Anti acoustic feedback devices

[dave2010]

Decades ago I read about circuits which could do frequency shifting, and their use to reduce acoustic feedback. I have been to several meetings of various amateur groups recently where acoustic feedback almost always appears as a problem - you know the sort of thing - "Can you hear me at the back?" Tap, tap, thump on the microphone. "Speak up, we can't hear you" from those at the back of the room who haven't turned their hearing aids on, or plugged into the loop system. More thumps, sometimes above the threshold of pain, then some bright spark boosts the output, and "WOW Schreech ... WeeeeeyooooowwwwwwWWWWWW ..." - absolute ear drum shattering noise as the acoustic feedback builds up rapidly through a wild crescendo.

Maybe anti-feedback circuits don't work perfectly, but I think there should be possibilities to improve things. The circuits I read about did frequency shifting, the theory being that any feedback would not be on the same frequency as the original signal. Clearly frequency shifting could be by a fixed frequency, which would probably make the source sound a bit odd, but maybe not much, or could be by frequency ratios which if done well would preserve harmonic relationships. It might also be desirable to apply adaptive filtering to the signal, as well as some careful limiting in order to give good results. My ideas here are rather speculative.

Although there could be circuits to do this sort of thing, it would probably now be possible to do everything pretty much in software, so diy effort could be spent on software design rather than circuit design.

One other approach, if the positions of the loudspeakers are known, is to put an amp plus additional loudspeaker(s) between the microphone(s) and the main speakers, and to use phase cancellation so that the microphone does not "hear" the sound from the main speakers.

Yet another approach would be to put delays in the line, or phase shift the input slightly - to tune out the main resonances, again which could be done using circuits (analogue) or simply in software.

Possibly a combination of all these techniques would be needed to come up with a really robust and easy to use solution to this problem.

Has anyone tried this kind of thing, and had any success with it?

[Rothchild]

Generally, in live sound, the first places to start to combat feedback are mic placement (and mic handing technique) and gain structure. Keep the mic close but not too close to the source, don't handle the basket (It'll change the directionality of cardiod mics) and keep the gain on the channel in a sensible range.

If you have a stage with foldback on it it (ie on-stage monitoring) the classic solution is to run a multiband eq on the feeds to the foldback. You then run around the stage trying to induce some squeal, judging what frequency the squealing is at and 'notching' it out using the eq (ie just turn down the feedback frequency in the monitors so it's harder to get it back in to the mic).

Depending what frequency the feedback is happening at (and how 'hifi' you need the pa to be) you could also just apply a notching eq to the FOH feeds at the feedback frequency and see if it's tolerable in terms of what it does to the sound of the voice being put through it.

[Oldpinkman]

It depends on the context you want to use them. Stopping the feedback by mic placement, or blocking (you can get studio mike sheilds which could be used in a "church hall" ) arrangement. Good mikes like the shure sm57 are reasonably resistant to feedback (unlike a lot of condensers). My acoustic guitar box has 2 feedback blockers - a notch filter and a phase shifter. Both can help in different circumstances. I couldn't define which - I just try it and see. Active noise cancellation techniques are usually for removing a fairly constant background noise rather than dealing with a feedback loop I believe, but I happen to know a leading expert on the subject, and I'll ask him if you are interested.