11.5 Feedback and echo control

Feedback in sound systems

Feedback is one of the most common problems in sound reinforcement. It happens when a loudspeaker's output gets picked up by a microphone, re-amplified, and sent back through the loudspeaker again, creating a self-sustaining loop. That loop produces the familiar howl or screech you've probably heard at live events. Controlling it requires understanding what triggers it and how much headroom your system actually has.

Causes of feedback

Feedback occurs whenever the amplified signal from a loudspeaker reaches a microphone at sufficient level to be re-amplified through the loop. Several factors make this more likely:

• Insufficient isolation between microphones and loudspeakers, whether due to proximity or poor aiming
• High gain levels in the signal chain, which amplify even small amounts of re-captured sound
• Reflective surfaces in the room that bounce loudspeaker output back toward microphones, effectively shortening the acoustic distance between them

The core issue is always the same: sound traveling from loudspeaker to microphone and arriving at a level high enough to sustain the loop.

Frequency response and feedback

Feedback doesn't happen equally across all frequencies. It tends to occur first at frequencies where the combined response of the microphones, loudspeakers, and room acoustics has peaks. Those peaks reach the feedback threshold before other frequencies do.

• The room itself contributes peaks through its resonant modes and reflective geometry.
• Microphones and loudspeakers each have their own response curves that add further unevenness.
• Equalizing the system to flatten these combined peaks is one of the most effective ways to push feedback further away.
• Choosing microphones and loudspeakers with naturally flat frequency responses reduces the number of problematic peaks from the start.

Gain before feedback

Gain before feedback (GBF) measures how much amplification you can apply to a system before feedback occurs. It's one of the most practical metrics in sound reinforcement design.

• A higher GBF means the system can deliver more reinforcement while staying stable.
• GBF depends on microphone and loudspeaker placement, their directivity patterns, the distance between them, and the room's acoustic properties.
• Every 3 dB increase in GBF roughly doubles the acoustic power you can deliver without feedback.
• Optimizing placement and directivity is often more effective at raising GBF than adding electronic processing.

Feedback threshold

The feedback threshold is the maximum system gain at which the loop gain equals unity (0 dB) at any frequency. Above this point, the system becomes unstable and feedback sustains itself.

• It's determined by the GBF and the specific characteristics of the system and room.
• In practice, you want to operate several dB below the threshold to leave a safety margin for changes in room conditions (e.g., audience size, temperature shifts).
• Monitoring system levels during an event and adjusting as conditions change helps maintain stable operation.

Ringing and feedback

Ringing is a sustained, tonal coloration that appears just before full feedback develops. It occurs when the loop gain at a particular frequency is very close to, but not quite at, unity.

• Ringing sounds like a narrow-band resonance layered on top of the program material.
• It serves as an early warning that the system is approaching the feedback threshold at that frequency.
• When you hear ringing, the correct response is to reduce gain at the offending frequency using an equalizer, or to lower the overall system gain.
• Addressing ringing promptly prevents it from escalating into full feedback.

Controlling acoustic feedback

Effective feedback control combines physical techniques (placement, directivity) with electronic tools (equalization, suppression). Neither approach alone is as effective as using both together.

Microphone placement techniques

Where you place microphones has a direct impact on GBF:

1. Move microphones closer to the source. The inverse square law works in your favor: doubling the distance from source to mic drops the direct sound by 6 dB, but the loudspeaker-to-mic distance stays roughly the same. Closer micing improves the ratio of wanted signal to unwanted loudspeaker pickup.
2. Keep microphones away from reflective surfaces like walls and tabletops, which can redirect loudspeaker sound into the mic.
3. Use only as many open microphones as necessary. Each doubling of open microphones reduces GBF by approximately 3 dB.
4. Angle microphones so their least sensitive axis points toward the loudspeakers.

Loudspeaker placement techniques

• Place loudspeakers closer to the audience and farther from microphones to reduce the level of sound that reaches the mics.
• Aim loudspeakers away from microphones and away from large reflective surfaces (especially rear walls).
• Distributed systems using multiple smaller loudspeakers can provide even coverage at lower individual levels, which reduces the sound pressure at any single microphone location compared to one large centralized system.

Directional microphones for feedback control

Directional microphones reject sound arriving from certain angles, which directly improves GBF when loudspeakers are positioned in the rejection zone.

• Cardioid mics reject sound from the rear (roughly 6 dB of rejection at 180°).
• Supercardioid mics offer maximum rejection at about 125° off-axis, with a small rear lobe.
• Hypercardioid mics have maximum rejection at about 110° off-axis, with a slightly larger rear lobe.

Choosing the right pattern depends on where the loudspeakers are relative to the microphone. A supercardioid mic, for example, works best when the loudspeaker is positioned at its null angle (125°), not directly behind it. Proper orientation still matters: a directional mic pointed the wrong way loses its advantage.

Graphic equalizers for feedback control

Graphic equalizers let you cut gain at specific frequency bands where feedback is most likely.

1. Identify the frequencies where ringing or feedback occurs. A real-time analyzer (RTA) can help pinpoint these.
2. Apply narrow cuts at those frequencies using the equalizer. Typical cuts are 3–6 dB on the problematic bands.
3. The goal is a flatter combined system response, which distributes the GBF more evenly across the spectrum instead of having a few frequencies that limit the entire system.
4. Avoid excessive equalization, which can degrade the tonal quality of the program material.

Automatic feedback suppressors

Automatic feedback suppressors use digital signal processing to detect and suppress feedback in real time.

• They continuously monitor the audio signal for the telltale signs of feedback: sustained, narrow-band tonal content that builds over time.
• When feedback is detected, the device places a narrow notch filter at the offending frequency, reducing gain just enough to break the loop.
• Good suppressors can react within a fraction of a second, providing a safety net during live events.
• They work best as a supplement to proper placement and equalization, not as a replacement. Over-reliance on suppressors can result in audible artifacts or a system riddled with notch filters that degrades overall sound quality.

Echo in acoustic spaces

Echo is a distinct, delayed repetition of a sound caused by reflections from room surfaces. Unlike reverberation (which is a smooth, blended decay of many overlapping reflections), an echo is heard as a separate, identifiable repeat of the original sound. Controlling echo is critical for speech intelligibility and musical clarity.

Causes of echo

Echo occurs when a reflected sound arrives at the listener more than about 50–100 milliseconds after the direct sound, with enough level to be perceived as a separate event.

• Large, flat, reflective surfaces (walls, ceilings, floors) are the most common culprits.
• Concave surfaces (domes, curved rear walls) can focus reflections into concentrated spots, producing especially strong echoes.
• The delay time depends on the extra path length the reflected sound travels compared to the direct sound. Longer paths mean longer delays.

Echo vs. reverberation

Both involve reflections, but they're perceptually and physically different:

• Reverberation is the accumulated sound field from thousands of reflections arriving in rapid succession, creating a smooth decay. It adds warmth and fullness.
• Echo is a discrete repetition that stands apart from the direct sound. It typically requires a delay of at least 50 ms and sufficient reflected level.
• Reverberation is usually desirable (within limits). Echo is almost always a problem, because it interferes with clarity and intelligibility.

Flutter echo in small rooms

Flutter echo is a rapid, repetitive echo that occurs between two parallel, reflective surfaces. Sound bounces back and forth in a "ping-pong" pattern, producing a buzzy or metallic coloration.

• It's most noticeable at mid and high frequencies.
• You can test for flutter echo by clapping your hands in a room with parallel walls; a rapid "zipping" sound indicates flutter.
• Treating even one of the two parallel surfaces with absorptive material or a diffuser is usually enough to break the cycle.

Calculating echo delay time

The echo delay time is the difference in arrival time between the direct sound and the reflected sound. You can calculate it from the extra distance the reflection travels:

$t = \frac{\Delta d}{c}$

where $\Delta d$ is the additional path length of the reflected sound compared to the direct sound, and $c$ is the speed of sound (approximately 343 m/s at room temperature).

For example, if the reflected path is 20 meters longer than the direct path:

$t = \frac{20}{343} \approx 0.058 \text{ seconds} = 58 \text{ ms}$

This is right at the threshold where echo becomes perceptible. A 40-meter additional path length would give about 117 ms of delay, which would be clearly audible as a distinct echo.

Perception of echo

The human ear perceives a reflection as a separate echo when the delay exceeds roughly 50–100 ms, depending on the type of sound:

• Speech and transient sounds (claps, percussive hits) have a lower echo threshold, around 50 ms, because their sharp onsets make repetitions easy to detect.
• Sustained sounds (held musical notes, continuous noise) have a higher threshold, closer to 80–100 ms, because the ongoing signal masks the reflection.
• The level of the reflection matters too. A reflection that's very quiet relative to the direct sound may not be perceived as an echo even with a long delay.
• Echo degrades speech intelligibility because the delayed repetition overlaps with subsequent syllables, causing masking and confusion.

Acceptable echo levels

What counts as "acceptable" depends on the space's purpose:

• Speech spaces (classrooms, lecture halls, conference rooms): Echo should be minimized. A common guideline is that any discrete reflection should be at least 10 dB below the direct sound level to avoid intelligibility problems.
• Music performance spaces: Some late reflections can add a sense of spaciousness and richness, but strong discrete echoes still harm clarity and ensemble precision.
• The acceptability of echo should always be evaluated in context. A reflection that's fine in a concert hall could be unacceptable in a courtroom.

Controlling echo in rooms

Echo control uses a combination of absorption, diffusion, geometric design, and (when necessary) electronic processing. The most effective approach addresses the problem at its source: the room surfaces that create the problematic reflections.

Sound absorption for echo control

Absorptive materials reduce the energy of reflections, lowering them below the threshold of perception:

• Porous absorbers (acoustic foam, fiberglass panels, mineral wool) are most effective at mid and high frequencies. They're the primary tool for treating flutter echo and strong wall reflections.
• Resonant absorbers (perforated panels, Helmholtz resonators) can be tuned to target specific low-frequency ranges, addressing low-frequency echo and problematic room modes.
• Strategic placement matters. Treat the surfaces responsible for the most problematic reflections first, typically rear walls and ceilings in lecture halls, or the back wall facing the stage in performance spaces.

Diffusion for echo control

Diffusers scatter reflected sound into many directions rather than sending a concentrated reflection back to the listener.

• By breaking up a single strong reflection into many weaker ones arriving at slightly different times, diffusers eliminate the perception of a discrete echo.
• Diffusers are especially useful when you want to preserve some acoustic energy in the room (for liveliness or spatial impression) while eliminating distinct echoes.
• Combining diffusers with absorbers gives you balanced control: absorption removes excess energy, diffusion redistributes what remains.

Critical distance and echo

The critical distance ($D_c$) is the distance from a sound source where the direct sound level equals the reverberant sound level. Beyond this distance, reflected sound dominates.

• Listeners seated beyond the critical distance are more likely to experience echo problems because the reflected energy is relatively stronger.
• Increasing the room's total absorption shortens the critical distance and reduces the level of late reflections.
• In large spaces where critical distance is unavoidably short relative to the room size, distributed loudspeaker systems help by placing a sound source closer to every listener, keeping more of the audience within the critical distance.

Room shape and echo

Room geometry is one of the most powerful factors in echo control:

• Parallel surfaces create flutter echo and strong specular reflections. Splaying walls by even 5–10° can significantly reduce these problems.
• Concave surfaces (domes, barrel vaults, curved rear walls) focus reflections into hot spots, producing intense echoes. These should be avoided or treated with absorption/diffusion.
• Irregular or faceted surfaces scatter reflections naturally, reducing the coherence of any single reflected path.
• Sloped ceilings and angled walls are common design strategies in performance halls and lecture rooms specifically to break up echo-producing reflection paths.

Electronic echo cancellation

When physical treatment isn't feasible or sufficient, electronic echo cancellation can help, particularly in communication systems.

1. The algorithm estimates the echo path between a loudspeaker and a microphone by analyzing the known output signal.
2. It generates an inverted copy of the predicted echo.
3. This inverted signal is subtracted from the microphone input, canceling the echo component.

Electronic echo cancellation is standard in teleconferencing and video conferencing systems. However, it has limitations: it can introduce artifacts, it adapts imperfectly to changing room conditions, and it doesn't improve the acoustic experience for people physically in the room. Proper room acoustic design and treatment remain the foundation of echo control, with electronic cancellation serving as a supplementary tool.