Concert hall reverberation

Why does the same orchestra sound thin outdoors and rich indoors when the playing has not changed?

Take an orchestra out of its hall and put it in a park. Same players, same instruments, same bowing. The sound goes thin, distant, effortful — the strings stop blending, the players push, and something that felt grand indoors now sounds like a lot of individuals working hard.

Nothing was subtracted from the playing. So whatever was lost belonged to the room, which means the room was contributing something to the sound rather than merely containing it. What, exactly?

Follow one note. It leaves the instrument in all directions and spreads out; energy released at a point thins as it travels, and outdoors that is the end of the story — the sound reaches you once, weakened by the distance, and departs. Indoors, the portion that headed for the ceiling, the side walls and the floor comes back. It arrives a few milliseconds later, having travelled further, and then those reflections reflect in turn, and their reflections do, until the returns are arriving so densely you cannot separate them and hear only a decaying wash.

Now notice the consequence, because it is an accounting one. Sound is still being poured into the room while earlier sound is still bouncing around it. The energy present in the room at any moment is what has accumulated minus what the surfaces have absorbed — so a hall does not just relay the orchestra, it stores it briefly. That store is what makes the indoor sound louder, more sustained, and blended, since each note overlaps the tails of the notes before it.

Which raises the design question: how much store do you want? Suppose you make the surfaces very absorbent. The tail vanishes, notes stop the instant the bow lifts, and the sound goes flat and lifeless — a room musicians call dead. Suppose instead you make them highly reflective and the room large. The tail runs on for many seconds, every note smearing into its successors, and a fast passage turns to porridge while a spoken word becomes unintelligible.

So there is a quantity to be chosen, and Wallace Sabine, working at Harvard in the 1890s, was the first to pin it down. He defined reverberation time as the interval for the sound to fall by 60 decibels after the source stops — RT60 — and found it obeys a simple relation: RT60 is roughly 0.161 times the room's volume in cubic metres, divided by its total absorption in square-metre units. Bigger room, longer tail. More absorption, shorter tail. Absorption includes the audience, which is why an empty hall sounds so different at rehearsal from the same hall full.

And the right value depends on what is being played. Speech needs the tail short enough that one syllable has cleared before the next arrives. Chamber music wants clarity with some bloom. A romantic symphony orchestra wants the sustain — the great halls sit around two seconds occupied. A cathedral, where the repertoire was written for slow-moving vocal lines, may run several times longer, and plainchant works there precisely because it was composed for a room that blurs.

Sabine's equation is genuinely predictive, and it is also an idealisation with visible seams. It assumes the sound field is diffuse — energy evenly spread and travelling in all directions — and that absorption is scattered fairly evenly around the room. In a very absorbent space, or an oddly shaped one, or one with all the soft material on a single surface, it overestimates the reverberation time, which is why later formulations such as the Norris-Eyring correction exist.

The deeper caveat is about the number's authority. It is tempting to read the history as science solving a problem, but the halls everyone still measures against were not designed from the formula: the Musikverein in Vienna opened in 1870 and the Concertgebouw in Amsterdam in 1888, both before Sabine had written anything. Boston Symphony Hall, opened in 1900, was the first designed with his advice, and it worked — yet the decades since have produced halls built with far more acoustic science that disappointed badly on opening, New York's Philharmonic Hall of 1962 being the standard example, rebuilt repeatedly thereafter.

The reason is that RT60 is one number describing a decay, and listeners care about things it does not encode: how soon the first reflection arrives after the direct sound, how much early energy reaches the ear from the sides rather than overhead, how clear a fast passage remains, how loud the hall makes the orchestra for its size. The narrow rectangular plan shared by the celebrated old halls is now usually credited with delivering strong lateral reflections — a property their reverberation times say nothing about. Modern practice measures a family of parameters instead, and still leaves room for judgement.

• Why lateral reflections, rather than overhead ones, are credited with a sense of being enveloped by the sound.
• How variable acoustics — movable panels, retractable drapes, coupled chambers — try to let one hall serve several repertoires.

Every term the collection defines is gathered in the glossary.