The reverberation time (RT) is one of the most fundamental room measurements. It gives a broad brush stroke description of the general acoustical behavior of an acoustically “live” space, paving the way for the use of other metrics to determine clarity and direct-to-reverberant ratios.
Wallace Clement Sabine was the first to formalize the RT into an acoustic metric. His method of measurement involved a program source, a stopwatch and a quiet room.
Sabine determined that 60 dB of room decay was audible in a quiet space, so he formalized the RT60 to describe it. His work provided the basis for the statistical RT formulas that are still in use today.
Further refinements came with the ability to measure the Room Impulse Response (RIR) (Figure 1). ISO3382 is an international Standard for “scoring” the RIR.
First, the squared RIR is backward integrated (energy summed from right-to-left, a.k.a., the Schroeder curve). The slope of the Schroeder curve is used to determine various decay times.
Two of the ISO 3382 decay metrics that are commonly used by sound system designers include:
Reverberation Time T30 – The reverberation time for 60 dB of room decay, based on a straight line curve fit between the -5 dB to -35 dB points on the Schroeder curve (Figure 2).
This time is doubled to be made consistent with the traditional reverberation time for 60 dB of decay (RT60). The T30 is a measure of the “persistence” of sound in the space.
Early-Decay Time (EDT) – Based on a straight-line curve fit between the 0 dB to -10 dB points of the Schroeder curve, extrapolated to 60 dB of decay (Figure 3). A strong direct field will produce an EDT that is less than (steeper slope) than the T30. This is a desirable attribute for increased speech intelligibility or music clarity.
Strong early reflections have a similar effect, which is why acousticians often specify band shells and acoustic clouds to enhance the stage sound in large spaces.
The T30 and EDT provide important information about the way that sound behaves in a room. Both are most applicable when a homogeneous reverberant field exists (i.e., medium to large rooms with low absorption). Their use in small spaces (i.e., control rooms, home theaters, and other rooms with high absorption relative to room volume) is questionable. Even so, it is a popular metric for spaces of all sizes, probably because its measurement is relatively simple.
Caveat: The determination of T30 and EDT according to ISO3382 seems straightforward, so one would expect to get the same result from any measurement program. But this is not necessarily the case. So much for Standards!
One reason is based on the intended use of the Standard. ISO3382 was not created for sound system evaluation, per se. It was created to be used for room acoustics testing.
For such tests, a low-directivity sound source is prescribed. This can be an actual impulsive source, like a balloon pop or starter’s pistol.
If a loudspeaker is used, it should have low directivity (i.e., a dodecahedron). When measured in this way, the -5 dB point on the Schroeder curve is typically beyond the early energy arrivals and falls on a straight segment of the curve.
When highly directional sound reinforcement loudspeakers are used, the steeper slope of the early Schroeder curve extends further into the RIR. In such cases the -5 dB start of the T30 calculation may be too early, producing a shorter T30 (steeper slope) than the actual RT. While strong early energy is a good thing for sound systems, it can fool the algorithm used to determine the actual RT.
Long before the existence of a standard, reverb times were determined from the manual placement of cursors on the Schroeder curve (Figure 4). A left and right cursor were used to segment a straight portion of the curve and the RT (RT60 in those days) was determined from these placements.
The left cursor was moved beyond the slope change caused by the early reflections and direct field (typically >300 ms for most rooms). Since ISO3382 defines the slope to be from -5 dB to -35 dB on the Schroeder curve, this can result in the “start point” being on a different slope of the curve than the “end point.”
Figure 5 (next page) shows how the 8 kHz octave band can trick a measurement program. This auto ISO3382 algorithm reports a 13-second, 8-kHz-octave T30! Manual cursor placements yield the correct RT.