Microphone Feedback: Troubleshooting and Principle Analysis - Blog

Microphone feedback, known professionally as "audio feedback," is one of the most awkward audio accidents in meetings, live streams, karaoke, and online classes: a normal speech is suddenly interrupted by a piercing high-frequency squeal, everyone covers their ears, and equipment is fumbled with in a panic. Many people think feedback means the device is broken or the microphone is poor quality, but feedback is actually a pure physical phenomenon — as long as certain conditions are met, any combination of microphone and speaker can produce feedback. Once you understand how it works, you'll find it is not only predictable but completely preventable and controllable.

Before you start troubleshooting, we suggest opening the Microphone Volume Test page first, speaking normally into the microphone, and observing the level. If you are using external speakers, feedback may be triggered as soon as the volume is slightly high; if you put on headphones, the problem often disappears instantly. This little experiment almost immediately confirms the feedback principle explained below. Once you understand the principle, you can systematically troubleshoot and solve the problem by following the steps in this article.

1. Understand the Principle First: How Feedback Is Generated

The essence of feedback is a positive feedback loop. Sound comes out of the speaker, is picked up by the microphone, amplified, and comes out of the speaker again, then is picked up by the microphone again… and so the cycle repeats. When the loop gain is greater than or equal to 1 and the phase condition is met, a signal at a specific frequency grows like a snowball, eventually forming a sustained high-frequency squeal.

The Three Necessary Conditions for Feedback

A loop exists: The microphone can "hear" the speaker's sound. If the microphone and speaker are in the same room with no physical isolation between them (such as headphones, a soundproof wall, or directional isolation), a feedback loop is formed.
Loop gain ≥ 1: On the path from microphone to speaker, the amplification factor (microphone gain + speaker volume + room acoustic gain) is greater than or equal to 1. In other words, more sound is sent out than is picked up, so it keeps growing.
Phase alignment: The phase corresponding to the loop delay is exactly an integer multiple of 360°, so the feedback signal adds in phase with the original signal. Different frequencies have different delay phases, so feedback usually concentrates at only a few frequency points rather than across the full frequency range.

Why Feedback Always Starts at a Few Fixed Frequencies

A room has its own frequency response characteristics — certain frequencies (called "standing wave frequencies" or "resonant frequencies") are more easily amplified in that space. At the same time, the microphone, speaker, and audio interface also have their own frequency response peaks and dips. When a frequency happens to satisfy all three conditions — "room resonance + device response peak + phase alignment" — it is the most fragile frequency point, and it will be the first to feed back when the volume rises.

This is why, in many cases, simply attenuating that specific frequency slightly with an equalizer can significantly raise the usable volume before feedback occurs — this is called "Gain Before Feedback."

2. Distinguish First: Which Type of Feedback Are You Dealing With?

Not all harsh sounds are standard acoustic feedback. Different manifestations correspond to different causes and solution directions.

High-pitched sustained squeal: Typical acoustic feedback. It usually appears suddenly when the volume rises to a certain critical point and disappears as soon as the volume is lowered. This is the most common type of feedback.
Low-frequency "hum" or "rumble": Usually caused by low-frequency feedback or a ground loop. Low-frequency feedback generally occurs when the microphone is too close to a woofer, or when the microphone's polar pattern is set to omnidirectional while the environment has low-frequency reflections.
Periodic "clicking" or pulsed squealing: Common when software automatic gain control (AGC) oscillates back and forth at the volume boundary. It is not physical feedback but instability in the digital signal processing algorithm. Turning off AGC usually solves it.
Only the other party can hear it, not you: This may be network echo (A's voice goes to B and then from B's microphone back to A), rather than local acoustic feedback. In essence, it is a loop formed between the other party's speaker and microphone and has nothing to do with your local setup.
Only occurs with speakers on, disappears with headphones: A typical physical feedback loop. Headphones isolate the acoustic path between the speaker and the microphone, the loop is broken, and the feedback naturally disappears.

3. Emergency Handling: What to Do First When Feedback Occurs

The most important thing at the moment feedback occurs is to stop the bleeding immediately to avoid harming listeners' ears and the equipment. Operate in the following order:

Lower the volume rather than mute: Quickly pull down the speaker volume first, or lower the microphone gain. Do not directly mute the microphone, because coming back from mute easily triggers it again.
Increase distance: Move the microphone away from the speaker, or move the speaker away. Every doubling of distance attenuates the sound pressure level by about 6dB, which can quickly push the loop gain below 1.
Change the angle: Point the microphone away from the speaker. A cardioid microphone has low sensitivity at the rear and sides, so pointing its back toward the speaker effectively breaks the loop.
Temporarily shield with your hand: Cup your hand near the microphone head to create physical shielding and quickly reduce pickup sensitivity in an emergency. Be careful not to block the sound intake ports with your fingers, as that would instead create a new resonance.
Put on headphones: If conditions allow, immediately switch to headphone monitoring to physically break the loop. This is the most thorough emergency solution.

4. Adjust the Gain Structure: Reduce Loop Gain From the Source

The core of feedback is that the loop gain is greater than 1, so reducing the gain at any point in the loop can delay feedback. The key is "which one to reduce" causes the least loss in sound quality.

Lower the speaker volume: The most direct and effective. Lowering the speaker volume by 6dB directly reduces the loop gain by 6dB, with no impact on recording sound quality (only on the live listening experience). As long as the audience can hear clearly, keep the speaker volume as low as possible.
Lower the microphone input gain: The second-best choice. Lowering the microphone gain also lowers the pickup signal, and you need to speak closer to the microphone to compensate. But gain that is too low makes the recording noise floor relatively louder, so you have to weigh the trade-off.
Avoid double amplification: A common mistake is pushing both the system input gain and the in-software input gain to 100% and then stacking "microphone boost" on top. Every stage of amplification increases the loop gain, so the total gain should be concentrated at the stage closest to the microphone.
Turn off system microphone boost: The "Microphone Boost" or "Gain" levels (+10dB/+20dB/+30dB) in Windows microphone properties contribute greatly to feedback. Avoid using them if possible, and compensate by getting closer to the microphone.
Use an attenuator (Pad): If the microphone itself has too strong an output (such as a condenser mic connected directly to an audio interface), you can add a -10dB or -20dB attenuator to the signal chain to reduce the preamp input level.
Turn off AGC/automatic gain: Automatic gain control automatically raises the level when the volume is low, which is exactly the link most likely to let feedback break through. For meetings and live streaming, it is recommended to turn off AGC and use fixed gain.

5. Microphone and Speaker Position and Angle

Position and angle determine the strength of acoustic coupling and are the most critical physical variables affecting gain before feedback. Proper placement can greatly reduce the probability of feedback.

Don't point the microphone directly at the speaker: Having the microphone diaphragm's central axis pointing at the speaker is the worst placement. At least place the microphone at a 90° angle to the speaker, and ideally have the speaker fall in the "null" or rear of the microphone's polar pattern.
Place the microphone behind or to the side of the speaker: In desktop meeting scenarios, place the desktop microphone facing the speaker and with its back to the laptop speaker. Laptop speakers are in front of the keyboard, so if the microphone does not point toward the keyboard, it picks up much less energy from the speakers.
Increase distance: The farther apart the microphone and speaker are, the greater the feedback path attenuation. When space allows, keep them at least 1.5-2 meters apart, using longer cables or a wireless microphone to arrange this.
Keep the microphone close to the speaker (the person speaking): The closer the microphone is to the mouth, the stronger the voice energy, the lower the input gain allowed, and the higher the gain before feedback. 10-20 cm is a common optimal balance point.
Avoid pointing the microphone at hard reflective surfaces: Large glass, whiteboards, and smooth walls reflect speaker sound back to the microphone, forming an indirect feedback path. You can hang curtains, place plants, or lay carpets on reflective surfaces to reduce this.
Cardioid/supercardioid is better than omnidirectional: In meeting and live streaming scenarios, cardioid or supercardioid microphones have low sensitivity at the rear and sides, pick up less from the speaker direction, and have significantly higher gain before feedback than omnidirectional microphones.

6. Use Headphone Monitoring: Physically Break the Loop

If conditions allow, having the speaker or host use headphone monitoring is the most thorough and simplest anti-feedback solution. Headphones deliver the played sound directly to the ears without going through the room's acoustic path, so the microphone "cannot hear" the speaker's sound.

Closed-back headphones are better than open-back: Closed-back headphones leak less sound, making it harder for the microphone to pick up sound from inside the headphones. Closed-back monitoring headphones are the first choice for live streaming and recording scenarios.
Moderate volume: When headphone volume is too high, sound leaks from the gaps in the headphones or through bone conduction in the head and is picked up by the microphone. Keep a reasonable volume that is just enough to hear clearly.
In-ear alternative: If closed-back headphones are inconvenient, in-ear monitors (IEMs) provide better isolation and leakage prevention and are a common choice for live streaming and stage performances.
Multi-person scenarios: In multi-person meetings, it is not realistic to have every speaker wear headphones. As a compromise, use a highly directional conference microphone placed in the center of the table, with the speaker placed at the far end of the room.

7. Software-Side Processing: Echo Cancellation, Noise Gate, Limiter

Modern meeting software and audio processing software provide a variety of digital processing methods to suppress feedback. Understanding their applicable scenarios helps you choose the right tools.

Echo Cancellation (AEC): AEC in meeting software specifically handles the "speaker → microphone" loop. By referencing the speaker output signal, it subtracts the same component from the microphone signal, fundamentally suppressing feedback. It is enabled by default in Zoom, Tencent Meeting, and Teams and is usually not recommended to be turned off.
Noise Gate: Only opens pickup when the sound exceeds a threshold and mutes when below it. It can completely cut off the microphone during gaps in speech to prevent a loop from forming. But improper settings can cut off the beginning of speech, so the threshold needs to be adjusted carefully.
Limiter: Limits the maximum signal value to avoid instantaneous peak stacking that causes feedback. But a limiter itself cannot stop sustained feedback; it can only slow the onset of feedback.
Automatic Feedback Suppressor (AFS): Professional conference systems and mixing consoles have built-in AFS that can detect and attenuate the frequency point currently feeding back in real time. DAW plugins have similar features, but their cost-effectiveness in home scenarios is not high.
Equalizer (EQ): The most practical manual tool. Applying a -3 to -6dB narrowband attenuation to the frequency point most prone to feedback (usually a narrow band in the mid-high frequencies around 1kHz-4kHz) can significantly increase gain before feedback. You need to find the specific frequency point first through a sweep test.

8. Precisely Eliminate Feedback Frequencies With an Equalizer

If the environment cannot be changed and the placement is already optimal, but feedback still occurs at a certain volume, you need to use an equalizer for frequency attenuation. This is the standard method sound engineers use to handle live feedback.

Step 1: Find the feedback frequency. Slowly raise the volume until it is about to feed back, and note the pitch that is about to squeal. You can use a spectrum analysis app on your phone, or a real-time spectrum tool in software to see the peak position.
Step 2: Narrowband attenuation. On the mixing console or a DAW's EQ plugin, find that frequency and apply a very narrow bandwidth (Q value 8-16) attenuation of -3 to -6dB. Do not use broadband attenuation, as that loses too much useful signal.
Step 3: Repeat to find the next frequency. Every room and every position may have more than one feedback-prone frequency. After attenuating the first one, continue raising the volume, and a second one often appears. Repeat the process.
Step 4: Control the total amount of attenuation. Generally, attenuating 3-5 frequency points can significantly increase gain before feedback, and the total attenuation should not exceed -15dB, otherwise the overall sound quality becomes muffled and unnatural.
Step 5: Verify the effect. After adjusting, gradually raise the volume again and confirm how much higher the new feedback threshold is compared to before. Usually it can be raised by 6-10dB without feedback.

9. Handling Recommendations for Special Scenarios

Different usage scenarios have different feedback mechanisms and priority measures. Here are specific recommendations for several common scenarios.

Video Conferencing / Online Classes

Wear headphones whenever possible to fundamentally eliminate the loop.
If speakers must be used, use the meeting software's AEC and enable the noise gate.
A laptop's built-in microphone and built-in speaker are extremely close together, making it the combination most prone to feedback. You can connect an external USB microphone or use a headset.
Avoid multiple participants using speakers at the same time, as it easily forms a multi-point feedback loop.

Live Streaming / Recording

The host must use closed-back headphone monitoring. This is a basic industry standard.
Add a noise gate and limiter to the microphone filter chain in OBS, with an EQ in between to suppress frequency points.
Use a highly directional dynamic microphone (such as Shure SM7B or SM58) with close-miking, which provides much higher gain before feedback than a condenser mic.
Avoid using laptop built-in speakers for monitoring, and avoid placing speaker amplifiers too close to the microphone.

Karaoke / Stage Performances

Use supercardioid microphones and keep them as close to the mouth as possible.
Place stage monitor speakers in the null direction of the microphone's polar pattern (usually 120°-150° behind and to the side of the microphone).
Use professional feedback suppressors or the AFS function of a digital mixing console.
Leave enough headroom in stage monitor volume, and avoid constantly pushing it higher just to hear clearly.

Multi-Person Meetings / Roundtable Discussions

Use ceiling array microphones or desktop omnidirectional mics, combined with room acoustic calibration.
Do not set distributed ceiling speakers too loud; distribute them as evenly as possible rather than concentrating them at one point.
Do a "gain before feedback test" before the meeting to find the ceiling of usable volume.
Use all-in-one conference devices with AEC and AFS (such as Poly, Yealink, or Logitech conference devices).

10. Microphone Feedback Troubleshooting Checklist

Check the following list item by item, and you can usually solve feedback problems thoroughly:

✅ Confirm whether the feedback is a high-frequency squeal (acoustic feedback) or a low-frequency hum (grounding/low-frequency feedback).
✅ Immediately lower the speaker volume or gain when feedback occurs, and do not directly mute the microphone.
✅ Check whether the microphone is pointing directly at the speaker, and adjust the angle or position.
✅ Increase the distance between the microphone and the speaker (at least 1.5 meters).
✅ Keep the speaker (person) as close to the microphone as possible (10-20 cm).
✅ Prioritize headphone monitoring to physically break the feedback loop.
✅ Turn off system "microphone boost" and software AGC automatic gain.
✅ Use cardioid or supercardioid microphones, not omnidirectional mics.
✅ Enable echo cancellation and noise gate in the meeting software or recording software.
✅ Use an EQ to find and narrowband-attenuate the feedback frequencies.
✅ Check hard reflective surfaces in the room and add sound-absorbing materials if necessary.
✅ In multi-device scenarios, avoid multiple speakers playing at the same time to reduce loops.

11. Summary: Understand the Principle to Thoroughly Avoid Feedback

Feedback is not a device failure but a physical feedback phenomenon formed jointly by the microphone, speaker, room acoustics, and gain structure. Once you understand the three conditions — "loop + gain ≥ 1 + phase alignment" — you will find that all solutions essentially break at least one of them: headphones break the loop, lowering gain breaks the gain condition, placement changes the phase, and EQ attenuation suppresses specific frequency points. With these principles mastered, feedback is no longer an accident you can only passively avoid, but an engineering problem you can actively prevent and precisely control.

If you want to further verify whether your microphone is prone to feedback, or quantify gain before feedback under different placements, we suggest opening the Microphone Volume Test page to observe real-time level changes, or using the Microphone Device Capabilities tool to view the current device's gain structure and effects status. After adjusting placement, gain, or EQ, return to the test page and compare the before-and-after performance, so you can intuitively judge whether the improvement is effective. We hope this guide helps you completely get rid of feedback problems, making every speech and performance clean, stable, and squeal-free.