Sprocketless Film Camera

[carllooper]

Have been experimenting with camera designs, purely in theory rather than practice, but it seems to me conceivable that one could design a film camera for film that need not have sprockets.

A sprocket system takes up a significant amount of a film's surface area, not just the hole's footprint but the space that must then be lost between the sprockets as well.

The issue, of course, is how one then otherwise solves the problem that sprockets solve, without them.

A sprocket participates in the solution to the overall problem of:

a. transport of the film
b. holding the film stationary during exposure
c. registration of frames to each other, in terms of time

So how might we otherwise solve these problems? The traditional design involves an analog to digital conversion of the motion supplied by a conventional motor. Such a motor provides a continuous rotation (analog motion) which is effectively digitised, ie. into the discrete step motion of the film (eg. 24 steps per second).

An alternative to this analog to digital motion system would be to use a stepper motor. It can provide the required digital motion directly to that section of film undergoing exposure, and can do so at precise programmable rates (CPUs rely on precise timings regardless of the reasons a camera might require such, so you get crystal sync rates for free). A separate stepper motor for take up, can be coordinated in relation to the exposure motor, to accommodate the variation in rate, between exposure rate and take up rate (as the take up reel expands in radius).

The accuracy of a stepper motor can be out by fractions of degree but it is consistent. For a given step it's rotational error will be fixed, ie. it doesn't vary. It is locked in during manufacture of the motor. It's the same error every time. But insofar as it translates to a positional error, it means one can't rely on one frame being exactly the same distance to the next. It will be out by some fraction of a millimeter. To compensate for this error one can encode the error directly into the film. During subsequent transfer of film to digital, the error is decoded and then applied to the transfer: encoded error minus decoded error equals zero error.

Encoding can be done by means of optically printing one or two cross hairs at the edge of the film during camera exposure. A small cheap laser diode can be used for this. Obviously this encoding would take away space from the area an image could otherwise use but it could take away less space than that occupied by a sprocket hole.

The only remaining problem would be how to physically move the film, ie. converting the stepper motor rotation into a grip on the film that moves it by the stepped motion. The grip could very well end up using the same amount of real-estate that a sprocket hole occupies. You'd want a firm grip. I don't have any firm ideas on this. Conceptually I think of how I'd do it with my hands imagining the film was, say, a foot wide. I'd grab the edge area of the film, with two fingers, and move it the required distance. I assume film sandwiched between two stepper driven rollers, (that grip the film by it's edge), would be able to do the job.

Stepper motors can be very cheap, very accurate, very programmable, very strong, very fast, and provide exactly the type of digital motion that traditional movie camera's otherwise aim at delivering to a roll of film.

Carl

[wado1942]

I asked some time ago about the plausibility of a capstan/pinch roller type transport to allow for doing away with sprockets. Without sprocket holes, one could shoot a Super-8 cart and get a superior image with better registration than the sloppy claw pull-down. If you want to shoot 16:9, you can do it without cropping as well since you could probably get a 7.5mm x 4.2mm image. You can use stepper motors or you can merely use a continuous DC motor and intermittent sprocket-driven capstan with something similar to the Geneva movement, except it's turning the capstan instead of grabbing the film itself. A stepper motor would probably be more precise, though. The rollers can make contact with the entire width of the film for more stability rather than just the edges. A camera like this would naturally be very quiet as well.

Ultimately, though, the world is forgetting about film. We're lucky to have Super-8 at all and I don't think anybody will be willing to spend money on a new film system.

[carllooper]

I asked some time ago about the plausibility of a capstan/pinch roller type transport to allow for doing away with sprockets. Without sprocket holes, one could shoot a Super-8 cart and get a superior image with better registration than the sloppy claw pull-down. If you want to shoot 16:9, you can do it without cropping as well since you could probably get a 7.5mm x 4.2mm image. You can use stepper motors or you can merely use a continuous DC motor and intermittent sprocket-driven capstan with something similar to the Geneva movement, except it's turning the capstan instead of grabbing the film itself. A stepper motor would probably be more precise, though. The rollers can make contact with the entire width of the film for more stability rather than just the edges. A camera like this would naturally be very quiet as well.

Ultimately, though, the world is forgetting about film. We're lucky to have Super-8 at all and I don't think anybody will be willing to spend money on a new film system.

The design is towards something I'd be building, so the last point isn't really an issue. In other words, it wouldn't be stopping me spending money on it. Not that I'm rich but I can see such a camera being put together quite cheaply. As to my sanity, that might be a different question altogether. As to film stock there must be somewhere in the film manufacturing pipeline, where one can just pull the film out prior to it being punched with sprocket holes. 8mm or 16mm film without the sprocket holes. Or are the sprocket holes put in prior to the emulsion coating? I believe it's afterwards. In the meantime I can just use sprocket film, even if the sprockets do use up real estate. Its the simplicity of the camera design that interests me more - and the registration mechanism (laser diode encoding of cross hairs).

I don't have much spare time, so it always takes a while for me to put things together. But I do put my money where my mouth is. I put $800 on some custom optical grade mirrors a few months back (two fringe accuracy), for a custom digital to film printing system I'm building. It's purely for the fun of it. I get a kick out of it. I haven't yet used the mirrors (they are kept safe in their delivery box) - but in the meantime they're great to look at - really sharp reflections - mind boggling ...

I've mentioned this in other posts recently but I take the position that film has been dead for three decades. This is purely a theoretical proposition (neither true nor false) but allows the last three decades to be be re-read according to a different narrative. Instead of reading film as something dying a slow death over the last thirty years (with a demise always about to occur next week), one re-reads the last thirty years in terms of how film has returned back from the grave, in various ways, over that same period. Instead of a narrative about aging diehards and habituated industrial setups dwindling in number until the last frame of film is shot, one speaks instead (or in addition), of those who have turned, or returned, to film, or never left, over the last thirty years, and do so for reasons other than those that have been historically the case, ie. in full knowledge of where "substitutes" (video/digital) would be and have been going. Film as something that is not a substitute for video/digital. And vice versa.

To develop an alternative narrative, of the emergence from zero, of film usage, over the last three decades, by those for whom film makes sense on a somewhat different tangent, regardless of what has otherwise taken shape in the larger marketplace. Film as something that has been rebooting for three decades, by anyone who uses it, for whatever reason they have, for whatever kind of result they deem fit. Indeed one could go back further in time. Film as something that has been rebooting since day one.

Anyway, back to the task at hand:

If the rollers were to make contact with the entire width of the film what sort of material do you think would be best for that? My concern was damaging the film in some way. I don't know how sturdy film is. I often play rough with film anyway but don't want to necessarily engineer that into a machine. But if there is some material that would be perfect for the job I'd much prefer to grip the entire width. I imagine some sort of rubber around the rollers. Squished onto the film from both sides, it would provide a largish surface area in contact with the film, reducing the required friction force per unit area of film. A couple of tight fitting rubber O rings around the rollers perhaps. Anyway - something to investigate ...

Less friction per unit area of film could be obtained by having a two dimensional contact area: a flat rubber belt gripping the film (over a number of frame lengths), rather than just the one dimensional contact a roller provides. On the other hand the belt drive could introduce uneven distribution of forces which might rub the film up the wrong way (so to speak). So perhaps a simple roller is the better idea. I imagine one might need to change the O rings often, to avoid build up of any emulsion residue from the film the rubber might capture. But rubber O rings are very cheap. I bought a bag of 50 a while back for a couple of dollars. I guess the thing is to just try it and see. To experiment. It's where half the fun is anyway.

Is rubber appropriate though? Might there be some build up of electrical potential that the rubber might transfer to the film causing issues?

Some reading up suggests there are all sorts of chemical resistant synthetic rubbers for O rings. While they are designed to resist chemistry they presumably have the reciprocal effect, where the chemistry is correspondingly unaffected by the rubber.

Carl

[wado1942]

Well, if this is purely for your own experimentation, I'd suggest 16mm instead of 8mm. In this day of artificially sharp high definition video, 16mm can go back & forth between the two worlds quite well. A 15 x 6.25 (or so) frame will give you a 2.4:1 ratio and leave some space between frames for registration marks. This is similar to what the MaxiVision system did, except a grid pattern was optically burned into it. Also yes, the perforations are one of the last steps. I believe the film is coated, dried, slit, then perfed right before being wound onto cores.
I was also going to suggest synthetic rubber (definitely not soft polyurethane) which should be good for years. You can take an idea from the world of tape recorders. BTW, I have a couple tape recorders from the early 70s that still have good pinch rollers.
Something similar to the mechanism shown here would work. The rubber pinch roller would contact the emulsion surface while the polished aluminum (or ceramic in some cases) capstan moves the film. The felt pad shown in the picture should not be necessary since you'd be using new film. The capstan/pinch roller would be below the gate instead of above it like the claw mechs in many projectors (I never understood that) to pull the film through the gate. They would be mechanically aligned to the film is pulled straight downward, so the pinch roller would need to be below and in front of the gate.

newcammech.gif

The pinch roller would also be engaged via solenoid when the camera is running.
You can use an optical system to burn in any simple mark like what they did with date/time for many consumer grade 35mm cameras. With this system, the changes to the film path itself would be pretty minor, even though the construction of the internal workings would be quite different.

[carllooper]

Well, if this is purely for your own experimentation, I'd suggest 16mm instead of 8mm. In this day of artificially sharp high definition video, 16mm can go back & forth between the two worlds quite well. A 15 x 6.25 (or so) frame will give you a 2.4:1 ratio and leave some space between frames for registration marks. This is similar to what the MaxiVision system did, except a grid pattern was optically burned into it. Also yes, the perforations are one of the last steps. I believe the film is coated, dried, slit, then perfed right before being wound onto cores.
I was also going to suggest synthetic rubber (definitely not soft polyurethane) which should be good for years. You can take an idea from the world of tape recorders. BTW, I have a couple tape recorders from the early 70s that still have good pinch rollers.
Something similar to the mechanism shown here would work. The rubber pinch roller would contact the emulsion surface while the polished aluminum (or ceramic in some cases) capstan moves the film. The felt pad shown in the picture should not be necessary since you'd be using new film. The capstan/pinch roller would be below the gate instead of above it like the claw mechs in many projectors (I never understood that) to pull the film through the gate. They would be mechanically aligned to the film is pulled straight downward, so the pinch roller would need to be below and in front of the gate.

newcammech.gif

The pinch roller would also be engaged via solenoid when the camera is running.
You can use an optical system to burn in any simple mark like what they did with date/time for many consumer grade 35mm cameras. With this system, the changes to the film path itself would be pretty minor, even though the construction of the internal workings would be quite different.

Thanks heaps Wado1942, sounds great. I'm going to go ahead with this idea at the first opp and see what eventuates. Purely for fun. Love the idea of solenoid for the pinch roller - makes the mechanics simple - which is what I'm after - something as simple as possible to create - and yet can give perfect results. I'm a software developer by trade so that side of things won't be a problem. Would have a small computer inside to run the stepper motors, etc. (Raspberry Pi, Arduino, etc).

Its basically a project inspired by someone on Cinematography.com building their own Super8 camera, but the difference from that one is that this one would use a sprocketless drive and use registration mark encoding. Indeed, if designed the right way could be a multi-format camera: Super8, Standard8/Ultrapan8 and 16mm. Or even 35mm for that matter, ie. if made in a modular way. But rather than go that far, for version 1.0, I might just make it a 16mm variable aspect camera, ie. can shoot 4:3, 16:9, cinemascope, or anything else one likes (10:1 anyone?) without having to sacrifice any emulsion at all. Indeed if plug and play gates were to be employed then gates for sideways exposure could be done, ie. for increased emulsion per frame area.

cheers
Carl

[wado1942]

Yeah, that's an interesting idea. I have an AMPEX tape recorder from the early 70s I use fairly regularly that lets you switch back & forth between 1/4" and 1/2" tape and all it takes is two screws to pull off the 1/4" head block and drop on the 1/2", then the guides pull and turn to rotate the 1/2" tape portion into the path. It can also be changed to 1" tape but that requires quite a bit more work. BTW, I really love this machine for its brilliant, modular design and care for quality.

[carllooper]

Yeah, that's an interesting idea. I have an AMPEX tape recorder from the early 70s I use fairly regularly that lets you switch back & forth between 1/4" and 1/2" tape and all it takes is two screws to pull off the 1/4" head block and drop on the 1/2", then the guides pull and turn to rotate the 1/2" tape portion into the path. It can also be changed to 1" tape but that requires quite a bit more work. BTW, I really love this machine for its brilliant, modular design and care for quality.

A friend of mine (a software developer as well) is very much into the older tape recording machines, for his music. He does plenty of digital work of course, but the analog work is a kind of special activity - a back to basics thing that keeps one grounded in many ways.

C

[carllooper]

The digital universe has engendered a renewed interest in mechanics, but one based on reprogrammable controls systems (small cheap computers) rather than hardwired analog controls (such as dials and push buttons).

It's a movement called "physical computing", examples of which are robotics and of course: 3D printers.

Film provides another outlet for physical computing ideas. This camera fits in with that movement. Digital control of film. It is based on the idea of film as a medium in it's own right (if one ever had any doubts), rather than as some substitute for some other medium, such as digital sensor imaging. But one that can be controlled by cheap digital control systems with otherwise intelligent software driving such.

The central idea here is simply a stepper motor, the same idea at the heart of 3D printing. Digital control over movement. The outcome is a programmable variable aspect film camera, minus a huge amount of complicated mechanics.

Is it, at least, conceivable? So far, yes.

$35

http://www.phidgets.com/products.php?ca ... _id=3317_1

Max. RPM (revolutions per minute): 904
Step Angle: 0.35 degrees
Rated Torque: 16.2 kg·cm

For torque comparison the recommended torque for Super8 film is about 6 g-cm, and this motor is 16,200 g-cm!

Assuming a drive roller of 10mm radius, based roughly on the radius of a 16mm camera sprocket drive, then

The roller circumference will be: 62.8 mm

This equates to: 8.4 frames (of 16mm film) per revolution of the roller

The maximum rate of the stepper motor is 904 revs per minute, or 15 revs per sec

This would equate to a maximum FPS, for 16mm film, driven by this stepper motor, being: 126 FPS (ie. plenty)

The step angle, for a 10mm radius roller, would give a corresponding positional resolution of 0.35/360 x 62.8mm = 0.06mm (1/123th of 16mm frame height)

Decreasing the radius of the roller will increase the positional accuracy of the transport and vice versa. But a resolution of 0.06mm would be fine. To obtain an exact match for 16mm film (ie. without any space lost between frames) we could just calculate a required radius for say 8 frames per rev

C = 2 pi r
r = C/2 pi
r = (8 x 7.49) / 2 pi
r = 9.53656419 mm

However there is a 5% error in the step accuracy (5% of 0.35 degrees translates to 0.003 mm in film positioning on a 10mm radius roller) so one can just leave a very small inter frame spacing to accommodate for this.

Now on further reading this particular motor has a gearbox and corresponding backlash error of 3 degrees at maximum drive, so a gear-less stepper motor will be a better idea:

$23

http://www.phidgets.com/products.php?ca ... _id=3308_0
Max. RPM (revolutions per minute): 2344
Step Angle: 0.9 degrees
Rated Torque: 8.5 kg·cm

The step angle is larger, by 3X. So this decreases the positional resolution by 3X to about 0.009 mm. No great loss
The better point is that it increases the RPM by 2.6X

So for 16mm aspect frame this gives a max frame rate of 326 FPS, ie. plenty

Running the motor at it's highest rate might induce some issues. The highest rate is just a statement of the motor's limit rather than a recommended usage rate. At normal rates, such as 24 FPS, I don't envisage issues. Not yet anyway.

The interesting thing about this camera idea is that the specs for 16mm film become almost irrelevant. One can just nominate one's own aspect ratio and digitally increase/decrease the frame pitch accordingly, with a corresponding change in gate. The concept of "Plug and Play" gates of various sizes can be used. Indeed, if the camera were to support lenses otherwise used on 35mm still cameras (for example), one could turn the 16mm film sideways (so to speak) and get some larger frame sizes, eg.

7.49 mm x 21 mm (2.8:1)
Max FPS = 45 FPS (first motor) or 117 FPS (second motor)

The important thing in all of this is ensuring that registration marks are encoded. It is this which makes the mechanics so relaxable and capable of supporting not just 16mm, but 8mm and 35mm as well. What would be interesting is a corresponding projector, one that used a real time mirror to re-register the projection based on the encoded registration mark. Can also double as part of film to digital transfer system, although the real time mirror becomes redundant as the image can be more easily registered digitally.


For slitscan photography this camera would also be useful. Perhaps the slitscan effects of 2001 were just a function of Trumbull otherwise working on a stepper motor camera.



A google search on slitscan gives some interesting results:



Carl

[carllooper]

Correction on second motor:

The positional rez would be about 0.156mm for a 10mm radius roller, or 1/48th the height of a standard 16mm frame height.
It is the positional error that would be 0.009 mm

Re. shooting sideways. From memory, spinning a wheel horizontally induces a force that wants to swing the wheel back into the vertical, ie. the force behind why moving bicycles don't readily fall down. But I doubt the film is spinning that fast to induce such a force. On a different front, I like the idea of a coaxial design (like Super8) in terms of ergonomics, albeit, at the expense of some more work on feed/take up guides.

What would be really interesting (if only for fun) is to design it for creation on a 3D printer (not the motor and lens obviously). One problem with plastic is it's transparency. Certainly white plastic is a problem. But probably black plastic as well. Will need to sus that out. Otherwise can coat it (paint it) in something. Just a small camera that runs on 25 foot loads. A proof of concept camera (for the stepper motor) but also a perfectly working one.

C

[MovieStuff]

If 8mm, just use a large diameter, rubber edge disk of about, say, 6 inches or more and wrap the film around that. Forget the pressure plate. The curvature of the disc will not be enough to affect focus and any minimal distortion will be either unnoticed or easily corrected in transfer to digital. The large disc will give positive traction and the disc could be direct drive off the stepper motor. Take the resolution of the motor's encoder and make the circumference of the film drive disc an equal division of that to ensure an equal and convenient progression/pitch of the film/frames. Use twin laser diodes to burn a single dot on each edge between frames each time the disc stops. Between the stepper motor and the take up, you'd have only two moving parts. Should run forever or until no film is available, whichever comes first. ;)

[carllooper]

If 8mm, just use a large diameter, rubber edge disk of about, say, 6 inches or more and wrap the film around that. Forget the pressure plate. The curvature of the disc will not be enough to affect focus and any minimal distortion will be either unnoticed or easily corrected in transfer to digital. The large disc will give positive traction and the disc could be direct drive off the stepper motor. Take the resolution of the motor's encoder and make the circumference of the film drive disc an equal division of that to ensure an equal and convenient progression/pitch of the film/frames. Use twin laser diodes to burn a single dot on each edge between frames each time the disc stops. Between the stepper motor and the take up, you'd have only two moving parts. Should run forever or until no film is available, whichever comes first. ;)

Cool. Thanks MovieStuff. I was literally drawing doodles of the same thing just five minutes ago - no pressure plate at all.

EDIT: just re-reading your post I realised what you are saying is even simpler than what I was thinking: ie. your 6 inch wheel is centred on the lens and presenting the film to the lens (and explains the size of the wheel and what you were saying about focus). That's a really interesting idea. In fact it's a brilliant idea. I'm going to do that for a first version camera, and put your name in the credits!

cheers
Carl

[MovieStuff]

Yes, I was using 6 inches as a minimum to prevent distortion or focus issues. The actual size would be based on an even division of the stepping motor encoder. The lens would be centered on the diameter of the disc, pointing at its edge. A flat belt stretched around the diameter of the disc would provide positive traction. The camera could also easily switch from regular mode to vistavision mode by turning the camera on its side and simply having the stepping motor move two increments instead of one.

Roger

[wado1942]

That's a pretty good idea and some professional tape recorders are driven the same way, like the ATR-102. My only reservation with this would be that the roller will shrink with age, reducing the distance the film travels per frame.

[carllooper]

Yes, I was using 6 inches as a minimum to prevent distortion or focus issues. The actual size would be based on an even division of the stepping motor encoder. The lens would be centered on the diameter of the disc, pointing at its edge. A flat belt stretched around the diameter of the disc would provide positive traction. The camera could also easily switch from regular mode to vistavision mode by turning the camera on its side and simply having the stepping motor move two increments instead of one.

Roger

Yes, for sure. Can make the diameter larger if need be.

I'll just make a really simple one first - on the 3D printer - and just burn some film through it and see what happens, in regular and vistavison mode! As long as the frame mask height corresponds to the length the stepper advances the film, that's all that need matter for a first test. For example, with a 0.9 degree stepper and a circumference that came out at say 500mm (for whatever reason), then the available frame heights the stepper could do would be in increments of 1.25mm (0.9/360 x 500mm) and I can just print a couple of frame masks at such sizes: 3.75mm, 5mm, 6.25mm ... 30mm

Will be interesting to see effect on focus in vistavision mode! I reckon it will look great however it turns out.

cheers
C

[carllooper]

That's a pretty good idea and some professional tape recorders are driven the same way, like the ATR-102. My only reservation with this would be that the roller will shrink with age, reducing the distance the film travels per frame.

I'm not too concerned with longevity of the camera. At my age (with my habits) my own mortality appears a sooner proposition than that of the rubber's demise. But a good point. Can work towards a more robust polished version in due course. By keeping it as simple as possible at the moment can get some photographic results sooner and get a better feeling for where to take it.

C