OT: Audio recording with 23.976

Okay, I think I'm right, but can't find any confirmation anywhere.

Feature shooting 23.976 RED.

Audio should be recorded at 23.976 48k right? not 48.048k.

48.048 would be used if this was a film based or true 24fps based right?

Or do i have it backwards?

Hi Tim,

I'm a bit tired explaining these things over and over again.
You guys over there in NTSC world are definitively poor guys in so far everything is strange.

Video is always recorded as full frames, audio is always recorded as full samples.
When you start a project you have to care about the editing environment. In case you use FCP - which I assume - you have to think about which kind of setup you will use. And you MUST keep this setup thru the whole process. And you must understand how FCP handles audio.

In real world a second of audio is a second of audio no matter what sample rate - same for video.
The next thing is the TC and TC rate. This is just a flag to a timestamp at the beginning of the file which may tell an NLE how to handle the file.
The more important thing is: how will the NLE handle and how does the user.
Recording 24 (NTSC or real world) does not have any influence on the audio, you record audio in whatever sample rate you need.
If you need to have audio in sync with video and want to match your system setup then you may record audio with a 29.97 DF flag for TC as most of the setups output to 29.97 DF.
You may also record with 480048 Hz or with 47592 Hz depending on your setup - most NLEs calculate based upon 48 kHz or other even kHz rates.
You may also flag the audio with a virtual sample rate (playback sample rate), which can be done with some audio recorders and some NLEs will look for that - FCP is one of of them if you got the right setup.

So conclusion.
If you use a SoundDevices or Aaton Recorder the TC flag doesn't matter as they always buffer back to a full second. If you use a Zaxcom recorder the TC flag can matter as they use the exact sample for the recording. Other devices may handle it in the one or other way.
The main thing is to setup your NLE(FCP) for the project and have a clear plan and an understanding of what is does and never change it thru the editing project.
Keep in mind: a second is a second.

Regards
Andreas

While it was a worthwhile explanation, and certainly has good information, it in no way answers the question, or even helps anyone understand any sort of workflow.

Feature shooting on RED 23.98... Is there any issue recording audio to 23.976 48k for sync purposes?

Hi Tim,

The answer is in there - it doesn't matter.

Tell us what kind of FCP setup settings you want to use and what kind of audio HDR and whether the audio is going to other systems as well. That is the question.
You can record audio with any setting you like.

Andreas

Well it does matter from a "dealing with the audio on a day to day basis as an editor" perspective.
The only reason people do the 48048 trick is because sync houses can only pull down audio speed for syncing to film by playing it back from MM8's and other bits and bobs of older technology. So if you were to shoot film you might go this route.
As for your day to day syncing of rushes or RED proxies, your methodology is correct.
Camera @ 23.976
Audio @ 23.976, 48k (or 96k)

Syncing in FCP should be a walk in the park. And working in FCP you'll be able to tell if you've accidentally slipped sync by the timecode display in the canvas (or the sync flags in the timeline if you've merged clips). Which is why I say it does matter what you shoot, and while Andreas is right that it will stay in sync no matter what you shoot, it's cleaner to have a 1:1 timecode based workflow.

ak
Sleeplings, AWAKE!

What matters is the workflow within you will use this audio.

As Andrew wrote 48 kHz and 96 kHz are the common ones.
You can modify any BWAV files play rate like with video - though audio does have way more 'frames' than video. You can create the pull down pull up while recording. But that doesn't make sense except you work with Avid. As I said earlier the 'play rate' can be changed in the header of the BWAV file and FCP will take that. Google for 'ConvertTo' that's a set of AppleScript files for the Avid but is useful for FCP as well.
If you stay in FCP only and use an Aaton or SoundDevice recorder the TC flags will always start at HH:MMS:00 and this will be the same with every frame rate NTSC or PAL.
Next and really important thing is that FCP will interpret the playback speed. This depends on the easy setup which was chosen either at startup of FCP5 or with FCP6 the setup which currently is selected. Be VERY careful with that especially if you got one project to setup the sync and another where you edit *.

So if I would open a 48 kHz 23.976 fps in a 25 PAL environment everything is fine - it doesn't matter - a second is a second (in PAL), it's a bit different with NTSC.
With NTSC and a chosen NTSC setup FCP will do a virtual resampling (means it will change the playback speed) and you may see a render bar for the audio depending on your system, the amount of tracks, quality etc.

Andrew confirmed the setup you described with your original request.
I said it depends on more parts of the workflow and that you can record what you want. So another 'standard' for 24 NTSC is to record at 29.97 fps 48 kHz as this matches global NTSC playback behaviour best and many people do it that way when they have to interchange files during post.

Andreas

*)If you stay in FCP it's good idea to 'marry' the AV files by exporting them to QT this way you will never lose sync with subclips or multiclips which may happen when you just link/merge them the FCP way.

Andreas