Audio sync problem question for C2V and others

[DPStewart]

Hey all,

The much respected C2V answered a recent post about a problem with externally recorded digital audio not syncing up.

The poster was recording audio with an Olympus LS-10 which is an outstanding small hand-held stereo field recorder (I have used it to record live music, not video)

C2V said that sync can get "off" when using these.

I need to know EXACTLY how and why. (So as to fix/compensate)

FIRST: we assume that sample rates are set identically between camera and audio device.

It has always been my understanding that digital audio would have identical timing when recorded at identical sample rates. Is this not true?

And if not, why?
I have a bit of audio experience so I can handle technical answers.

I intend to buy a CANON 7D to add shallow DOF and lens options, and was planning to get an LS-10 or the new ZOOM recorder that has the XLR inputs for the audio capture.
Are there going to be "GOTCHA"s?

Thanks all,

~Dane

 

[j1clark@ucsd.edu]

Hey all,

The much respected C2V answered a recent post about a problem with externally recorded digital audio not syncing up.

The poster was recording audio with an Olympus LS-10 which is an outstanding small hand-held stereo field recorder (I have used it to record live music, not video)

C2V said that sync can get "off" when using these.

I need to know EXACTLY how and why. (So as to fix/compensate)

FIRST: we assume that sample rates are set identically between camera and audio device.

It has always been my understanding that digital audio would have identical timing when recorded at identical sample rates. Is this not true?

And if not, why?
I have a bit of audio experience so I can handle technical answers.

I intend to buy a CANON 7D to add shallow DOF and lens options, and was planning to get an LS-10 or the new ZOOM recorder that has the XLR inputs for the audio capture.
Are there going to be "GOTCHA"s?

Thanks all,

~Dane

Even though the device is 'rated' at producing a 48K sample data rate. The various components have tolerances, such that two devices may differ, say device 1 is running at 48005, that is just a 'small' 0.01 percent error... say device 2 is running at 47995, again a 'small error', at 10 seconds, they are about 100 'counts' different, at 100 seconds (just over a minute and a half) they are 1000 counts different...

I don't recall the exact amount of time that one can begin to detect differences in lip sync, but I think it may be on the order of 3-4 '24 fps' frame times, which is about 1/8 second, so, at 48K, in the above example 6000 counts off would be approx 3 24 fps frametimes, and would be after 600 seconds, or about 10 minutes.

To compensate for this difference, systems were built that use a 'common clock', feed both recorders, or one recorder acting as clock master for the second, or phase locked loops, from a 'really precise' frequency source, that 'adjusted' one or the other device to match more exactly the speed of the other so that the data output is lock step synchronized.

 

[DPStewart]

Great reply - THANKS.

As long as the timing differences are built into the device I should have no problems.

Just put a metronome in front of the camera and record 10 minutes of it with both devices, then calculate the percentage of difference, then just use that percentage every time - a no brainer really.

Thanks again.

~DPS