Digital Sound Tracks and Work Prints

[Konton]

For all those crazy people like me still interested in making a film to be edited and kept on film, let me ask you something. Have you ever considered buying a Workprinter to make a digital work print of your film for editing, and then use the information to edit the original film? And then make an optical print with your Workprinter?

As many of you know, not only can the Workprinter be used to transfer your film frame-by-frame digitally, but it can also be used to make optical prints by simply connecting your Super 8mm camera to the trigger on the Workprinter and using a brighter external light to display the image (I'm sure Roger can explain better than I can).

With these two available functions, it seemed to me that it would be great if you could also make a digital transfer of your film containing FrameCode. FrameCode would be the feet and frame information of the film being displayed. As you know, Super 8mm film contains 72 frames per foot. A Super 8mm gang synchronizers keeps count of the feet, as the wheel itself shows what frame is being counted. Thus with a digital work print that contains FrameCode, you could sync your audio digitally, edit your work print on the computer, and then use the information to edit the original film. After the film is edited you can using your digital work print as a guide (the FrameCode being the reference) and the gang synchronizer to locate the proper frames.

I contacted Jeff about the Dotcap program and he think it would be possible to add the option to use FrameCode if people wanted it. If any of you are interested in this idea please let me and him know!

Last of all is the issue of synchronizing the audio. Even if sound stripe and sound stripers were easily available and not prohibitively expensive, how do you make sure the audio is transferred at the right speed?

To me this is where Roger's Synchronizer P1008 comes in handy. With the properly modified projector, it will keep the running speed in sync with a 1khz burst track. The projector will not even start until it hears the first burst. So burn a CD of you audio with the 1khz bursts on one track and you audio on the other. Then connect the CD player to the Synchronizer and you're good to go. All you need to do now is know what frame to cue with.

Alright, just wanted to toss those thoughts out there. Let me know what you all think.

[MovieStuff]

To me this is where Roger's Synchronizer P1008 comes in handy.

Um, er. I think you mean Pedro's Synchronizer. I haven't built a synch unit.....yet!

Roger

[mattias]

Have you ever considered buying a Workprinter to make a digital work print of your film for editing, and then use the information to edit the original film? And then make an optical print with your Workprinter?

yes, i have tried this as a test (it worked), and i'm going to do it with my current festival short when (not if, mind you) ;-) somebody buys it so that i can afford it. the thing is that i don't use a wp, but have a lab do the optical stuff for me.

i rank transferred the footage to dv and edited that, and i will cut the original by counting frames, just as you suggest, but by using a simple ruler and count from the previous flash frame, easily identifiable on both film and video. my camera, like most super-8 ones, doesn't produce real flash frames, but it's still not a problem with reversal where you have the visual feedback too. i then blow this up to 35mm on an optical printer for theatrical release. et voila. this whole process could be done just as easily or more easily with a workprinter, and of course back to s8 intead of 35.

i also use a technique called skip splice printing, which gives me cleaner cuts by overlapping every edit by 8 frames and skipping those when printing. this could be hard to do on a wp since it's not computer or even mechanical counter controlled, menaing you would have to manually advance frames while keeping count, but such a thing could easily be built, right roger? ;-)

/matt

[MovieStuff]

i also use a technique called skip splice printing, which gives me cleaner cuts by overlapping every edit by 8 frames and skipping those when printing. this could be hard to do on a wp since it's not computer or even mechanical counter controlled, menaing you would have to manually advance frames while keeping count, but such a thing could easily be built, right roger? ;-)

/matt

Actually, I already have such a device on the drawing board and shared the idea previously with both Andreas and Konton. The unit is a small interupt circuit that goes between the WorkPrinter and the device being controlled; either the mouse or a film camera set for animation as an optical printer. You would use a standard Ciro splicer but, instead of using the clear tape normally employed, one would use the silver sensing tape often used for security systems or for masking slides. This tape has a highly polished chrome surface on one side and adhesive on the back. When editing your film, you find your cut point and add five frames. This means that there is always five frames at each end of your selected film clip. Together, that leaves ten frames seperating each targeted selection of film.

A small sensor goes on the WorkPrinter just before where the film enters the projector gate, exactly five frames from the gate opening. A small LED and sensor will pick up the silver tape and a counter will disengage the trigger circuit for exactly ten frames. This will let the splice clear the gate so that you don't get any jumping of the image. Then the circuit re-establishes contact with the capture device and the copying resumes. The film is effectively cut together "in camera" with no visible splices.

The beauty of this system is that you can make an automatic trial run by copying to your computer first. Once you are sure that you like where the edits are, all you do is load the WorkPrinter with your original, use a film camera instead of a video camera, and copy your footage frame by frame to another reel of film. The "Auto-Cut" sensor will automatically skip the pieces of film you do not want and only allow the camera to copy the pieces you do want.

Anyway, that's the idea and it's guaranteed to work but, like so many things these days (such as sleeping), I find I have very little time.

Back to the salt minds, er, I mean mines....

Roger

[Pedro]

Konton:
the most simple way to interface between NLE and real film is the following:
1. you transfer your footage to digital by any method, maybe workprinter. Transfer quality is NOT an issue. If transfering and EDITED film, no frame counting is required, only a sync start frame or standard sync leader.
2. you create your sound track in your NLE system crop it that the sync start frame is the very first frame of the whole stuff and export it to wav-file.
3. you load your film into the projector and place the sync start frame into the gate, prepare your crystal time base to 24 or 25 fps.
4. you prepare your wav file for playback at exactly 24 fps or 25 fps.
5. you start the crystal time base and the wav file at the same time (klick mouse button and switch on the crystal) and the audio will be transfered in sync with the film, without any problems, in stereo quality.

alternativly you can create an artificial burst track in the NLE system and place it to one stereo chanel. This burst track will be able to auto-start the projector during transfer and keep it synced. Each burst of the burst track must be in parallel with one frame.

When having the possibility to superimpose frame numbers during the transfer of the footage, you can pre-edit the film in NLE and only touch your original after finishing the NLE copy. For superimposing frame numbers, you could use my counter C1008 or C1008D, placing it in a way that the lens of the camera can see the display (5 digit red LED). Or overlaying it with a sheet of glass between the camera and the transfer screen in 45 degrees.
After editing the work copy, you can splice your original footage according to the frame number refernce. This time you connect the C1008 counter to your editor and it will count physical frames in both directions. (editor modification kit available for Erno EM1801 and Goko MM-1)
After striping, you transfer the sound as described above.

Pedro

[S8 Booster]

Without studying the details here and not knowing the details on the WP either it still appears to me that a small industrial PLC (Programmable Logic Controller) could be a device to easily make a "foolproof" and flexible control system for all funtions described above.

Easily detects frame/numbers and keeps control of the frames and can run the overlap process automatically. Very easy to program, either by on-board programmer or using a PC (Easier but not nessecary).

Interlock auto frame/camera advance is very easy to set up.

Many has built in small high capacity relays, solid state switches etc.
There is absolutely no limit!

Sample to mention one:
Click here!

R

[Pedro]

That industrial programming systems, as well as c-control basic 1-chip computer of http://www.conrad.com would require a lot of additional hardware in order to interface to the specific film equipment. The pure counting & synchronizing funcionality in my synchronizer is performed only with two standard TTL up/down counters. All the rest is input amplifying/interfacing, output D/A transforming etc. So, why spend a lot of money for an (expensive) industrial system when beeing able to integrate the required funcionality of this system (up/down counting) with only two cheap integrated circuits for about € 1 each? Not to forget the programming&testing affort, and the doubts, if such a system is able to perform two asynchonous processes at one time (counting up and down without loosing any singular units)? The micro-processor controlles Gebuhr synchronizers contain a 16 bit microcontroller, which is running at its limits.
Pedro

[S8 Booster]

Sure your system works well Pedro but the industrial systems are very cheap these days and has bulletproof reliability and does anything required for simoultaneus processing. Basically unlimited.

Programming and simulation is extremely simple. Most of the hardware required here is built in and, they have proven reliability in all sorts of environments.

Also add flexibility for upgrading/improvement.

Your descripion of PLCs seem to belong back in the ´70s? BBCs?

R

[MovieStuff]

Your descripion of PLCs seem to belong back in the ´70s?

But then again, so does super 8! Nothing wrong with using tried and true simple technology.

Roger

[S8 Booster "OFFLINE&]

Your descripion of PLCs seem to belong back in the ´70s?

But then again, so does super 8! Nothing wrong with using tried and true simple technology.

Roger

Agree but here it is a question amongst others about kHZ performance (today Mhz, GHz). The PLCs of the 70`s if one programmed to use 80% of full capacity the processing speed dropped like if you shot a racehorse in the middle of a race. (actually modern computers do the same if stressed, check how much processor capacity your computer use for average work. Not much. Pumping graphics is a different story)

Analog: Drain a car engine for its lube oil and see how far it will run without. Tune your watches. Actually it could be dangerous to stress the mini PLCs of those days for critical applications.

Nothing wrong about much of the old technolgy. Actually very impressive when one take into consideration that did not really have any computer tools to trial and error a billion times.

Once I think, if I remeber correctly, I saw a TV progrmme about teh early space trips, Gagarin, Moon Shuttle (Apollo).

The Russians used a mechanical washing machine like timer to control Gagarins trip and if I remember correctly any portable current scool calculator woud outperform the computers in the Apollo moon travellers computer so all respect for the old stuff.

R

[Pedro]

Technology of the 1970ies???

I am afraid, the basic technology we are dealing with is from about 1895 or somewhat like that.
Back in that times, where the smallest counting unit for registration of moving pictures had been
ONE FRAME per period of time.
And now? We are still using the same kind of film, perhaps in different dimensions, but with the same
kind of basic advance mechanisms like over 100 years ago. While new software and nwe computers normally are just
overpassed when finaly beeing released, we can be happy to enjoy a technology of nearly unlimited compatibility!

As any movement in film is based on singular frames per period of time, which are resolving these movements,
the smallest existing counting units of film are frames. Expressing any film position in frame numbers defines
this particular position absolutely exactly.
So ANY synchronization with audio or other medias MUST be related to the actual frame positions of the film.
To achieve a exact relation, COUNTING is required. You had to count, back in the 1970ies, where the first digital
synchronizer devices rised up, and you have to count nowadays, where we are trying to implement our complete live into
Bill Gates virtual world.
There is no change in digital synchronization prinziples when dealing with film:
You must count the film frames and you must count the equivalent time periods of the audio and see if one counter
is overpassing the other in order to define what kind of speed regulation you need to apply.
And you must compare and regulate again and again, with every frame. As the film is frame based,
any other source must be converted to frame units. Therefore they invented full coat and perfo tape.

How you resolve, that two counters can run at one time and give a regulation result "online" at any frame time, is up
to economics and production number. It´s relatively easy (but not primitive) with TTL or CMOS hardware. Just take a
up/down counter with two clock inputs (one for up, the other for down), set it to the middle of it´s counting range
and clock one input by the projector and the other by the audio. Keep the on-time of every clock shorter than the
on-time of one frame, and you won´t loose any frame, even if the two clocks should appear at the same time. In this
case, the result keeps unchanged, exactly what we want. But if the projector decides to go a little bit faster, the
counter result will drift away and we can use this result to create the neccessary analouge correction signal for the projector.
This is working very reliable and frame-accurate, again, exactly what we need.
Well, we can sofisticate the process, trying to resolve it in software. The first problem to resolve is the asynchronous
counting process. I just had considered a software solution but abandoned the idea due to additional costs and
additional development time.

Finally, I want to make FILMS and not SOFTWARE.

And, after resolving it in software based systems, there is still a huge amount of interfacing hardware, which
cannot be realized in software. The counter itself consists of only two chips. The advantage of software would be
the possibility, to implement more comfort options, f.e. direct frame access, big frame memory etc.

Another part of a digital editing system is dealing with mechanics and sensor technology for that mechanics.
F.e. you MUST be able to registrate not only the number of frames passing thru your film editor or perfo/fullcoat tape,
but also the moving DIRECTION.
I had considered that as easy, but it became one of the most time intensive tasks to
develop a foreward-backward direction recognition without mechanical switches (only a 2-chanel opto coupler) and
without loosing/adding any frame during change of direction. Now I am using direct hardware quadrature encoding
to create the correct fwd/rev signal. This would have been nearly impossible with software. Together with the asynchron
countings, there are too many processes running in parallel.

So why not use hardware? There is nothing wrong with hardware!
Continuing with interfacing the regulation of projectors, motorized viewers and tape decks or even CD-players.
Simply no chance for easy standard solutions.

So, after over a year of developing a complete digital editing/projecting system, where I can do anything with film and
audio without or with computer interfacing for NLE, I will not change anything, only to be "up-to-date".
I will finally return to edit my accumulated inumerous film rolls and enjoy a working system!

Pedro