The Vienna Musikverein, the perennial benchmark of concert hall design, measured an unoccupied early decay time of 3.1 seconds, the longest of six leading European halls tested by Aalto University (Lokki et al., JASA 2016). Yet in blind listening tests the same hall did not stand out in listener preferences, and the largest listener group found it too reverberant at balcony seats (Lokki et al., Proceedings of the Institute of Acoustics 2015). Hamburg’s Elbphilharmonie, by contrast, publishes an occupied T30 of 2.3 s and a clarity C80 of just 0.3 dB (Nagata Acoustics, Elbphilharmonie data sheet), and a 2025 study found that non-experts’ preferred reverberation ranges were up to 83% broader than those of experts (Hidaka, Omoto and Nishihara, Acoustical Science and Technology 2025). We aggregated data from the Journal of the Acoustical Society of America, the Audio Engineering Society, Aalto University, Nagata Acoustics, the ISO 3382-1 standard as published by BSI, and a dozen other peer-reviewed sources. For the audience side of the same story, see our classical music concert attendance statistics, and for the firms that design these rooms, our architectural acoustics consulting statistics.
TL;DR
- Best-rated halls have occupied reverberation of 1.7 to 2.0 s; lower-rated halls 1.5 to 1.7 s (Beranek, AES Heyser Lecture 2007).
- Elbphilharmonie Grand Hall: T30 2.4 s unoccupied, 2.3 s occupied, 2,100 seats (Nagata Acoustics, Elbphilharmonie data sheet).
- Suntory Hall loses 0.5 s of reverberation when full: 2.6 s empty, 2.1 s occupied (Nagata Acoustics, Suntory Hall data sheet).
- Musikverein early decay time 3.1 s vs 1.6 s at Cologne Philharmonie, both unoccupied (Lokki et al., JASA 2016).
- Hall geometry explains up to 47% of the variation in C80 at 250 to 1000 Hz across 58 halls (Basaran and Oktav, IJMO 2025).
- Berlin Konzerthaus early lateral fraction 0.27 vs 0.11 at Berlin Philharmonie, both at 11 m from the orchestra (Haapaniemi and Lokki, Psychomusicology 2015).
- The largest listener group (43%) found the Musikverein too reverberant at balcony seats (Lokki et al., Proceedings of the IOA 2015).
- Experts preferred 1.2 to 2.0 s of reverberation for piano and 1.8 to 2.4 s for violin (Hidaka et al., Acoustical Science and Technology 2025).
- The ISO-listed C80 just noticeable difference is about 1 dB, but trained listeners needed approximately 3 dB (Vigeant et al., JASA 2015).
- 43 participants (23 acousticians, 20 architects) used measurably different vocabularies to judge four halls (Llorca-Bofi et al., JASA 2026).
1. Reverberation Time in the Benchmark Halls
Reverberation time (RT60, or T30/T20 when extrapolated from a shorter decay) is the oldest number in room acoustics, and the famous halls cluster in a narrow band. The best and better halls have occupied reverberation times between 1.7 and 2.0 s, while those of lower subjective quality sit between 1.5 and 1.7 s (most recent available data: Beranek, AES Richard C. Heyser Memorial Lecture, 123rd AES Convention, 2007). The gap between “great” and “merely good” is a few tenths of a second, which is why consultants fight over the last 0.2 s.
The newest landmark sits right at the top of that band. Hamburg’s Elbphilharmonie, opened in 2017 on a 1963 warehouse and tuned with a 1/10-scale physical model, publishes a T30 of 2.3 s occupied, slightly above the classic shoebox halls (Nagata Acoustics, Elbphilharmonie Grosser Saal data sheet).
| Metric | Value | Source |
|---|---|---|
| Occupied RT, Vienna Grosser Musikvereinssaal vs Boston Symphony Hall | 2.0 s vs 1.9 s | Beranek, AES Heyser Lecture 2007 |
| Occupied RT, Royal Festival Hall London vs Carnegie Hall New York | 1.45 s vs 1.8 s | Beranek, AES Heyser Lecture 2007 |
| Elbphilharmonie Grand Hall T30, unoccupied / occupied | 2.4 s / 2.3 s | Nagata Acoustics, Elbphilharmonie data sheet |
| Helsinki Music Centre T60 | 2.5 s | Lokki et al., Proceedings of the IOA 2015 |
| Boston Symphony Hall: construction began / opening concert | June 1899 / October 15, 1900 | Boston Symphony Orchestra, The Building of Symphony Hall |
| Boston Symphony Hall total cost of building and land | $771,000 | Boston Symphony Orchestra, The Building of Symphony Hall |
Context note: the Boston Symphony Orchestra states that Symphony Hall “rates among the top three music halls in the world,” alongside Amsterdam’s Concertgebouw and Vienna’s Musikverein. All three are shoeboxes, and all three predate the formal science that now measures them: Wallace Sabine, a Harvard physics professor, consulted on Boston without payment and developed a formula to calculate reverberation time. The Royal Festival Hall value predates its renovation.
2. Volume, Seats and the Audience Effect
A hall’s published reverberation is meaningless until you know whether the seats were full. People are the largest absorber in the room, and modern halls are designed so that the drop from empty to full stays small. Suntory Hall in Tokyo falls from 2.6 s unoccupied to 2.1 s occupied, a 0.5 s drop (2.6 minus 2.1) (Nagata Acoustics, Suntory Hall data sheet), while the Elbphilharmonie in section 1 falls only 0.1 s (2.4 minus 2.3). Seats that absorb almost as much empty as occupied are the reason rehearsals and concerts can sound alike.
Volume per seat explains much of the rest. Using the published figures in the table, the Musikverein offers about 8.9 cubic meters per seat (15,000 / 1,680) and the Elbphilharmonie about 11.0 (23,000 / 2,100). The same physics drives the reverberation problems of worship spaces covered in our church acoustics and sound systems statistics.
| Metric | Value | Source |
|---|---|---|
| Vienna Musikverein volume and seats | 15,000 m³, 1,680 seats | Lokki et al., JASA 2016 |
| Elbphilharmonie Grand Hall volume and seats | 23,000 m³, 2,100 seats | Nagata Acoustics, Elbphilharmonie data sheet |
| Tokyo Metropolitan Art Space (1,999 seats, 25,000 m³) RT unoccupied / occupied | 2.6 s / 2.1 s | Nagata Acoustics, Tokyo Metropolitan Art Space data sheet |
| Kyoto Concert Hall (1,833 seats, 20,000 m³) RT unoccupied / occupied | 2.2 s / 2.0 s | Nagata Acoustics, Kyoto Concert Hall data sheet |
| New World Center, Miami Beach (748 seats, 14,200 m³) RT unoccupied / occupied | 2.4 s / 2.1 s | Nagata Acoustics, New World Center data sheet |
| Extra sound absorbed per person at Munich Philharmonie (900 people per 500 m²) vs Amsterdam Concertgebouw (1,200 per 500 m²) | About 33% more | Beranek, AES Heyser Lecture 2007 |
| Share of reverberation time variation at 125 to 500 Hz explained by geometry alone, 58 halls | Up to 47% | Basaran and Oktav, IJMO 2025 |
Context note: the occupied/unoccupied pairs come from the design consultant’s own data sheets (Tokyo Metropolitan Art Space, Kyoto Concert Hall, New World Center). The machine-learning study (Basaran and Oktav, International Journal of Modeling and Optimization 2025) compared 28 algorithms on 10 geometric features across six octave bands: more than half of the variation still comes from materials, seating and detailing, not shape.
3. Clarity (C80) and Early Decay Time
Reverberation time describes the tail; early decay time (EDT, the first 10 dB of decay) and clarity (C80, the early-to-late energy ratio in dB) describe what the listener actually perceives during running music. At the Berlin Philharmonie, a vineyard hall, C80 drops from 2.4 dB at 11 m to 0.2 dB at 19 m, while the shoebox Berlin Konzerthaus stays negative at both distances (Haapaniemi and Lokki, Psychomusicology 2015). Negative C80 means more late than early energy: lush rather than crisp.
The EDT spread among the famous halls is wider than their occupied RT values suggest. Vienna’s unoccupied EDT is almost double Cologne’s, and that gap, not the 2.0 s occupied RT, is what makes the Musikverein sound “full.”
| Metric | Value | Source |
|---|---|---|
| EDT (500/1000 Hz, unoccupied): Musikverein / Concertgebouw / Berlin Philharmonie / Cologne Philharmonie | 3.1 / 2.4 / 1.9 / 1.6 s | Lokki et al., JASA 2016 |
| Elbphilharmonie EDT / C80 / centre time Ts | 2.3 s / 0.3 dB / 135 ms | Nagata Acoustics, Elbphilharmonie data sheet |
| Berlin Konzerthaus C80 at 11 m / 19 m | -1.5 dB / -2.3 dB | Haapaniemi and Lokki, Psychomusicology 2015 |
| Berlin Philharmonie C80 at 11 m / 19 m | 2.4 dB / 0.2 dB | Haapaniemi and Lokki, Psychomusicology 2015 |
| Mid-frequency EDT range across nine recorded halls | 1.49 s to 2.60 s | Lokki et al., JASA 2012 |
| Share of C80 variation at 250 to 1000 Hz explained by geometry, 58 halls | Up to 47% | Basaran and Oktav, IJMO 2025 |
Context note: the Lokki et al. 2016 dataset, the most recent available data for these four halls, was measured unoccupied, so its EDT values run higher than occupied figures. The 2012 nine-hall study reported halls only by code letters.
4. Strength G, Lateral Energy and Envelopment
Strength G measures how loud a hall makes a source relative to the same source at 10 m in free field; lateral energy measures how much of the early sound arrives from the sides. Across nine recorded halls, the early lateral fraction ranged from 14% to 31% (Lokki et al., JASA 2012), and that spread maps onto shape: narrow shoeboxes send early reflections from the side walls, while vineyard and fan halls send them from the front and above.
Loudness and envelopment tend to travel together. Vienna leads both strength and calculated listener envelopment (LEV) in the table below, while the Berlin Philharmonie, the archetypal vineyard hall, sits below zero on envelopment. Most recent available data: measurement campaigns published between 2007 and 2016.
| Metric | Value | Source |
|---|---|---|
| Mid-frequency strength Gmid, Vienna vs Boston | 6.5 dB vs 4.2 dB | Beranek, AES Heyser Lecture 2007 |
| Calculated LEV, Vienna / Amsterdam / Boston / Berlin Philharmonie | 2.0 / 1.4 / 0.3 / -0.2 dB | Beranek, AES Heyser Lecture 2007 |
| Strength G (500/1000 Hz, unoccupied), Musikverein vs Helsinki Music Centre | 4.1 dB vs 1.4 dB | Lokki et al., JASA 2016 |
| Elbphilharmonie Grand Hall strength G | 5.4 dB | Nagata Acoustics, Elbphilharmonie data sheet |
| Early lateral fraction (JLF) at 11 m, Berlin Konzerthaus vs Berlin Philharmonie | 0.27 vs 0.11 | Haapaniemi and Lokki, Psychomusicology 2015 |
| Late lateral sound level LJ, Berlin Philharmonie vs Konzerthaus | 2 to 3 dB lower | Haapaniemi and Lokki, Psychomusicology 2015 |
| Early lateral fraction range across nine recorded halls | 14% to 31% | Lokki et al., JASA 2012 |
Outlier note: the two Vienna G values in this table (6.5 dB and 4.1 dB) differ by 2.4 dB because they come from different campaigns, one an occupied-era compilation and the other an unoccupied loudspeaker-orchestra measurement. The 2012 Aalto paper warns its own G and LJ values are relative only, because its sources were not omnidirectional as ISO 3382-1 requires. Compare G only within one dataset.
5. Listener Preference Studies
The most consistent finding in two decades of blind listening tests is that there is no single “best” hall: listeners split into taste groups. In Aalto’s six-hall test, the Musikverein “does not stand out in the indicated preferences,” and the largest group, 43% of listeners, found it too reverberant at balcony seats as measured unoccupied (Lokki et al., Proceedings of the Institute of Acoustics 2015). A second group (29%) preferred the less reverberant Philharmonies in Cologne and Berlin.
The emotional data point the other way. Aalto’s 2016 JASA paper, titled “Concert halls with strong and lateral sound increase the emotional impact of orchestra music,” recorded finger skin conductance while listeners heard the same Beethoven excerpt at two positions in six European halls (Aalto University news, 2016). A follow-up concluded that emotional impact “follows the degree of the dynamic responsiveness” of the hall (Patynen and Lokki, Proceedings of the IOA 2018).
| Metric | Value | Source |
|---|---|---|
| Preference groups in six-hall paired-comparison test | 43% / 29% / 29% | Lokki et al., Proceedings of the IOA 2015 |
| Nine-hall study preference groups (17 assessors) | 7 vs 10 | Lokki et al., JASA 2012 |
| Expert preferred mid-frequency RT, piano / violin | 1.2 to 2.0 s / 1.8 to 2.4 s | Hidaka et al., Acoustical Science and Technology 2025 |
| Non-expert preferred mid-frequency RT, piano / violin | 0.9 to 2.1 s / 1.6 to 2.7 s | Hidaka et al., Acoustical Science and Technology 2025 |
| Listening-test subjects and mean age, dynamic responsiveness study | 28 subjects, 39.6 years | Patynen and Lokki, Proceedings of the IOA 2018 |
| Skin conductance measurements (12 stimuli x 27 usable subjects) | 324 | Patynen and Lokki, Proceedings of the IOA 2018 |
| Participants in architects vs acousticians hall assessment | 43 (23 acousticians, 20 architects) | Llorca-Bofi et al., JASA 2026 |
Context note: the Hidaka, Omoto and Nishihara study used binaural impulse responses from 34 positions in 18 symphonic halls and found non-experts’ preference ranges 50% (piano) and 83% (violin) broader than experts’, concluding that expert judgment is more reliable for design. The newest study (Llorca-Bofi et al., JASA 2026) found architects and acousticians disagree on which objective parameter tracks subjective balance and timbre. In Aalto’s 2012 work, “none of the standardized objective room acoustical parameters could explain the proximity and preference data” (Lokki et al., JASA 2012).
6. Measuring Halls: ISO 3382-1, JNDs and Uncertainty
Every number above depends on how it was measured. ISO 3382-1 defines the parameters, and its Annex A gives the just noticeable difference (JND) used to decide whether two halls really differ. The ISO-listed C80 JND is about 1 dB, but a Penn State study with extensively trained listeners found approximately 3 dB for both the concert hall and the chamber music hall tested (Vigeant et al., JASA 2015). If that holds, several of the C80 differences in section 3 sit close to the threshold of audibility.
Instrument repeatability is far tighter than perception. The standard itself is BS EN ISO 3382-1:2009, published 31 August 2009 and listed as “Current, Under Review” by BSI. For the same measurement discipline applied to the quietest rooms, see our anechoic chamber acoustic testing statistics.
| Metric | Value | Source |
|---|---|---|
| C80 JND listed in ISO 3382:2009 vs measured with trained listeners | About 1 dB vs approximately 3 dB | Vigeant et al., JASA 2015 |
| EDT JND used for perceptual relevance (ISO 3382-1) | 5% | Haapaniemi and Lokki, Psychomusicology 2015 |
| ISO 3382-1 parameters analyzed for minimum decay range (EDT, T20, T30, C80, D50, TS, G, ST, IACC, LF, LFC, LG) | 12 | Hak, Wenmaekers and van Luxemburg, Acta Acustica 2012 |
| Repeatability of 10 reverberation-room reference measurements over four years (Mid rating) | Standard deviation 0.2 dB | Wenmaekers and Hak, DAGA 2014 |
| Round-robin G standard deviation at 1000 Hz (Lundeby et al., as reported) | 0.2 dB | Wenmaekers and Hak, DAGA 2014 |
| Test hall for seat-absorption simulation study / receiver positions | 394 seats / 16 | Lee et al., Applied Acoustics 2026 |
| Best RMSE reduction from ensemble post-processing of simulations | Up to 14.6% (EDT) | Lee et al., Applied Acoustics 2026 |
Context note: Hak, Wenmaekers and van Luxemburg (Acta Acustica 2012) showed that under adverse conditions the decay range alone can shift parameter values by more than one JND. Hidaka (Acoustical Science and Technology 2005) concluded that the minimum number of positions given in ISO 3382 should be exceeded for some parameters. The 2026 simulation study (Lee et al., Applied Acoustics 2026) found no single seat-absorption model was best across all parameters, and Wenmaekers and Hak define the standard “Mid” single-number rating as the 500/1000 Hz average.
Summary: Concert Hall Acoustics by the Numbers
| Metric | Value | Source |
|---|---|---|
| Occupied RT, best halls / Vienna / Boston | 1.7 to 2.0 s / 2.0 s / 1.9 s | Beranek, AES Heyser Lecture 2007 |
| Elbphilharmonie T30 occupied / C80 / G | 2.3 s / 0.3 dB / 5.4 dB | Nagata Acoustics, Elbphilharmonie data sheet |
| Suntory Hall RT unoccupied / occupied | 2.6 s / 2.1 s | Nagata Acoustics, Suntory Hall data sheet |
| New World Center RT unoccupied / occupied | 2.4 s / 2.1 s | Nagata Acoustics, New World Center data sheet |
| Kyoto Concert Hall RT unoccupied / occupied | 2.2 s / 2.0 s | Nagata Acoustics, Kyoto Concert Hall data sheet |
| Musikverein EDT vs Cologne Philharmonie EDT (unoccupied) | 3.1 s vs 1.6 s | Lokki et al., JASA 2016 |
| Berlin Philharmonie C80 at 11 m / 19 m | 2.4 dB / 0.2 dB | Haapaniemi and Lokki, Psychomusicology 2015 |
| Early lateral fraction, Konzerthaus vs Philharmonie | 0.27 vs 0.11 | Haapaniemi and Lokki, Psychomusicology 2015 |
| RT and C80 variation explained by geometry, 58 halls | Up to 47% | Basaran and Oktav, IJMO 2025 |
| Largest preference group in six-hall test | 43% | Lokki et al., Proceedings of the IOA 2015 |
| Nine-hall study: reliable attributes retained | 60 of 92 | Lokki et al., JASA 2012 |
| Expert preferred RT for violin | 1.8 to 2.4 s | Hidaka et al., Acoustical Science and Technology 2025 |
| Non-expert range broader than experts, piano / violin | 50% / 83% | Hidaka et al., Acoustical Science and Technology 2025 |
| Skin conductance measurements in dynamic responsiveness study | 324 | Patynen and Lokki, Proceedings of the IOA 2018 |
| Architects vs acousticians study participants | 43 | Llorca-Bofi et al., JASA 2026 |
| C80 JND, ISO-listed vs measured | About 1 dB vs approximately 3 dB | Vigeant et al., JASA 2015 |
| Round-robin G standard deviation at 1000 Hz | 0.2 dB | Wenmaekers and Hak, DAGA 2014 |
| ISO 3382-1 publication date | 31 August 2009 | BSI, BS EN ISO 3382-1:2009 |
Methodology and Sources
- Beranek, L. “Aspects of Concert Hall Acoustics,” AES Richard C. Heyser Memorial Lecture, 123rd AES Convention, New York, 2007: occupied RT, Gmid and calculated LEV for rated halls, audience absorption.
- Nagata Acoustics, Elbphilharmonie Hamburg Grosser Saal data sheet, plus the Suntory Hall, Tokyo Metropolitan Art Space, Kyoto Concert Hall and New World Center data sheets.
- Boston Symphony Orchestra, “The Building of Symphony Hall” exhibit.
- Lokki, Patynen, Kuusinen and Tervo, “Concert hall acoustics: Repertoire, listening position, and individual taste of the listeners influence the qualitative attributes and preferences,” JASA 140(1), 2016.
- Lokki, Patynen, Kuusinen and Tervo, “Disentangling preference ratings of concert hall acoustics using subjective sensory profiles,” JASA 132(5), 2012.
- Lokki et al., “The secret of the Musikverein and other shoebox concert halls,” Proceedings of the Institute of Acoustics 37(3), 2015.
- Haapaniemi and Lokki, “The preferred level balance between direct, early and late sound in concert halls,” Psychomusicology 25(3), 2015.
- Aalto University news on Patynen and Lokki, JASA 139(3), 2016 and Patynen and Lokki, “Dynamic responsiveness in concert halls as source of emotional impact,” Proceedings of the IOA 40(3), 2018.
- Hidaka, Omoto and Nishihara, “Listener’s quality in subjective evaluation of concert hall acoustics,” Acoustical Science and Technology 46(2), 2025.
- Llorca-Bofi, Heck, Au-Yeung, Infantes-Lopez and Vorlander, “Seeking differences between architects’ and acousticians’ concert hall assessments,” JASA, 2026.
- Basaran and Oktav, “Predicting Concert Hall Reverberation Time and Clarity from Geometric Features Using Machine Learning,” IJMO 15, 2025.
- Lee, Lee, Hong and Lee, “Seating absorption modeling in concert hall geometrical-acoustics simulations,” Applied Acoustics, 2026.
- Vigeant et al., “The effects of different test methods on the just noticeable difference of clarity index for music,” JASA 138(1), 2015.
- Hak, Wenmaekers and van Luxemburg, “Measuring Room Impulse Responses: Impact of the Decay Range on Derived Room Acoustic Parameters,” Acta Acustica 98(6), 2012.
- Wenmaekers and Hak, “Reference level in ISO 3382 parameters: G, ST, L and STI,” DAGA 2014.
- Hidaka, “Supplemental data of dependence of objective room acoustical parameters on source and receiver positions at field measurement,” Acoustical Science and Technology 26(2), 2005.
- BSI, BS EN ISO 3382-1:2009, Acoustics: Measurement of room acoustic parameters, Performance spaces.
- Data watch: concert hall acoustics is a slow-moving field, and many benchmark figures are older than three years (the 2007 AES lecture and the 2012 to 2016 Aalto campaigns). Occupied and unoccupied values are not comparable, and loudspeaker-orchestra measurements differ from omnidirectional-source surveys, which is why the two Vienna G values in section 4 disagree by 2.4 dB. The Concertgebouw has no single agreed reverberation figure: a 2014 Meyer Sound press article states 2.8 s empty and 2.2 s full for a 1,974-seat hall without naming a measurement source, while the 2016 Aalto table lists 2,040 seats and an unoccupied EDT of 2.4 s. The Nagata figures are published by the halls’ own acoustic consultant, not an independent survey, and the New World Center sheet gives both 748 and 756 seats. ISO 3382-1 is listed as under review, so its JND table may change.
- Last updated: October 18, 2026. We update this roundup quarterly, and the next refresh is expected when the revised edition of ISO 3382-1 (currently under review) is published.