Workflow for multitrack audio in film

[Stephen Mark]

I’m fascinated by your comment that working with several production audio tracks is “always done” in the professional world. I edit professionally and never want more than one production track per source clip unless there is no way to get a reference mix of the multi-track master audio. Patrick’s intuitive sense that manipulating 8 tracks of production audio while getting to picture lock is cumbersome is absolutely right by me. And since I checkerboard production audio while editing, using 8 track source would require a 16 track timeline. What I want is a single audio track that references via timecode or actually links to the multi-track original. That’s the track I would synchronize to my video. I’ve been a professional editor for 40 years, but what I’ve only recently found myself having to do is work with multi-track production audio in Final Cut Pro. And it is indeed a mystery to me how to import the multi-track production audio, sync it, and then create those single audio track working clips that I can manipulate until it is time to hand the project to my sound designer (or send it to an audio editing program). If there are no professionals other than me who work this way, I’d be amazed to find it out. On the other hand, if there is indeed someone who knows the best workflow for creating the kind of source material I want to work with in FCP, I’d very much appreciate having that information.

[Michael Gissing]

A recent program I post supervised was done with a Sound Devices 788t audio recorder and the Sound Devices PIX240 video recorder. Twice a day the sound recordist would momentary jam sync to free run code from the camera which was also the code source for the PIX.

The recordist was creating discreet tracks for boom and radio mics and sending a rough mix back to the PIX via a radio link so the editor had no syncing to do, just take the DNxHD files from the PIX and cut. We organised that both audio and video files were stored in folders for each day. Both picture data wranglers and audio used the same folder name with the addition of an A in front of the audio folder names. Files were just numerical increments and names for audio and picture did not need to match. Timecode and reel number did hence the same name for folders (plus the ‘A’).

During the edit some of the audio files were added by the editor where he needed to dig into individual tracks to separate dialog in some complex scenes. They were easy to sync via the embedded timecode which matched the picture.

After the edit we generated an EDL of the audio tracks after managing a copy of the sequence and deleting non sync audio elements (FX, music, narration etc). In a text editor we added the A back into the reel name (simple find & replace). In the Fairlight we ran the EDL pointing the system to the audio folders as reels. Fairlight did the auto conform. A few things were weird but mostly the audio reconformed cleanly and quickly.

A few gottchas. There was a two frame delay from the Sony F3 camera to the PIX240 recorder so all the audio had to be slipped to compensate this was done before the EDL was generated. There was also one grab used where the recordist didn’t roll but the transmitted audio mix back to camera was there so the editor used this not knowing there wasn’t a co-responding polywav to match.

Most devices like the 788t record separate files plus a stereo mix (same mix that was transmitted back to camera) so the transmit was nice but not essential. This method meant that timecodes were common so clapper boards were not required and a clean auto conform path via the humble EDL was possible.

[Stephen Mark]

I appreciate your comments, Michael. You were clearly able to find work with people who gave thought to post-production while shooting. They understood that having a post supervisor is a good idea. They saw the benefit of relating picture and audio timecode on set. The sound man had the decency to generate a rough mix of his discrete tracks. And if the editor was using DNxHD files, I’m guessing he or she was working with Avid. This is the kind of professionalism I’m used to as well. I come to a forum like this when in moments of weakness I agree to work for folks whose budgets are in inverse proportion to their ignorance about editing workflows. The project that prompted my question here was underway before I arrived and its organization left a lot to be desired (by me, anyway). The audio had been shot onto a 24 track master. Whoever had been working on the edits before me had done something or other to reduce the working tracks — perhaps a mix down, perhaps an export and reimport — and since I wasn’t going to have to deal with tracing the work track back to the original sound, I didn’t ask whether that process is problematic. I did, however, come upon a case where the cleanest audio I could find for one moment of the show lived on track 24 of the master audio. It was not audible in the mix downs (which were already occupying about 8 tracks of timeline — way more than I want to be working with} I am not an FCP expert. I kept fiddling with that 24 track audio clip but could not for the life of me figure out how to get the sound I wanted into my sequence. (In Avid I would have just patched it to one of the pre-existing audio tracks in the timeline.) I kept thinking there must be a way to do this and a better general way to work with multi-track recordings within FCP, but despite clicking on everything I could think of while holding down option and whatever other modifier keys I found lying around, I never did figure how. Help sites and tutorials said nothing. And so I posted my question here. If anyone knows the trick of taking a few frames of good sound from track 24 of a 24 track master audio clip and getting it onto track 1 of an FCP audio timeline, please reply. Thank you.

[Michael Gissing]

The 24 track is probably a broadcast poly wav. Sound Devices have free software called Wave Agent that will let you extract individual or other groupings of tracks from a poly wav. You could try downloading that software and extract jut the track you want and then import that into FCP.

Wave Agent maintains timecode so the mono track will have the correct timecode reference as the parent file in case it needs to be rematched later. FCP does allow track patching but perhaps it behaves differently with a big poly wav.

[Stephen Mark]

Thanks Michael. That sounds promising — especially if there’s a Mac version (which I guess there must be if you’re telling me this on an FCP forum). I will definitely check out Wave Agent — at least for the next time. Had to turn that other project over so made it someone else’s problem. And I’m inferring from your suggestion that I look to a third party product that there is indeed no way to get that track cut in using FCP native functions. (Yeah, I know I could create 24 timeline tracks, splice the piece I want across, shift it up to one of the lower number tracks, delete the extra timeline tracks… but while I was doing all that, I would feel as if I’d gone mental.)