24pn Sync Sound Question

T

TheLastTzar

New member
Hi,
I am an assistant editor who is trying to find out if our workflow is correct. We are shooting on HVX200s and recording at 720pn24. Our sound recorder is recording to hard drive. We are editing on Final Cut Pro.

My question is if we are shooting in 24pn (native) mode, should our sound recorder be recording at 30fps NDFTC in order for our video and sound to sync in Final Cut Pro on a 23.98 timeline?

I'd really appreciate any insight.

Thanks
 
Whew...long post warning...got a little carried away here : )

Since audio is not actually recorded in frames, the frame question, in practice, is more or less moot. Digital audio is a continuous bitstream that will remain in sync with the corresponding video just fine as long as we can assume that it starts at the right moment and that both devices (the HVX and the audio recorder) are properly clocked internally at whatever rate you choose (48KHz/16 bit being the native format of the HVX). This is a pretty safe assumption, given the audio clock stability of most modern devices.

In post, you can use the tried-and-true slate clap method to visually line up the audio start from both the HVX's onboard mics (which are always being recorded) with your audio recorder's tracks in FCP. The only *tiny* wrinkle to remember then is that sound travels at about 1 foot per millisecond, so if one mic is close to the slate and another is 40' away (however unlikely) there will be a 40 ms time delay (of the slate clap) in the distant mic. One frame = 33ms so this means the audio sync would be improved by advancing the distant mic's audio start by one frame relative to the close mic's slate pop. It's certainly easy enough to do in FCP.

If the audio is being recorded into something like Pro Tools, Digital Performer or another professional computer-based DAW you can certainly set the audio timeline to be 24 frames or whatever, but the audio itself will remain unaffected. Bitrate is bitrate is bitrate. Only the graphic timeline in the DAW will contain frame-related information, and that won't make the trip over to FCP unless you're going OMF (see footnote 2 below).

Going a little further in the DAW direction, there is another process that might facilitate your workflow considerably. If your audio recorder or DAW can both generate and chase SMPTE time code (as, for example, with a MOTU 8pre or 828MKII and Digital Performer running on a Mac laptop), you could generate audio TC from the 8pre and record it onto one of the line inputs on the HVX as you shoot. Then, in post, you could feed that HVX TC track back to the 8pre, set DP to external TC mode, and it would immediately chase-lock your DP file to the correct frame (in practice this is accurate down to a small fraction of a frame).

As long as the devices are properly word-clocked together (I haven't yet tried it, but connecting the the optical output of an 828 or equivalent to the optical in on a Mac should enable this), then the audio could be transferred digitally back into FCP with no quality loss.

Otherwise, the audio could be transferred in using analog connections with relatively insignificant loss. While you could use the Mac's built-in analog audio input for this, you'd be better off using an external device, like another 828 for example. What you'd get in either case would be near-instant spotting of the external recorder's audio, at the expense of having to do a real-time transfer.

This is the type of workflow I'm planning for bigger shoots, although like I said I haven't yet tested it out. It *should* work fine... : )

----------

Footnote 1: A device like a MOTU 8pre, as its name indicates, has 8 mic preamps, enabling you to record 8 simultaneous mic or line sources. Pretty neat on a large set to have several lavs, shotguns, cardioids, music playback, etc. all captured at once in perfect sync. You can stack up additional MOTU boxes for even more channels if you want : )

Footnote 2: OMF can be used to transfer audio between FCP and DAWs like DP or Pro Tools, especially with the utilities by Automatic Duck:

http://www.automaticduck.com/Mercha...PROD&Product_Code=FCPSuite&Category_Code=BNDL

They're expensive ($800 for the import/export bundle), but should allow extremely fast and seamless audio transfers while retaining all audio edit and volume information.
 
good 24pn info

good 24pn info

Thanks Carlone and Ted. Lots of great info. We did a 8 min take test for 29.97 non drop audio with our HVX200 shooting 24p native, and it's sync was perfect. So I guess that would mean the HVX200's 24p and 24pn settings are actually recording at 23.98 or 23.976 if you want to get picky. Thanks again.
 
The audio is clocked separately at its own rate, completely irrespective of any frame rate indications. So in the case of the HVX the standard video rate for audio of 48KHz is the only meaningful number where time is concerned. That means that 48000 audio samples will be recorded every second, and will play back identically. An audio device's GUI may read out in any of the many frame rates that exist, but ALL of them will produce identical results with 48KHz (or any other given sample rate) clock-time wise.

The only way to change that (or screw it up) is to change the clock rate of the audio on playback, not the "frame rate" (which again, doesn't exist at all "under the hood" as an audio spec). So for example if you recorded the audio on a separate recorder at 44.1 KHz (which is the CD audio standard) and imported it into FCP without rate-converting it, it would play about 10% too fast, which is 10% shorter in duration. It would get wildly out of sync very quickly. You certainly don't want to make that mistake, and it's a very easy one to make. Just double-check your recorder's sample (and bit) rate prior to each recording.

One last point to consider is that for long clips, the difference between one device's version of 48KHz and another's can differ, usually very slightly, but enough to matter. If you're shooting a 2 hour concert non-stop, for example, it's likely that by the end of the show an independently recorded audio capture could be out of sync by a noticeable amount. Imagine starting two quartz stopwatches simultaneously and finding that two hours later they disagree by a quarter second. Same idea. That's 6 frames though, which is going to be a problem. The SMPTE time code lockup method I described in the previous post would solve that.

For clips of a few minutes in duration though (as you found in your 8 minute test) it shouldn't be a problem. In a pinch, you could also apply a few well-placed nips and tucks to the audio in post to keep it lined up. Use the HVX's built-in audio mic tracks as a reference for that, unless your audio recorder's mics are far away from the camera. In that case remember to apply a 1 millisecond per foot speed-of-sound delay correction factor. 41.7ms = 1 frame at 24fps, so if a (presumably B) camera is 125 feet from the stage/mics, the on-camera mics' audio would be late by 3 frames.

Fortunately, audio-to-video sync is much less critical than audio-to-audio sync, where even a small fraction of a frame's difference can be clearly heard. Very strict lockup schemes are used in audio to address this, but in video a frame or less of "slop" is usually at least forgiveable. Some programs aired on highly respected networks (PBS notably) often have far worse audio/video sync problems than that.
 
Last edited:
You must log in or register to reply here.
Back
Top