Recommended strategy for multicam event and separate audio

Hi,

I haven't really cut anything since Final Cut 7. I have Premiere but haven't learned it and I'd really like to learn Resolve instead because the Adobe "subscription" may become tiresome and prefer to purchase outright.

I have an event that was recorded with three cameras. All the cameras shot the same frame rate, codec, color profile, and have scratch audio that I hope to sync up with clean audio tracks.

1. I have a clean audio output straight from the board with the mike and music for the moderator.
2. I have clean audio recording from a lav mike on a speaker.

The problem is only one of the cameras was locked down and on for the entire event duration. Two of the other cameras were mobile and the operators made clips by starting and stopping the recording. So my "master audio" tracks are much longer (say 20 minutes) than many of the video clips from the two mobile cameras. Their clips range from 1 minute to 5 minutes.

I was "hoping":

1. I could simply sync all the clips using the waveform method like it was a two-stage setup.
2. And then use the multicam feature to match up shots for "scenes" of the event speakers.

But I am wondering...

1. Is that the best approach?
2. Will the sync feature be able to recognize and match a shorter clip on a longer audio master?
3. Will I be able to manually adjust sync?
4. I should probably use Resolve 14 for this project just to avoid any errors from the beta. Right? So will version 14 be up to this?

Looking forward to advice / recommendations.

I was never really happy with the waveform sync in Resolve (neither in other NLEs).
Try PluralEyes, it rocks. I think they have a fully functioning demo for 10 days or so.

Hey,

1/ The one which will resolve 100% of your needs
2/ This I don't know for I didn't try, but it's designed for it so give it a try
3/ When synced, yes, you can fine tune (try . and , after selecting clips)
4/ Though I don't think major problems of the beta versions hide in the multicam fileds, I philosophically wouldn't use betas for serious projects anyway

Actually you should try the audio way sync in Resolve first, though I doubt what you need will work as a one step action (selecting all clips, including the long master audio, and click sync). But this is what will work in Plural Eyes, just throwing every clip in the main window and finally click "synchronize"...
(I've been cutting tons of multicams for years and years, I don't bother trying multiple options anymore and go to the more relevant and efficient straight forward now, hence the PE suggestion)

Now, you may also try making a multicam file with cams which sync, then "open in timeline" the multi file and manually add another angle there (playing with markers takes time but has often good results in the end). If it works, put this well featured multicam file on the TL and now add your clean audio output on another audio track.
Don't forget to kick your cuts addict operators btw...

Conclusion : if waveforms doesn't work, try syncing just what can be synced and add files in the open multi file and/or use markers, but above all... use Plural Eyes.

Thanks.

I tried Audio Sync in the latest beta before I posted this. It wasn't too far off on some clips, but it wasn't quite right. So then I tried marker sync. The problem with setting the marker was that I had a clear audio cue in the waveform but since the markers are frame-based I could not drop the marker at exactly the same spot on the waveform in the three different camera angles. So they were close but I could still see a difference. And then over time the audio sync drifted on the locked down camera that was on for the duration of the event.

So reading about Plural Eyes and hearing your feedback it sounds like it will work more easily for me and that's worth the price of admission. I'll give the evaluation period a try.

On similar note, even if wave form sync hasn't worked all that well for me, I've been real happy with what I've seen so far in Resolve.

Two of the other cameras were mobile and the operators made clips by starting and stopping the recording.

Step 1, kick their asses! (I'll help.)

over time the audio sync drifted on the locked down camera

That's...not normal. What make and model?

The camera was a Panasonic Lumix GX-8. Let me clarify. I don't think the camera's audio has drifted out of sync. I think the audio from the board and mike feeds that I tried to sync up with the camera drifted.

It may also be a function of using a marker to sync because the waveform match didn't hit it exactly right.

I'll have to look into it a bit more. This was my first attempt to play with audio syncing and multicam clips.

I wouldn't be surprised if a GX-8 has some drift, but PluralEyes can handle even that (may just need a bit longer).
And I don't subscribe to that ass-kicking, having been in similar situations. With less than pro equipment (and money) available in the field of arts, like concert, theater or performance, you are often limited on media and/or power. So it's quite normal to switch off cameras which don't have a great picture at that moment.

Well, I see you'll succeed in the end, and it's good for you...
About drifts, this may be connected to the pace of different angles recordings (including fps of course) but it's not normal at all and should be checked first.
About markers now, whenever I use that technique, I can tell I prefer trusting my eyes looking at the show more than looking at waveforms (like feet hitting the floors, positions of arms etc) : I have the feeling I'll find common pics to sync more precisely and easily this way (imho).

Now, I just finished cutting 3 multicams of dance shows in theaters. In one of them, a cam was spotted near the stage (bottom left) - and was recording on a external HD but that's not what's at stake - and its sound was just for the sync in post. If this very audio was the one used to be the master one, the global multicam sync would show a couple of frames drift, just because the second audio was recorded by a central cam spotted 30 meters away from the first one... A question of speed of sound probably... It's always the same in this theater btw.
So even if audio sync is the more accurate in the beginning, I never skip the check step consisting in watching the video sync very carefully.

Now, just a word about that kicking joke : it was a joke in the first place and it still is (obviously that's not a way to treat a co-worker). However, if the recording will end up in a multicam edit for sure, never cutting cams is the very first thing I tell everybody, the very first... No matter long minutes of garbage useless footage you will shoot, it's a question of cool work in post : once correctly synced, it is for the whole show.

I may be cursing their names later, but I can't fault the camera men. They were used to shooting events where you wanted "clips". So I imagine in their shoes that I might have inadvertently pressed the button out of habit.

Plus they had two directives.

1. Leave the cameras running for sound sync under ideal circumstances,
2. But watch remaining space on the cards to avoid having a card change come up during key moments. The nature of the event was fluid in length, so you could plan for a length of x to y, but you don't want to miss key moments of a, b, c.

From the raw footage it appears they got me what I asked for.

Plural Eyes
So I ran my first test through Plural Eyes last night and I exported video clips with the new audio just to see what it had done. It appears to have things right on.

My follow-up questions are:

1. Do you use the XML export and pull the clip sync data into Resolve? And if so what XML format do you use since Resolve is not listed as one of the export options? Any other workflow options I am missing?
2. OR Do you use the option to export the video clips with Audio Track replaced? And if so I hope Plural Eyes is somehow not recompressing the MP4 video stream and instead is just replacing the audio and repacking the original video stream in the same container? (I plan to check their documentation.)
3. And if Plural Eyes IS recompressing the MP4 stream then I think I should transcode my original clips to something like DNXHD before running them through Plural Eyes? I was not going to transcode everything and was instead relying on the setting in Resolve to transcode only what was needed for performance reasons.

Use the XML and the original files. XML for Premiere or FCP 7 (which is the same) should work, but I think Resolve 15 is even taking some of the newer versions.

Yep, the right stuff :
- Step 1 : FIRST put your clips in the place (HD/folders/...) they won't leave until the full end of the project > then Plural Eyes and fcp7 (Premiere) xml export > afterwards import of this latter in Resolve > leave this stacked files TL alone for now but copy the entire TL files
- Step 2 : make any fake multicam file with any files in your project > "open in timeline" > erase these fake files and paste what you copied at the end of step one > close (double click the multicam name at bottom left)

You have your multicam file synced by Plural Eyes... Though you may open in TL again to finally add other audio sources for example, better choose to add them in the final TL than in the multicam file, you'd avoid useless issues.
Enjoy your edit now