How to merge DV from two cameras with seperate audio?

[lemmy999]

I have DV video from two seperate cameras, and then audio on a DAT tape. Is it possible to get them in a tool like Premiere an get them synced up, then mix the two cameras and use the audio from the DAT source? This is footage from a musical concert. So the audio has to be in near perfect sync with the video. I know that it is possible to pull in two videos and audio and mix, but I do not know how hard it will be to get everything synced up. BTW, I have Adobe Premiere 6.0. Is there any other tool that could do this function with MPEG files? Thanks.

[bugster]

I have DV video from two seperate cameras, and then audio on a DAT tape. Is it possible to get them in a tool like Premiere an get them synced up,

Yes, it can be done, but will require a lot of time and patience.

BTW, I have Adobe Premiere 6.0. Is there any other tool that could do this function with MPEG files? Thanks.

I would not reccomend trying to do this with mpeg files, use the raw DV captured from the cam via firewire, it is much easier to edit than mpeg. Convert to mpeg only when you have finished editing.

[lemmy999]

Thanks.

I guess I will just have to get both videos and the audios in sync and one point, then it should be in sync everywhere.

Is there any way to record differently to make this easier short of having a time code generator fed in to all of the cameras?

[Richard_G]

I do just this but using Ulead Mediastudio Pro rather than Premiere but I believe the process should be very similar, if not the same. I shoot stage productions using 3 cameras but take a composite video feed from one camera and feed it into a VHS recorder. I then take a clean audio feed directly from the sound mixer and feed this into the same VHS recorder. This means I end up with 3 DV video streams from the cameras with audio that includes audience sounds and background rubbish and a further video stream (captured as DV using one of the cameras as a passthrough) with identical video to that from one camera and a 'clean' audio stream.

I place the 4 streams on the timeline and sync them roughly using the video as a guide (any stage pyrotechnics are ideal for this). By displaying the audio from each stream as a waveform, I can then adjust the relative positions of each stream until they are approximately in sync. I then select a pair of streams (disabling the other 2) and pan the audio from one fully left and the other fully right. Playing from the timeline while listening on headphones, I can hear which audio stream is ahead. Adjust one or other on the timeline until they are properly synced with each other. Disable one of these streams and enable another and do the same. Repeat this process until all 4 are in sync then pan all of them back to centre. It takes time, but once done should sound fine (even 1 or 2 frames out is very noticable), it also means that the video is perfectly synced too.

Once this is done, I have 4 streams that are all syncronised to within a frame and I can adjust the relative levels to give decent quality audio (by using 60 - 70% from the clean feed) but keeping the audience sounds there (by using 10 - 15% from each of the cameras). I then have to split the audio from the video on each stream so that I can edit out chunks of video (to cut from one shot to another) without cutting the corresponding audio. Finally, I delete the video that came from the VHS recorder just leaving the audio.

In your case it should be much simpler. Place the 2 video streams on the timeline and get them in sync using the audio (by panning one fully left and the other fully right). Then, when these are correct, put the DAT audio on the timeline and sync it to one of the video streams. Assuming Premiere will do what MSP will do, you can then either cut the audio from the two video streams completely, or adjust the levels so you have good quality audio but still retain some audience sounds.

It will take time to get it absolutely spot on, and you'll suddenly realise it's well into the early hours of tomorrow morning, but it's worth it in the end!

[Richard_G]

Just to add an extra bit and reinforce what bugster says. Do all of this with DV avi files, don't even think about converting to mpeg until you have finshed all of the editing. Only when you have a final finished video do you want to convert to mpeg.

[lemmy999]

Thanks for the detailed response. That is a really good idea to use the audio in the R & L channels to get it synced.

[themichael]

If you are going to be shooting more then do what hollywood does.

Use a clapboard. Or use something that produces a short sharp sound and short visual cue. A light flash and chirp will work. Start all the camera and point at the cue, fire away, and then DONT STOP ANY CAMERA.

When editing, just line up the cue and select the best audio channel and remove the other. Using A/B editting just cut out the parts you don't want.
Simplified version

If using more cameras then use multi-track editing. You still have to fine tune the tracks but this method will get you into the ballpark and almost to home (or the goal)

[Cornucopia]

If you're using consumer-type DV cameras (as opposed to Pro DVCam or DVCPro) you may get audio drift. And you will have that much harder of a time getting things in sync (and keeping them that way).

To help avoid this problem, do a clapper mark at the start AND at the end of your recordings. If both the start and end are shown to be in sync, then even if there is some segment that drifts in the middle, you know it corrects itself sooner or later. This is similar to the 2-pop and Tail-pop that are used for film presentations.

Scott

[Richard_G]

themichael makes a good point. Use a clapperboard if practical (it isn't for me, that's why I said pyros are good) and once the cameras are running, do not stop them!

The stage shows I shoot consist of two performances, each of two 45 minute acts. I use a DV tape per camera, per act. I use Firewire and transfer each act as one continuous DV avi file, so I end up with 4 files, each around 10Gb. The important thing here is that for everything to remain in sync from start to finish the file lengths must be absolutely identical. The only tolerable number of dropped frames is ZERO. Admittedly, with Firewire DV transfer to avi you shouldn't get any dropped frames unless you do something stupid or your system is really poor.

Once the edit is finished, I save the final output still as DV avi. I supply copies of the finished video on both VHS tape and DVD. I produce the VHS tapes by playing the avi files out of the pc to a VHS recorder. I only convert to mpeg when I get to the author stage for the DVD version. I create the DVD as an iso and then simply burn as many copies as are required from that.

One other thing you may find, although I have only ever come across it once, is that the run time of the audio on DAT differs from the run time of the video. I shot a band concert a while ago and recorded a clean audio feed from the mixing desk as wav on a laptop. This was then burned to CD. At the same time, a second audio recording was made of the audience noise and applause using a minidisk recorder. When the two were played back, the minidisk ran fractionally slower so by the end of the gig was almost half a second out. Not a lot but it sounded absolutely horrible! I ended up slicing a few milliseconds out between numbers to keep the sync in over the whole length.

[Cornucopia]

That's where the mark at head and tail would have helped.

You could have synced up both start points, and calculated the difference between endpoint A and endpoint B. Then you could have done a time-stretch/compress to one of the files to get both A's & B's length to be equal. Done, no editing needed! (just machine number crunching time).

Scott

[Richard_G]

That's true, I could have done. But, it has only happened once (when using the laptop and minidisk recorder method) and hasn't done it with the usual gear. The other problem would have been trying to use a clapperboard in front of an audience of 800 drunken students! At the start would be impractical, at the end would be nigh on impossible (and the cameras/recorders wouldn't be able to pick up a clapperboard over the other noise anyway)!

[themichael]

You don't have to use a clapper. There were some toys that flashed and beeped at the same time. You only need something very short burst to sync with. Ideal would be 1 frame.

One thing to watch out for is the speed of sound and the speed of light are not the same. If one camera is farther from the "clapper" the sync signal will be off for the editing. I don't mean a couple of feet (.27 m) either.

I haven't run into any crowd problems, yet. But then I haven't done much work lately.