Jump to content

Spotting 29.97 files


Tom Duffy

Recommended Posts

Slight rant...

 

29.97 is not what most people think it is.

 

The original definition ties it to the color oscillator frequency, and the math comes down to

30 fps * (1000/1001).   The actual calculation being 5×7×9/(8×11) MHz  * 2 / 455 / 525

 

For word clock, we often talk about "0.1% pull down", so many engineers and software writers probably assume that 29.97 is 0.1% pull down from 30fps, which is  30 fps * (999*1000)

 

Yes, that means there is a difference.    It's almost exactly 1ppm,    good enough for Hollywood?

 

For Time Of Day Timestamping, a recording made in the afternoon or evening could be off by 2 or 3 frames just by the assumptions made here.

Does the recorder calculate its Timecode to Broadcast wave time stamp using 29.9700000 or 29.97002997 math?

Does the Post software spot the broadcast wave file back to timecode using 29.97000000 or 29.97002997 in its calculations.

 

For a workflow where these match up, the editor is happy.   Where they don't, does the assistant editor get blamed, does the recordist get blamed, or does the equipment get blamed?

 

All the public documenation for these software talk about 29.97, but there's no way to know if they use the simplified version of the historical accurate version.

 

Some recorders boast 0.1ppm clock accuracy, but a 1ppm mistake in post wipes that out...

Link to comment
Share on other sites

For Time Of Day Timestamping, a recording made in the afternoon or evening could be off by 2 or 3 frames just by the assumptions made here.

Does the recorder calculate its Timecode to Broadcast wave time stamp using 29.9700000 or 29.97002997 math? Does the Post software spot the broadcast wave file back to timecode using 29.97000000 or 29.97002997 in its calculations.

 

No one cares. A frame or two drift in one day is not a big deal. All dailies people who are even halfway competent still check the slate clap no matter what the timecode says. I've often reminded producers, "it takes less than 1 minute to sync up one take, so if there's 200 takes in one production day, it's roughly 3 hours of work. It takes another few hours to duplicate files, set up bins, make viewing copies, and other chores. It's part of the job." Not a big deal.

 

What baffles me are neophyte filmmakers who don't understand why dailies need to be synced at all. You'd be amazed by the number of people who believe sound should always be recorded in camera. They seem unaware of the necessity for multitrack in many, many production environments, and also the huge limitations in camera sound recording. Just the fact that cameras can't record sound in high-speed/slo-mo is a big problem they don't consider.

Link to comment
Share on other sites

It could be, and sometimes is, much worse.

 

Back in the twilight of analog, I used to work for the largest commercial shoot/post house in New England: two stages, film xfr, 4 CMX rooms, 2 24-tk rooms, early CGI, giant staff. They'd shoot video when requested, but preferred to shoot super16 and then transfer to tape for the edit. 

 

They'd shoot film at 24, and record field audio on N-IV with timecode. Then they'd transfer at 23.976 (more or less, with a 3:2) to one videotape, play the audio on a Nagra T with TC locked to house sync onto another videotape, and sync using TC for rough location and slate for precise start.

 

Naturally, sync would drift. By the time you got 30 seconds past the slate, sound was almost 1 frame early. 

 

Being a good little soundie, I pointed out that it wouldn't be much hassle at all to use one of our spare Adams Smith modules and make the field tape play at -.1% speed for the transfer.

 

Being a good little employee who was paraded out in front of clients, I shut up when the boss said "Nobody will notice. If the client complains, we'll make an edit... and have a couple more billable minutes between the mark, preview, and execute with all the pre-rolls."

 

---

 

Bottom line: I'm awfully glad for digital audio and video, and the demise of tape. If two files are going to drift a frame or two by the end of a day, I'm quite happy to make the edit. 

Link to comment
Share on other sites

It could be, and sometimes is, much worse.

Back in the twilight of analog, I used to work for the largest commercial shoot/post house in New England: two stages, film xfr, 4 CMX rooms, 2 24-tk rooms, early CGI, giant staff. They'd shoot video when requested, but preferred to shoot super16 and then transfer to tape for the edit.

They'd shoot film at 24, and record field audio on N-IV with timecode. Then they'd transfer at 23.976 (more or less, with a 3:2) to one videotape, play the audio on a Nagra T with TC locked to house sync onto another videotape, and sync using TC for rough location and slate for precise start.

Naturally, sync would drift. By the time you got 30 seconds past the slate, sound was almost 1 frame early.

Being a good little soundie, I pointed out that it wouldn't be much hassle at all to use one of our spare Adams Smith modules and make the field tape play at -.1% speed for the transfer.

Being a good little employee who was paraded out in front of clients, I shut up when the boss said "Nobody will notice. If the client complains, we'll make an edit... and have a couple more billable minutes between the mark, preview, and execute with all the pre-rolls."

---

Bottom line: I'm awfully glad for digital audio and video, and the demise of tape. If two files are going to drift a frame or two by the end of a day, I'm quite happy to make the edit.

That part about the boss "no one will notice" is classic. Create a problem then charge to fix it.... Wow.

Sent from my iPhone using Tapatalk

Link to comment
Share on other sites

<snip>

 

The original definition ties it to the color oscillator frequency, and the math comes down to

30 fps * (1000/1001).   The actual calculation being 5×7×9/(8×11) MHz  * 2 / 455 / 525

 

 

Tom,

 

Sorry, late to the party.  I'd be interested to know what all the different components of your calculation represent.  525 is obviously lines in the NTSC raster, but not sure what the rest would be.

 

Thanks,

Link to comment
Share on other sites

They'd shoot film at 24, and record field audio on N-IV with timecode. Then they'd transfer at 23.976 (more or less, with a 3:2) to one videotape, play the audio on a Nagra T with TC locked to house sync onto another videotape, and sync using TC for rough location and slate for precise start.

 

We had zero drift with this method... but you have to hose 59.97 video sync reference into the Nagra T, and pull down the machine (with a one-switch setting) to 29.97. I've had 20-minute takes stay in absolutely perfect 0-drift sync before, no problem. All they gotta do is know what they're doing. We used this quite often for sitcoms in the 1980s, literally laying down the soundtrack for the entire show, duplicating the tape 4 times for 4 cameras, and then dropping picture-only into the videotape. So it can "theoretically" work.

 

We also had a little preset so we could deliberately make the edit a frame early or a frame late, whatever was necessary to compensate for an overall offset during the shoot. Not a big deal, easy to do. The only thing that counted was that the clap was dead on.

Link to comment
Share on other sites

We had zero drift with this method... but you have to hose 59.97 video sync reference into the Nagra T, and pull down the machine (with a one-switch setting) to 29.97.

 

 

I had zero drift when I did a similar thing (though using the A-S module) on a Studer 810 in my audio suite... the few times they let me do transfers in audio. The film transfer room billed slightly higher, and always required an edit suite afterwards... so you can guess which method the company preferred.

Link to comment
Share on other sites

  • 4 weeks later...

Tom,

 

Sorry, late to the party.  I'd be interested to know what all the different components of your calculation represent.  525 is obviously lines in the NTSC raster, but not sure what the rest would be.

 

Thanks,

Wikipedia article on NTSC.   There are no references to primary sources, but it makes sense that the other numbers derive from multiples of small primes, which were the only possible multiplier ratios used with tube-era electronics in use when TV was being established.

Link to comment
Share on other sites

Join the conversation

You can post now and register later. If you have an account, sign in now to post with your account.
Note: Your post will require moderator approval before it will be visible.

Guest
Reply to this topic...

×   Pasted as rich text.   Paste as plain text instead

  Only 75 emoji are allowed.

×   Your link has been automatically embedded.   Display as a link instead

×   Your previous content has been restored.   Clear editor

×   You cannot paste images directly. Upload or insert images from URL.

×
×
  • Create New...