Jump to content

16mm pc film scanner used as a telecine


Jonathan Bryant

Recommended Posts

  • Premium Member

Kodak actually had a low cost desktop telecine for Super-8 film, the Kodak Supermatic VP-1 videoplayer. Used a flying spot scanner, gave pretty good quality, and only cost a few thousand dollars. Introduced in August 1973, not many were sold.

 

http://tig.colorist.org/archives/public/ti...une/000935.html

Link to comment
Share on other sites

The question is that if it so obvious, then why isn't it already on the market?

 

Scanning at a decent quality level at practical speeds for motion picture work with enough steadiness is not as cheap to build as you think, and if sales aren't high enough to recoup development and manufacturing, no one is going to build it... or at least, not at the price levels that some of you expect.  Something limited to a niche market will naturally be higher.

 

I mean, if it cost $15,000, and still was lower in quality than a modern telecine, how many of you would buy it?

Speed wouldn't be such a factor for me. At our production house we have lots of different editing stations where we will be editing on one,capturing on another, and rendering on another,etc..

 

You could leave it scanning over night and even during the day since if it could be a dedicated machine.

 

I think you could somehow scan the registration holes and let the software do some stabilization. Just like the Avid plugin where you tell it to track something still and make the rest of the shot steady.

 

As for color just shoot a color chart and have a color correction tool where you have a color picker and tell it which swatch it is. I think that would give you decent color. Of course I would never discount a talented colorist.

 

I am not saying that I am anywhere near smart enough to know how telecines work or for that matter build one. Just wishful thinking is all.

 

John, Thats kewl to find out that Kodak actually had a low cost Super 8 telecine. So how many people would have to be interested in buying something like this for a company to try and develop it?

 

Here in the AtL Spirit Telecine rates are 650 an hour Gulp :huh: http://www.cinefilmlab.com/services_telecine.htm

 

Then there is these guys with the Rank Turbo for only about 135 an hour

http://www.posthouse.com/

 

How much better is the Spirit in the right hands?

 

This low cost solution called the Movie Stuff Sniper http://www.moviestuff.tv/16mm_telecine.html

was tested on the same footage as the Rank and the results weren't that much different.

http://www.pcmag.com/article2/0,1895,1819174,00.asp

Link to comment
Share on other sites

  • Premium Member

Hey johnhollywood,

 

I can see that your going down the same path as me. Here's the direction I'm going. I need telecine for workprint only. So, I shoot an XL2 into the gate of my Mitchell. I'm in the process of getting my Mitch modified by Ken Stone to have a see through pressure plate and a gate plate cut out to show the edge numbers as well. The XL2 has an EF adapter with a standard Canon 50mm lens. This configuration covers the film plane including the edge numbers. When shooting 16:9 I get all the pic info and frame number matched visually. The XL2 does both 3:2 and 3:2,2:3 pull down in camera. When the workprint is cut I match up a Kodak DSC SLR with macro lens to shoot into the gate of the Mitch. I'm still working out a cheaper intervalometer. With a little software tweaking, I can use a computer to coordinate the IV and Kodak DCS as a film scanner. This should give me 5K Bayer pattern scans.

 

This system is full of compromises compared to big budget lab gear. At the same time, it's a damn sight cheaper than big budget lab prices. I'm still waiting for the chuckhole that makes this impossible, but I haven't hit it yet.

 

Is this any use to you?

Link to comment
Share on other sites

  • Premium Member

There's more, JH,

 

I got your que through your computer layout. I'm setting up a render farm. I have six computers that act as departmental work stations by day and render farm by night. I'm using PCI SATA ports to bypass the high cost of SATA arrays. I have gotten the SATA drives from Tiger Direct at 200Gig for around $0.50 per gig. THe Motherboards all have Gig-LAN on board and I'm routering them together. XP can handle data flows. Get a buddy in school to get your software on educational discount. You can get MAYA 6.5 for under $400.00. It usually goes for around $6,200.00. You can also snag Adobe Video Collection the same way. I'm already in school, so, it was an easy deal.

 

The only trick I haven't found a passable solution for is film recording. I'm working on that as we speak.

 

I admire your inventiveness and determination to get things done in the face of overwhelming and prohibitve production costs. Best of luck to you.

Link to comment
Share on other sites

  • Premium Member

Hi,

 

It's offtopic, but the one thing to watch with those cheap PCI SATA RAID controllers is that most of them offload a lot of the grunt work onto the host system. It's probably a good way of working through data in a scanning sort of situation, where disk bandwidth isn't the limiting factor, but if you end up trying to record 2K onto them at 24fps, you'll completely thrash the processor on the host.

 

The only plug-in cards I know of that don't do it this way are 3ware's 9000 series, which are several hundred units of currency apiece. They have to be, as they're required to do the number crunching onboard. Other than that, as you've probably surmised, you're into fiberchannel to a big expensive array.

 

Not really necessary for a scanhost, though. Any competent single drive should be able to write 12Mb 2K files twice a second.

 

Phil

Link to comment
Share on other sites

  • 2 months later...
Guest dokworm
Richard,

 

Are you familiar with MovieStuff's telecine machines? I think they're about as close to what you're describing as we're likely to see.

 

-Bon

 

Makng your own HD telecine rig isn't hard at all.

 

Start off with a Super8 or 16mm projector that has the driveshaft running to the back of the machine.

eg. like the Eumig 610D.

Then you can just change the thumbwheel type knob on the back of the projector, to one shaped like a cam. (i.e. a wheel with a 'bump' on it).

No you can trigger the computer every time the shutter is open - once per frame of film.

Then replace the projector light with some white LEDs and a piece of Opal Glass as a diffuser. You now have a cold light source that cannot hurt the film, and is more than bright enough for telecine.

 

Remove the projector lens and fit a CCD camera (with suitable lens) to the front of the projector, pointing at the film gate.

You can use any of the multitude of ultra hidef camera backs in USB2 or firewire (up to 20 megapixels!) or use a simple webcam like the Philips toucam or the new logitech one with true 1024x768 sensor if budget is a problem.

 

So now you have a hidef camera focused directly on the film frame, and is connected to the PC via firewire or USB2.0

You also have the PC triggered by the microswitch mounted on the projector. Use cinecap software (free) or similar, and away you go.

You should be able to run the projector at around 5 fps with a decent camera.

It runs itself after that point, the PC captures and saves an image each time the shutter makes one full revolution - so you can capture a 10 minute film in under an hour.

 

It can all be done with off the shelf products, and you end up with a very high quality, reliable system.

 

It is an easy job to widen the gate on most projectors so you can capture the full frame.

 

So with a bit of work, you can build a better unit than a sniper or workprinter for a lot less bucks.

Even if you want to go to better than HDTV resolution, you are only looking at around $1200 for the camera back, and as low as $100 if you want SD resolutions.

 

There is a site here with some pics

http://users.telenet.be/ho-slotcars/s8_index.htm

He also has another using a higher resolution camera here.

http://users.telenet.be/ho-slotcars/s8_telecine_imaging.htm

Edited by dokworm
Link to comment
Share on other sites

  • Premium Member

Hi,

 

I'm not sure if I'd quite trust a $100 camera in this application, at least not without a double exposure with variable lightsource to capture more dynamic range.

 

But yes. Doing this on a flatbed scanner is nuts; the hard bit is the pulldown, so you're much better off shooting a camera into the gate of a projector.

 

Phil

Link to comment
Share on other sites

Guest Ian Marks

Interesting thread!

 

I've been tinkering with an idea for a cheapo digital film scanner. I have most of the bits and pieces (except for the digital camera body) but haven't put anything together yet. It may work, or it may be a complete bust, but here's the basic idea:

 

For a film movement, I'll use an old Kodak Cine Special camera with the taking lens and shutter removed (the shutter becomes superfluous with this set up). The Cine Special is a spring-wound camera which provides a 1:1 shaft and does single frame. It uses a detachable film magazine which incorporates its own pull-down claw and gate (the gate will need to be ground out a bit to expose the entire image area for super-16).

 

There are single frame animation motors around for the Cine-Special, but I was thinking I would drive the camera with a low-torque motor with a slip clutch which will apply a slight constant rotational force to the drive shaft. The shutter will be tripped via a wire pulling down on the shutter release (on this camera you push up for continuous running, and pull down to expose a single frame). To start, the wire can be actuated by just about anything - a cogged wheel mounted to a slow AC synchronous motor, a solenoid, a foot pedal, whatever. Obviously I'm describing a very low-tech arrangement.

 

Most of the magazines for the Cine Special are for 100' foot spools, which obviously are too short, but I have a magazine which has a 400' Mitchell magazine grafted to the top. Some kind of torque motor will be required for film take-up too - not a big deal.

 

The Cine Special mags, including the one I'll be using, all have a swing-out door, and with the door open and the 100' spools removed, it should be possible to use an LED light source behind the film, or possibly piped in from the side and directed towards the back of the film path using a small 45-degree mirror. The removable pressure plate will have to be replaced with something which presses against the film (gently, of course) immediately above and below the picture area, but which leaves the image area open so that it can be illuminated. If this sounds like an old fashioned rotoscoping set-up, you've got the idea.

 

The Cine Special will be mounted to a heavy base opposite a Nikon digital SLR with a bellows-mounted macro or enlarging lens attached. A couple of years ago I picked up a macro lens which someone had constructed by mating a reversed 13mm Switar lens to a Nikon lens mount. This has worked extremely well for shooting 35mm stills, and I'm hoping it will work well on the scanner. If not, I'll try using a Schneider Componon-S enlarging lens.

 

The digital camera rig will be mounted to an micrometer optical bench thingamajig I picked up off of Ebay which permits the camera to be repositioned along any axis with great precision and then locked off. It will aimed directly into the open lens port of the Cine Special. Images will be shot out to a hard drive via a USB cable and saved as numbered files. Scanning a 400' roll would not be quick (several hours or more), but I'm not a lab and don't need to do this in anything approaching real time. If I can set the process in motion and then walk away, that's okay by me.

 

In the past, I have taken numbered digital stills and pulled them into After Effects to form a "film clip," so in principle it should be possible to assemble the thousands of stills from a 400' roll of 16mm. I realize that the registration on a 1930's-era camera leaves much to be desired, but I believe I can use After Effects' tracking feature to stabilize the footage should this be necessary. Of course, After Effects can trim, apply effects, crop, and reformat the footage, providing (I hope) a poor man's path to a digital intermediate. As a bonus, I can perform a 2:3 pulldown in After Effects to output 24fps film to video should I choose that route.

 

If initial tests are encouraging, there are a number of refinements I could add to improve the machine. Obviously, some kind of computer control to vary the interval of the exposures (which will depend upon the selected resolution of the frame, the speed of the hard drive, etc.), and maybe some kind of rig to hit each frame with a puff of compressed air just prior to exposure. Also, I thought the entire contraption could be encased in a plexi box with some kind of ionic dust catcher inside. I've done plenty of digital "dustbusting," and it's no fun.

 

Any thoughts?

 

 

I shouldn't have said the "shutter would be tripped" on the Cine Special - more accurately, the film would be advanced by one frame by pulling down on the release button. The only shutter in this set up would be the one in the digital SLR.

Link to comment
Share on other sites

Hi Ian,

 

The topic of a home-built telecine is an interesting one. It?s motivated by the realization that DI (digital intermediary) is essential to the indie filmmaker who wants to compete on a world level and that DI is a very expensive process.

 

More often than not, DI is too cost prohibitive to be considered. Our local film lab charges $2.50 per 2k frame scan. If I shot my indie feature on 35mm at a tight ratio of 5:1, I?m looking at a book rate of $1,764,000. And sure, the lab will be more than happy to cut me one heck of a good deal on the bulk. But even with an unlikely 99% discount it?s still more than I can afford.

 

You might argue that the lab down the street will transfer my film to 24p HD, but geez?that?s expensive too and what am I going to do with the HD tape? I still have to get it off to a hard disk. Renting or worse, buying a deck and an HD edit system is out of the question. Let?s not even talk about the fact that even at 1080i, it?s still ¼ the resolution of a decent 2k scan. Probably less?due to the interlacing factor.

 

It?s a problem that has dogged me for the past few years, but one that I?ve only recently started to think about seriously.

 

The system I devised is quite similar to yours. It uses an old Mitchell 1000? magazine, two film sprocket assemblies, and a film gate with a pressure plate, a high power strobe and a Nikon D2 DSLR on a macro bellows.

 

The strobe projects diffused light through the film negative and into the D2 camera at the end of the bellows.

 

The sprocket assemblies hold the film steady and registered at the head and tail of the film gate/pressure plate. Additionally, the film gate/pressure plate also contributes to holding the film steady and registered (thus no need to stabilize in post). The film gate also acts as a mask?keeping in mind that the aspect ratio of the frame being shot by the D2 is twice as wide as the film frame on the negative (you could get real creative and ?scan? two frames simultaneously and use a Photoshop batch process to separate them, thus reducing your scan time by half but increasing your post-processing time).

 

A 1.8 degree (or a 0.9 for finer control) stepper motor is attached to the Mitchell?s take up reel and drives the system. The stepper motor is connected to a PC via the parallel port. A very simple BASIC program sends pulses to the stepper motor (or, depending on the power requirements for the motor to a relay and then the motor). To determine the number of pulses per frame, you just need to do a little initial experimenting (you could try to calculate it out, but man?math!). Obviously, it?ll be dependent on the gearing between the stepper motor and the mag.

 

Let?s say it?s 200 steps per frame. The program, upon having sent 200 pulses to the motor, pauses a moment and then sends a pulse to a relay connected to the 10 pin cable of the D2. The relay closes the trip release circuit of the D2?automatically firing a frame and thus simultaneously firing the strobe. Again, there is a built-in programmatic delay to accommodate for the time needed for the camera to buffer and transfer the data down the 10 pin cable and into a hard drive. The program then repeats itself.

 

At a generous 3 seconds per frame (the time I estimate you would need to advance the frame, take the shot, advance to the next frame, recharge the strobe and transfer the frame downstream), with the D2 configured to 10 mega pixels, you?d be able to scan 1000? (16,000 frames) every 1.5 hours. Not bad.

 

Using PhotoShop and its batch processing, you?d create a batch that automatically crops the wide frame to correct size, inverts the negative and saves it to a new location. The speed of this process will be dependent upon your processing power and hard disk space.

 

Calibrating the alignment of your frames is a simple process?just simply turn the take-up wheel on the mag by hand until the film is correctly aligned in the gate. Set this as the Zero position in the software and blast away.

 

The biggest problem in this scenario rests in color calibration. How can you ensure the D2 sees the film negative the same way intended as the camera that shot the negative in the first place? Perhaps the easiest way would be to film a full gray card at the head of each new scene during production. Even a few frames of correctly exposed gray card would suffice. The gray card will provide you with the white, black and gray points needed to calibrate the histograms of the D2 to ensure that you?re pretty spot on. You could also shoot a Macbeth color chip chart along with the gray card. During your post process, you could use the chip chart to fine tune the color correction.

 

However, will the quality of such a system be high enough for VFX work and film delivery? The system isn?t really that much different than an optical printer or a slide/negative duplicator. But I wouldn?t expect Spielberg to be my new best friend. That being said, the quality should be decent enough to be more than acceptable to the vast majority of indie filmmakers on a budget.

 

Comments? Ideas?

Link to comment
Share on other sites

Guest Ian Marks

TOMR -

 

What you've done is just about exactly what I have in mind, although your ideas are well ahead of mine in terms of sophistication. I don't seen why you shouldn't be able achieve very good results with your set up, especially with all of the image enhancement tools available in the digital realm.

 

I'm hoping that I can put a similar rig together using mostly bits and pieces I already own (although I hate to tear up my old Cine Special). The weak point seems to be the digital SLR's shutter. Are you locking yours open and using only the flash for expsure? Can you lock your camera's shutter open without damage (or battery drain?).

Link to comment
Share on other sites

  • Premium Member

Hi,

 

You'd do much better I think with one of those monochrome machine vision cameras. However, the cost is high for what they are - at least two grand a shot for something reasonable, and eight or ten for one that'll shoot 3K frames. They are very fast, though, and you could get a full 4:4:4 scan.

 

I want a digicam with a mono sensor. I think it'd almost be cheaper to get hold of a pocket digicam and rip the guts out of it since you'd not be using so much of an expensive DSLR.

 

Calibration is a complicated issue; you'd almost certainly end up doing a lot of work on the stripped-down RGB channels from a RAW image to establish a LUT that would make things approximately normal. To establish the linearity of the sensor, you'd have to either expose a known linear ramp from fully over to under, or take a lot of successive exposures with progressively larger amounts of ND and interpolate between them. With these techniques, you could establish the linearity of the camera; then, do the same with a software-controlled lightsource, and you have the information to create a two-pass scanner capable of automatically assembling two varying exposures for very wide dynamic range.

 

Phil

Link to comment
Share on other sites

  • Premium Member

Hi,

 

If you did that, you'd probably end up backlighting it with a xenon strobe. This is good for several reasons - it's very brief, so you'd have the best chance of getting away with continuous movement. It's also very blue, which helps with the orange mask, and it's probably bright enough.

 

Phil

Link to comment
Share on other sites

  • 4 weeks later...
I recently saw an internet page where the guy make a DIY telecine with a Eumig ST 610 D super 8 projector. I was finding but I can´t find. Someone knows the link to this page?

 

thanks

 

Hello everyone, my very first post here. Nice to meet you guys.

 

Back to our topic: not sure if this is the website you meant.

 

http://www.moviestuff.tv/

 

;)

Edited by Siah Leng New
Link to comment
Share on other sites

Guest uoliwils

I already tried using using a DSLR camera to scan super8 film.

 

I didn't modify the projector to run at a slower speed, I'm just moving the film manually. The lens comes from a B&W darkroom enlarger, it's a Schneider Componon f/4, which is a very good lens.

I choose this one because I thought these lenses were made to give the least possible distortion and maximum sharpness. Also, it's the only one so small to get really close to the gate of my projector.

 

The results were decent, but still not as good as a rank transfer (I compared the same frames transferred by Andec).

 

I tried to get the best possible focusing, but nothing changed.

 

I suspect that because the camera (nikon d70s) doesn't have the mirror lock-up feature, the frames will never look sharp.

 

Now my device is disassembled because I need my camera for its main purpose (taking stills...), anyway I'd be really curious to see a similar setup with a different camera (the 350d for example has the mirror lock up feature).

 

post-3745-1134914857.jpg

 

Andres, this may be link you're looking for: http://homepage.mac.com/onsuper8/diytelecine/

 

David

Link to comment
Share on other sites

I already tried using using a DSLR camera to scan super8 film.

 

I didn't modify the projector to run at a slower speed, I'm just moving the film manually. The lens comes from a B&W darkroom enlarger, it's a Schneider Componon f/4, which is a very good lens.

I choose this one because I thought these lenses were made to give the least possible distortion and maximum sharpness. Also, it's the only one so small to get really close to the gate of my projector.

 

The results were decent, but still not as good as a rank transfer (I compared the same frames transferred by Andec).

 

I tried to get the best possible focusing, but nothing changed.

 

I suspect that because the camera (nikon d70s) doesn't have the mirror lock-up feature, the frames will never look sharp.

 

Now my device is disassembled because I need my camera for its main purpose (taking stills...), anyway I'd be really curious to see a similar setup with a different camera (the 350d for example has the mirror lock up feature).

 

post-3745-1134914857.jpg

 

Andres, this may be link you're looking for: http://homepage.mac.com/onsuper8/diytelecine/

 

David

 

yes you´re right. Thanks a lot. ;)

regards

Link to comment
Share on other sites

I was the guy who built the Eumig based telecine device as I figured there must be a cheap way of getting my Super 8 stock into DV for editing. The system works well for scene by scene transfers (which is what I do) but not particularly fast (no better than 6fps). The advantages are its cheap and results in very stable images, but it could be improved by doing away with the lens and using a fixed camera set up and macro lens focusing on the frame. Here's a sample image (Super 8 Kodachrome 40 shot on Nizo 801) which I dont think is too bad for less than $100 and a little time and effort.

 

Giles P.

www.onsuper8.org

onsuper8.blogspot.com

 

Everything Super 8 in today's digital world

 

 

post-1488-1135117680.jpg

Edited by onsuper8
Link to comment
Share on other sites

  • 9 years later...

I cannot help but notice that instead of answering the question that was asked, you pointed out the folly of even asking such a question. How is that helpful?

 

To answer the question, right now there is no consumer-grade unit that allows you to feed film in and scan it frame-by-frame. I also wish there was. Most folks either use Telecine (expensive) or the DIY solution, flatbed scanning and software to assemble.

 

Best of luck.

 

For 24 fps material, that's 14,400 individual frames you'd have to scan just for ten minutes of footage.

Even if you only took ten seconds to scan each frame individually by hand, it would take 40 hours -- basically two days straight -- to scan that ten minutes of footage.

Link to comment
Share on other sites

A better flatbed will fix this problem. Instead of scanning at low numbers and then using interpolation to get your frame size, you need to scan at higher resolutions, 9600 being a good starting point. If your flatbed scanner cannot scan at this resolution, it's time for a new scanner.

 

Also, using Photoshop is way too tedious. There are software solutions that automate this process. Cine-to-Vid comes to mind.

 

 

If you use a flatbed you can get about 15 frames in one scan, and cut them in photoshop. Also a very long job.

But a flatbed will give you troubles with sharpness because 16mm needs so much enlargement, even for SD.

Link to comment
Share on other sites

Take a look at the date on this thread - 10 years ago almost to the day! Things have changed a bit in the past decade, so it seems unfair to criticize someone for something they said back then...

 

That said, while scanning motion picture film on a flatbed is doable, it's a needlessly tedious process. There are much better ways to handle this kind of thing in 2015 using simple, cheap motion control systems, LED lighting and industrial machine vision cameras, scanning frame by frame, and avoiding the need for all all that post-scan processing work.

 

Also, telecine for 8mm and 16mm films is basically done at this point. There are several scanners on the market now that can do 2k or higher resolution scans for small gauge film, and all do an excellent job. Anyone using a telecine for this is pretty much limited to HD at the most, unless it's heavily modified, and even then, it's subject to all the limits of that kind of a system.

Edited by Perry Paolantonio
Link to comment
Share on other sites

Yeah I have looked into their line of machines. Interesting concept but, they are telecines and you can't do S16 with them. I'm speaking of a mid-range scanner where you scan the whole frame, wheather 16 or S16 and not lose information by going into a camera. Where you are not dumbing down the frame by the camera's limits. Your only limit would be drive space.

 

Richard Mills

Cinematographer, Photographer, Editor

Phoenix, AZ

Yes you can do S16 with Moviestuff units. The new Retro16 does Reg/S16 or the Retro Universal units can do multiple formats. They are super easy and very cheap compared to high end machines, not to mention the quality is excellent. They scan at 10-15fps frame by frame and output image sequence files. The next step up from Moviestuff would be the new Blackmagic Cintel that runs around $30,000.

Edited by Anthony Schilling
Link to comment
Share on other sites

Create an account or sign in to comment

You need to be a member in order to leave a comment

Create an account

Sign up for a new account in our community. It's easy!

Register a new account

Sign in

Already have an account? Sign in here.

Sign In Now
×
×
  • Create New...