-
Root cause of 99.9% audio sync drift problem?
I’ve encountered the “99.9%” problem whereby 2nd system audio drifts slightly slower than 29.97 fps video on long takes. Amount of drift: video is shorter by 99.9%, alternatively audio is longer than video by by 100.1%
In this situation video was 720×480 29.97 fps from an old JVC GZ-MG155, audio from a Korg MR1000, 16bit 44.1kHz. CS 5.03. Project was DV 41.1Khz, 30 fps drop frame.
There’s lots of discussion about the 99.9% sync drift problem, but no definitive explanation of the root cause (that I understand). Of course I’d like the recommended workflow to fix it, but am also interested in understanding the underlying technical cause.
E.g, if video both records and plays back at 29.97, whether drop frame time code or not, 1 hr of recorded video = 1 hr of elapsed time. E.g, 1 hr of 60 fps video if played back at 60 fps takes 1 hr elapsed time.
Likewise whether audio sample rate is 44.1Khz or 48Khz, provided it plays back at the recorded rate it will stay in sync with elapsed wall clock time (plus or minus drift precision).
This situation isn’t drift from lack of timebase precision — audio drifts slower than video by exactly 100.1%. This is reproducible and consistent. The amount of drift is the exact proportion of 29.97 to 30.00 so this implies the cause, but I don’t understand why.
Could someone help explain the root cause of this problem?