• Wondering which camera, gear, computer, or software to buy? Ask in our Gear Guide.

Syncing audio -- discuss your method.

Okay, so in a different thread, syncing audio was brought up. In writing about it, I thought it likely that there is probably a great deal of variation in how each of us does this. I think it'd be good to hear the different methods, and of course I'm particularly curious to hear from those who do this professionally. As far as I know, I'm doing it bass-ackwards.

Antihero was the first time I'd ever sync'd audio. I edited visuals first, using in-cam audio, then sync'd. I tried Pluraleyes, but decided that it's really only helpful if you sync all audio, for every single shot, before editing. Editing before syncing makes Pluraleyes pretty much worthless.

So, I ended up not even using my slates. Almost every shot was synced with dialog -- every time someone spoke a P, T, K, etc. I'd go frame by frame, barely watching, but mostly listening to the very beginning of that hard consonant. Drop in timeline, and give it a listen, with both in-cam audio and external audio playing. About half of the time, I nailed it on the first try, and the other half, it was almost always within a single frame.

This workflow went rather quickly. It was tedious, and I hated every minute of it, but I got it done in a couple of (very) long days. This is an 83-minute movie, with relatively fast-paced editing.

I feel like this method worked pretty quickly and effectively, but of course I'm not going to be offended if anybody more experienced points out that my method is ill-advised. I think everyone who has done this should weigh in with their experiences -- noobs and pros, alike.

Cheers!
 
I really like Plural Eyes, I sync everything in advance in one timeline then bring the clips as I need in for editing. It took about 10 hours to do about 25 hours of footage across three cameras and an audio recorder. Let it roll overnight then was good to go in the AM.

When you shoot a lot of coverage and you're using a bit from this take and a bit from that, it seems like it would take a lot of time to do it after it's edited. Granted you may want to try to use audio from a different take, still this sounds like a monster extra step.

That's my audio sync workflow.
 
I shoot w/ a T2i and use a Zoom H4N for audio. I usually preview each take raw off the cam, decide which ones I will use in the edit. My editor ( Cinelerra on Ubuntu Linux ) does not like the audio straight from the cam, so I must open the video file in audacity ( I know it's an AUDIO program, but it will rip the audio from the .MOV file. ) I then save the audio file as an aiff file, then I place the video on my timeline, match the cam audio up w/ the video file. It always is the same length. I match it up by the frame. I then I drop in the Zoom audio and visually synch it with my slate peak then I delete the cam audio and VOILA synch.







.
 
Last edited:
VIOLA synch.

This is a viola:

VIOLA.jpg


I think that you mean "Voilá!"
 
Plural Eyes for the win. Shot a pilot on a 5D and 7D with a Zoom audio recorder and worked flawlessly. I was a little confused in the workflow department and how to speed up productivity. Basically, Drop each scene into it's own timeline with all of its takes in sequential order, add the clean recorded audio in sequential order a track down, then pop open Plural Eyes and make sure the "sequential order" box is checked and you're good to go. Beats the old school way of finding the slate. Most interviews I cut are slated with a hand clap and Plural Eyes.
 
I bitch and moan and complain to the director and/or producer until the editor sends me an OMF with everything in sync.

Many of the editors on low budget indie projects are so disorganized I am never given an EDL, audio clips are not time-stamped and audio clips are not properly labeled. So if I am given only a guide track and thousands of un-time-stamped, unlabeled audio clips (boomed mic, several lavs, plant mic, up to eight tracks) - and the editor won't do the sync work - I charge full hourly rates until all of the audio sync is done. It can take dozens of hours. Most directors/producers learn their lesson the first time. I also let an intern do it; I'm certainly not going to weed through all that crap!
 
Last edited:
The times I've had to synch audio (generally dialog from one clip and video from another -- as I record audio in camera), I've used the waveforms to set a marker on each clip, then just line up the markers. Dead on every time... (also using the hard consonants to see on the waveform).
 
Synced my first 5D/H4n short project by hand using slates and it was a pain in the ass. Have used Pluraleyes for every project in the two years since and I couldn't imagine going back to doing it manually now. I just sync all my footage up front and do all my editing with the synced clips. It used to take a little manual work to create the synced clips but even that's been automated in the latest versions, it's now literally a process of hitting go and coming back to everything ready to start editing. You'll occasionally get clips where the camera audio is too different from the recorder for Pluraleyes to sync (i.e. when the camera is far from the main subject) but I'd much rather manually sync a few clips than all of them.
 
I film the scenes with audio on the camera at the same time I have another mic plugged into my computer to capture the audio using Cool Edit Pro. Then I narrow down the takes to the ones I think would work best. After that I bring the video/audio from the camera and the audio I recorded on my computer into Premiere. I film the scene using a clapboard (sometimes just a clap if I don't have it with me) and in the time line you can see where that sound matches up and sync them together. Then unlink the audio from the video and delete it and use the audio you recorded. Not sure if that's the proper or best way to do it but its been working for us.
 
I don't have an audio editor (yet) for my puter, so I record my sound from an external mic into the camera and basically match up/edit right there on the editing field (I use Adobe Premiere Elements). Other sources I just upload and again, "manually" as it were sync up my audio. (Fortunately my shorts so far have all been silent ones really, so it's a bit easier :P :lol:)
 
Il a un accent grave - c'est "voilà" :)

HEY come on I'm TEXAN we mangulate Englush, Freench and that thar Spanglish languagez real goode.

I will not confuse viola with voilà
I will not confuse viola with voilà
I will not confuse viola with voilà
I will not confuse viola with voilà
I will not confuse viola with voilà
I will not confuse viola with voilà
I will not confuse viola with voilà
I will not confuse viola with voilà
I will not confuse viola with voilà
I will not confuse viola with voilà
I will not confuse viola with voilà
I will not confuse viola with voilà
I will not confuse viola with voilà
I will not confuse viola with voilà
I will not confuse viola with voilà
I will not confuse viola with voilà
I will not confuse viola with voilà
I will not confuse viola with voilà
I will not confuse viola with voilà
I will not confuse viola with voilà
I will not confuse viola with voilà
I will not confuse viola with voilà
I will not confuse viola with voilà
I will not confuse viola with voilà
I will not confuse viola with voilà
I will not confuse viola with voilà
I will not confuse viola with voilà
I will not confuse viola with voilà
I will not confuse viola with voilà
I will not confuse viola with voilà
I will not confuse viola with voilà
I will not confuse viola with voilà
I will not confuse viola with voilà
I will not confuse viola with voilà
I will not confuse viola with voilà
I will not confuse viola with voilà
I will not confuse viola with voilà
I will not confuse viola with voilà
I will not confuse viola with voilà
I will not confuse viola with voilà
I will not confuse viola with voilà
I will not confuse viola with voilà
I will not confuse viola with voilà
I will not confuse viola with voilà
I will not confuse viola with voilà
I will not confuse viola with voilà
I will not confuse viola with voilà
I will not confuse viola with voilà
I will not confuse viola with voilà
I will not confuse viola with voilà
I will not confuse viola with voilà
I will not confuse viola with voilà
I will not confuse viola with voilà
I will not confuse viola with voilà
I will not confuse viola with voilà
I will not confuse viola with voilà
I will not confuse viola with voilà
I will not confuse viola with voilà
I will not confuse viola with voilà
I will not confuse viola with voilà
I will not confuse viola with voilà
I will not confuse viola with voilà
I will not confuse viola with voilà
I will not confuse viola with voilà
I will not confuse viola with voilà
I will not confuse viola with voilà
I will not confuse viola with voilà
I will not confuse viola with voilà
I will not confuse viola with voilà
I will not confuse viola with voilà
 
Syncing audio? I'n my limited experience, just crack on and do it! Sync everything first for music videos, for narrative just make an edit with the camera audio then find your proper audio and link it!
 
Not sure about other cameras, but the 5D has a known issue of starting 2 frames off when you use Plural eyes. Also ANY time you have two different sound sources they will drift from each other.

In which host application? I've personally never seen the 2 frame offset in FCP.

There was a well known drift issue with FCP and the 5D early on, but it was a bug in FCP that was easy to correct for once it was known. It also only occurred with true 30p material, so once the 5D firmware was updated to shoot at 29.97 it was no longer an issue. While it's generally true that any two recording devices are going to drift, in practice it's rarely an issue with narrative projects because you're shooting fairly short takes - especially considering the 12 minute limit on Canon DSLRs. I've run tests with my 5D and H4n that showed less than a frame of drift at the 12 minute limit so I've never really worried about it.
 
In which host application? I've personally never seen the 2 frame offset in FCP.

There was a well known drift issue with FCP and the 5D early on, but it was a bug in FCP that was easy to correct for once it was known. It also only occurred with true 30p material, so once the 5D firmware was updated to shoot at 29.97 it was no longer an issue. While it's generally true that any two recording devices are going to drift, in practice it's rarely an issue with narrative projects because you're shooting fairly short takes - especially considering the 12 minute limit on Canon DSLRs. I've run tests with my 5D and H4n that showed less than a frame of drift at the 12 minute limit so I've never really worried about it.

I just saw people going nuts about it on another board. Like I said, we sync eveything manually to the slate unless it's a small cut out from another take, and then we use peaks like hard consonants. I never even knew such a thing as pluraleyes existed until about a month ago, and don't really see the point of it.
 
I've officially been on a set with separate sound recording... they're going to have a nightmare synching it as the slate guy and the person keeping the records weren't always on the same page. They had some repeated takes, some misspoken takes and some misnumbered takes. I'm glad I'm not doing the synch on that... I'll stick to my in camera recording.
 
Back
Top