RED based 2k/4k film scan

[J. Bernard Vallon]

Has anyone else thought of this?

A RED camera attached to a projector with a nikon macro bellows would give you a really nice 2k/4k transfer machine.

Way cheaper than the monsters they use now, and very high quality at 12 bits.

Are there any practical problems with this that i'm not thinking of? When I was in film school i built one of these:

http://www.jkcamera.com/digital_printer.htm

It was kind of a pain, but I couldn't afford an HD transfer from S16 at the time, and it worked very well and very VERY slowly.

[Brook Willard]

Well it'd be great for poor-manning it at the very least...

[Thomas Mathai]

Ummm no, you don't need a motion picture camera, just a very good digital SLR.

I worked on one long ago using a Hasselblad medium format camera with a digital back on an optical printer bench. It was slow of course, but worked well enough for what was needed. It was a pain calibrating for motion picture negative and the scans weren't of any standard resolution, but it was for inhouse work.

If you want good scans, you want to use your negative instead of your print, and it should be a pin registered movement.

The link you provided looks similar to the setup I worked with. It's way cheaper than just the Red camera body alone.

[J. Bernard Vallon]

I dont like using a DSLR for this kind of stuff. When i did it for a short film it put 100k cycles on my S3, which is basicly the lifetime of the camera, seeing as how its only tested to 150k. It still works fine, but i wouldnt want to do it often. Also, the transfer took weeks to finish.

Doing it real time, or close to real time, is way better. Why exactly would a RED based transfer be inferior? As long as the film is within the dynamic range of the camera, and the lens is good and sharp, and film is pin registered, it should work fine in my mind. I dont know enough about telecine machines, if anyone wants to enlighten me.

[Cail Young]

The main problem would be syncing the shutters (or at least the pin lock on whatever's drawing the film across a gate with the RED shutter). Once that's solved, the rest is optics.

[Jim Arthurs]

John, that link to the JK products was a trip down memory lane for me... as I was quite the optical printer geek in my youth. I trained on an ACME 103 printer doing restorations, and I still own enough components to cobble up the mechanics of a scanner if I wished. I have a couple old ACME printer heads with half a dozen old B&H type one shuttles laying around.

For that matter, I've got all the gear to do old school Ray Harryhausen single frame process projection and re-photography from stem to stern. But the question is why would I? I bring this up, because to a degree, that's the same question you should be asking about wanting to build a film scanner.

I was on a mission to do this, maybe five years ago, and researched all the available imagers. The RED would be a very good solution and better than the DSLR's for the reasons mentioned, but not as good as a single b&w sensor triggering three successive RGB exposures via LED illumination. The reason for this is to overcome the orange masking designed in all negative and interpositive stocks to control contrasting in duping. A single mixed color light source just can't provide the same purity to the individual colors that individual exposures can. Understand that I'm talking about the "indy.. i.e. inexpensive" solutions... obviously with enough resources and testing and money you could work into a single light source solution, but at that point all the money-time-effort would have been better spent doing something else, IMO. My advice is to leave film to tape transfer to those who specialize in it, and besides, with the RED, who would WANT to mess with film anymore, anyway? The RED is sort of an end-around the whole mess in the first place.

[Jim Arthurs]

The main problem would be syncing the shutters (or at least the pin lock on whatever's drawing the film across a gate with the RED shutter). Once that's solved, the rest is optics.

Ah, but the devil is in the details... what kind of pin registered movement will he find that can run full speed and what kind of mechanical chain (pulleys/sprockets/reels) would be needed? At full speed, you're basically engineering a live action CAMERA that just happens to have a gate with a hole in it. There's a reason that most pin-registered transfers are done non-realtime. Anyone remember the Stedigate on the Rank? I even remember seeing a pinned gate system from Sony at NAB years past and it never got off the ground... there's just too much chance for film damage at that speed... especially in a home-brew indy system.

Big companys spend lots of money to solve these problems, and it isn't easy for them. The potential value for an indy to solve these problems and stand as a scanning service becomes less and less each year as the need for film scanning will decrease. I certainly wouldn't trust an indy to man-handle my negative... too much risk...

And lastly, what's the point in the RED in the first place? Film resolution without needing to shoot and scan film... then to want to use the RED to scan film... :)

[Thomas Mathai]

I dont like using a DSLR for this kind of stuff. When i did it for a short film it put 100k cycles on my S3, which is basicly the lifetime of the camera, seeing as how its only tested to 150k. It still works fine, but i wouldnt want to do it often. Also, the transfer took weeks to finish.

Doing it real time, or close to real time, is way better. Why exactly would a RED based transfer be inferior? As long as the film is within the dynamic range of the camera, and the lens is good and sharp, and film is pin registered, it should work fine in my mind. I dont know enough about telecine machines, if anyone wants to enlighten me.

If you do real time, you're not going to get pin registered film.

The idea of pin registering is to hold the film rock steady in the gate while it is scanning to cut out the inherent weave you'd get from real time scanning.

The most you can get with a pin registered system is probably a few frames a second.

The of course, are you going to be using print or negative? If you're using a negative, then you need a good optical printer set up that is scratch free otherwise your negative will get damaged.

If you use print, you're not going to get the same quality as from your negative. It's like scanning a photo print instead of the negative.

If you're fine with just using a print, then it's like doing a kinescope, so you might as well get one of those beam splitter setups they use for super 8 transfers. The dynamic range benefits will be lost too using a print.

There is software that triggers digital SLRs remotely, and some can just take the feed from the camera and capture purely through software.

If you use a macro program like QuicKeys you can trigger the software and advance the film in sync, but that takes experimenting.

Telecine is designed for real time transfers to tape or data, while film scanners are for slower but steadier transfers to data, at usually higher quality.

The Spirit is closer in quality to a film scanner, but I haven't tested it myself to know the details.

[Stephen Williams]

If you do real time, you're not going to get pin registered film.

The idea of pin registering is to hold the film rock steady in the gate while it is scanning to cut out the inherent weave you'd get from real time scanning.

The most you can get with a pin registered system is probably a few frames a second.
.

Hi,

I don't see why you couldn't use a Mitchell camera to advance the film. You could run at any speed within reason.

Stephen

[Jim Arthurs]

Stephen, I wouldn't ever trust original negative back through it... too many things can go wrong at full speed.

Yes, there are modified mitchell gates in use on process projectors, and a few very custom rigs built around actual mitchell bodies, but then you have to deal with the extra thickness of film splices, etc. all running at speed. I wouldn't ever trust original developed negative back through it... unexposed camera negative is one thing, as it doesn't have the accumulated "worth" of processed approved negative...

I've done LOTS of bi-pack work in the day... both with Mitchell and B&H mags... which is another case of running exposed negative back through a gate, this time in contact with raw stock. Hell, I even have a QUAD pack rig for B&H mags... You never escape these processes without a certain amount of dirt, micro scratches, etc.

Regards,

[Thomas Mathai]

Hi,

I don't see why you couldn't use a Mitchell camera to advance the film. You could run at any speed within reason.

Stephen

I'm sure someone has modified a Mitchell movement to use in an optical printer or projector, but I'm betting it's not cheap.

Just a quick web search mentions that a Mitchell movement goes for like $10k.

Then you have to know how to do all the modifications without damaging anything.

If this is meant to be a low cost alternative, then it'll be cheaper buying one of the JK Camera products.

It's easier to just get a beam splitter to project and shoot off of.

I would use a Mitchell camera for shooting an high quality LCD monitor to go back to film.

I missed out on the optical printer days. My experience started with the digital scanners like the Kodak Genesis. The only optical printer I saw was sitting in closet.

[Noah Kadner]

It's a cool one of those what if's but when you think about the expense of the gear involved and the hassle just to get something usable(ntm the potential risk to your negative) vs. the cheap costs of high quality telecine and scans these days it seems a bit of one step forward two back.

[Stephen Williams]

I

Just a quick web search mentions that a Mitchell movement goes for like $10k.


Hi,

They seem to sell on Ebay for about $350. Complete cameras about $2000 +

Stephen