Position of optical sound on film in Switzerland and France

[Simone Corelli]

Hi, this is my first post here. Happy new year to all! And sorry for poor English.
I'm here asking you to check my schematics about the sound on film, AND to tell me if you know somewhat about a different standard in Switzerland and France... I remember having read something about 21 frames instead of 20... thanks to all!

[Steve Guttag]

No. There is no difference by country on sound sync. Physically, there is a 21-frame separation. However, on the leaders, the sound start is printed at 20-frames for optical sound to give a normal thread up a 1-frame lead. I can only presume, historically, this was done to ensure that in the large theatres of the day that the sound didn't appear to have excessive delay. 1-frame is about 42ms of time. Using "kitchen-physics," 1ms ~ 1ft so leading the sound puts dead sync 42-feet into the theatre. Most cinemas, back in the day, were in excess of 100ft long and often had a stage that ate up the first 10-feet or so. If you had a balcony, it was not unusual to have 120-150-feet from last row to screen. So leading the sound seemed prudent and only the very front rows may notice an issue versus the bulk of any theatre where most sit in the back third of the room. If you change your lower-loop by just 2-perforation and you've cut that compensation in half to just 21-feet so it works if for a more modern/shorter theatre.

One thing digital brings is the ability to move dead-sync to where ever it makes the most sense, including dead sync or even lagging (though there shouldn't be a need for that). A down side for digital is that it is all too easy for the extra processing of the video image to introduce latency and cause a lip-sync issue if the audio is not similarly delayed to keep in sync with the picture. I have definitely received tapes and even DCPs where lip sync was poor and on-the-fly adjustments had to be made.

[Marcel Birgelen]

Since most modernish audio processors have a delay line, I always kept it at 20 frames and worked it out from there if necessairy. I've never heard about a difference between Swiss or French optical sound tracks compared to the rest of the world, but like Steve indicated, there are many sloppy mixes out there and the problem of "dead sync" extends well into the digital realm...

Ironically, the problem with "lip sync" is often more prevalent for sound effects like a door slamming shut or an explosion than for actual lip-sync and if you happen to play a dubbed mix (not uncommon in France), lip-sync is all-over-the-floor anyways.

[Simone Corelli]

Physically, there is a 21-frame separation. However, on the leaders, the sound start is printed at 20-frames for optical sound

thank you for the reply. But I don't understand: I have checked a Dolby film test for dolby digital surround EX and the POP was printed exactly 20 frames before the corresponding visual POP (sound frame..19 frames..visual pop). I also know that projectionists modify the latham loop so that the distance in projection is 19 to compensate for big rooms. Where did you read 21? Thanks

[Carsten Kurz]

The spec for the print offset is 21 frames (+/- 0.5).e.g. ISO 2939. So, the print has to come from the lab with that offset.

The spec for the actual projector/loop setup is 20 frames. The one frame difference is to assure sync for the middle of a typical traditional auditorium. If you know what you are doing, you can adjust that as needed.
So, 20 vs. 21 do not work against each other, but complement each other.

Test films may be created with a different offset, because they may not be intended for auditorium checks, but to be used with booth-equipment, like oscilloscopes, etc., where the size of the auditorium is irrelevant.

[Simone Corelli]

The spec for the print offset is 21 frames (+/- 0.5).e.g. ISO 2939. So, the print has to come from the lab with that offset.

Thank you Carsten! Can you kindly provide me with a little screenshot of the schematics in the ISO2939 just to be sure if they count including both the extremes or not? My email is simonecorelli@mac.com
It is very important for me to have a direct documented source.

[Marcel Birgelen]

I'm not familiar with ISO2939, but the relevant IEEE/SMPTE documents for this should be ST 40:2002 and ST 301:2005. Please notice that you require a valid subscription/pay the appropriate fees to access the documents.

[Simone Corelli]

I'm not familiar with ISO2939, but the relevant IEEE/SMPTE documents for this should be ST 40:2002 and ST 301:2005. Please notice that you require a valid subscription/pay the appropriate fees to access the documents.

Can you kindly confirm that you have ST40:2002 (which is very expensive) and that it is CLEARLY stated that you have the sequence: 1 sound frame, 20 frames in between, 1 visual frame related to the sound frame? Thanks a lot.

[Marcel Birgelen]

I can't publish the entire document here, although it's just 3 or 4 pages and yes, they're asking ridiculous fees for access to those documents...

But yes, I do have the document and below are two screenshots from the relevant sections.



smpte40-screen1.png


smpte40-screen2.png

[Simone Corelli]

This answers definitively my question. I'm grateful! Thanks! Strange that Dolby decided to put the sound 20 frames in advance instead of 21 and that in so many books people write 20 without specifying that this is the distance for normal projection between sound head and center of frame projected, not the PRINTED sound. WOW

[Marcel Birgelen]

No problem. A lot of stuff runs on assumptions... I guess that's part why stuff is always in some sort of state of disrepair.

I guess Douglas Adams hit the nail on the head: The most misleading assumptions are the ones you don't even know you're making.

[Simone Corelli]

Now the question shift to the position of MAGNETIC sound recorded: 27? 28? 29? frames after the projected frame? And what about the position of the magnetic head? 26? 27? 28? 29 or 30 frames after the projected frame?
Look at this and tell your opinion, please: Fig 27 .jpg

[Simone Corelli]

No problem. A lot of stuff runs on assumptions... I guess that's part why stuff is always in some sort of state of disrepair.

I guess Douglas Adams hit the nail on the head: The most misleading assumptions are the ones you don't even know you're making.

Ouch! Dolby states that "Dolby Digital data is recorded 6 frames in advance of the analog sound track, which is itself placed 20 frames in advance of the picture gate". NOT 21. I'm going crazy! What's your opinion?

[Marcel Birgelen]

For the exact specs, you would need to contact Dolby, as many of those specs are only available under license, but maybe someone on this forum who has worked for Dolby can confirm the number. I'm pretty sure it follows 26 frames after the picture. I don't know what the source of the number "20" is in this regard. Is that line you quoted from an actual specification or from some manual? In the latter case, it could just be a mistake in the manual.

Keep in mind that, in practice, during presentation, this matters very little, because of compatibility reasons with all the different projector models out there, the placement of the Dolby Digital reader is pretty flexible (both "above" and "below" the projected frame is possible) and you're supposed to configure the static delay in perforations inside the audio processor. You verify and adjust the whole thing while running a test-loop.

As for magnetic sound, the standard for 35mm is 28 frames after the projected frame. The fact that magnetic sound follows the picture limits the placement of the magnetic sound head. Those analog systems all date from a time where delay wasn't as flexible as it is today. Nowadays, delay can be created using some rather cheap digital buffer, analog delay requires analog delay lines, which are pretty expensive for large delays and also significantly impair signal quality for longer delays.

[Simone Corelli]

For the exact specs, you would need to contact Dolby, as many of those specs are only available under license, but maybe someone on this forum who has worked for Dolby can confirm the number. I'm pretty sure it follows 26 frames after the picture. I don't know what the source of the number "20" is in this regard. Is that line you quoted from an actual specification or from some manual? In the latter case, it could just be a mistake in the manual.

Yes Marcel,
Dolby wrote me an email saying that:
"Hi Simone,
If you are talking about encoded Dolby Digital data on a film reel, then this data starts 6 frames in advance of the analogue soundtrack in order to allow for any decoding latency and 26 frames in advance of the corresponding picture frame. Unfortunately with Dolby no longer producing any products that support film, documentation has become quite difficult to source, but I am attaching here what is the most relevant documentation I can find on the subject for you. Along with a technical paper on Dolby Digital encoding I have attached the Dolby DA20 manual which has some useful information on film setup. I hope this answers all of your questions but if not, then please let me know and I will answer them as best I can."

On the DA20 manual you can read this, that confirms what Dolby says in the email:
"Dolby Digital data is recorded 6 frames in advance of the analog sound track, which is itself placed 20 frames in advance of the picture gate. If the Digital data is read 40 frames above the picture gate, for instance, the data must be delayed by 40+20+6 = 66 frames, or 264 perforations, or about 2-1/2 seconds."

I don't think it is a mistake.
Note also that in SMPTE is not specified if the distance is from the center of the visual frame to the BEGINNING or to the CENTER of the related sound... I don't understand how a standard can be so ambiguous!