Beamforming microphones to isolate individual performers in a group

Question

A lot of conference room telephony gear has built-in beamforming microphones. They use an array of microphones and then process the recorded audio, incorporating the phase to isolate sound from one particular direction at a time. Since there is an array of microphones, this direction is effectively steerable in software.

I am curious how well this works in practice, and is it a technique being used to record a group? One scenario I envision this being useful for is making easy recordings of groups.

Suppose I need to record a small choir of 12 people, standing in a single arc. I could set up my small microphone array in the middle and record on a portable recorder. Later on when editing the recording, I might find that choir member #9 always sings out of tune, and it might be desirable to lower his level in the mix a bit. If I recorded with just a stereo pair, this would be a difficult problem. If with my beamforming microphone array, I could extract a single channel for the general direction each performer was standing in, then I could mix in a traditional way and adjust performers individually, all without mic-ing them up individually in the first place.

Is this scenario practical today? Is there hardware and software for doing this today? The only equipment I've seen that does beamforming is for telephony, so I am wondering if there is a quality limitation with the beamforming algorithms.

Answer 1

Yes, regretfully the sound will suffer tremendously from this, which only makes it usable for telephoning and such. There are, however, as far as I know variants of this technique. Schoeps has a digital version of the CMIT-mic that allegedly can do just this, but for a single sound-source. I haven't tried it out myself, so I can't really say how well it works, but evidently there are uses for phase-controled sound even in film.

However, extreme directivity can have its downsides as well. Myself, I always use cardio-microphones in interiour places because I want the fullness of the voice. The more directional the mic, the more it will focus on a single characteristic in the voice, though cardios is much harder than super-cardios and lobars to fend off against acoustics. When it comes to music, and then especially choirs, extreme directivity can sound downright awful. It's a little hard to explain exactly in what way, but to give some comparison: Imagine you wanna make a painting exactly replicating for example Rembrandt (http://www.ibiblio.org/wm/paint/auth/rembrandt/1630/nicolaes-tulp.jpg), which is, no matter what one think of the style in general, highly dynamic, rich in colour and tone, and vivid. But you end up with an extreme closeup in WAY too bright colors that constantly change as the subject moves relative to the microphone. In vivid choirs, extreme directivity with lots of microphones would have it sound like several random recordings of noses and chests whereof most of them are misdirected due to the subjects changing positions continously unconsciously (standing absolutely still for extended periods of time can be very wearing on the body if not trained properly), if ever so slightly, and resulting in horrible phase-distortions, then pasted together as a whole with virtually no room save for some early reflections and a heavily hollow rear. Depending on taste and opinion (as I said, myself I prefer fullness and natural timbre even in film), this doesn't have to be a problem in film as sound in movies are mixed differently, but in any kind of music not using things like this as a tool (I am also have a history as an Industrial musician, so I actually did use extreme phase cancellation as a tool on voices when needed though), with any kind of classical or otherwise organic styles in particular, it sounds horrible and unnatural.

Yet again! In styles like Industrial, some kinds of Dubstep and otherwise experimental and Avant Garde-styles recording choirs like this can be a lot of fun! But ONLY if you re VERY aware of what you are doing, and that it will sound anything but natural and beautiful.