A microphone does not hear a voice. It hears a voice plus the room's answer to that voice, arriving a few milliseconds later from every hard surface within reach. By the time the capsule converts anything into voltage, most of what the audience will experience has already been decided by the architecture.
The mechanism is reflection and decay. Sound leaves the performer, strikes plaster, glass, timber, upholstery or bodies, and returns attenuated. How much comes back, and how long it keeps coming back, is the room's reverberation time. The measurement is standardised: ISO 3382-1 defines how the decay is captured and reported, most commonly as T30, extrapolated from the first 30 dB of decay to the time a sound would need to fall by 60 dB after the source stops. That standard comes from an international standards body with no product to sell, which is worth stating because most other numbers in this field do not.
The physical driver is straightforward. Sabine's relationship holds that reverberation time rises with the volume of the room and falls with the total absorption in it. Double the cubic volume without adding soft material and the tail lengthens. Fill the same room with an audience in winter coats and it shortens, which is why a dress rehearsal in an empty house never sounds like the performance.
Why the number matters more for speech than for music
Consonants carry meaning, and consonants are quiet. The plosives and fricatives that separate bat from pat from fat hold a small fraction of the energy of the vowel beside them. In a long reverberant tail the vowel of one syllable is still decaying when the consonant of the next arrives, and the loud thing masks the quiet thing. Sung legato survives this comfortably; a Baroque aria can sound better in a long room than a short one. Recitative, dialogue and comic patter do not survive it at all.
This is measurable rather than impressionistic. IEC 60268-16 defines the Speech Transmission Index, which scores a transmission path by how well it preserves the amplitude modulations that carry speech information. Reverberation and background noise both degrade the score. Again the source is a standards body, not a vendor.
Room targets exist where comprehension is the whole point. ANSI/ASA S12.60, the American classroom acoustics standard, sets a reverberation limit of 0.6 seconds for small learning spaces. No equivalent binding standard governs a parish hall or a converted chapel used for a play, which is precisely why so many of them defeat the productions staged in them.
Adding gain does not repair it
The instinct when words are lost is to raise the level. It does not work, and the reason is geometric rather than electrical. A loudspeaker excites the same room the performer does. Raising its output raises the direct sound and the reflected sound together, leaving the ratio between them unchanged.
That ratio has a boundary, called the critical distance: the point away from a source at which reflected energy equals direct energy. Closer than that, a listener mostly hears the source. Further away, a listener mostly hears the room. Critical distance shrinks as reverberation lengthens, so in a very live hall it can fall well inside the seating, and every row behind it hears essentially the same wash regardless of how much power is delivered. Beyond that boundary a larger system makes the problem louder rather than clearer.
What actually moves the ratio
Three things shift the balance toward direct sound, and only three.
Shortening the distance from source to capsule. This is why a head-worn microphone at the hairline behaves so differently from a float at the stage lip; the companion article on gain before feedback works through the arithmetic.
Narrowing the coverage of the loudspeaker so its energy lands on absorptive audience rather than on reflective wall and ceiling. A directional source excites less of the room to begin with, which is the design intent behind column and line arrangements.
Absorbing the room itself. This is the slowest and most expensive lever, and the only one of the three that also helps unamplified performance. Drapes, an audience, and soft goods hung at the rear wall all count toward it.
The limits of the fix
None of this recovers late energy once the room has been excited, and none of it helps a performer who cannot hear themselves. A hall that flatters a choir will fight a play, and the same building can be right for one production and wrong for the next. The decisive choice is usually made when the venue is booked, months before anyone opens a flight case.
What we cannot verify
Reverberation figures quoted for named venues in trade press are frequently published without stating the measurement method, the source position, or whether the house was occupied, and figures gathered differently cannot be compared with each other. Manufacturer literature on intelligibility is produced by parties selling the remedy, and we treat it as advocacy rather than evidence. Nothing here is a measurement of any particular room: a room's behaviour is established by measuring that room, occupied, from the seats people will actually sit in.
The short version
- A microphone captures the room's answer along with the voice.
- Reverberation time is standardised by ISO 3382-1 and usually reported as T30.
- Consonants are quiet and are the first casualties of a long decay.
- Past the critical distance, more gain cannot improve the direct-to-reverberant ratio.
- Capsule distance, loudspeaker directivity and absorption are the only three levers.
- The largest decision is which room, and it is made before any equipment arrives.