Conference Room Echo Control: Mic Pickup and AEC Tuning

Echo in a conference call happens when the far-end voice leaves the room speakers, re-enters the room microphones and travels back to the remote party. Fix it in this order: place microphones correctly, set gain structure, then tune the acoustic echo canceller (AEC). Raising AEC aggressiveness to cover a poor layout produces clipped speech and does not remove the echo.

This reference is written by Singapore-based AV and IPTV integrator Prestige Solutions for AV consultants and facility teams who commission boardrooms with our professional audio service. It gives pickup distances, a gain-before-feedback calculation, AEC parameter ranges and a tuning order we use on site. Figures are planning values as of 2026; the microphone and DSP manufacturer documentation remains the final reference for any specific product.

How far should a microphone be from the talker?

Keep the talker within about 0.5 to 1 m of a table microphone, and treat ceiling microphone coverage as a smaller radius than the datasheet suggests. Pickup quality falls roughly 6 dB each time the talker distance doubles, so distance trades directly against the gain you need.

Why the difference matters: the level at a microphone from a talker 2 m away is about 12 dB below the same talker at 0.5 m. The preamp has to make up that 12 dB, and it amplifies room noise and speaker leakage by the same amount.

What is gain before feedback and how do you estimate it?

Potential acoustic gain (PAG) is the most gain a system can add before it rings. Estimate it as PAG = 20 log10(D1 x D0 / (Ds x Dd)) minus a safety margin, where Ds is talker to microphone, D0 is talker to listener, Dd is speaker to listener and D1 is speaker to microphone, all in metres.

Worked example, planning arithmetic only. A talker sits 0.6 m from the microphone (Ds). The nearest listener is 3.0 m from the talker (D0) and 2.5 m from the ceiling speaker (Dd). The speaker is 2.0 m from the microphone (D1).

The lesson in the last row: moving the microphone away from the talker removes all usable gain in this layout. In a conferencing room the same acoustic path also feeds the far-end signal back into the microphone, so a layout with little PAG also gives the AEC the hardest job. The levels from microphone to amplifier are set using the method in the end-to-end multi-zone audio gain structure reference.

What does an acoustic echo canceller actually do?

An AEC models the path from the speaker output to the microphone input with an adaptive filter, then subtracts the predicted echo from the microphone signal. It needs a clean reference of what was sent to the speakers, which is why the reference signal must be taken after any processing that changes level or tone.

Rule of thumb: tail length should be at least as long as the room's RT60 energy that reaches the microphone. A boardroom with an RT60 near 0.6 s needs a longer tail than a treated 0.3 s room, but the filter still cannot cancel late reverberation, so add absorption before adding tail length. Reverberation measurement for this is covered in the RT60, noise and STIPA measurement reference.

Where does the echo come from in the signal path?

Most echo faults are layout or routing faults, not AEC faults. Check these causes before touching any canceller setting.

The mix-minus rule matters for rooms with several sources: the far-end signal must never be returned in the signal sent to the far end. If a laptop feeds the far end and also plays the far end in the room, the system is a loop, not a conference.

What is the tuning order that works on site?

Tune in a fixed order so that each step explains the next. Skipping to the canceller first leaves a room that passes the test only at one volume.

1. Fix the physical layout. Position microphones at least 1.5 m from the nearest ceiling speaker where the room allows, and keep table microphones clear of notebooks and laptop fans.
2. Set gain structure. Set each microphone preamp so normal speech peaks at about -12 dBFS at the DSP input with no clipping, then set the speaker output level for comfortable speech at the farthest seat.
3. Verify the reference. Confirm the AEC reference signal moves with the volume and the room EQ so the canceller sees what the speakers actually emit.
4. Set tail length. Start at 128 ms, increase in 64 ms steps while a far-end test call still shows a residual tail, and stop when it clears.
5. Run a double-talk test. Two people speak at the same time, with the far end also talking. Reduce NLP until both sides remain intelligible.
6. Test all microphone and room states. Repeat with partitions open and closed, all microphones active and the room at full speaker level, then save each as a preset.
7. Record the settings. Log the final tail length, NLP setting, gains and reference tap so a replacement unit can be configured the same way.

The last step is the one most often skipped. A recorded configuration turns a firmware update or a replaced DSP from a re-tune into a restore. Additional routing and room-control work around these presets is listed on the Prestige Solutions home page.

How should the room be checked after handover?

A short repeatable test call beats a one-time commissioning pass. Agree a five-minute check that facility staff can run when users complain: one far-end test call, one double-talk exchange, and a walk of the table to confirm every microphone shows signal and the mute indicators match the actual state.

Keep the acceptance record next to the room: microphone model and position, speaker positions, the reference tap, tail length and NLP values, and the date of the last test. When a complaint arrives, the first question is whether the layout or the settings changed, and the record answers it.

FAQ

Can a hardware DSP fix echo caused by a bad room?

Only partly. The canceller handles the direct and early part of the echo path. Strong late reverberation, speakers aimed at the microphones and moving furniture all exceed what an adaptive filter can remove, so absorption and layout changes give more improvement than a longer tail setting.

Is a laptop's built-in echo canceller enough in a boardroom?

Usually not when a room system is present. Running one canceller in the laptop and another in the room DSP means two adaptive filters in series, and each changes the signal the other relies on. Use one point of echo cancellation, normally the room DSP, and set the laptop to use the room as a plain USB audio device.

How many microphones can be open at once?

It depends on the mixer, not on a fixed number. Every open microphone adds noise and picks up the speakers, so the automixer should limit the number of active channels and attenuate the rest. Check that the automixer reduces gain as channels are added, rather than leaving every table microphone at full level.

Recommended next step

Contact Prestige Solutions to review your site drawings and current operating pattern. Call +65 8010 2337 or +65 9825 8218, message us on WhatsApp (+65 8010 2337), or email sales@prestigesolutions.com.sg. You can also browse the full product range before the site walk.