Misc. topics

Dilemma: Trying to make rehearsal room acoustics similar to concert hall stage acoustics

For a violinist playing in a symphony orchestra, rehearsing in a small rehearsal room, what is the proper duration and level of the reverberant sound?

If RT is similar to the hall where the music is to be performed, e.g. RT=2.0 s, the sound level /support level would be too high.

A. In a room with dimensions 7m*5m*3m and average absorption coefficient a=0.10, we predict (at 1m distance): T30=1.2s;  EDT=1.1s;  STearly=1dB;  STlate=-1dB. Even if RT is much shorter than in the 2.0s preferred in a concert hall, the acoustical response would be much louder, in terms of ST parameters being some 10 dB more than recommended and some 15dB more than in many high ranked halls.

If support is to be similar to support on the stage where the music is to be performed,
e.g. STearly= -13dB, the RT must be very short, like in the following example.

B. In a room with dimensions 7m*5m*3m and average absorption coefficient a=0.5, we predict (at 1m distance): T30=0.2s;  EDT=0.15s;  ST_early=-10dB;  ST_late=-40dB. While ST_early is close to recommended , the late stage response, ST_late , is very low, indicating the perception of a completely dead space.

C. Commonly, individual practice rooms are even smaller than the examples A and B above, making stage-like acoustics even harder to achieve. Many musicians practice in rooms with floor area less than 10 square meters. Let us consider a realistic 9 sqm practice room with dimensions 3.3m*2.7m*3.0m and average absorption coefficient a=0.17: T30=0.5s;  EDT=0.47s;  ST_early=0dB;  ST_late=-9dB.

EDT is assumed to describe listeners perceived reverberation, i.e. reverberance, but on a concert hall stage at 1m from the source, EDT is in the order of 0.1s, which usually would be regarded as inaudible reverberance.

It may seem impossible to achieve acoustical conditions in a rehearsal room similar  to those on stage in a large concert hall. Is it—and if so, what are the preferred acoustical properties?

You may judge for yourself in the reverberant space demo.

Stage acoustic parameters approach

International consensus currently assumes performers reverberance to be described by late support, ST_late, [parameters], while early support ST_early is suggested to describe ensemble conditions. While late support is the energy level of reverberant sound of 100-1000ms delay, early support is the energy level of reverberant sound in the interval of 20-100ms. Both parameters are measured on an empty podium, relating their levels with reference to direct sound 1 meter distance from an omnidirectional source. Measurements commonly take the energy in the interval 0-10ms interval as reference. In practice, the first reflection from the floor can lead to underestimating ST levels by up to1dB. In rehearsal rooms less than 20m2 floor and less than 70m3, this underestimation increases with decreasing room size.

Late support may be estimated by Barron’s Revised Theory from the following formula

                 ST_late ~ 10·log (312T/V) - 6/T (dB),

Where V is the room volume (m³) and T is the reverberation time T30 (s).

The 280m³ group rehearsal room in the  reverberant space demo has ST_late= -15dB, which is not very different from what can be measured on a concert hall stage. If late support is an adequate criteria for rehearsal room acoustics, the criterion needs to be corrected for the different sound power from a different number of musicians.  One could suggest a criteria for the group rehearsal room acoustics on the form

                 ST_late ~ 3 -10·log(N+1)  (dB),

where N is the rehearsal group size. This would be consistent with podium acoustics proper for a symphony orchestra in a full-size concert hall with T=2.0s. For a quartet rehearsing in 70m³ the criteria would imply T~0.7s, while a single player rehearsal room with T~0.7s would need 30m³. The term 1 added to N is introduced to prevent excess loudness in the individual rehearsal case (in individual practice, musicians will skip the pauses occurring in polyphonic music, thus individual rehearsal would on average emit more power per instrument than ensemble rehearsals).

In the full-size concert hall, a preferred balance between early and late support would be close to
ST_early = ST_late + 3dB. Now in a small rehearsal room, at least those of V<300, the build-up of the reverberant sound field would be close to exponential, implying

                 ST_early ~ 10·log (312T/V) - 3/T (dB), and the early to late balance ST_early = ST_late + 3/T

For T close to 1s, this balance would be close to 3dB in the small rehearsal room just like on the podium of a full-size hall. In contrast, for T<0.5s, the early sound would dominate because
ST_early > ST_late + 6dB. This illustrates how the early-to-late balance can change dramatically with reverberation times less than 1s.  In small rooms of T<1.0s we face again the conflict between proper loudness and proper reverberance as introduced above. In order to avoid excess loudness one would most likely prefer a proper loudness dimensioned by choosing ST_early with respect to the number of sources and let the reverberance suffer, i.e. allow early-to-late balance to be more than the preferred 3dB. Again, letting n denote the number of musicians, demanding consistence with full-size podium acoustics, loudness priority would lead to the following recommendation:

                 ST_early ~ 6-10·log (N+1) (dB),

To sum it all up in a recipe, rehearsal rooms should have reverberation times close to 1s or more in order to maintain proper early-to-late balance and reverberance. Room volume and room height in particular is important in order to accommodate for this. T and if possible V should be chosen to satisfy

                 10·log (312T/V) - 6/T - 10·log (N+1) = 3dB

For small to medium rehearsal rooms with T<1.0s, to avoid excess loudness, T and if possible V should be chosen to satisfy

                 10·log (312T/V) - 3/T - 10·log (N+1) = 6dB

Running Reverberation in the discussion of musician’s perceived reverberance

Running Reverberation (RR) is assumed to be important to the musician in a rehearsal room. Can a short and loud reverberation tail be perceived equivalent to a long and weak one?  How loud is my Reverberation?, D Griesinger asks in an akuTEK publication. Listening tests with solo instrument played with reverb of varying level and reverb time were performed, testing for match of perceived reverberation levels. “The observation that curves (level-time curves, editorial note) made with a 160ms integration time tend to cross at about 160ms if two spaces have similar musician self support leads to proposal that musician self support can be measured with RR160”.

                 RR160 = E(0-160)/E(160-320)

Here E(0-160) and E(160-320) denotes the energy of the impulse response in the two intervals 0-160ms and 160-320ms respectively. RR160 is thus the energy ratio of the 160ms intervals before and after the time 160ms of the impulse response. During solo play, perceived reverberation level tended to be equal whenever RR160 values were equal.

Level matching of perceived reverberation during quartet play and orchestral music led to different results than the level matching during solo play, simply because circumstances are different.

Running Reverberation, or reverberance, i.e. perceived reverberation during play or speech, is not a property of the room alone, but a property of the interplay between source and the room.

Griesinger suggested a measure for the Running Reverberation during orchestral music based on the Schroeder curve, similar to the original EDT measure, evaluating the slope of a straight line from the top point to the 350ms point of the Schroeder curve. Denoting the top point level S0 and the 350ms point level S350, the running reverberation in orchestral music is

                 RR(Orchestral) = 0.35s∙60/(S0-S350)

During orchestra play, perceived reverberation level tended to be equal whenever RR(Orchestral) values were equal.

According to Griesinger, It is possible to come up with similar measures of RR for speech with very little scatter in the data.

Do not confuse solo response with ensemble response

In the examples A, B and C above, we compared soloist response in practice rooms with soloist response on a concert hall stage, and found that these can never be the same. Still we know that apparent trade-off acoustics like example C, with much shorter and much louder soloist response than any preferred concert hall stages, are often reported to be satisfactory practice spaces. How come?

Musicians’ preference for practice room acoustics becomes easier to explain if we instead of demanding the practice room to provide acoustics for the soloist similar to concert hall podium acoustics, demand that the practice room acoustics simulate the sound of the ensemble on a concert hall stage. For example, the violinist playing in a small practice room, e.g. example C above, plays along with her own image sources. This view is especially relevant in the ’hard case’ (see below), where the absorbing ceiling and reflecting walls allowing only the horizontal image sources to ‘play’.

Tilted walls, flutter-echo, energy modes and standing waves

How much should room surfaces deviate from parallel to avoid unwanted acoustic effects? Angles in the order of 6-12 degrees have been suggested, but is the answer always angle-dependant? What if there is very short distance between the pair of walls? Follow link to this prediction demo.

A common case—the hard cuboid with soft ceiling

Controlling standing waves and flutter-echoes: Small Room Acoustics - The Hard Case (paper) (presentation)

Suggesting a new room-acoustical cross-over frequency in rooms dominated by a 2-dimensional sound field, typically a horizontal sound field, dominated by horizontal modes and a lower modal density: Schroeder Frequency Revisited (paper) (Presentation)

AKUTEK Article: Two-dimensional room acoustics

First published May 2007

Latest change 12.08.2019