Acoustic Feedback (Larsen Effect)
Definition
Acoustic Feedback (Larsen Effect)
Acoustic feedback occurs when amplified sound from a loudspeaker re-enters the microphone at sufficient level and correct phase to create a self-sustaining oscillation. The resulting tone rings at the frequency where loop gain first exceeds unity (0 dB). SonaVyx detects feedback-prone frequencies using spectral peak tracking and Q-factor analysis in real time.
How Feedback Is Detected
SonaVyx identifies feedback by analyzing the real-time spectrum for narrow peaks with very high Q-factor (greater than 10). Feedback peaks grow rapidly over successive FFT frames and maintain constant frequency, unlike musical content which varies. The detector tracks peak prominence — the ratio of peak amplitude to surrounding spectral level — and flags frequencies where prominence exceeds the threshold, enabling preemptive notch filter placement before full oscillation develops.
Practical Example
During soundcheck, a vocalist moves closer to the wedge monitor and feedback begins building at 2.5 kHz. SonaVyx detects the rising peak in the spectrum within 200 ms and highlights 2,500 Hz in red on the RTA display. The engineer applies a narrow notch filter at 2.5 kHz (Q=8, -6 dB) on the monitor output, increasing gain-before-feedback by approximately 6 dB at that frequency without audibly affecting vocal quality.
The Nyquist Stability Criterion
Feedback occurs at any frequency where two conditions are simultaneously met: the loop gain (speaker-to-mic coupling minus processing losses) equals or exceeds 0 dB, and the total phase shift around the loop equals a multiple of 360 degrees. The Nyquist stability criterion from control theory governs this behavior. In practice, feedback always starts at the single frequency where loop gain is highest — this is the frequency most sensitive to the speaker-mic geometry.
Gain Before Feedback (GBF)
GBF is the maximum system gain achievable before feedback onset. It depends on speaker directivity, mic pattern, speaker-to-mic distance, room absorption, and frequency response shape. A system with high GBF can operate at louder levels without ringing. The theoretical GBF for an omnidirectional mic 3 meters from an omnidirectional speaker in a reverberant room is approximately -10 dB, meaning the system can only amplify 10 dB before howling. Directional mics and speakers significantly improve this figure.
Ring-Out Procedure
The traditional method for increasing GBF involves slowly raising system gain until feedback begins (ringing out), identifying the feedback frequency, applying a narrow notch filter at that frequency, then continuing to raise gain until the next feedback frequency appears. This process repeats until adequate GBF is achieved, typically placing 5 to 8 notch filters. SonaVyx automates frequency identification, making the ring-out process faster and more precise than manual methods.
Prevention Strategies
Effective feedback prevention combines multiple approaches: using directional microphones (cardioid or supercardioid) to reject rear-hemisphere sound from speakers, increasing the distance between speakers and microphones, orienting speakers away from microphone pickup patterns, using in-ear monitors instead of wedge monitors, and adding acoustic absorption to reduce room reverberance. Each strategy increases GBF independently, and their effects combine additively in decibels.
Try It Now
Eliminate feedback step by step — free guided workflow in SonaVyx