money.
There's nothing wrong with experimentation, though as long as you don't get so wrapped up your ideas that you fail to see the big picture.
Kodak triacetate BASE used for 35mm and 70mm print films had a nominal thickness of 0.0056 inches (142 microns).
------------------
John P. Pytlak, Senior Technical Specialist
Worldwide Technical Services, Entertainment Imaging
Eastman Kodak Company
Research Labs, Building 69, Room 7419
Rochester, New York, 14650-1922 USA
Tel: 716-477-5325 Cell: 716-781-4036 Fax: 716-722-7243
E-Mail: john.pytlak@kodak.com
Web site: http://www.kodak.com/go/motion
When I calculated small reels like trailers on cores or anything under about 1/2 full on a 15 in. reel my numbers were off by about 10%. It seems like I have just found where the problem lies!
Thanks, John!
