Jump to content


Photo

Reliability of Neg Cut according to Keycodes?


  • Please log in to reply
9 replies to this topic

#1 Marc Roessler

Marc Roessler
  • Basic Members
  • PipPipPipPip
  • 264 posts
  • Other

Posted 09 May 2008 - 04:35 PM

In principle, cutting your telecined or scanned film in AVID/FCP and then doing the negative cut according to the image with the burned in keycode (or even just the EDL!) seems to be bullet proof...

But how reliable is this, actually?

I've met several guys who have the opinion that keycode should not be relied upon, and some people even seem to do an additional pos conform before cutting the negative.

I've now had this problem myself: keycode prefixes not changing on roll changes, and keycode prefixes changing mid-roll - great! Did you ever encounter this? If so, under what circumstances? How did you catch it, how did you fix it, and how did you change your workflow to catch this in the future?

After all, the good old flat bed doesn't seem soo bad: immediate pos conform, shuttling without rendering, good feeling for the material, no crashes... :rolleyes: (if only sound wasn't such an issue and the trim bin wouln't be so time consuming...)

Greetings,
Marc

Edited by Marc Roessler, 09 May 2008 - 04:40 PM.

  • 0

#2 Robert Houllahan

Robert Houllahan
  • Sustaining Members
  • 1584 posts
  • Industry Rep
  • Providence R.I.

Posted 09 May 2008 - 05:26 PM

If the Keycode reader is properly setup with the telecine and decks and the film is properly punched on each "k" frame and the keycode reader is run right the Flex file and neg cut list from Avid or FCP can be highly accurate. There are allot of "IF's" in there. I have two Aaton Keylink keycode readers and we always punch every first "K" frame of each roll then we always run the Keylink as a Pass-1 with tabs and then a Pass-2 synced to the film, this is the most accurate and reliable way to run a keycode job on the Aaton.

As an example of what can go wrong I have seen people run a one-pass with the aaton and that can cause multiple rolls of keycode to only have code from the last roll on the flat when wound back to the head. There are many TK shops that do allot of TV Spots or Episodic tv where the accuracy of keycode reading is not as important and they will often do incremental timecode for each tape but this becomes a problem for a feature where the timecode will repeat as there are more flats of film that 24 or 48.

As always good post planning saves money in the long run..


-Rob-
  • 0

#3 John Sprung

John Sprung
  • Sustaining Members
  • 4635 posts
  • Other

Posted 09 May 2008 - 06:26 PM

After all, the good old flat bed doesn't seem soo bad:....

Even back in the days of cut workprints and synchronizers, negative cutting wasn't perfect. Any time you have human beings doing a dead boring repetitive physical task, there will be errors. Cutting to tape with a lock box only made it harder to get right. I've had the good fortune not to have to deal with physical negative cutting for a couple decades now. In episodic TV, it's a thing of the distant past. It's dangerous, dirty, boring, barbaric, I don't miss it at all. ;-)



-- J.S.
  • 0

#4 Brian Drysdale

Brian Drysdale
  • Basic Members
  • PipPipPipPip
  • 5070 posts
  • Cinematographer

Posted 09 May 2008 - 06:44 PM

I had a neg cutting problem with one of my shorts on which we used an EDL. A film editor friend had to run the 35mm print on his Pic Sync to find out where the problem was. It turned out that some black spacing in the opening titles was too long.
  • 0

#5 Robert Houllahan

Robert Houllahan
  • Sustaining Members
  • 1584 posts
  • Industry Rep
  • Providence R.I.

Posted 09 May 2008 - 07:01 PM

It's dangerous, dirty, boring, barbaric, I don't miss it at all. ;-)



-- J.S.



Those gang sync's can sneak right up on you and they bite!

-Rob-
  • 0

#6 Dominic Case

Dominic Case
  • Basic Members
  • PipPipPipPip
  • 1357 posts
  • Other
  • Sydney Australia

Posted 11 May 2008 - 02:55 AM

The cold hard fact about negative cutting is that you can't uncut the negative. If you use conventional overlap splicing you inevitably lose a frame when you cut and splice. If you cut the wrong frame, you can't have it back. Even with the Hamman mitred splices, you need great care to separate a splice (which shouldn't be possible if it's a good one) if it's wrong.

The pos conform approach is just a trial run before you cut the negative itself - it's wise insurance.

Apart from those human or mechanical errors - mis-copied numbers, black spacing wrong length, wrongly set-up flex files, or even faulty keykode, you have major problems unless you have transferred your images frame-for-frame on telecine (regardless of camera speed. In other words, at 25fps for PAL or 29.98 for NTSC tape. (If you go straight to a digital file format with no videotape, then you will have frame-for frame matching.)

At ANY other speed, the telecine inserts phantom fields and frames that don't correspond to real film frames, and it becomes physically impossible for a perfectly accurate neg match that also maintains sound sync. This isn't carelessness or incompetence or sloppy software. It's basic maths.

Using the burnt-in timecode from the tapes eliminates some of these issues (eg flex files) but the burn-in is also subject to the sorts of IF that Robert points out. And STILL, cutting picture on the correctly matching frame of film can introduce small but potentially cumulative sound sync errors if you have transferred at a non 1:1 speed.
  • 0

#7 Marc Roessler

Marc Roessler
  • Basic Members
  • PipPipPipPip
  • 264 posts
  • Other

Posted 11 May 2008 - 02:57 PM

The material was transferred frame-for-frame, of course.

In the meantime I found out what went wrong by carefully inspecting the negative.

Two of the original reels were loaded as 100 feet daylight spools (ISO 100), and one of those had both edges (about 1 mm on each side) fogged (this was N16, just FYI... posted it to the "35" forum because keycode is a universal issue). Keycode was still readable by eye but seemingly the reader had a problem with it.

Ironically, the affected daylight reel was loaded at a darker place than the non-affected daylight reel. Fogging on both reels was about constant in density for the whole reel.

I knew that 16mm daylight reels are not a good idea for high ISO S16 projects, but wouldn't have thought it is also an issue with ISO 100 and keycode...

Greetings,
Marc

Edited by Marc Roessler, 11 May 2008 - 03:00 PM.

  • 0

#8 Chris Keth

Chris Keth
  • Sustaining Members
  • 4427 posts
  • 1st Assistant Camera
  • Los Angeles

Posted 12 May 2008 - 12:28 AM

The material was transferred frame-for-frame, of course.

In the meantime I found out what went wrong by carefully inspecting the negative.

Two of the original reels were loaded as 100 feet daylight spools (ISO 100), and one of those had both edges (about 1 mm on each side) fogged (this was N16, just FYI... posted it to the "35" forum because keycode is a universal issue). Keycode was still readable by eye but seemingly the reader had a problem with it.

Ironically, the affected daylight reel was loaded at a darker place than the non-affected daylight reel. Fogging on both reels was about constant in density for the whole reel.

I knew that 16mm daylight reels are not a good idea for high ISO S16 projects, but wouldn't have thought it is also an issue with ISO 100 and keycode...

Greetings,
Marc


If the fog density was constant for the whole roll, you should be checking your camera, especially if you will be using it again soon.
  • 0

#9 Paul Bruening

Paul Bruening

    (deceased)

  • Sustaining Members
  • 2858 posts
  • Producer
  • Oxford, Mississippi

Posted 12 May 2008 - 12:33 AM

The easiest and cheapest solution to this problem that I could come up with was to shoot the work print with my XL2 in 24fps, letting it do the pull down in-camera. While it shoots the frames in sync, it also shoots the Kodak edge code through a slot that I had cut in the NC's aperture plate. That way, the actual in and out frames in the work print are physically verifiable. Since the edit points in a copied, telecined work print, EDL can be moved left or right the edge code can be double checked on both sides of each edit. A hand made scan list with accurate count from head marker on each roll can be achieved. I can use the numbers from the edge code and do the math on a report for each roll or I can reset the timecode for each telecined roll to a standardized system. Unless I've overlooked something, my GUI, scan controller should capture in 5K only those frames needed to match the work print. Sure, this is slower than automated methods. But it was much cheaper and a good bit more idiot proof. I like idiot proof.
  • 0

#10 Marc Roessler

Marc Roessler
  • Basic Members
  • PipPipPipPip
  • 264 posts
  • Other

Posted 12 May 2008 - 07:04 AM

Chris, that was what I suspected first, too... but one reel was shot on an optical bench, in an otherwise pitch
black room. In case it had been a camera light leak, there shouldn't have been any fogging on this reel.
  • 0


Paralinx LLC

CineLab

Tai Audio

Glidecam

Abel Cine

FJS International, LLC

Broadcast Solutions Inc

Opal

The Slider

Ritter Battery

Metropolis Post

Wooden Camera

CineTape

Gamma Ray Digital Inc

Technodolly

Media Blackout - Custom Cables and AKS

Willys Widgets

rebotnix Technologies

Visual Products

Aerial Filmworks

Rig Wheels Passport

Paralinx LLC

CineLab

Ritter Battery

Wooden Camera

Technodolly

Glidecam

Broadcast Solutions Inc

Visual Products

Rig Wheels Passport

Gamma Ray Digital Inc

rebotnix Technologies

Aerial Filmworks

Abel Cine

Opal

Tai Audio

Metropolis Post

Willys Widgets

CineTape

The Slider

Media Blackout - Custom Cables and AKS

FJS International, LLC