How is off-line film edited

Mark Wilson · October 7, 2006

I had always been under the impression that fully computerized off-line editing of film was a similar procedure to off-line editing of videotape.

That is, after the edit decisions has made on the computer screen, an automated machine would find the relevant frames by reading the timecode or keycode information on the negative, and then do the cut and splice automatically. At least, that was what a Kodak rep told me back when keycode was launched. Admittedly, he did go to some trouble to emphasize that keycode had something to offer even to diehard traditionalists, since reading the keycode numbers was likely to be faster and more accurate than traditional negative matching. I still have the free magnifying glasses they were handing out!

But I have been told that this is not the case, that all editing is still done by hand, although it does make use of the frame codes, but only in the form of an edit decision list printed out by the editing computer.

I find it hard to believe that nobody has been able to come up with a machine that could do this, or is it more a "horses for courses" situation.

Also, when editing film, is it normal to overlap frames, or do they make a butt splice?

Dominic Case · October 8, 2006

an automated machine would find the relevant frames by reading the timecode or keycode information on the negative, and then do the cut and splice automatically.

l nearly finished composing an answer to this, when my thumb hit the wrong key, and the entire message was deleted. Unlike most test editing software, there is no "undo" button. So I have to start again.

Similarly, there is no "undo" button when it comes to negative cutting. Any mis-cut stays mis-cut. And that's the key to this answer. Nothing can possibly go wrong . . go wrong . . go wro . . .

There are so many processes applied to the numbers in an edl - some relying on human data entry, some on complex (and not precise) corrections for varying frame rate and other issues, the adjusting to remove non-cutting events from edls and so on - that the chances of the final cutting list being 100% perfect without human intervention are slim indeed.

A machine that went ahead and cut negative by itself would be dangerous.

Also, while a fully automatic negative conform might seem attractive in principle, it's unlikely that it would be vey practical for any but the smallest productions. Presumably it would have to assemble the negative in edited sequence: if a sequence of shots came from - shall we say - camera rolls 23, 45, 23, 46, 23, 9, 39, 9, 23, 45, etc., you would be forever changing camera rolls. (By comparison, when you do an on-line video edit you can lay down all the shots from each source roll in sequence).

Finally, while it may be possible to make a machine that could splice negative automatically as well and as consistently as a good cutter can do by hand, (scraping just enough, right amount of cement applied etc), it would be a sophisticated and very costly machine. With a fairly small potential market.

I actually did some work with a company some years ago who were considering making such a system. (Your informant at Kodak might have been aware of this!). They even trademarked a name for it. But market research told us that it wouldn't sell because:

a... it would cost far more than any negative matcher could afford to pay

b... there was no perceptable benefit to anyone in making it or using it

c... no neg matcher would trust it

d... no neg matcher believed any of their customers would allow them to use such a machine on their negative. It would be a quick path to ruin.

As for overlapping frames - to be more precise, conventional negative splices work with an overlapping join. The emulsion is scraped off one surface, the other surface is cleaned and slightly roughened, and the two surfaces are cemented or welded together with a solvent. Because there is an overlap, you lose the first frame of the unused section of the negative, which therefore can't be included in another shot. Also, pertaining to the first question, if you want to rejoin the original shot, you have lost a frame by virtue of the cut.

There is a newer type of splice that many cutters now use, in which the two pieces of negative are cut at an angle. So it's a cross between a butt splice and an overlapping splice. Much neater, no double-thickness splices in the neg, but requires even more ar to make it well and make it strong. As there is no overlap it's possible to use the next frame - though neg cutters often use that frame for handling purposes if it's not needed in the film.

Michael Most · October 9, 2006

There are so many processes applied to the numbers in an edl - some relying on human data entry, some on complex (and not precise) corrections for varying frame rate and other issues, the adjusting to remove non-cutting events from edls and so on - that the chances of the final cutting list being 100% perfect without human intervention are slim indeed.

Which is also the reason there is no such thing as a "straightforward" digital intermediate conform.

Phil Rhodes · October 9, 2006

Hi,

I have to admit that when I was first exposed to EDLs and data management in the world of DI, I was completely horrified at frankly how disorganised and spur-of-the-moment it all is.

Most of this stuff is still being made up as people do it.

Horrifying.

Phil

Mark Wilson · October 9, 2006

l nearly finished composing an answer to this, when my thumb hit the wrong key, and the entire message was deleted. Unlike most test editing software, there is no "undo" button. So I have to start again.

Don't you HATE that! With me it usually happens with cheap keyboards with a very narrow surround such as you often find in Internet Kiosks. My worst experience of that kind was when I was holidaying in Vanuatu, when I was paying a fortune for Internet access and either the generator would cough and splutter and crash the computer, or I'd accidentally delete everything the way you have. I have an original IBM PC keyboard, and that never used to happen with those; IBM put an enormous amount of R&D into their keyboard designs. Unfortunately, it won't work properly on a modern PC.

Anyway, thanks for the information. This is starting to sound like one of those never-ending "Yes, but what actual PROBLEM does this SOLVE?" issues :)

Just one other question if you don't mind:

Do the negative matchers make any use of the key/timecodes at all, or do they still do that the traditional way as well?

Dominic Case · October 9, 2006

Do the negative matchers make any use of the key/timecodes at all, or do they still do that the traditional way as well?

Neg matchers build a database of each roll of negative, relating roll number, footage, keykode and timecode (which relates the neg itself to the digital images used in editing).

This is either captured during telecine transfer, or as a separate pass by the negative matcher. Or a bit of both.

In some instances, the keykode data is fed into the nonlinear editing system, and it can produce a cutting list based on keykode numbers rather than a timecode-based edl. However, this doesn't tell the neg cutter where each neg roll is located.

In other instances the neg cutter receives an edl which is then processed into a cutting list by the neg cutter's own software (that manages the database).

During cutting, the neg cutter can wind camera neg through a barcode reader which will find the required keykode and emit a beep. Or, more often, they simply wind through the negative until they visually read the required keykode. In either case the exact frame to cut is identified by a count or offset of a number of frames away from the actual keykode on the edge of the film, as they only appear once per foot.

If all the numbers have worked, they will cut on the right frame. On some productions there is a "pos conform" whereby a cutting copy is made by editing selected work print takes to match the edl first. It acts as a template for cutting the original negative itself, and if the cutting copy is correct, it is easy for the neg cutter to visually match the actual frames to conform to. It's like the old carpenter's adage "measure twice, cut once".

When it comes to pulling shots for scanning for DI work, it's a little more haphazard still. (as Phil points out) While some houses have developed software to convert edls into pull lists into select take logs into scanning lists to control the film scanner, they are mostly custom programs: I'm not aware of any universal approach to this.

So the answer is yes, neg cutters DO use keykodes AND timecodes, but only so far as they are helpful.

Michael Collier · October 10, 2006

In a follow-up, how is keycode utilized for a telecine session?

I ask because I plan to do this with my film (and at the moment, filmscribe on my avid isn't working for some reason)

If I get the keycode burned into the first telecine, can I create my own pull list by hand? I should know what roll (and also what lab roll, as they will be grouped into flats) and I can organize the whole list and create a pull list, with this list does the telecine scroll to the right frame and begin capturing? Its a short, so I don't mind making one by hand, or conforming by hand (though I would prefer filmscribe to quit dorking out and make things easier)

Does this workflow make sense? What is the possibility of the wrong peice getting transfered, if I include a pull list and a DVD copy of the finnished cut (with keycode burned in as well). This is assuming I don't screw up when creating the pull list (tripple check before I send it out of course)

Dominic Case · October 10, 2006

Not quite clear what you are intending to do. A pull list would normally be to make up a short roll of select takes for re-transferring (with a full grade), for a video finish. The complicated part here is that once you assemble the select takes, that roll obviously has different footages and therefore timecodes from the original camera rolls and flats. Neg matchers have software that can translate that.

It is probably best to talk to your neg cutter and your telecine facility to see what each of them needs. There are so many minor variations on a theme in terms of workflow, and if you end up doing something even slightly differently from what they are used to, there will be confusion somewhere along the way.

Michael Collier · October 10, 2006

I am trying to finnish to HD cheaply. I would like to skip a negative cutter all together. If its possible, I would like to create a pull list or cut list so the telecine operator can transfer only what was used in the offline cut.

I would like the footage to be matched back to the original automatically. If that option is too costly for whatever reason, I would not mind taking the footage and matching it back by hand. Its only a 15 minute movie, and all the editing equipment is in house, so all it means is a few more late nights. I would make sure keycode is burned in the upper window so I can double check the pull list (or whatever you would call it in this situation)

I would like to keep the negs from being cut or printed during the post to save on lab costs. Right now all I want (and can afford) is an HD finnish. Festival receptions would dictate any future work.

I am transfering on a Spirit to a D5HD with a dub to DVCPRO-HD or HDCAM (depending on local post houses setup.) Offline cut would be made from a onelight transfer to DVCPRO with keycode burned in the upper window. I figure combine the keycodes, with the roll number shown on the slate, and the lab roll number from my lab report, and the lab should have all the info they need to find the exact frame to start on. Maybe provide 2 seconds pad on each shot to allow for minor tweeks. Then arrange the clips in order they are in on the lab rolls and it should be easy to find the shots with accuracy.

Would that work?

Michael Most · October 10, 2006

figure combine the keycodes, with the roll number shown on the slate, and the lab roll number from my lab report, and the lab should have all the info they need to find the exact frame to start on.

You're taking something that's usually pretty simple and making it complicated. In most cases - the primary exception being telecine of print takes only, usually done only on high end television and some high end features - telecine facilities will transfer lab rolls with continuous time code, starting on a punched frame at the head of the roll. Each lab roll will be assigned a specific starting time code, usually noted on a lab report and on the box containing the roll. When retransfers are needed, the roll is put back up and the time code initialized to the original transfer code. All retransfers are then cued by time code. There is no need to do fancy databases, hand written pull lists, or anything else. That would be needed if you were going to a film scanner to create individual frame files for a DI conform, but it is usually not needed for retransferring on a telecine, even if the code on the original was 30 frames and the retransfer is 24 (the codes are reconciled at even seconds, and every 4 frames thereafter).

Michael Collier · October 11, 2006

Thanks for the info. I suppose I was making it a bit overcomplicated. I am just trying to wrap my head around the workflow to make sure I don't paint myself into any expensive corners.

I talked with the lab today, and they said basicly what you just said, so I am feeling a bit more comfortable and confident in the post workflow. As I said (or maybe I haven't) this is my first movie shot on film, and I don't want to miss anything. The more involved I am in the proccess, the more I will learn.

Thanks for all your help.

Dominic Case · October 11, 2006

Sorry if my answers only served to confuse the issue. Iwas thrown by misunderstanding your reference to a pull list. As Mike points out, if you keep the neg intact (and that is vital), then you can find any shot accurately again by simply resyncing the telecine to the punched frame at the start of the roll.