Reverberation Time Optimization for Concert Halls

By James Hartley · April 30, 2026

1. Project overview: what, where, who, and why

In late 2024, SonusGearFlow was brought in to optimize the reverberation time (RT) of the Grand Marlowe Concert Hall, a 1,260-seat venue in a mid-sized U.S. city that hosts a mix of symphonic concerts, touring chamber ensembles, amplified jazz, and spoken-word events. The hall had a strong reputation for programming but a growing reputation among visiting engineers for being “unpredictable”: beautiful on some orchestral nights, overly washed out for speech, and occasionally harsh in the upper strings depending on seat location.

The client team consisted of the venue’s technical director, the facilities manager, and an executive sponsor from the arts foundation that funded the hall’s original construction (circa 2008). Our project team included an acoustical consultant (lead), a systems engineer focused on measurement and modeling, and a project manager coordinating fabrication, access, and scheduling around rehearsals and ticketed events.

The primary goal was not to “make it dead.” The goal was to bring the room’s mid-frequency reverberation into a tighter, more controllable range and improve consistency across seating zones—without compromising the hall’s identity as a concert venue. A secondary goal was to reduce stage-to-house ambiguity that was affecting both performer clarity and mix translation for visiting FOH engineers during amplified events.

2. Challenges and requirements at the outset

From the first walkthrough, several constraints shaped the project:

No major architectural reconstruction. The board would fund treatments and modest carpentry, but not structural changes to the ceiling or balcony geometry.
Limited access windows. The hall’s calendar allowed only two dark periods: 12 days in January and 9 days in May. Any work affecting seating, rigging, or stage had to be planned around those windows.
Multi-use programming. The hall needed an RT target around 1.6–1.8 s for symphonic work (500 Hz–1 kHz) but closer to 1.2–1.4 s for speech and amplified events. The solution needed variability.
Existing finishes were visually protected. The architectural wood panels and plaster details were considered “signature elements.” Any visible treatment had to be color-matched and approved by the design committee.

Initial measurements confirmed the complaints. Using a combination of swept sine and interrupted noise methods, we measured mid-frequency RT60 values averaging 2.05 s at 500 Hz with significant scatter: 1.75 s in the front stalls and up to 2.35 s under the rear balcony and upper seating. Early decay time (EDT) tracked similarly, with the worst zones producing the impression of “slow clarity” even when RT60 was not extreme.

Speech transmission index (STI) in an unreinforced test with a talker position at center stage was 0.46–0.52 across many seats—adequate for some concert announcements but below what the venue wanted for lecture rentals (target was 0.60+ without relying on heavy reinforcement).

3. Approach and methodology chosen

We chose a three-track method: (1) a measurement campaign to establish a baseline with seat-zone granularity, (2) predictive modeling to test treatment placement and variable-acoustic options, and (3) staged implementation with verification at each milestone.

Measurement was conducted using a 32-bit float recorder (Sound Devices MixPre-6 II), a calibrated measurement microphone (Earthworks M50), and analysis in Room EQ Wizard and EASERA. For sound sources, we used a dodecahedron loudspeaker (NTi DS3) with an NTi PA3 power amplifier for repeatable omnidirectional excitation, supplemented by a small full-range point source (d&b audiotechnik E8) for stage-to-house directivity checks. We selected 12 receiver positions: four in stalls, four in the loge/balcony, and four in the upper seating, with two additional positions under the balcony overhang where intelligibility complaints were most common.

For modeling, we built a hybrid workflow. We created a simplified 3D room model from provided CAD and on-site laser distance checks, then ran ray-tracing simulations to evaluate early reflections, energy distribution, and the effect of proposed treatments on RT and clarity metrics (C50/C80). The goal was not to match every diffuser angle in the room but to confidently predict directional energy and the broad impact of absorption changes.

Finally, we required that any solution include a variable acoustic component. Rather than a full motorized banner system (costly and maintenance-heavy), we designed a combination of retractable absorptive curtains and selective, hidden broadband absorption behind existing woodwork in targeted zones.

4. Step-by-step execution narrative

Week 1–2: Baseline measurement and stakeholder listening sessions. We started with two overnight measurement sessions to avoid HVAC cycling and external noise. In parallel, we held a listening session with the resident orchestra’s principal players and two visiting FOH engineers. We played calibrated program material and conducted quick A/B tests using temporary drape panels on portable frames to demonstrate what a 0.2–0.3 s change in midband RT felt like in the room. This helped align expectations: reducing the “wash” did not have to mean losing envelopment.

Week 3–5: Concept design and mockups. Modeling indicated two dominant issues: (1) excessive late energy build-up in the upper volume due to high reflectivity of rear wall surfaces and ceiling coffers, and (2) a lack of beneficial early lateral reflections in certain mid-stall seats, contributing to weak clarity and inconsistent image. The second point was critical: if we only added absorption, we risked making the hall less exciting without improving definition.

We proposed three interventions:

Variable absorption curtains installed high on the rear wall and under-balcony rear surfaces, retractable for orchestral concerts.
Hidden broadband absorbers behind selected decorative wood panels on the rear side walls and upper rear wall, using acoustically transparent backing and maintaining the existing visual design.
Selective diffusion on the rear wall in zones where flutter echoes were observed, using shallow-profile quadratic residue diffusers (QRD) integrated into millwork.

Week 6–8: Detailed design, procurement, and scheduling. The key to staying on timeline was pre-fabrication. We worked with a local scenic/millwork shop to build modular absorber frames sized to fit behind existing panels without reengineering the wall. Material selection centered on 100 mm (4 in) mineral wool (Rockwool/ROCKBOARD 60 equivalent) with a 50 mm air gap where possible to extend absorption lower in frequency. For curtains, we specified a 24 oz/yd² IFR velour with a measured absorption coefficient suitable for mid-high control without making the room dull when deployed.

January dark period (12 days): Installation phase 1. The first window focused on the rear wall and under-balcony treatment. Crews removed preselected panel sections (documented and labeled), installed absorptive modules, and reinstalled panels with an acoustically transparent scrim behind perforated elements to prevent fiber shedding. Curtain tracks were installed at the rear upper wall with manual pull lines routed to discrete access points. Manual was chosen over motorized due to maintenance concerns and cost; the trade-off was operational discipline (staff training) to ensure correct presets for each event type.

Verification 1: Post-install measurements. Immediately after phase 1, we repeated the measurement set with identical source and mic positions. RT60 at 500 Hz dropped from 2.05 s average to 1.78 s with significantly reduced scatter in the rear seating. Under the balcony, EDT improved noticeably, correlating with the subjective impression that speech “started and stopped” more cleanly.

Week 9–14: Adjustments, diffusion integration, and operational presets. With the largest RT gains achieved, we focused on refining clarity without stripping warmth. We added diffusion to a rear wall section that was still producing a distinct slapback in the upper seats (measured as a strong reflection around 95–110 ms, depending on seat). The diffusers were 200 mm deep shallow QRD modules, finished to match existing panel stain and installed above audience sightlines.

We also worked with venue staff to create three acoustic presets:

Preset A (Symphonic): curtains retracted, standard seating, HVAC at low-noise mode
Preset B (Chamber/Jazz acoustic): rear curtains 50% deployed, under-balcony curtains deployed, modest stage shell adjustment
Preset C (Speech/Amplified): curtains fully deployed in rear and under-balcony, stage shell opened, PA tuned for lower room contribution

May dark period (9 days): Installation phase 2 and final commissioning. The second window was used to add additional hidden absorption in upper sidewall cavities and complete remaining diffuser integration. We performed final commissioning measurements, then held a rehearsal-based verification with the resident orchestra and a spoken-word presenter to evaluate real program material.

5. Technical decisions and trade-offs made

Variable curtains vs. permanent absorption. Permanent absorption would have simplified operation, but it would have forced a compromise RT that didn’t serve all programming. Curtains allowed event-by-event control. The trade-off was ensuring durability, training, and consistent use. We addressed this by documenting presets, installing clear position markers on pull lines, and training two staff members per shift.

Hidden absorption behind panels vs. exposed acoustic panels. Exposed panels would have been cheaper and easier to tune, but the venue’s design constraints made them unlikely to pass review. Hidden absorption added labor and required careful detailing to avoid rattles and maintain fire and safety requirements. We used rigid frames, isolation tape at contact points, and verified that panel reinstallation maintained tight fit without buzzes at higher SPL.

Diffusion only where measurable problems existed. Diffusion is often treated as a “nice-to-have,” but in this hall it solved a specific slapback and improved subjective spaciousness in seats that became slightly less lively after absorption was added. We avoided over-diffusing the room; too much diffusion can reduce helpful specular reflections that support clarity and presence.

Measurement repeatability over “one perfect measurement.” Rather than chasing a single high-resolution dataset, we prioritized repeatable positions and consistent excitation level. The most useful comparisons came from identical pre/post procedures, not from increasing complexity.

6. Results and outcomes with specific details

After final commissioning, the key metrics moved into the target range:

RT60 (500 Hz): reduced from 2.05 s average (range 1.75–2.35 s) to 1.62 s average with curtains retracted (range 1.50–1.72 s), and 1.32 s average with curtains deployed to Preset C (range 1.25–1.40 s).
EDT (500 Hz): improved from an average of 1.95 s to 1.55 s (Preset A) and 1.25 s (Preset C), with the largest improvement under the balcony.
C80 (music clarity): increased by 1.5–2.5 dB in the mid-stalls and under-balcony seats, reducing “blur” without making the room feel dry.
STI (unreinforced talker test): improved from 0.46–0.52 to 0.58–0.63 depending on seat, primarily due to reduced late reverberant energy in problematic zones.

Subjectively, the resident orchestra reported improved definition in fast passages and more consistent hall response across rehearsals. Visiting engineers reported that amplified mixes translated with less corrective EQ and less reliance on aggressive gating or overly dry vocal levels. One touring FOH engineer noted that he could run 2–3 dB less vocal level for the same intelligibility, because the room was no longer “fighting” consonants in the 2–4 kHz range.

Operationally, the venue adopted the preset system within two weeks. The technical director reported that staff reliably deployed the correct curtain configuration after an initial learning period, aided by a laminated quick-reference sheet kept at stage management and in the lighting booth.

The overall project timeline ran from kickoff to final commissioning in just under six months, with on-site work concentrated into the two dark periods. Total installed cost (materials, fabrication, labor, lifts/scaffolding, and commissioning) landed at approximately $214,000, below the initial not-to-exceed budget of $240,000.

7. Lessons learned and what could be done differently

Documenting existing construction saved time. The hall’s as-builts were accurate in broad terms but missed several cavity details behind decorative panels. We avoided surprises by opening a small number of “exploratory” panels early, then updating the install drawings. If we did it again, we’d schedule that exploratory work even earlier—before concept design lock—so modeling assumptions match reality sooner.

Staff operations are part of the acoustic design. Variable acoustics succeed or fail based on consistent use. The decision to avoid motorization was correct for this venue, but only because the operational workflow was addressed with training and clear presets. In a venue with frequent staff turnover, motorized systems with control integration might be justified despite higher upfront cost.

Don’t over-correct based on a single audience complaint zone. Under-balcony intelligibility was a real issue, but it would have been easy to over-absorb that area and create a noticeable acoustic discontinuity. We treated it, measured it, and then balanced the result by addressing rear-wall reflections and adding diffusion where needed.

8. Takeaways applicable to other projects

Start with zone-based measurements. Averaging the room hides the seat-to-seat variability that drives complaints. Pick repeatable source/mic positions and compare before/after using the same procedure.
Balance absorption with reflection management. Reducing RT can improve clarity, but it can also reduce engagement if you remove supportive early energy. Use modeling and targeted diffusion to preserve musical “life.”
Variable acoustics don’t need to be complex. A well-designed curtain system and hidden absorption can provide meaningful RT flexibility without full architectural reconstruction.
Plan around the calendar. For active venues, the schedule is the real constraint. Pre-fabricate, label everything, and design modules that install quickly during dark periods.
Define success in numbers and in use-cases. Targets like RT60 and STI matter, but so does how the hall functions on a rehearsal day, a touring show day, and a lecture rental day. Build presets around real programming.

Reverberation time optimization is rarely about chasing a single “ideal” number. In multi-use concert halls, the win is controllability and consistency: knowing how the room will behave, being able to shift it toward the event’s needs, and doing so without sacrificing the hall’s musical identity. This project’s best outcome wasn’t just a shorter RT—it was a hall that behaved predictably for the people responsible for making it sound good night after night.