2009 April 11th, 07:35
Just wondering if I can do this with my HV20:
I'm exploring the possibility of transferring 16mm film footage frame by frame into a PC using the HDMI output of my HV20 (or maybe utilising the camera's stills option). I'd have to automate this -15 minutes of film transferred at a frame/second would take 6 hours!- using software like Stop Motion Pro. Someone's done this with a DSLR and a modified Mitchell 35, but rather than use an intermittent film movement, I was thinking of using a continuous film movement, a flash, and an HD camera (a bit like Flashscan).
2009 April 11th, 22:37
We have a couple of threads on the subject: here's one, I'm sure a search will turn up others ...
2009 April 12th, 08:32
Well, it would be nice if the links in those threads worked, particularly that DIY one everyone keeps referring to. From what I can glean it involves using a Eumig Super-8 projector.
However I'm talking about 16mm (inc Super 16), and not using an intermittent transport that damages the sprocket holes, but instead uses a flash as a means of grabbing the frames. I thought it was maybe a bit of a wild idea until I saw the Flashscan. Someone's already doing it - but at £25,000 for a S8 unit!
So... can anyone think how one might grab, and dump to PC on the fly, successive images illuminated with a flash? They use DSLRs on stop-motion animated films these days. But can the HV20 do it..?
Can it focus down to 16mm ( 0.404 by 0.295 in (10.26 by 7.49 mm)) and Super16 ( 0.493 by 0.292 in (12.52 by 7.41 mm)) with no vignetting etc.? Super16 is widescreen, so there'd be no loss of image area.
Anyone have any ideas on what flash system the Flashscan is using? Most photographers' flashes have an error pop-to-pop of over 10%! Cheap ones 50%. Clearly Flashscan has gotten round that - even at 50 fps! My guess is they use proprietary LEDs, but then I thought those had poor spectral distributions...
I'm wondering if the transport from an old magnetic tape reel-to-reel could be adapted...? Maybe swap out the rollers for 16mm ones cannibalised from a bust projector? Or even use a 16mm projector and take out the intermittent transport, but retain the gate?
But how would I trigger the flash and camera/PC? Some kind of sensor looking at the sprocket holes?
One worry is digital noise. I've heard this from a number of sources. Essentially digital cameras aren't designed for this - read the Bride thread below. But the HV20's HDMI output should be ok..?
But my biggest worry here is dust. How do you clean moving film without scratching it? Flashscan say they use a light diffuser to get around that. Can anyone tell me how that might work..?
The potential here is to make a cheap telecine that would substantially cheapen the production of low budget Super16 productions.
'White' LEDs http://ledmuseum.home.att.net/ledwht.htm
Image stabilisation: http://toolfarm.com/plugins/index.ph..._Bullet_Steady
Using a DSLR on Corpse Bride: http://www.stopmotionworks.com/artic...dstrpdbare.htm
Stop Motion software: http://www.stopmotionworks.com/stopmosoftwr.htm
Single perf Super 16:
Last edited by Karel Bata; 2009 April 12th at 08:46.
2009 April 12th, 09:34
Sorry that post went on a bit - kind of like thinking out loud. Still, you get the picture..?
2009 April 12th, 20:59
The Canon certainly CAN shoot to those specs if that's what you want to do. You just need the right lens attachments, although you'd be shooting with the HV as a still camera rather than a video camera. However, to be honest, I think you're kind of going about it backwards and trying to make something that's very simple and 19th century more complex and expensive than it needs to be.
Originally Posted by Karel Bata
First, the way those scanners work is by reading the entire frame and using the sprocket holes as a cue as to when to scan/grab the image into memory, so you're looking at a sophisticated integration of photo-optical, mechanical and image processing system... and let's not even mention that most of the better units work with the film in a fluid environment to fill in the imperfections of the surface of the film (ie scratches, puckered sections due to moisture, etc,) and make multiple passes to obtain YCbMs (yellow, Cyan, Magenta color separations) of the image.
The key thing to remember is that systems like that don't make the transfer any better; and the best machines are still built to go very slowly, one frame at at time. The systems you're talking about are built to be FAST, and all that additional complexity is really only worth the cost to commercial labs that have to turn around hundreds of hours worth of transfers a week, usually overnight. The old adage holds true: you can have it done cheap, you can have it done fast and you can have it done well... but you can only have two at any given time (at best.) Because the cost of telecine is only a tiny part of the budget of most commercial projects, professional transfer houses have opted for the latter two, speed and quality, and try to work off the cost of the equipment over multiple jobs.
Second, what causes damage to sprocket holes & film isn't the transport mechanism per se, but rather the rapid repeated impact of the pull down/release/pull down cycle of a traditional "Maltese Cross" mechanism, coupled with the intense heat that builds up in the film gate area that can fade, burn or buckle the film. (You also have to deal with the friction of the film's surface against the projector gate.) However, if you slow the process down, the physical wear on the film drops down to practically negligible, while dropping the amount of light you're shoveling through the film minimizes the heat damage. Trade-off: you lose speed, but you gain cheap and maintain good. if you want speed, suck up and acknowledge that if you want both SPEED and GOOD, CHEAP goes out the window, and the least expensive way to get the work done will be to hire a professional service. Believe me, I've had a lot of 35mm and 16mm transfers done in the past and a very good team of engineers analyzed all options and even for one project with 35 hours of 16mm prints, it was still cheaper to hire it out.
However, if you're trying to DIY, I'm guessing that you don't have thousands of hours of footage to transfer. If you have HUNDREDS, I recommend you look at this link:
While still pricey by amateur standards, Moviestuff's workprinters compare quite well to a Rank Cintel transfer and is probably the cheapest semi-pro product on the market. Even if you don't want to shell out the cash for one of those, they still make a good model to use as a base. It's a pretty standard process:
Buy a cheap 16mm projector off e-bay.
Rig a way to copy the film at a 1:1 ratio.
Rig a way to advance the film frame by frame.
Rip out the lamp assembly and put in a cooler light source.
Take shots of the film one frame at a time, load the files into aftereffects or whatever and stitch it back together.
The easiest way to do all of the above is actually surprisingly simple and counter-intuitive: set your camera up to shoot the film gate from the INSIDE of the projector, rather than through the lens barrel. Basically, just take a junk projector and completely remove both the lamp assembly and the back part that covers the motors. Next, remove the blades off the three bladed shutter mechanism (5 blades if it was a NTSC telecine projector and 2 blades, IIRC, if it was a PAL telecine projector, no blades if it was an analytic projector.) if projector had a "heat shield" for single frame and slow motion, usually just a metal screen punched with holes, pop that out as well. If any of the above can't be unscrewed easily, just cut them off with a hacksaw or tinsnips; this projectors days "projecting" are over. Now, with the lamp removed, you should have an unobstructed view of the film gate area from the back, although most who do this mod will go the extra step of filling the gate area out to cover the full width of the film. This will let you set your camera up to shoot the film plane directly, though in some cases it may be necessary to place an achromat/magnifying lens and/or small front surfaced mirror at a 90 degree angle to the gate in order to get the best view. For the light source, just aim the projector lens at an even white subject like a piece of illustration board. For frame advance, go into the back of the project, disconnect the motor from the drive shaft and instead use a slow turning motor (like from a barbecue grill) to slowly advance a frame at a time (or if you're really being cheap, a hand crank.) Clamp everything together so there's no movement and you've actually got a pretty impressive film copying device for less than $50.00. At this point it's still manual, and how you go about automating the process will vary depending upon the projector and your camera's external control capability. However, given that any projector will have a large number of turning gears that move in precise synch with the advance of each frame, you should have a number of places to tap into.