TLDR: Why Meeting Rooms Sound Terrible on Video Calls
The post-pandemic meeting room is fundamentally different from its pre-2020 predecessor. Before COVID, meeting rooms served in-person groups. Now, 78% of meetings include at least one remote participant, per Gartner's 2025 Digital Workplace Survey. This changes the acoustic requirements dramatically: a room that works adequately for face-to-face conversation can be completely unusable for video conferencing.
The core problem is that video conferencing systems transmit the room's acoustic signature along with the speaker's voice. In a face-to-face meeting, the brain processes reverberation unconsciously — we have evolved to extract speech from reverberant fields. A digital codec cannot do this. It transmits the reverberant tail as signal, creating echo, coloration, and reduced intelligibility for remote participants who experience the room acoustics overlaid on their own listening environment.
The solution is straightforward: RT60 below 0.5 seconds for rooms under 30 m², absorptive treatment on ceiling plus two wall surfaces, and background noise below 30 dB(A). These targets are stricter than traditional meeting room standards because they must satisfy both human listeners in the room and digital systems transmitting the signal remotely. The treatment cost is typically 3–7% of the AV equipment budget — a fraction of the investment, delivering the majority of the audio quality improvement.
The Global Bank That Wasted £50,000 Per Room
In 2023, a global investment bank completed a £12 million refurbishment of their Canary Wharf headquarters, installing premium video conferencing equipment in 20 meeting rooms. Each room received a Crestron Flex system with ceiling microphone arrays, 85-inch displays, and dedicated network infrastructure. The per-room AV cost averaged £50,000.
Within six weeks, internal surveys revealed that 18 of the 20 rooms were rated "unusable" or "poor" for hybrid meetings. Remote participants reported persistent echo, hollow sound quality, and difficulty understanding speakers seated more than 2 metres from the microphone array. In-room participants heard their own voices returned with a 200 ms delay through remote participants' speakers — the classic echo feedback loop.
An acoustic survey measured the problem precisely. Every room had RT60 exceeding 1.0 seconds at mid-frequencies. The largest boardroom measured 1.6 seconds. The rooms featured floor-to-ceiling glazing on two walls, plasterboard on the remaining walls, and a plasterboard ceiling with recessed downlights. Total absorption in a typical 40 m² room was approximately 6 m² Sabins — against a requirement of approximately 25 m² to achieve the 0.5-second target.
The bank spent an additional £280,000 on acoustic treatment across all 20 rooms. The project manager's observation was blunt: "We spent a million pounds on equipment and forgot to treat the rooms. The £14,000 of acoustic panels in each room made more difference than the £50,000 of AV kit."
How Acoustic Echo Cancellation Actually Works
Understanding why reverberant rooms defeat conferencing systems requires understanding acoustic echo cancellation (AEC), the signal processing algorithm at the core of every video conferencing endpoint.
The AEC Problem
When a remote participant speaks, their voice emerges from the room's loudspeaker. That sound propagates through the room, reflects off surfaces, and arrives at the room's microphone as an echo. Without AEC, the remote participant would hear their own voice returned with a delay equal to the room's acoustic round-trip time — typically 100–300 ms. This is profoundly disruptive to conversation.
AEC algorithms model the room's impulse response — the transfer function from loudspeaker to microphone — and subtract the predicted echo from the microphone signal. Per Benesty et al. (2001), modern AEC implementations use adaptive filters with 128–512 ms tail lengths. This means the algorithm can suppress echo components arriving within that window.
The RT60 Threshold
Here lies the critical interaction with room acoustics. When RT60 is 0.3 seconds, 95% of the echo energy arrives within the first 300 ms — well within the AEC filter window. The algorithm converges rapidly and suppresses echo effectively. When RT60 is 1.0 seconds, significant echo energy arrives beyond the filter window, creating residual echo that the algorithm cannot cancel.
| RT60 (s) | Echo within AEC window (%) | AEC Performance | Subjective Quality |
|---|---|---|---|
| 0.3 | 98% | Excellent | Natural, clear |
| 0.5 | 90% | Good | Clean, occasional artifacts |
| 0.7 | 75% | Degraded | Noticeable coloration |
| 1.0 | 55% | Poor | Echo, choppy speech |
| 1.5 | 30% | Failed | Unusable for remote participants |
Furthermore, reverberation reduces the direct-to-reverberant ratio at the microphone, making it harder for the AEC algorithm to distinguish the desired local speech signal from the loudspeaker echo. This triggers double-talk detector failures — the algorithm cannot determine whether the microphone signal is local speech or echo, causing it to either suppress local speech (making the speaker inaudible) or pass echo through (creating feedback).
Model your meeting room acoustics. The AcousPlan room calculator shows RT60 across all frequency bands and recommends the minimum treatment to reach video conferencing targets.
Acoustic Design Targets for Video Conferencing Rooms
Platform-Specific Requirements
Microsoft, Zoom, and Cisco publish acoustic specifications for their certified room systems. While the documents differ in format, the targets converge:
| Parameter | Microsoft Teams Rooms | Zoom Rooms | Cisco Webex | ISO 3382-2 (general) |
|---|---|---|---|---|
| RT60 (< 20 m²) | ≤ 0.4 s | ≤ 0.4 s | ≤ 0.4 s | ≤ 0.6 s |
| RT60 (20–50 m²) | ≤ 0.5 s | ≤ 0.5 s | ≤ 0.5 s | ≤ 0.8 s |
| RT60 (> 50 m²) | ≤ 0.6 s | ≤ 0.6 s | ≤ 0.6 s | ≤ 1.0 s |
| Background noise | ≤ 30 dB(A) | ≤ 35 dB(A) | ≤ 30 dB(A) | ≤ 40 dB(A) |
| NC curve | ≤ NC 25 | ≤ NC 30 | ≤ NC 25 | ≤ NC 35 |
Note that every platform specifies targets 0.2–0.4 seconds stricter than the general ISO 3382-2 recommendation for the same room size. This reflects the additional demands of digital audio transmission.
STI Requirements
For video conferencing, STI at the microphone position (not at listener positions) is the critical metric. Per IEC 60268-16:2020, STI of 0.60 or above ("good") is required for reliable speech recognition by digital codecs. In a room with RT60 of 0.5 seconds and background noise of 30 dB(A), STI at 2 metres from the talker is typically 0.65–0.70 — sufficient. At RT60 of 1.0 seconds, STI drops to 0.45–0.50, crossing into the "fair" category where codec performance degrades noticeably.
Treatment Specification for Meeting Rooms
Wall Treatment
Meeting rooms require absorptive treatment on the ceiling plus at least two wall surfaces. The priority order is:
- Ceiling: Full absorptive ceiling tile (alpha_w ≥ 0.85). This is non-negotiable.
- Wall behind display: Often forgotten because the screen is mounted on it. Treat the exposed area around and above the display.
- Rear wall: Absorb reflections from the far end of the table back toward ceiling microphones.
- Side walls: Treat at least one side, preferably the wall opposite any glazing.
Glazing
Floor-to-ceiling glazing is the most common acoustic problem in modern meeting rooms. Glass has an absorption coefficient of approximately 0.03 at speech frequencies — it reflects 97% of incident sound energy. Options:
- Acoustic curtains: Heavy drapes (>500 g/m²) with 50% fullness provide alpha of 0.40–0.60 at mid-high frequencies when deployed. Cost-effective but require occupant action.
- Acoustic blinds: Vertically hung absorptive fabric blinds. Less effective than curtains but always deployed.
- Fritted or laminated glass: Marginally better absorption (0.05–0.08) but insufficient as the sole treatment.
- Secondary absorptive screens: Free-standing or wall-mounted panels positioned in front of glazing.
Background Noise
Meeting rooms for video conferencing require background noise below 30 dB(A) — significantly quieter than general office standards. This typically requires:
- Dedicated air supply with attenuated ductwork (not tapped off open plan supply)
- Fan coil units sized for low-speed operation during meetings
- Acoustic seals on doors (drop seals and perimeter gaskets)
- No shared plenum with adjacent noisy spaces
Common Mistakes in Meeting Room Acoustic Design
1. Spending the AV budget without treating the room. A £50,000 conferencing system in an untreated room will sound worse than a £5,000 system in a properly treated room. The room is the instrument; the equipment merely amplifies what the room does to sound.
2. Treating only the ceiling. Meeting rooms are typically low-ceilinged (2.4–2.7 m) with relatively small floor areas. The wall surfaces are proportionally more significant than in larger rooms. Ceiling-only treatment in a 4 m x 6 m room achieves RT60 of approximately 0.7 seconds. Adding two wall surfaces brings it to 0.4 seconds.
3. Using acoustic foam instead of tested absorbers. Thin acoustic foam (25 mm melamine or polyester) provides good absorption above 2 kHz but minimal absorption below 1 kHz. Meeting room treatment requires broadband absorption — Class A or B products per ISO 11654:1997 tested to ISO 354:2003. Foam wedges from Amazon are not equivalent to engineered acoustic panels.
4. Ignoring the microphone-room interaction. Ceiling microphone arrays are designed with specific pickup patterns and echo cancellation algorithms calibrated for certain room characteristics. Placing absorbers asymmetrically or leaving one reflective wall creates an asymmetric impulse response that degrades beamforming performance. Treatment should be balanced across the room.
5. Not testing after treatment. Post-installation RT60 measurement per ISO 3382-2 takes 30 minutes and confirms whether the treatment has achieved the design target. Without measurement, the project is specifying blindly. A handheld RT60 meter costs under £500; a professional measurement per ISO 3382-2 costs £300–600.
Summary: The 90/10 Rule for Meeting Room Audio
Meeting room audio quality follows a 90/10 rule: 90% of the perceived quality is determined by room acoustics, and 10% by equipment. A properly treated room with a mid-range conferencing system will consistently outperform an untreated room with premium equipment. The treatment cost — typically £1,500–3,500 for a standard meeting room — represents a fraction of both the AV budget and the organisational cost of failed meetings.
The design process should always begin with acoustics: calculate RT60, specify treatment to achieve platform targets, verify background noise levels, and only then select AV equipment. Start with AcousPlan to model your meeting room and verify compliance with Microsoft Teams, Zoom, and Cisco acoustic requirements before specifying a single piece of equipment.