How to prevent audio feedback

Harman engineer Brandon Graham offers his top tips on setting up and EQing a system to avoid this common problem, as well as explaining the best way to use suppressors.

We’ve all probably been exposed to far more feedback than we’d like (even one exposure is one too many).

As performers and live sound engineers, it can be frustrating when we’ve done all that we can and still encounter problems with feedback. This article is intended to help you prevent it, and get the most out of feedback suppressors.

There are many things you can do to passively prevent feedback, and it is important to address these issues first before you start maxing out gains and setting filters.

Passive Solutions

Check the polar patterns of your mics. Make sure that your monitors are aimed at the nulls in the pattern to minimise feedback loop gain. Keep your mics off-axis from your speakers when possible, because the gain is lower there (especially for high frequencies).

Keep as much distance between your mics and speakers as possible. The feedback gain reduces with distance.

Less is more. The fewer microphones and speakers you have in your system, the fewer feedback paths. These feedback paths can build on each other, so you may get more gain-before-feedback by using less equipment.

Watch out for room reflections. They can be strong enough to cause feedback.

Now we reach two conflicting goals: achieving the ideal tone for your performance and maximising gain-before-feedback. If you are still experiencing feedback problems after doing your best with passive reduction, you will have to start using EQ.

EQ Solutions

For maximum gain-before-feedback, EQ for a flat response coming from your mics. One broad EQ boost can result in several feedback problems. Better to solve the EQ problem now than to deal with feedback later.

A room full of people has a different response than an empty room. This is one reason why a feedback suppressor is useful, because you can never anticipate all feedback problems ahead of time.

About Automatic Feedback Suppressors

As a signal processing engineer at Harman, I contributed to the development of the latest dbx feedback suppression products. During that time I had the opportunity to test feedback suppressors from all major brands, and one thing became clear – everyone has improved their algorithms significantly since their first generation of suppressors. They are better at distinguishing between feedback and music, faster at eliminating feedback, and more careful about how much they notch out of your audio.

In some ways, we often ask automatic suppressors to do the impossible. We can imagine what our high-gain setup would sound like without feedback, so we think that it should be possible to maintain perfect sonic integrity at those levels. This is like expecting a refrigerator to preserve your food without allowing it to lower the temperature. There may be some tricks it could try, but when it gets really hot something is going to spoil. If we understand the limits of automatic suppressors, we will be more capable of using them to their full potential.

Note: if you operate near maximum gain, be aware that you are adding strong comb-filter-like EQ to your sound. Well-placed feedback suppressor notches can cancel some of that and help you get closer to your desired response.

Your feedback suppressor should be used for two things – revealing and eliminating potential feedback before your performance, and quickly eliminating any feedback during the performance.

Before the Performance

Every feedback suppressor has some form of ‘Setup Mode’, which is intended to help you address your main feedback problems before the performance starts. Use it! It is far better to deal with it now than during the show. Here are some useful tips to make your suppressor setup more effective.

Rerun your setup mode for every show you do. Don’t expect to save a preset and have it be totally effective the next time you set up. Feedback frequencies are dependent on many factors, including mic and speaker positions, where the performers and audience are, and even the room temperature.

As you ring out your system, have the performers stand in their positions (with ears covered) if possible. Feedback often occurs as a performer nears a microphone or speaker.

You can tap the micropones or clap to elicit feedback. Be aware that in setup mode, feedback suppressors don’t distinguish well between feedback and music. Ensure your performers don’t talk or sing into the mics, as you may get some filters set on them.

If all of your notch filters are lumped together in a small group, this is an indicator of a larger EQ problem. Do a second round of EQ to address the problem and then try again. It’s likely that the first handful of filters will address the worst problems and give the most benefit. After that, you may find that you can’t increase the gain much before running into several new feedback tones. This is a good indicator that you are nearing your maximum operating gain.

During the Performance

Every feedback suppressor has a ‘Performance Mode’ where it works hard to distinguish between music and feedback. This is meant to monitor your performance and only kick in if feedback happens. If you’ve set up your system correctly and aren’t pushing the limits with your gains, this should only be an issue when the unexpected happens (ie, when the vocalist points the mic at the monitors).

For the fastest response, use the widest filters possible. This ensures that each filter tackles the broadest potential feedback region possible. If you’re worried about tone, use the narrowest filters and operate at the lowest practical gains.

You can also set your filters to auto-remove based on a timer. This ensures that they are only active for the short time they are needed during a feedback emergency.

For more on feedback prevention, a detailed paper on the topic can be found here.