The AMPS Technical Committee received the following letter from Nick Lowe in connection with their work on frame rates and it was decided to publish it here.
I am writing to gripe about the elementary blunders that are still being made in feature film sound production that have irritating and time-consuming consequences in post production. In all fairness I have to say that most of the bungling can be laid firmly at the door of whoever is responsible for the production planning, and also that these problems are rarely encountered with material that has come from UK sound mixers who tend to have a good knowledge of the technical side of their trade.
What it all comes down to is taking expert advice when the subjects are formats, frame and sampling rates, and off-speed picture takes. All of these anorak topics have a profound influence as to whether the post production will go smoothly or not. The thorny question of how the production sound is synchronised to picture is an issue too big and tedious to go into here but my preference is for the picture department to handle this as they have always done.
So here are some basic ideas concerning 24fps 35 mm film:
1) If the picture is to be cut on film the production mixer must -
a) Record the roll number in the 8 user bits that are available to him / her. And say which ones they are using to denote said roll number.
b) Record timecode that is continuous on each roll - obviously it is only possible, apart from using user bits, to absolutely identify 24 hours worth of recording but the idea is never to repeat a particular timecode frame within a sound roll. Ironically, the only production mixers who can ignore this are the ones still using 1/4 inch Nagra pilotone or FM pilot. All they have to worry about is whether they use 50 or 60 Hertz pilot.
c) If there is any possibility that any dialogue editing will be done on a digital system, ie not on 35 mm film, then the transfer bay must be instructed to use the 35 mm triple track format and record the audio on track 1 and if necessary track 2, and the timecode on track 3. Analogue diehards who are still convinced by the superb quality of Stefan Kudelski's mike amps need only ensure that the transfer bay can resolve and transfer their material to the aforesaid 35mm format whilst simultaneously making a copy DAT bearing the same timecode and user bits as are routed to track 3 of the 35mm recorder. This is how a dialogue editor gets an EDL - a way of knowing which bit of synch sound goes where. The only other question here, apart from off-speed picture takes, is whether to use 25 or 30 NDF timecode. The answer is - are you going to post - produce in the UK or the USA? For UK use PAL 25fps for sound as there are still hidden costs and complications when using 30 NDF / 29.97 NDF NTSC. For the USA always use 30 NDF. Never mix the frame-rates, and never ever under any circumstances have anything to do with 24 fps timecode.
2) If, as is common these days, the picture is being cut on Avid or Lightworks, decide early on whether to use the PAL or the NTSC standard. Never try to mix the two and don't change your mind halfway through. How an EDL is derived is largely down to the way the action and sound is synchronised and that is another issue. At least identify each roll number somewhere specified in the user bits (mostly the last three) and keep the timecode vaguely continuous within each roll.
3) Off-speed camera takes (ie anything other than 24 fps) can be a severe nuisance and are relatively easily dealt with. If the idea is to record any interesting or usable sound that is likely to make it to the dubbing theatre then some consideration has to given as to how this material will be processed to keep it in synch. This is dead easy and I won't go into it here but suffice it to say that the 'stretched or 'compressed' takes should be then transferred to another DAT sporting a new set of timecode and user bits before they are synched up to the relevant action. In this way, an EDL of the right audio can be derived without endless research into key numbers and monotonous resynching.
Of course file compatibility such as the much-heralded OMFI and the lately touted AAF will make all these gripes redundant. And then pigs might fly....
NICK LOWE