How to record a sing-along choir track without a pile of headphones?

February 7, 2007 9:10 AM

Lo-fi group recording: how can I record a 6-8 person sing-along choir cleanly if I don't already own a bunch of headphones and splitters?

Imagine this: I have a track of a song-in-progress, and I'd like to add to that a track of half a dozen or so people singing a part. (See also: clapping a part.) It's an untrained group of friends and family—they don't know how to work as a group, and may have essentially zero musical training, but they can sing along to something with a little practice.

So I can't hum a reference note, wave a baton, and get an a capella performance out of them: they'll need to hear the track they're singing to.

But I don't have 6-8 pairs of headphones, nor the splitter hardware to run them through.

I could play the song into the room through a monitor, but I don't want the backing track bleeding significantly into the choir recording.

So: clever tricks? Alternate routes? Good bleed-management compromises, if I monitor? And should I just pick up a bunch of super-cheapo headphones, and if what's a good route for splitting out to 8 or so pairs of 1/8" jack phones?

posted by cortex to Media & Arts (18 answers total)

Response by poster: I dimly recall reading about a hack in TapeOp where someone would monitor a track into a room and record the choir, and then, uh, reverse the phase on the track and do it again, and merge the two tracks and the room sound of the out-of-phase backing tracks would cancel each other out.

It seems plausible, and yet sci-fi: and in an acoustically uncontrolled environment (my apartment living room, possibly), would that sort of thing be likely to work?
posted by cortex at 9:17 AM on February 7, 2007

If you use ear buds you'd only need 3-4 pairs. You'd still need splitters, but those are only ~$3 at RadioShack. You can get cheap ear buds at most dollar stores; they'll be uncomfortable as hell, and the sound won't be great, but it will work for your purposes. If you need the people to be farther apart, you can carefully separate the two wires that make up the headphone cord pretty much to the jack without problems.
posted by JeremiahBritt at 9:25 AM on February 7, 2007

I don't know, but you're already miles ahead of me on your RPM project I see.....;)

I suppose you could always track people individually (or in pairs if you have two pairs of headphones but not 6-8) and then combine the tracks into a submix to sound like a cohesive unit. Although I imagine it would take some pretty creative mixing skills to get it to sound right in the end (I've never tried this, so I guess it might not be that helpful)...

I'll keep thinking.....
posted by dan g. at 9:32 AM on February 7, 2007

If you're not really picky about quality, I've dealt with situations like that by putting the mikes right next to the sound source (so I guess the people's mouths) and playing the backing track through the monitors just loudly enough for them to hear; with enough futzing, I've been able to hit a balance that didn't bleed through a lot. The key is to get the choir to dominate the mikes.
posted by COBRA! at 9:38 AM on February 7, 2007

Response by poster: I don't know, but you're already miles ahead of me on your RPM project I see.....;)

Haven't actually rolled tape yet. Plan, plan, plan some more, etc.
posted by cortex at 9:39 AM on February 7, 2007

In theory, there are phase reversal manipulation things you can do, but I'm pretty sure you'd need a tightly controlled environment.

You're going to need to get them headphones of a sort.

Usually the easiest way to do this sort of endeavour is to record the amateurs first, then edit it to sound alright. Then you get the pros to record to it.

Could you maybe get someone to play an acoustic guitar, or something equally inoffensive, and get them to sing along to that? You could sit it directly in front of them, use a monodirectional (known as a cardioid) mic to stop it being recorded and then EQ out the reflections?

Personally, I'd try playing the backing track softly. Insulate around the mic in the direction of the source, and maybe it'll be ok. Remember that humans have the cocktail party effect, so your choir should be able to tune into it, while hopefully the mic will be overwhelmed by the choir.

Those were just some things that came off the top of my head. I haven't actually tried them myself.
Good luck!
posted by Magnakai at 9:40 AM on February 7, 2007

Can you sing the "choir" part a capella? If you do sing it, can your friends and family sing along? Why not just lead the choir yourself?
posted by nebulawindphone at 10:02 AM on February 7, 2007

Perhaps you could record them one at a time, overdubbing each over the next?
posted by eschewed at 10:40 AM on February 7, 2007

What's so bad about a little bleed? Set your mic to hypercardioid if you can; put the monitor behind the microphone; put sound dampening material behind the choir; have the monitor play as quietly as you can get away with. Then you can do the out-of-phase trick already mentioned. I can't say I've ever done it, but I don't see how an acoustically uncontrolled environment would matter. You record the choir with the monitor bleed, then record just the monitor bleed at the same volume and with the mic in the same place, reverse the phase of that track, and add it to the mix.
posted by ludwig_van at 11:12 AM on February 7, 2007

Response by poster: Good ideas all around. This sort of thread really helps me get past the trapped-in-a-box duh factor. Keep 'em coming.

I'm less inclined to go the overdub route; it works, but (1) I've done it before, (2) it feels very inorganic during and after, and (3) it takes a long time to get enough layers to fake a crowd and massage the tracks convincingly. None of these are showstoppers, but those are the things that have me inclined to go the live full choir route.

I guess I just find the idea of the group of people really reacting to and feeding off each other in one take to be very attractive, at the moment. I'll experiment with the phase trick (thanks for verifying my sanity, ludwig_van) and with controlled bleed, and consider an earbud/Radio Shack-splitter contraption as well.

(What's "wrong" with a little bleed is the fear that I'll significantly change the backing track after recording the choir, and have bleed of stuff that I don't like. I may be recording this choir well before I've finalized the mixes, or even major tracking decisions, of some of the tracks, so it could be a bit dicey. Heh.)

For overdub context, here are some examples of the overdub-based crowd fakes I've done before: me, 8 times, two helpers, several times, two helpers, several times. Obviously it gets some of the feel of a real crowd, but it's not what I want to do this time.
posted by cortex at 11:19 AM on February 7, 2007

Got a hundred bucks to spare? 4 of those = 9 amped outputs. Make yer people bring their own headphones. And contribute 10 bucks each to the cause. To prove they really love you.
posted by nanojath at 11:23 AM on February 7, 2007

Well, could you wait to record the choir until more of the backing track is finalized?

And I wouldn't buy any gear unless it was a real studio-quality headphone amp that would be useful down the line.
posted by ludwig_van at 11:25 AM on February 7, 2007

Also, the tapeop messageboard (which seems ot be down right now) is a great place for recording questions.
posted by ludwig_van at 11:30 AM on February 7, 2007

Also, it probably goes without saying, but you could probably minimize bleed by massaging the track that you're going to send to the monitors for the "choir" to listen to.
I'd imagine you'd want to keep anything really percussive pretty low in the mix for example...
(on the other hand, maybe percussive stuff is easier to eliminate with that little phase trick, and more ambient stuff will really be tough to isolate and will become that intangible "something" that causes your awesome choir part to not sit in the mix.)
Food for thought anyway...
posted by dan g. at 11:50 AM on February 7, 2007

on the other hand, maybe percussive stuff is easier to eliminate with that little phase trick

I'd be inclined to think the opposite; that the most regular, periodic waveforms will be most effectively cancelled out.
posted by ludwig_van at 11:51 AM on February 7, 2007

If you go for the phase trick, be careful with your mic settings and placement. You want a cardioid mic - not hypercardioid - with the monitor directly behind the mic, facing directly down the shaft of the mic (assuming dynamics here, Shure 57 or 58 or comparable). You want the mic perpendicular to the monitor's cone for maximum cancellation; if you've got it tilted, you're going to get leak and it's going to be out-of-phase, so it'll sound extra bad. You'll want to do this in the biggest room you can, and make sure that the wall behind your chorus is angled so that you get fewer reflections of the monitor signals back into the mics.
posted by tylermoody at 3:34 PM on February 7, 2007

Oops, not hypercardioid. I forgot about the little rear node in that pattern. But hanging up blankets on the walls tends to do a decent job killing of reflections.

I'm not sure what tyler means about the leakage being out of phase though.
posted by ludwig_van at 5:58 PM on February 7, 2007

Response by poster: Despite the excellent collection of suggestions, I never got around to actually trying any of this during the February recording project. If I get a chance to do so in the future, I'll follow up.
posted by cortex at 9:49 AM on March 5, 2007