Feedback is the bane of any audio engineer's existence. So learning how to smite it is a must for live sound. If feedback always seems to creep up on you, read on to be rid of it once and for all!
Contending with Feedback
It has been said the best compliment an audio engineer can receive after a show is no comment at all. Rarely does an audience or band critically pay attention to the sound itself. It's all about the music or keynote speaker. But when feedback shows up, boy does everyone take notice!
We need to be at the ready to fix feedback at a moments notice during the show. However, preventing it in the first place is always a better approach! One of the easiest and most efficient ways to do this is by peaking or ringing out the room.
What Does It Mean to "Peak" a Room?
The idea behind peaking a room is simple. Bring up the microphone gain until feedback occurs, and then knock down the problem frequencies. By doing this, you hit the peak amplitude allowable for the room. Think of it as finding your limitations.
Sometimes we are able to get our gain really high before the room even begins to ring. Other times it might feed back at even the slightest hint of gain. No matter how many times you may peak out the same room, remember that things change so always start from square one each time.
Prepping to Peak
Before we go about actually peaking a room, we need to set ourselves up so we don't peak at the wrong reference level. We also need to account for where in the room the microphones will or could end up being. Never leave anything to chance!
One of the biggest problems with peaking is the gain staging of the setup. If the monitor amplifiers are way too hot, then it will seem like you are feeding back really easily. On the flip side, if the amps are way too soft, then it might lull you into a false sense of security.
Here is a safe procedure to get you started:
- Make sure your speakers are turned all the way down.
- Place either pink noise, or a test tone (most likely 1kHz) on a track.
- Bring up the gain until you see a solid level on your mixer just like you would for a mic. Somewhere between 12 and 3 o'clock is a safe range.
- Next, raise both the track and master faders (or AUX/submix in the case of monitor mixes) to unity. Remember you should not hear anything yet.
- Now, slowly bring the gain up on the speakers until you reach your desired RMS amplitude. Use a sound pressure meter from a reasonable distance to measure the RMS during this process.
- You should now have an even gain stage throughout the whole system.
The trick here is determining what your RMS should be. Obviously a rock band is going to be much louder than a presenter at a conference. I recommend giving yourself just a little extra headroom, as inevitably people always want things louder, and louder means more chance of feedback!
Worst Case Scenario
Live sound engineers traditionally use cardioid or hyper-cardioid microphones on stage. There are of course cases for hemispherical and omni, but they are rare. Why? Because a directional mic like a cardioid rejects sound better from behind.
This, of course, is paramount when trying to tame feedback. Since the stage monitors are going to be aimed right at the musicians (and subsequently their microphone) we want to prevent as much sound getting to that mic as possible. Except for when we are about to do this...
- Turn down any track faders and gain you may have up on your mixer. Also turn down any AUXs you are using if you are trying to ring out your monitor wedges.
- Make sure your master fader/AUX and speakers are left untouched from the gain staging you just performed
- Plug in a microphone on stage and place it roughly at head level as if a vocalist was singing. If you have a podium for a speaker, place the microphone there.
- Turn this microphone around and aim it right into the monitor wedges or out into the audience
While this might sound like sonic suicide, remember that the microphone currently has no gain. By setting up our microphone to face into the sound source, we can account for the worst case scenario. You could of course put the microphone literally right on the speaker, but this is highly unrealistic and going to cause more problems than benefits.
With the gain staging set and the microphone in a worst case scenario position, we can finally peak out our room. To do so, you are going to need to ride the fader very gently and use a 31-band graphic EQ. Why a graphic EQ with so many bands? Because we need to notch out very specific frequency ranges so we can leave as much signal unaffected as possible.
- Begin by deciding whether you are ringing out the main monitors or the stage monitors first.
- Grab the corresponding fader or AUX potentiometer for the mains or stage monitors. This will be your primary gain riding control.
- With your fader or AUX all the way down, bring your microphone gain up to roughly 2 o'clock. If you know your vocalist/instrument is very soft bring it up even higher.
Very slowly bring up the tracks fader or AUX until you start to hear a little bit of feedback sneak in.
- Once you hear the feedback bring your fader/AUX down just below where you started to hear the feedback until it disappears.
- Go to your graphic EQ and once again very slowly bring the gain up and back down on each individual band. This will show you which frequencies are feeding back.
- Once you find a frequency that feeds back, bring that frequency down on the EQ.
- If you barely touched a particular band before it fed back, then you will need to bring down that band quite a bit. However, if you had to raise that particular band a lot to make it feedback, then you only need to turn it down a little.
- With your feedback now suppressed, very slowly bring up your fader/AUX once again.
- If you encounter feedback again, repeat the same process above until your fader/AUX reaches unity (0dB) or even a little above.
If you do all of the above and still find you can't reach unity, you might have reached your system's limitations. Remember, as we turn up our systems gain, the more and more gain reduction we will need in our EQ.
Occasionally, a particular microphone in a certain place may feedback so easily that even -15dB of gain reduction is still not enough for how loud the venue needs to be. Bar bands are notorious for this.
If this should arise, you have the following options:
- Accept the lower overall volume, and turn down your monitors so the fader/AUX can reach unity.
- Tell the band or venue owner that the system cannot go louder, and should be turned down. (Good luck!)
- Daisy chain two 31-band EQs together.
While the last option sounds ideal, you may not always have access to additional spare graphic EQs. This setup also induces more noise and phase smearing. Lastly, if a venue needs to be that loud, then you might be violating noise laws and should not be that loud anyway!
Phew! As you can see, peaking out a room can be a time-consuming task. While it might sound like a lot of steps, it really isn't once you try it yourself and understand the procedure. As you peak out rooms more and more, you will get better at finding problem frequencies by ear, and will not need to check every EQ band every time.
The trick is to be gentle with the feedback, and not haphazardly raise and lower gain all over the place. Just because a band is feeding back does not mean you need to drop it by 15dB immediately. If you had to raise that band by 15dB to make it feedback then you probably only need to drop it by 1 or 2dB. If you drop almost all 31-bands by similar amounts then all you did was lower the system's gain!
Also, be sure to do the same procedure with different mics. If one vocalist is using a 55SH and the other a OM2, then these microphones will have very different feedback characteristics. Also, be sure to ring each monitor and the mains separately.
Thanks for reading, and happy peaking!