I'm not sure if I understand your intent, but a delay of the audio will not prevent feedback. If your desire is to create a bunch of layers of vocal sound and have it continue to build and you want the choir to be able to hear what is building and react to it, then I would either go "in ear" monitoring or exclude the mic input into your monitoring system altogether. The mic output could certainly come back through post effect without problem like any other instrument. A few aux send/returns on your mixer should do the trick unless I'm missing something.

I suppose the intent is a little different to what you might have understood from my first post. The choir are both the performers and the audience. So the mic input must go into the monitoring or, to put it another way, the monitor and main PA are one and the same. This is so that the choir can experience an extraordinary increase in the sound of their own voices.

I suggested a phased delay which might perhaps break up the positive buildup of waves though this is uninformed conjecture. A notch filter is another technique but I wondered if there might be something more elegant.

You will not easily be able to create sound levels above those that are achieved by conventional foldback monitor systems before feedback sets in. For foldback monitors taking the form of floor wedges, stage sidefill types or overheads, the level of reproduced sound at the performer is governed by the microphone gain from that point and not by the type or position of the monitor.

The only way I have been able to get an equivalent sort of system to work is to separate the mouth and the ear acoustically. To do this, you have to fit each performer with an omnidirectional headset microphone on a boom that positions it about an inch away from and directed towards the performer's mouth. With care, uncomfortably high sound pressure levels are achievable without feedback, as long as the performer's head remains stationary.

Take great care with this project - hearing damage from sudden unintentional feedback is a real risk. No amount of delay, phase manipulation or other processing in the signal loop can remove the possibility of feedback while still leaving acceptable sound quality.

Thanks for the feedback.

I think part of my answer is in Boswell's caveat "while still leaving acceptable sound quality". We will be running a drone session with this, so high frequencies may not be necessary. If we use a strong LP filter or a heavy notch, perhaps we can reproduce much of the voice very loud whilst completely damping feedback.

Does this sound plausible?

No, not with conventional microphones. It will simply influence the frequencies at which feedback happens.

The only way I know of to achieve what you are describing is without any PA at all. You need to find a warm resonant smallish hall. Then you set the choir up partially or as a whole to "drone" at the resonant frequency of that hall. You will have a natural build up particular standing waves which then a small section of the choir could vocalize around. In ancient times, certain cave grottos were utilized in this fashion.

I know of no way to accomplish this phenomenon with a PA system unless you utilize in ear monitors-which diminishes the effect of course.

Choir

You would be much better off getting them situated so they can blend themselves. Adding monitors complicates so much and tremendously reduces the quality of the capture of their actual voices.

It sounds like you're trying to be some sort of live-art installation. Speakers pointing inward toward the mics will ALWAYS create a problem. Even if you EQ all of the problematic high & mid frequencies out you'll still have a new batch of challenges. Your low frequency drone notes are the ones that will be most excitable in the room's natural resonance.

Thinking outside the box for a minute, maybe what you need is the equivalent of a number of discrete sound systems. Rather than amplify every mic through every speaker you might try something like this...

If you geometrically find a position where the null of every mic was pointed toward an opposing speaker and positioned the choir to effectively block the remaining feedback sources, you might get to a certain volume level - again the room size and shape would be very crucial in pulling this off. I think in a smallish room you are doomed, but in a large enough room something like this might work. The dimensions of the performance space will dictate how loud the overall volume will get before it all goes horribly - horribly wrong.

The AKG D5 is one of the most feedback resistant mics on the market. A super-cardioid would be at the bottom of the list of mics you would typically use to mic a choir or ensemble of any kind, but mics with high gain before feedback would be ok for close micing each vocalist like suggested in the diagram.

Ways around it

Thanks again for all your advice.

I am still hopeful of finding some way that we can do this or some bastard version of it. One thing on our side is that the voices do not have to remain intact once emitted by the speakers.

So if we put some kind of effect in the audio path, surely we can use that to break the feedback loop? If the speakers are not outputting exactly the same thing as the mics are receiving priot to that output, then feedback will surely not arise? Some processing would allow for that surely? At least that's my feeling.

Then it's a question of finding the effect/process which leaves as much as possible of the original whilst breaking down the feedback loop. Latency, shifting delay, filters, modulators, anything counts, as long as we can get a feeling that the drone is being amplified.

Or perhaps I'm being too optimistic/fundamentally misunderstanding the problem?

DVDHawk:

Yours is an interesting proposition. We could probably achieve this with a stereo setup too.

What you are misunderstanding is that anything coming out of a speaker goes straight into a microphone pointed in it's direction. This is what causes the feedback and not the human voices singing.

Hawk's sketch is about the only way to do this with a PA (or multiple systems in this case).

A stereo setup won't work because you haven't separated the mic's individually. Combining the mic's into a stereo feed to the monitors negates the point of the sketch.

yes, i didn't mean stereo, i mean two speakers in an arrangement like 1 and 5.

TheJackAttack

"What you are misunderstanding is that anything coming out of a speaker goes straight into a microphone pointed in it's direction. This is what causes the feedback and not the human voices singing."

However, if that signal is being processed at each 'cycle' of a feedback loop, the speaker is emitting a different signal to that which it received and no stable oscillation emerges. Is this just mistaken? I'm seeking to break the loop without losing all of the mic's input.

But you can't break the cycle. If it comes out of the speaker it will go right into the mic. Any silly monkey business you apply to the already transmitted signal isn't going to stop that. By the time you notched out the room's natural standing waves and all the usual human voice freq's and everything else you would end up with a sound that was so unpleasing and unnatural as to defeat your goal in the first place. And it would still feedback after a certain volume.

Now by utilizing mic 1/speaker 1 and mic5/speaker5 (of Hawk's diagram) you could generate a bit of extra sound. How much will depend in large part on the microphones and number of bodies between the speaker and the mic's etc. A hypercardioid mic has the best rejection from it's rear side or null but it also has the narrowest pickup pattern at the front. At some point this scenario will also feedback but it is your best option. By utilizing 8 points you can keep overall volumes lower but impact on the choir at a hopefully decent psycho-acoustic level.

What about in ear monitors fed by a distribution system?

The speaker/mic setup 1 & 5 in Hawk's diagram would work from the feedback point of view, but not give you what you want. It would be like an intercom - you hear the other party but not yourself in the speakers.

As I mentioned earlier, I have had a similar (but not identical) setup to what you describe working by using headset boom microphones for each participant. The sound pressure levels could be got unbearably loud, but would have been damaging if feedback ever started.

Do you still want speakers in the room with the mics? or headphones only?

I'm not sure what it is you're trying to accomplish overall. Are you trying to cleanse my chakra, rattle my nutsack, or just create some weird sounds?

I'm just reallly baffled.

"the choir can experience an extraordinary increase in the sound of their own voices."

This is called feedback!