Question on RED. Shooting music driven film.

[Brandon Rice]

Two questions mainly.

#1: We have many musical scenes in this film, and we'll need playback on set for the actors. What would be the best method for playback, to maintain as good a sync as we can with the video footage?

#2: Last feature I had to manually sync sound from the production sound mixer. Is there any way to make this more automatic? Also, does RED have somewhere to record "scratch track"?

Thanks for any help on this!

 

[Tom Visser]

Music playback presents a different challenge than typical sync sound jobs in that you'll want to somehow roll timecode from the music clip rather than a free running timecode clock as is normal. I have recommended using the Ambient wireless digital smart slate, the ACD301RFA for such jobs. The production mixer or whoever is responsible for playback can playback sound and have timecode that is embedded on the file wirelessly update the Ambient slate... so that 00:00:50:00 on the music playback is what you will see on the slate. If the playback system is not timecode capable (field recorder with timecode output or ProTools system with Sync IO or Motu Timepiece), you can burn audible timecode onto one of the channels of audio and send signal to the Ambient wireless transmitter.

The other option is to just run free running timecode and then figure out where your cue markers are going to be for playback (beginning of song, 2nd verse, 3rd bridge, etc...) and place 2 pops in the appropriate place. Then in editorial, just use a timecode calculator to figure out where to spot the clips based upon the free running timecode value and the location of the 2 pop.

It would definitely be prudent to input scratch sound to the RED for an "all else fails" option.

 

[Stephen Pruitt]

We've shot several low-budget music videos and just had the music playing (loudly) in the background and then worked off a manual sync in post, which, of course, you are apparently trying to avoid. It isn't elegant, which Tom's suggestions certainly are, but it works pretty darn well. As to getting scratch-track sound into the RED, it's as easy as strapping a scratch mic to the camera, pluggin' 'er in, and setting the levels correctly. (And making sure 48V is on or off as appropriate for that microphone.)

Stephen

 

[JPSendall]

Once you have recorded your separate tracks you can either use PluralEyes for FCP or PrPro to auto sync the tracks and I think PrPro 5.5 now has an auto sync function built in. I'd also recommend a loud clap, even off camera, as it helps PluralEyes do its thing when examining the two sound tracks.

Also I think there is a Ipad App that functions as a clapper but also will play music after the clap at any point in the music track. You can feed that into the camera and feed it into the playback system at the same time.

 

[Marc Wielage]

I agree with Tom Visser above -- Tom definitely knows what he's talking about.

Having done a bunch of music projects in the past (too many music videos to remember), the key to me is to use a timecode slate showing wireless playback timecode, and shoot it prior to each shot. Make sure the music is playing back at the correct speed -- I assume you're shooting at 23.98fps, so the sound needs to be prepped in advance so that it's done at this same timecode rate (played from a CD or Pro Tools or wherever it came from).

Another major issue is if you have dialogue going on before or after the music scene. There have been complex systems using wireless earwigs and transmitters, "thumper" tracks, and other options. The key issues to me are: 1) make sure your post supervisor and/or editor can work with whatever method you choose, and 2) hire a good sound person to make sure this is done right. The music has to be totally off while dialog takes place, unless you plan to replace it all in post with ADR.

There is no way to "automatically" sync with the Red or anything else during editing, but you can get close to it with PluralEyes. A scratch track to camera can be extremely useful. A common trick is to have two audio machines on set: a playback timecode deck just for the music, and a recorder to record time-of-day timecode, dialog on the set, and also dedicate one track to the music (even mono). Assuming the sound recorder has the identical timecode as the camera, you've got a working copy of the music for a rough cut right there.

I don't like the idea of playing back from an iPad or an iPad because it's not playing back with any kind of crystal sync reference and it will drift over time. If it's a 3-minute music video, it may not matter enough to screw up a shot. Still, to me, this is not a kosher way of playing back music and timecode.

 

[Keith Putnam]

Any decent NLE will auto-sync video and audio which has the same timecode. Even if it doesn't, using timecode to sync audio and video manually is trivial. Timecode is your solution, so make sure your video and audio sources are locked to the same clock.

 

[Marc Wielage]

Any decent NLE will auto-sync video and audio which has the same timecode. Even if it doesn't, using timecode to sync audio and video manually is trivial. Timecode is your solution, so make sure your video and audio sources are locked to the same clock.

The problem here is that the playback audio will not have the same timecode as time of day, especially if it's the same song, over and over again, for 12 hours. This will not autosync.

I have seen some high-end productions (like Dreamgirls) use playback sound timecode as user bits, but this may not be an auction for modest-budget shorts. A TC slate works very quickly and easily -- it's maybe 2 hours of work for an assistant editor to sync up 5 or 6 hours of footage, at most. But that's assuming good slates and solid timecode playback. Varispeed shots present their own issues, but that's par for the course.

This brings back bad memories of old 1980s and 1990s music videos where camera operators missed slates and we were syncing up stuff in the bay by eye. I bet 10% of the Shine a Light concert was synced this way -- we got very good at getting precise lipsync based on Mick's position and body actions.