Topic: audio sync, on a Mac

[cleantone]

Hey. What do you guys do? I'm new to having a camera. I've some DVD's in the past. Sometimes audio syncs pretty well and otehr times not so much. I would even like to think about doing something more professional in the near future. Does anyone know how to sync a camera and audio equiptment at a concert? I know about Time Code but not everything there is to know of course.

In the mean time what are folks doing for single camera footage and audio from another source?

One method that makes sense to me is to extract audio from the footage and load it into a session (multitrack) with the audio I plan to replace it with. I then have the drifting problem you might have with two audio sources from different word clocks.

I know with two camera sources it is easier because you cn cut footage around the audio. With one source you need to cut audio, or time stretch/compress it to work right.

Anyone on a Mac have a good workflow for that? I have digital performer, peak, and final cut pro (which I'm now to).

Any input?

[TNJazz]

Lots of ways to approach this.  The right way is to generate both sync and timecode to all devices in the chain.  This involves using a blackburst generator to send sync to the camera, sync to the Timecode generator and ALSO convert that signal to word clock for the audio devices (Rosendahl Nanosyncs is what we use for this).  In addition, a Timecode generator to send time code to the cameras and also to record a SMPTE track on the audio device.  Digidesign Sync I/O or the Motu MIDI Timepiece A/V will do this.

Most lower end cameras don't have the ability to accept these inputs though.  You'll probably only find this capability on pro grade cams. 

Most of the stuff we do involves lining up things in post.  If there's drift, I often will cut tracks in the middle of song breaks (rather than right up against the next song).  This gives you the ability to line up each track individually to the reference audio (the camera audio).  Most of the time the drift increases with the audio segments, so they need to be moved to the left on the timeline to stay aligned with the video.  Having some dead space at the front and back of each track allows for individual songs to be overlapped (a few hundred ms at most in most cases) and the result with a crossfade is nearly inaudible.

For songs with segues it's a bit more difficult but it's possible to apply a small stretch or constriction to a segment of audio to keep it in line until the next overlap segment.

[cleantone]

Thanks TN,

What you so perfectly spelled out is about what I know. The top of your post is what I might someday like to get to. It is not feasable right now. My camera is not very high end, like a Viper or something. It is sort of low end professional. Some might argue that it is not proffesional. Those are people working with a million dollars work of camera gear of course. I have a Sony HDR-Z1U. I know it deals with timecode but have so far ignored those functions as I do not have a way to generate or distribute it anyway. I've only been using the camera for a couple of months. In the field only for a few hours all in all.

In terms of options for post with things I have shot and will shoot single camera. I know I want to do what I can in the furture to print decent audio to the DV tape. The first outing I left it at automatic adjustment and the audio is totally blown out. The next night I set it to let me trim the input. Keeping the meters below peak the audio is still driven at the internal mic input. On the third outing I hooked up a single mic, Neumann KMS105 (all I had to spare) and the audio is not "great" but not distorted. The camera does have XLR in's and phantom. When I am using a tripod and not handheld I can try to route a submix to the camera. I have thought about wireless for when I am going hand held.

The bottom line is that I for the most part will always have a multitrack or matrix of whatever I am shooting. So for the few shows I have shot, I have multitrack audio that I still need to mix. So me current problem is getting my audio to sync with the video. Like I said, I've done this before. Sometimes the audio will stay in sync for 20 minutes or more. That is pretty easy to deal with in the ways you mentioned. Other times it will not stay in sync for more than a couple of minutes. This is more where I need help.

I am used to the technic you talked about above. This does not always work. Espessially when your needing to make hundreds of edits mid song to make it stay in sync. I want to figure out a good work flow for the time stretch/compress method. I haven't done a lot of experimenting yet. Does anyone have a good method for this on a Mac?   

[hummat]

In FCP you can alter the speed of the tracks. I think it's right-click > speed or duration.  Not in front of my Mac right now.  I've had success with this sometimes.  Other times, not so much. but it's worth a try.

Oh, and I adjust the speed of the audio not video.

[willndmb]

right click??
how could i do that with a mac mouse? click and hold it?

as for matching the cam audio and another source, that doesn't always work either
its hard to get it perfect and there could still be drift
also
depending on how far away you are (just like making a sbd/mic mix on the fly) the words/music might not match the mouth/drums or other "keys" from the camera and the other source

[cleantone]

In FCP you can alter the speed of the tracks. I think it's right-click > speed or duration.  Not in front of my Mac right now.  I've had success with this sometimes.  Other times, not so much. but it's worth a try.

I thougth there was a feature like this. It might be my best bet. I am a totaly FCP newb. My first question would be, how can I calculate the alteration so I do not have to trial and error with 60min long files. I only have an 800mhz G4 right now so processing can take a bit. THIS is something I really need to learn more about. I'm pretty damn sure Digital Performer can make this happen as well but I need to know how to calculate and not spend days guessing. Any info would be great. Thanks.

how could i do that with a mac mouse? click and hold it?

depends on your mouse. I can right click with my logitech optical and do it all the time.

[willndmb]

how could i do that with a mac mouse? click and hold it?

depends on your mouse. I can right click with my logitech optical and do it all the time.

hmm i have the reg old mac mouse that came with my dome base imac
i tried clicking and holding it but that didn't work
because i have been trying to figure out how to stretch some audio recently as well

[cleantone]

Not to derail my thread too much but isn't there an "option" click, "ctrl" click, or "apple" click for a one click mouse. I forget which because it's been a long time since I've had a one click mouse.

[willndmb]

ctrl click
looks like you are right i did a search at apple.com and got a few replies about this
+T

as for your topic, again i think the overall best way is to do the cut and move in between tracks
however there are times (vhs a lot) you have dropped frames causing thwe audio drift and that doesn't always work

i am in that boat right now, thats why i want to try the time stretch/compression

[cleantone]

Thanks.

i am in that boat right now, thats why i want to try the time stretch/compression

yeah, me too.

[wilsonedits]

i don't see why you would need to jack with the duration of the audio... if anything you just have to do a couple of fade outs or remove a frame from the audio....
if you want to do that you can always just export the camera audio and match it up compress in the audio program of your choice...

and if you have more than one angle you should be syncing based on lights or movement...if you try to go by sound your always off

[cleantone]

i don't see why you would need to jack with the duration of the audio... if anything you just have to do a couple of fade outs or remove a frame from the audio....

I am trying to avoid removing audio or making any edits 100 times to the audio to make it sync.

if you want to do that you can always just export the camera audio and match it up compress in the audio program of your choice...

THIS is what I am trying to master. This is what I need to figure out. I want to be able to calculate to edit to avoid trial and error. I want to find the best sounding way to accomplish this as well. I do not want to muck the audio up with the processing. Does anyone have a good workflow for a time compression/expansion that doesn't wreck the pitch or quality??? A way to figure out how much or a ratio or percentange that would need to be changed? PLEASE SHARE.

and if you have more than one angle you should be syncing based on lights or movement...if you try to go by sound your always off

Yeah totally. But with one angle using the sound works okay. Maybe it is not the best way, that is why I made this post.

[guitard]

I don't use a Mac, but I'm sure the same features are available in FCP that are in Vegas Video.

For a single cam shot, I just drop the video and its audio into the timeline, and then I add the alternate audio in there.  As a rule, I never mess with the video's audio (snipping here and there, etc.) because that is my basis for how I need to correct the alternate audio.

Even though it's audio, within an NLE, its length is still based on framerates, so essentially, you're working with little chunks that are 1/30th of a second long.

I line up the audio at the beginning of a song, and then go to the end of it and see what the difference is.  Typically, I just look for a significant sound in the audio - drum beats are usually easy to work with.  If the alternate audio is 13 frames shorter than the video's audio, I stretch it 13 frames.  Forgot to mention - you have to split the audio at the end of the song - but only on the alternate audio timeline - not the video's audio.  I then recheck it and see how it compares.  Typically, it still won't be in synch, but it will be much closer than it was before.  I might do another stretch/reduction, and then if it's still off by a frame or two, I'll look for a spot in the song where I can split the audio and slide it over to overlap for a frame.  Or if it needs to be stretched, I'll split it, slide it over so there's a one frame gap, and then drag to audio segment on the left over two frames so that it overlaps by one frame.  This generally is undetectable to the ear if you do it anytime except on a drum beat.  You can do it during a long note (guitar, organ, vocal, etc.), and of course, it's perfect if you can do it during crowd cheering.