Live music video question

[Rick03079]

Hi all,

I work in a small bar in Ma. where I do the house sound and lights for the bands that play there. Recently we had in the band Believer, an OZZY Tribute band who is very good. I recorded the band 16 track on DP and recorded the video on my Sony HD AVCHD Camcorder. I have FCP and what I want to do is combine both audios from the 2 sources. I was thinking of mixing the tracks down and just using the completed audio with the video but now I'm thinking I would like to "enhance" the video's audio track. I am new at the whole video thing when it comes to music. It is something new I am trying at the bar to see if I can make a few extra bucks. I know I can open the video in Dp. Since I am new at this and with a quick read of the manual I believe I can sync the video with the audio tracks and bring up the guitar or vocals while adding effects where they need to be and also keep the audience reactions. The dry plain audio track I think would be too "sterile".
I know this was a long post and I'm sorry but I'm new and a little confused .

Thanks to everyone in advance for putting up with me.
Rick

[crduval]

I'm no expert when it comes to the video side of things (never used FCP), but here is a couple ideas for what they're worth...

Using FCP, you should be able to export the audio from the camera as an aiff or other standard stereo audio file. Once you do this, you can import the file into DP using the soundbite window. Set up a new stereo audio track in DP in your "Ozzy Rules!!" Project and import the camera soundbite.

You can now align the exported camera audio with your other 16 audio tracks. If the Camera audio started before the DP audio, then you should use the camera track as the reference and move the other 16 tracks to line up with the camera track. This will help you later.

Now you're ready to mix to your heart's content - once you are finished, bounce your mix using Bounce to Disk to create a stereo master track. When you select the tracks for bounce to disk, make sure to use the time range of the camera track to bound your selection. Make sure to bounce to a format that FCP can import (check FCP documentation for file formats).

Now, FCP should be able to import your final mix. Import the new audio, replace the old, and you should be good to go. If you use the original audio time range, it should sync right up to the video. If not, I'm sure there are tools there to help...

Note that if you can export SMPTE from your camera along with the audio, that would help too - just make sure your frame rate in the DP project is set to the rate your camera used - then you can maintain sync with the video perfectly.

Also, DP can directly import QT movies - so if you output QT video with your audio (make sure the audio is uncompressed), you will be able to see the results in DP (lo res) before your final video edits in FCP.

Hope that helps you!

BTW, there is a new Ozzy autobiography just out ("I am Ozzy") - check it out if you are a fan!

[MIDI Life Crisis]

Actually it's even easier.

Option 1: In DP open the video in the movie window and from the drop down select "import audio into sequence." Once imported, go to the movie window and mute the sound (reduce the volume in the lower left corner to zero). Import your DP audio if it's not already in the sequence and mix to your hearts content. From there you can either export the audio for FCP (if you're adding titles, FX, etc.) or bounce the movie right from DP if you use option 2.

Option 2: I'd probably edit the video footage first and import the finished movie with titles and FX as a movie to DP (but I'd also import the original movie sound first for use in DP as in opt 1) and then edit the sound to my edited video in DP rather than the reverse. That's basically how I did my demo reels.

[mrbillet]

What they said. It'll work. Maybe a bit of nudging about with the audio because they weren't clocked together and the speed might be slightly different between the devices. You can have a slight phase shift happen throughout the show. I think this is less of a problem when blending a distant camera mic with more direct mics off the board.

For even more fun get another camera for a angled medium shot of the stage. Then you can have something to cut to. Even a crappy web-cam over the bands shoulder, looking into the crowd will ad some real intensity (assuming they're getting the job done onstage).

I love the Multi-clip editing feature in FCP for live music video. If there is any flash photography happening you'll have plenty of sync points in the video to make sure everything is in its right place.

I used to love doing these multi-track audio/video concert shoots. Kind of wore me out though lugging all that gear around and no forseeable money coming in.

[Don T]

Hello,
I have one some more things to add. Edit your footage with the audio stripped from the video. I am assuming that the band wants a complete song. In this case the camera audio can act as the "master". After editing (it will take a multi-camera shoot to make it exciting unless you want to edit in, non timeline footage).
Export the movie as a compressed file if you need to conserve CPU power. Maintain frame rate at 29.97 so your audio edits will be frame accurate.

Open the movie in DP. In the mini menu extract audio to a stereo track. Mute the movie audio. Do this because Quicktime will have a latency during play back and you want to line up your audio tracks to be phase accurate.

The trick here is to keep the camera audio locked to the clips used and the timeline so onscreen action matches audio. "B" roll cut ins will have to be as synced as you can make it by eye if you only use one camera. If using multi-camera footage it all can be synced.

Don

[newrigel]

I'd just mix the song to taste in DP and export the file as AIFF and import all of the files into FCP and then you have a guide track to line up the audio to and mix in the cameras audio for the audience etc. pretty strait forward really. I don't see why you'd even need to import the video into DP... your not spotting anything to a hit point (unless your adding something that wasn't played live... but then it wouldn't be a live video). It's all linear so it's strait forward, mix in DP export file to FCP and nudge to get it tight... done. Any titles you can add to the project any way you want because you can slide the video and audio assets anywhere to make room for intro credits etc. If you really start producing the mix in DP then the band usually performs to the finished track on the video unless of course it's a true live video.

[Rick03079]

Thanks for all your answers everyone. I have recently picked up a second camera and will be getting more in the near future, and am using the multi-cam function in FCP to do the cuts. It's really sweeeeeeet . Since it is all "true live" and I am recording 16 tracks at the bar I think my best bet is just to remix in DP and import into FCP. I have done that with a CD recorded live off the board but I want more control over the mix. Anyone who has mixed in a small club knows that your not just mixing the band your mixing what's coming off the stage as well and sometimes the guitar or bass is pleny loud enough, lol. It's just a matter of figuring out what to charge for such a service to the bands to have a multi camera shoot with remixed synced audio.

[dp-user]

I've done this quite a bit with my own band, and can tell you that the biggest hassle has always been keeping the multiple camera and audio sources in sync. The clocks on cameras are generally really bad, and unless you have a common clock, they will all drift apart. Without common timecode and a common clock (which are basically not options for sub-pro video gear), synchronizing them in post is the only option. When it comes to that, you need to think about the two primary components of sync fixing: syncing start time, and fixing drift.

For start time, I usually focus on a click. With bands this is usually easy, as a drummer will often start the song with a stick count. Align the sharp transients in the waveforms (shifting as appropriate for camera and mic distance, if you want to be extra-perfectionistic about it...) of the first (or loudest) stick click. This is easy to do in DP or in FCP, although in FCP you do not get sub-frame accuracy very easily (and even when you've figured out how to do it, you still only get 100th of a frame, not sample-accurate). Even if you're not using the camera audio, keep it, to help you sync the FOH audio to the camera video. If you're only using the FOH audio feed (i.e. not using the camera audio as well), then the sync only has to be close (i.e. within a frame), but if you're also mixing in the camera audio, then phasing will quickly rear its head as they drift, unless the sync is very good.

As for drift, if you can keep the sequences very short (i.e. one or two songs), then it tends not to be an issue, but if you want a single sequence for the length of the performance, the drift can accumulate fairly quickly. And since most cameras have very flaky clocks, the drift can be inconsistent, so you may find that even if you change the speed (resample) the audio such that the start and end line up, they may be apart at various points during the performance. Again, this will not be as much of an issue if you're only mating the FOH audio to the camera video, but if you're also mixing in the camera audio (for a bit of ambience), then things get variably phasey as the clocks drift apart.

So how I've dealt with the drift in the past it to largely avoid it... I used a stereo mic suspended above FOH (the bar, in your case) pointed at the stage for ambience, and recorded that to the same device as the FOH mixer feed, so all the audio is second source, to the same device. I then only use the camera audio for positional sync. I then either keep the sequences short (using transitions, and crowd noise between songs, to meld between sequences), or line up the start and end of the sequence at known-good sync points (stick clacks or a bottle being thrown away behind the bar would work well), and resample the audio to match the video timing to those points (since subtle video retiming always looks terrible).

In the latter case, I use the sample times of the differences between the start and end "event" points in both audio sources (camera and mixer feed), to form a delta fraction, and that fraction is the desired rate change. If I'm using FCP, I then apply this rate change as speed adjustment to the audio, and it does the trick fairly well, although not very precisely. If I'm doing it in DP, I then apply that rate change as a multiple to the sample rate (for example taking 48000 to 48097), and either use SoundHack to hack the headers in the sound files to this new value, or (more recently) use DP's built-in sample rate correction feature (new to DP7, IIRC, I've only used it once...). In either case, the result is that the sound is resampled such that the overall length (between the two key beginning and ending "event" times) of the FOH audio now matches that of the camera audio (which has remained unchanged, to obviate needing to retime the video), and the two can be wedded without any drift (other than internal drift mentioned earlier).

Of course, if this is for a fixed installation, and you can spring for decent video gear (who can, nowadays?), then getting a common clock for all devices to slave to (both timecode and clock, so blackburst or house to the video and WC and MTC/LTC to the audio) would be best, so you can avoid these sync-in-post headaches...

And yes, as others have mentioned, using multiple cameras really sharpens up the production value significantly, and gives you a better product to give to the bands (and thus the option to charge more for the service). FCP's multi-clip (the term for multi-cam in FCP) feature let's you easily edit the angles and do jump-cuts or transitioned switches, to taste. Again, though, more cameras means more sync hassles, and since this would involve syncing multiple video tracks to each other, you're best off keeping the sequences short (if you cannot do common clock)...

Hope this helps...