Appendix 8: Electronic Acoustic Enhancement
←
→
Page content transcription
If your browser does not render page correctly, please read the page content below
Appendix 8: Electronic Acoustic Enhancement
8A. Introduction (Patrick Everett) Electronic enhancement has attractive potential for
improving our sanctuary acoustics, including the promise of adjusting reverberation and presence
to suit the occasion, and of being adjustable in real time. The sound can also be directed into
spaces that are starved of sound by the room geometry and absorbers. Improvements can be
readily verified by switching them in and out. Use of digital signal processors can give great
control over the acoustic properties, and allow them to be changed at will.
We have already pointed out (Appendix 3) that the exponential decay RT60 model is an
imperfect description of the passive decay of sound in a space, with or without electronic
enhancement. For the first 50ms or so, the RT60 concept does not lead to an accurate picture of
what is happening. Yet we know that "presence" is highly dependent upon such short-term
reverberation. To really know what is happening during that important initial period, we should
measure it as the intensity profile over time rather than employ a single exponential parameter to
describe a more complex function. This applies with or without electronic enhancement. A good
microphone and oscilloscope should be part of the standard equipment for measuring such short-
term behavior so as to understand the signal-to-noise characteristics of the sound that is
presented to our ears. To test ideas, and broaden our understanding, it would be good to have
things set up such that interested parishioners can do such measurements.
Electronic enhancement could open up possibilities for controlling the early reverberations to
add presence without adding noticeably to the longer-term decay times that contribute to
muddiness in faster-moving music and reduce speech intelligibility. We could achieve control of
the acoustics in a way that cannot be achieved when limited to purely exponential decay. We
could control the longer-term decay to better match the music selection, as occasion demands.
Examples of how we might achieve this are cited in the later sections of this Appendix. Here we
are taking a more general look at electronic enhancement.
We are aware of three options for electronic enhancement:
OPTION 1: Microphone input, without electronic time-dither (as attempted in the past).
OPTION 2: Microphone input, with electronic time-dither (as with LARES).
OPTION 3: Input from an accompanying electronic organ playing same notes (as at Allin
Church).
The first option has been proven far from ideal in the past. The other two are now practical as a
result of recent developments. In reading the following, please keep in mind the possibilities of
combining Options 2 and 3. This possibility will be revisited at the conclusion of Appendix 8E.
8A.1 OPTION 1: Microphone input, without electronic time-dither
This is the classical approach. A source produces sound that is picked up by one or more
microphones. The sound signal is manipulated in a simple way by electronic processing, and
then transformed back into sound by one or more loudspeakers.
Generally there is inherently some distance between source and microphone(s), which provides a
way for ambient sound to enter the microphone along with the intended input from the singer,
speaker or instrumentalist. Even if the distance is quite small, some of the room reverberation
will enter the electronic system. The more amplification and reverberation there is, the greater
Appendix 8 – page 1the possibility of some of this feedback being positive and causing the whole system of room and
electronics to oscillate in one or more acoustic modes of the room.
Feedback can be positive or negative, depending on frequency and time delay in the electronics.
Positive feedback can cause oscillation and the resultant unpleasant squealing. Negative
feedback suppresses the affected frequencies.
With a distributed sound source, such as an organ, it becomes difficult to have all distances
between sources and microphones small enough, because the pipes are so spread out. The reader
must surely have been aware of electronic speaker systems going into a squealing state, and
technicians then rushing in to make adjustments. The higher the gain and reverberation in the
system and the greater the gap between source and microphone(s), the more serious is the
potential problem.
With such systems it has proven difficult to obtain sufficient volume increase or enhancement
without these oscillation problems; or at least producing an unwanted coloration of the music
from amplified room modes, even if they do not go into oscillation. This has led to electronic
enhancement getting a bad name.
8A.2 OPTION 2. . Microphone input, with electronic time-dither (as with LARES)
A great step forward has been made in the last decade by electronically dithering the very small
time delay present in the electronics between microphone and loudspeaker.
Those who have studied resonators will recognize this as dithering the phase of the resulting
sound, so it becomes temporally incoherent. Hence the feedback becomes negative as often as it
is positive, so there is no longer any net constructive interference with the original sound. Thus
the "modes" are no longer exaggerated, relative to the other frequencies by the repeated selection
and amplification. This is the approach developed by David Greisinger and applied in the
LARES equipment. It is described in more detail in Appendices 8B and 8C. One of its first
prototype applications was in our neighboring 51 Walden Street! See Appendix 8D. Since then it
has been demonstrated, with much success, in various concert and opera halls, and places of
worship around the world, as described in Appendix 8C.
8A.3 OPTION 3. Electronic input (as at Allin Church)
If the sound is already in electronic form, then no microphone is needed. Hence the problematic
gap between the source and any microphone is absent. Consequently there is not the same
opportunity for the "acoustic coloring" from the room to enter the enhancing electronics and lead
to distortions or oscillations. The Allin Church has coupled a new electronic organ with their
existing pipe organ, matching the notes very carefully. The new digital console plays both organs
simultaneously. The same notes are played simultaneously on both organs. The pipe organ
delivers its sound at one end of the sanctuary. The electronic organ delivers its sound through 12
loudspeakers distributed around the periphery of the sanctuary.
The Allin Church in Dedham has serendipitously demonstrated the power of this approach,
which we discuss further in Appendix 8E.4.
8A.4 Combine OPTIONS 2 and 3
The possibility of combining OPTIONS 2 and 3 will have the benefits summarized here, with
further details discussed in Appendix 8E.4. By simply copying what the Allin Church did, there
Appendix 8 – page 2would be enormous advantages as far as the sound of our organ is concerned, with the following
added benefits.
− a spare organ for emergencies
− a new digital console
− new pipe sounds available in the electronic organ without extra cost
− MIDI keystroke recording capability for accurate reproduction
The latter should be very helpful indeed to the organist in evaluating how the organ really sounds
to the listeners in the different parts of the church. It would also help in the fine-tuning of the
sanctuary acoustics, and possibly of the organ itself.
In this first phase we would continue to use the present system for the spoken word. We might
extend this system to allow feedback to the choir, firstly from their own singing, and perhaps
some from the organ since at times they are apparently unable to hear the organ sufficiently
loudly. These tasks are likely to be within the capabilities of conventional electronics.
A second phase would add adjustable electronic reverberation control to the electronic organ.
This would be relatively inexpensive because it would use classical electronic boxes which are
available competitively. There would be no problem with acoustical feedback entering
microphones, because there would be no microphones. This would give the long reverberation
times which have been sought for the organ in the past. The RT60 time would be selectable by
the organist to be optimal for the music being played.
A third phase would add microphone input capability for congregational singing and other
instrumental inputs. The old acoustic feedback problem would be avoided by feeding the signal
from the microphones through a LARES electronic time-ditherer into the electronic organ
system. This input could also go through either LARES or conventional electronics to add
adjustable reverberation or presence to the extent desired. Again this could be adjusted as needed
for each requirement, even during a service.
A fourth phase could couple the preaching into the system in the same way, if this were thought
desirable. The presence and reverberation aspects for the different channels could be
independently controlled before going to the final amplifiers.
These phases could be spread over a number of years.
The details of how all this would be done, and how much it would cost, would depend upon the
FP expertise that can be brought to bear, as well as the cooperation of Steinmetz and the LARES
(or equivalent) people. These factors would determine whether we are followers of what other
people have already proven (and pay that price) or whether we would help shape the way things
are done.
If sufficient FP talent could be brought to bear on this, then this approach could well be more
cost effective than a pure LARES-type approach
8B. Electronic Acoustic Enhancement – The LARES Approach (Dave Kelch)
This falls under OPTION 2 of Appendix 8A – Microphone input with electronic dither.
What do the San Diego Symphony, the Deutches Staatsoper in Berlin, the Tsai Center at BU, the
Central Synagogue in NYC, the Piedmont Community Church in California and our own 51
Walden have in common? They all have an electronically enhanced performance space designed
by LARES Associates in Belmont, MA. Also, they all had performance spaces that were less
Appendix 8 – page 3than ideal acoustically and had to accommodate different kinds of performances with conflicting
requirements -- speech, choirs, orchestra, organs, ballet, opera, soloists, etc.
Electronic acoustics have come of age in the past ten years. Operas and churches seem to have
been on the forefront of this development because each has needs that are difficult to meet in a
single room. Operagoers want the full rich sound of the orchestra accompanying the singers but
they also want to be able to hear every word intelligibly from the singers. Churches want to hear
the minister and also have a good performance space for music - choir, congregation and organ.
These conflicting requirements are difficult to meet in one room and so compromise is
necessary. With electronic acoustics, it’s now possible to change the room’s characteristics for
each type of performance. For example, at the Piedmont Church the reverberation time (RT) can
be set to over two seconds for the organ and immediately changed to less than one second for
speech by pressing a single button on the control panel. There are other single button settings for
the choir and congregational singing, which are selected as needed during the service.
For operas, the frequencies of the orchestra are treated to produce a rich all-enveloping sound
throughout the hall. The higher sound frequencies necessary for understanding the singers’
words are enhanced in another way so that the audience gets the best of both. The critics seem to
like the overall effect and musical experience. These LARES systems produce their positive
effects without coloring the sound. Audiences just say the music sounds better -- the worse the
hall acoustics were before the electronic system, the more the improvement is noted.
How is it done? It’s based on the way humans hear sound. Our perception is based on sounds
that come directly from the source to our ears. This direct sound is augmented by reflected
sounds from surfaces in the room that give the sound richness and color. Some reflections are
pleasant and add to our enjoyment, some are not and detract (the so called "cocktail party"
effect). Organ music sounds best when the room has a long reverberation time and one gets the
feeling of being immersed in the sound. This same room would not be ideal for speech because
sound from previous words would interfere with immediate words. The LARES system creates a
sound field with predictable characteristics. It can add reverberation time to a room that’s
acoustically dead due to sound absorbing carpeting and upholstery. It can amplify speech and
make it intelligible throughout the room. It can pick up immediate sound from an orchestra string
section and project it toward the winds so the wind players can hear and thereby blend better
with the string players.
At 51 Walden, the system is set up for both the Concord Orchestra and the Concord Players. The
room was intentionally deadened (RT60 of 0.6 seconds) to make it "ideal for speech". When the
orchestra is playing from the Walden Street end of the room, the LARES system recreates the
acoustics of Symphony Hall with a reverberation time of about two seconds. The system also
projects some sound back to the orchestra so the players can hear each other and produce a better
blend. When the players are performing, the system is reconfigured (by the flip of a single
switch) to enhance the sound from the drama stage in a low RT60 environment.
Now we get to the technical stuff. The secret of the LARES system is its ability to create virtual
acoustic spaces. Doing this requires lots of loudspeakers and electronic time delay modules. In a
system for a space the dimensions of the First Parish sanctuary, one would need about thirty
recessed speakers in the ceiling in a 10’ center to center grid spacing. Each speaker is 9" by 15"
with a cloth or finely perforated metal front painted to match the ceiling. In addition, twelve or
so recessed speakers would need to be mounted in the ceiling under the balcony to provide a
horizontal component to the sound field. The speakers are driven by sixteen or more separate
Appendix 8 – page 4sound channels. They are derived from multiple variable time delay units that electronically
create sound reflections, which produce the reverberant effect. The time delay units are
programmed to build up the sound field needed to produce the desired acoustic environment.
One of the unique characteristics of the LARES system is its ability to add a lot of sound energy
to the room without producing unwanted feedback. This is done by randomly slightly varying the
amount of delay in each time delay unit. This has the effect of removing most of the correlation
in the reverberation and prevents unwanted build up of sound energy at room resonance points.
From a practical point, this allows great freedom in microphone placement. In fact, just two
small microphones suspended eight feet over the orchestra provide sufficient sound pickup for
the room reverberation enhancement system.
The other physical requirement is space for electronics system -- approximately two standard six
foot racks, about four kilowatts of electric power and ventilation to remove the unwanted heat.
The final piece of data is the cost -- in the neighborhood of $150,000 - designed and installed.
A lot of money to be sure. However, it would probably cost more to achieve the same effect with
traditional architecture and construction. With the electronic acoustics one has the advantage of
flexibility. Imagine, organ prelude with cathedral grandeur, sermon with improved speech clarity,
choir anthem with Jordan Hall acoustics, and congregational singing with Wellesley Hills
Congregational Church liveliness. UU Nirvana!
8C. ‘Recent Experiences with Electronic Acoustic Enhancement’ by David Griesinger,
Lexicon Corp
RECENT EXPERIENCES WITH ELECTRONIC ACOUSTIC ENHANCEMENT IN
CONCERT HALLS AND OPERA HOUSES
David Griesinger
Lexicon
3 Oak Park
Bedford, MA 01730
dg@lexicon.com
www.lares-lexicon.com
ABSTRACT
This paper gives a brief summary of acoustical theory based on human perception. It then uses
this theory to discuss the design and performance data of electronic acoustic enhancement
systems installed in a number of opera houses and concert halls. The installations include the
Deutches Staatsoper in Berlin, the Hummingbird Center in Toronto, and the Adelaide Festival
Center Theater in Adelaide, Australia. Solutions to the problems of maintaining optimum clarity
of the singers while providing optimum envelopment for the orchestra are given.
INTRODUCTION
Electronic acoustic enhancement of spaces for music performance has frequently been long on
promise and short on performance. The major problem has been uncontrolled acoustic feedback
between the microphones and the loudspeakers in the enhancement system. This feedback
induces an artificial metallic coloration into the system. Avoiding the coloration has meant
operating the system at loop gains that provide little overall benefit. Enhancement systems that
use multiple time variant reverberators have essentially solved the feedback problem, allowing
the system designer to create the necessary acoustic fields without artificial coloration.
Appendix 8 – page 5The technology is available – but how do we use it? We are faced with the problem of
determining just what are the major acoustic difficulties of a particular space. Assuming we
identify these correctly, how can microphones and loudspeakers be installed to solve them
without breaking the budget? This paper will present a brief update on our research into the
perception of musical acoustics, and show how this knowledge can be applied toward solving
problems in real spaces.
RECENT RESEARCH INTO PERCEPTION
In a previous paper (1) we wrote that acoustic descriptors can be divided into four categories:
descriptors of localization, spaciousness, intelligibility, and reverberation. Since that time we
have made some progress in understanding how humans perceive sound, and feel it is possible to
revise this list.
We find that there are several key processes in sound perception, each working at a successively
higher neural level. The low level processes, such as the separation of sound into different
frequency bands, and the detection of localization through Interaural Intensity Differences (IIDs)
and Interaural Time Differences (ITDs), occur early in the perception process, and in general
have very short time constants. Higher-level processes can take substantial amounts of time to
complete. The time constants are vital to the way we perceive room acoustics.
LOW LEVEL PROCESSES:
1. The analysis of incoming sound pressure into frequency bands. This analysis takes place
on the basilar membrane and is fundamental to the hearing process. At the lowest level
we hear sounds in separate frequency bands.
2. The detection of rapid increases in level in individual frequency bands. This "rising edge"
detection occurs early in the neural process and is the first step in the detection of the
starts of individual foreground sound units. In speech these individual sound units are
called "phones". In music they are called "notes".
3. The detection of interaural time and level differences in each frequency band. These
interaural time and level differences are gated with the "rising edge" data to determine the
azimuth of the sound source. When the "rising edge" data is absent – when the sound is
continuous – IID and ITD still determine azimuth if they are stable and consistent. For
example, when a sound occupies several critical bands the IID and ITD should be the
same for each band.
4. The determination of the average uncertainty in the IID and the ITD. In an acoustic
environment these uncertainties are primarily due to fluctuations in the both the IID and
the ITD. These fluctuations are caused by interference between the direct and reflected
sound. Fluctuations in IID and ITD that occur during the rise time of a sound event
broaden the source image. Fluctuations that occur later can be interpreted as room sound
or as envelopment, depending primarily on when these fluctuations occur relative to the
sound events.
Of these low level processes the most basic is the analysis of sound into frequency bands. In all
the processes that follow this separation has already occurred. Thus when we speak later of
localization or envelopment we are not assuming these perceptions to be independent of
frequency. Localization can be sharp at high frequencies, and at the same time it can be poor at
low frequencies. If a particular sound event includes both high and low frequencies the sharpness
of localization in the different bands can be separately perceived – although in overall impression
Appendix 8 – page 6the most accurately localized bands will dominate. The same frequency selectivity applies to
envelopment. The frequency dependence of intelligibility, localization, and envelopment is
particularly important to musical acoustics.
HIGHER LEVEL PROCESSES, IN APPROXIMATELY THE ORDER THAT THEY OCCUR:
1. The parsing of sounds into individual units, the phones and the notes. To perform this
parsing the hearing process must find where one sound event ends and another begins.
Thus detecting the ends of sound events is often as important as detecting the beginnings.
2. The determination of the direction and timbre of individual sound events.
3. The organization of groups of sound events into foreground streams. In speech the
phones from a particular speaker are organized into phrases and sentences. In music the
foreground streams consist of musical lines from individual instruments or sections.
4. When there are several speakers talking at the same time, we organize the sound units
from each of them into separate streams. Likewise there can be several simultaneous
foreground sound streams.
5. The stream formation process sorts individual sound events using all available clues, such
as direction, timbre, and pitch. Thus the azimuth of a particular event can help assign it to
a stream.
6. The formation of a "background stream" that contains the sounds perceived between
elements of the foreground stream. The background stream contains room noise,
reverberation, etc. While there can be several foreground streams, there is only one
background stream.
We wish to emphasize that the formation of sound streams is a vital part of our sonic
perception. We perceive localization, timbre, and reverberation much more strongly in a
series of connected sound events than we do in individual, isolated sound events. For
example, when we hear the reverberation from a loud chord that is followed by silence,
the reverberation becomes a foreground sound event – we can apply the entire analysis
power of our brains to it. When the reverberation is heard during the gaps between
syllables of speech an entirely different neural process is involved in its perception. This
perception is not of an event, but of a continuous stream of sound. If the reverberant level
is strong and there are significant fluctuations in the IID and ITD during these gaps we
will perceive significant envelopment. During and after the stream formation process
several further actions occur:
7. The assignment of meaning to the various foreground streams.
8. The inference of source distance from the relative strength of the foreground and
background sound streams.
9. The interpretation of the fluctuations in the IID and ITD as either a "room" impression, or
as envelopment. This interpretation depends on the time delay between the end of the
sound event and the reflected energy that produces the fluctuation.
Of the higher level processes, the parsing of sounds into individual events is by far the most
critical. Speech comprehension drops very rapidly when noise or acoustic conditions prevent the
reliable detection of the ends and beginnings of phones. This is why the modulation transfer
function of an acoustic channel is a meaningful measure of speech intelligibility.
Appendix 8 – page 7From the nature of the event detection process we can see that:
1. The effect of early lateral reflections on localization will depend on the rise time of the
sound events used as a sound source.
2. Where sound events have rapid attacks the sharpness of the sound image (the apparent
source width) is determined by the presence or absence of reflected energy that arrives
during the rise time of the sound event. The gating of localization with the "rising edge"
data gives a significant advantage to us as a species. The rise time of sounds is usually
not corrupted by reflections. Speech phones can rise quite rapidly – in under 10ms. Thus
speech can be accurately localized even in small spaces. The same is true of musical
sounds from many solo instruments. Perceptual experiments show that lateral reflected
energy arriving later than about 10ms has little effect on the source width of such sounds.
3. Legato music for a large string section tends to have long rise times for individual notes.
When we do a perceptual experiment using such music, we expect that the image will be
broadened by lateral reflected energy with delay times of 50ms or more. The expected
broadening is easy to confirm.
4. Intelligibility of either speech or music will be reduced if reflected energy reduces the
ability of the hearing mechanism to detect the starts and ends of phones. Phones in rapid
speech come as frequently as every 150ms. Normal speech can be somewhat slower. The
gaps between phones are typically 50ms or greater. From this data we can immediately
infer that reflected energy arriving between 50 and 150ms after the ends of a sound event
will be particularly detrimental to intelligibility.
5. The perception of reverberation and envelopment will depend on the presence of gaps
between phones or notes where reverberation can be heard, and on the ability of the
hearing mechanism to separate the sound in these gaps from the foreground sound events.
This separation process takes time – at least 100ms after the end of the sound event must
elapse before the sensitivity to background sound is at a maximum.
6. Thus the perception of reverberation and envelopment depend on:
a: the "transparency" of the musical material
b: the strength of the reverberant sound at least 100ms after the ends of notes.
For speech, solo music, and thinly orchestrated music, lateral reflections arriving in various time
ranges have the following properties:
0-10ms – these reflections make the sound event louder, change the timbre, broaden the
source image, and/or cause image shifts.
10-50ms – these reflections cause a "room" impression that is not enveloping, but desirable if
the reflections do not exceed the energy of the direct sound. Reflected energy in this time
range also increases the loudness and affects the timbre of the sound event. We call the
spatial impression created by these reflections "early spatial impression" or ESI.
50-150ms – these reflections produce some sensation of envelopment, but the primary effect
of energy in this range is to reduce intelligibility.
150-400ms – these reflections contribute to the background sound stream.
The background stream is highly audible. It produces the major perception of "support" for a
solo musician, and envelopment for an audience member. The strength of the envelopment
Appendix 8 – page 8perception depends on the absolute loudness of the reverberation. The louder the musician plays
the stronger the envelopment perception will be. We call this envelopment perception
"background spatial impression" or BSI.
For legato strings and continuous thickly orchestrated music, the beginnings and endings of notes
are not easily detected, and the effect of lateral reflections becomes much less dependent on the
delay time. Reflections in the following ranges have the effect of:
0-10ms – these reflections affect loudness and timbre. They can also widen the sound image
and shift the azimuth, but they do not affect envelopment or "room" impression.
10-50ms affect loudness and timbre, and broaden the sound image. They also contribute to a
form of envelopment we call "continuous spatial impression" or CSI. CSI is less audible than
BSI, and depends on the direct to reverberant ratio, not on the absolute level of the
reverberation.
50-150ms – these reflections affect the musical intelligibility, can broaden the source width,
and contribute to CSI
150-400ms – these reflections contribute to CSI. With very legato sources these reflections
can also affect source width.
Notice for continuous music nearly all lateral reflections affect envelopment. In fact, it is the
ratio between the total medial and the total lateral energy that will determine the amount of
envelopment. This is particularly true at low frequencies. At least in opera houses the reflected
energy tends to be medial at low frequencies, since it comes primarily from the front. Low
frequency envelopment tends to be low in such spaces, even for continuous music.
With the development of this perceptual theory, we can narrow our list of acoustic properties.
For perception the vital sound properties are intelligibility, localization, "room impression" (ESI)
and envelopment (BSI).
The importance of each of these perceptions may depend on the type of sound and on personal
preference. For speech everyone agrees that intelligibility is of primary importance. For music
many if not most listeners believe envelopment becomes much more important, and a substantial
degradation of intelligibility and localization is acceptable to achieve adequate envelopment. A
lack of "room impression", ESI, is perceived as a lack of distance and blending between the
listener and the sound source. This distance or blending is nice to have – but in our experience it
is much less important than intelligibility and envelopment. The distinction between ESI and BSI
is important, since traditionally it is a lack of early reflections that is blamed for most acoustic
problems, and yet it may be possible to augment late reverberation much less expensively than
early reflections.
APPLICATION OF THE THEORY TO ACOUSTIC ENHANCEMENT
Modern halls come in all shapes and sizes. In spite of an enormous range of audience capacity
and internal volume these halls are expected to sound good with a wide range of sound sources.
The problems of designing a hall with good acoustics for speech are well known and will not be
extensively discussed here. For speech one wishes to minimize the reflected energy that arrives
50ms or more after the direct sound. Reflected energy or energy from a reinforcement system
that arrives earlier than 50ms can be helpful as long as it is primarily medial.
Appendix 8 – page 9It is well known that if a hall is to be used for music it should have a longer reverberation time
than a hall designed for speech, and it is widely believed that a two second reverberation time is
optimal for music. Unfortunately the acoustic properties of a hall depend both on the
reverberation time and the volume of the hall. A small hall with a two second reverberation time
sounds very different from a large hall with a two second reverberation time.
One reason is that in natural acoustics the reverberation time, the reverberant level, and the
distance from the sound source for a given direct to reverberant ratio are all linked. You can not
alter one without altering all of them. In general a small hall designed for a two second
reverberation time will have much too high a reverberant level for most purposes. The critical
distance – the distance where the direct sound and the reflected energy are equal – is too small.
Another difference is that in large halls the early reflections in the time range of 50-150ms can
be lower in energy than in small halls of the same reverberation time. In many large halls the
reverberation decay is not immediately exponential. There may be a few early reflections from
the stage house, but then there is a little less energy than one might expect before exponential
decay begins. This characteristic is particularly noticeable in a hall without a stage house, such as
the Concertgebouw in Amsterdam. The result is a sound that has both high intelligibility and an
strong sense of envelopment. A small hall cannot achieve this sound. If we make it reverberant
enough to supply envelopment, the energy in the 50-150ms range is too high. Intelligibility,
localization, and timbre are all compromised.
With electronic enhancement the reverberation time and the critical distance do not have to be
linked. The energy in the time range of 150ms and beyond can be altered without excessive
energy in the early field. When using multiple time variant reverberation systems it is not
necessary to use a large number of microphones, and we recommend that two to four
microphones be installed as close as possible to the sound source consistent with uniform
coverage. With such an array the feedback can be minimized, and the designer has much more
control over the reverberant level.
All enhancement systems are not equal in this regard. Some current enhancement systems place
the pickup microphones at distances greater than the reverberation radius from the sound source.
With these systems the enhancement system acts as negative absorption. Any increase in level
from the enhancement system must increase the reverberation time through acoustic feedback.
EXAMPLES
Deutches Staatsoper, Berlin
The Berlin Staatsoper is typical of a great number of European opera houses. It has a horseshoe
plan with four rings, and a seating capacity of about 1500. The stage house and proscenium are
small by modern standards, which gives the house an intimate character very well suited to many
types of opera, particularly Mozart. The occupied reverberation time is below 1 second at most
frequencies, but speech intelligibility, a vital component of the dramatic connection between an
actor and the audience, is very good throughout the hall. Although the acoustics were excellent
for drama, they lacked envelopment for the orchestra, particularly for the music of Strauss and
Wagner. The lack of envelopment was blamed (as usual) on a lack of early reflections, although
attempts to improve the situation with reflectors had failed. Like many other opera houses the
Staatsoper had minimal funds available for any improvements to the hall acoustics, which had
been judged "good enough" for many years.
Appendix 8 – page 10Through the efforts of Albrecht Krieger, the tonmeister, and Daniel Barenboim, the music
director, a Lexicon Lares system was temporarily installed in 1996 for a series of performances
of Wanger’s "Das Rheingold" and "Die Walkure". Installing the system in an historic building
presents particular challenges. The theater had already installed a set of 10 loudspeakers in the
back wall of each ring. Eight additional loudspeakers were installed in a circle around the domed
ceiling, and a pair of subwoofers were installed over the proscenium. The ceiling speakers and
the speakers in the rings were driven by separate Lares frames, so each could be balanced
separately under computer control. The Lares systems were driven by two hypercardioid
microphones installed high over the orchestra pit.
The ring system and the ceiling system were separately equalized for flat feedback transfer
between the loudspeakers and the pickup microphones using the built-in calibration programs in
the Lares system. The overall balance was then adjusted for uniform coverage through the hall. It
turned out to be possible to achieve a ±1.5dB uniformity.
During rehearsals for "Das Rheingold" we learned that the adjustment of the system was critical,
and that the ability to separately adjust the performance of the system at different frequency
bands was vital. We found that the intelligibility of the singers must be preserved at all times. At
first, this meant reducing the system level to the point where there was little improvement for the
orchestra. However, we realized that the frequencies that convey the most information in speech
and singing lie between about 700Hz and 4000Hz, and the majority of the orchestral energy lies
in the fundamentals of the musical tones. These fundamentals lie chiefly below 500Hz. Thus in
theory it is possible to increase the envelopment for the orchestra without compromising the
acting. The system needs to be frequency dependent. With the help of the artistic staff, including
the music director, we found an equalization that did the job – about a 6dB reduction in
reverberant level above 500Hz.
With the equalization the orchestral sound was greatly improved. All the instruments became
richer, and the sound spread out from the pit and surrounded the audience. The measured
reverberation time rose to 1.7 seconds at 500Hz, somewhat less above and somewhat more
below. Compared to the original house these changes are enormous – but to an untrained listener
the sound was completely normal. The following performances brought critical praise –
particularly for the orchestral sound, and no complaints.
A permanent installation was completed in March of 1997, and has been in continuous operation
on every performance since that time. In the permanent system some of the loudspeakers in the
rings were replaced with cardioid loudspeakers above the doorframes, a solution that solved
some problems with hot spots directly in front of the earlier speaker positions. Two independent
subwoofers were installed at opposite sides of the ceiling dome. The system was installed
entirely by the house staff, using available equipment were ever possible. In spite of (perhaps
because of) the low budget approach the system fits the hall well. The system is particularly
beneficial for ballet, where the reverberation time is raised to 2.0 seconds, and less equalization
is used. Critical reception continues to be excellent.
The relative simplicity and low cost of this system depend on several factors unique to the
Staatsoper. One is the skill and dedication of the resident sound staff under Albrecht Krieger.
Another is the excellent speech intelligibility everywhere in the hall. Thus no conventional
acoustic modification was needed to achieve this vital goal. The relatively low level of early
lateral reflections in this case contributed to a sense of intimacy and connection between the
performers and the audience. The small proscenium opening and the very high use of the theater
Appendix 8 – page 11also contribute. Because there is always a new production in the wings, the upper stage house is
always full of absorbent curtains and sets. The natural reverberation time of the stage house is
thus usually quite low. An actor can move from far downstage to far upstage with only a
moderate change in the reverberant quality of the voice. This simplifies the pickup problem. A
single pair of microphones can successfully capture both the actors and the orchestra.
Hummingbird Centre, Toronto
The Hummingbird Centre – formerly the O’Keefe Centre – is the home of the Royal Canadian
Opera Company, and the site of many music and ballet performances. It is a very large hall, with
3200 seats. In spite of the large hall volume the reverberation time is low, about 1.2 seconds. The
unaugmented sound in the stalls is lacking in envelopment and reverberation, while the sound in
the balconies is weak and muddled.
A Lares system was installed in the spring of 1998. The system uses four Lares mainframes, and
four B&K cardioid microphones as pick-ups. The 312 loudspeakers are hidden in the
proscenium, in the diffusing elements along the sidewalls, in the ceiling of the hall, and under the
balconies. The hall is divided into four separate time delay zones for the reverberation and direct
sound reinforcement.
In the Hummingbird there are two problems. The envelopment needs to be augmented in the
stalls and under the balconies. But at the same time the loudness and intelligibility needs to be
raised throughout the house. Here is where the use of close directional microphones is helpful.
By using these microphones it is possible to reinforce the direct sound without increasing the
reverberant level in the hall through feedback.
The Lares software allows a direct sound reinforcement with some feedback reduction to be
mixed in with the later reverberation. This feature was used in the proscenium and in the later
time zones in the Hummingbird with good effect.
The equalization we found useful in the Staatsoper turned out to be equally useful in the
Hummingbird – and in several other installations. Since this equalization is based on the
properties of human perception, it is likely to be generally applicable.
Once again critical responses to the hall has been very favorable. The system is in use for all
music performances (except for modern musicals and operas that include electronic
amplification of all the instruments and voices.)
The Circle Theater, Indianapolis
The Circle Theater was originally a Vaudeville house, converted to a concert hall for the
Indianapolis Symphony. It seats about 1800. The natural reverberation time is low. A Lares
system was installed with Jaffe Holden Scarborough Acoustics. It consists of three Lares frames,
one dedicated to early reflections in the front of the hall, and two to the overall late
reverberation. The reverberation under the balconies is controlled by one system, and in the stalls
and over the balconies by another.
The primary purpose of the system is to augment the later reverberation, thus increasing the
musical support and envelopment. During the final adjustment I spent some time with Paul
Scarborough listening to the orchestra in many seats in the hall, while turning on and off the
augmentation of early lateral reflections. These reflections are chiefly responsible for increasing
a sense of distance and blend in the early sound of the orchestra. We both concluded that the
augmentation was worthwhile. However for me the effect was not essential for the enjoyment of
Appendix 8 – page 12the music. Without the early reflection augmentation the sound of the orchestra was a bit too
direct or "in the face." With the early reflections the sound was a bit better blended and more
gentle. However without it I doubt it was the later reverberant energy supplied by the system that
was essential. Sound throughout the hall was richer, more enveloping, and better balanced with
the system on. Once again the system has achieved critical acclaim. The major orchestral critic in
Indianapolis immediately noticed the system and wrote a glowing review of both the orchestra
and the hall.
Adelaide Festival Center Theatre (AFCT) – Adelaide, Australia
The AFCT is an example of what can be done with a large hall when there is adequate funding.
The hall has 2200 seats, and an occupied natural reverberation time of 1.2 seconds. As in many
opera houses the stage house was too reverberant and the audience house was too dry. About 350
sq. meters of absorption was added to the stage house to control the reverberation there, and 250
sq. meters of absorption was added to the stalls to correct low frequency problems. Additional
absorption was added to the rear walls of the rings to reduce focusing of sound back to the stage.
Carpets were removed and replaced with wooden flooring. The result was an articulate room
with low natural reverberation time.
The Lares system consisted of five Lares frames, two for early energy throughout the hall, two
for late energy, and one for reverberation to the stage. Six B&K cardioid microphones were used,
and 288 loudspeakers. The adjustment and calibration of this system was performed by the local
staff with the help of my colleague, Steve Barbar. The flexibility of adjustment is very high,
putting a large burden on the ears and the knowledge of the personnel making the final selection
of settings.
The system has been in use for an entire performance series of the Wagner "ring", and for several
purely orchestral performances. The control system allows easy selection of settings of the
system for opera, ballet, or orchestra. The critical response to the system has been very good.
Glowing reviews appeared in newspapers in both Australia and London, with several reviewers
commenting that the acoustics were better than any venue in London.
CONCLUSIONS
Electronic enhancement of spaces for musical performance has finally come of age. Multi-
channel time variant reverberation technology has made it possible to design practical systems
that produce very significant improvements with no artificial artifacts. The success of these
systems depends, as always, on careful consideration of the real acoustic needs of the hall in
question, and the resulting design and installation of the enhancement system. Through the use of
frequency dependence, and by controlling the undesirable reflections in the range of 50 to
150ms, it is possible to create acoustics that combine both high intelligibility and high
envelopment. The result is musically very effective, particularly for opera.
REFERENCES
1. Griesinger, D. "The Psychoacoustics of Apparent Source Width, Spaciousness and
Envelopment in Performance Spaces" Acta Acustica Vol. 83 (1997) 721-731
2. Griesinger, D. "Design and Performance of Multichannel Time Variant Reverberation
Systems" Proceedings of the Active 95 conference, Newport Beach CA 1995 p1203-1212
Note: There is much additional information on electronic acoustic enhancement on the web at
http://world.std.com/~griesngr/
Appendix 8 – page 138D. Notes from LARES on Houses of Worship
The following is some written material from the LARES company describing their approach to
church acoustics. It’s admittedly designed to sell their systems, but it’s still interesting to read
how they approach the broad range of demands placed on a church's acoustical environment.
“Houses of worship require acoustics that enable the spoken word to be heard and clearly
understood, provide the support and envelopment required for organ, musicians and choir, and
create a sense of intimacy for the congregation. The problem is that each of these goals
directly conflicts with the others even before architecture enters into the picture
LARES patented processing provides an enormous increase in feedback rejection, allowing
microphones and loudspeakers to be placed in close proximity to one another without the
consequences of coloration or howling. Hence, we can place microphones and loudspeakers
above the congregation and deliver energy throughout the sanctuary, which improves the sense
of intimacy and connection with a choir that is distant. Houses of worship require acoustics
that enable the spoken word to be heard and clearly understood, provide the support and
envelopment required for organ, musicians and choir, and create a sense of intimacy for the
congregation.
The problem is that each of these goals directly conflicts with the others even before
architecture enters into the picture!
In smaller churches, attaining reverberant characteristics necessary for a good musical program
often means increasing the cubic volume of the building, and hardening surfaces to increase the
strength, density and duration of reflected sound in the space. Even if raising the roof or
moving walls is practical, such an exercise is a construction project that will be costly and
require closing the building for some time.
LARES can provide a substantial improvement in overall acoustical balance in smaller spaces.
Although speakers need to be mounted throughout the venue, this exercise is much simpler
than reconstruction, and most often can be accomplished while primary worship services
continue. A LARES system can be designed to simply add energy for choir and organ -
improving spaciousness, warmth and envelopment, and the feeling of intimacy within the
environment.”
LARES provides information on a number of installations of their system in various houses of
worship at http://www.lares-lexicon.com/installations.html.
8E. Examples of Electronic Enhancement Investigated by ATF
8E.1 51 Walden St. (George King and Dave Kelch)
George King spoke with Bill Smith concerning the electronic reinforcement system (ER) system
at the 51 Walden St. hall in Concord. Following is George’s report:
Bill was one of the small group which put together the system there ~1993 for ~$25K. The
system was designed by Tom Horrol and built by Lexicon, Inc. It is a dual-purpose system 1) to
provide presence and reverberation for the orchestra and band playing at the Walden Street end
of the hall and 2) sound reinforcement for stage productions at the FP end of the hall.
The primary reason for installing the system was so that orchestra members could hear
themselves playing.
Appendix 8 – page 14The ER system utilizes a dual digital processor feeding 12 signals to 24 loud speakers located
strategically throughout. The hall has been deadened in order to have good amplified speech
intelligibility from the stage The longer reverberation time provided by the ER system has
proven to be very satisfactory to the musicians and the audience. Musicians to whom Bill has
spoken said the acoustics are wonderful for them. The system can also be reconfigured to allow
the orchestra or band to play at the FP end, or to be used as a general PA system.
When the orchestra rehearsals recommence for the next season, it should be possible to have a
useful demonstration of the system. A number of members of the ATF have either played or
attended (or both) 51 Walden St without being aware that electronic acoustic enhancement was
active!
8E.2: Piedmont Community Church - LARES System (Dave Kelch)
Steve Main, organist and choir director at the Piedmont Community Church in Piedmont (near
Oakland) California, spoke with Dave Kelch about their LARES acoustic enhancement system. .
The Piedmont church, similar to First Parish in size, is a historic 1900’s building with beautiful
redwood beam work which seats about 500. Their organ and choir are at the front of the church.
In 1999 they installed a LARES system and he’s extremely happy with it. It improves the organ
tremendously. It enhances the choir and, most important, congregational singing. He says the
most noticeable feature is how it brings out the "quiet, intense moments" in the choir’s
performance. It is now is possible to do choir pieces with sustained sections and plain chant. The
system has four preprogrammed settings that can be selected during the service: hand bells, soft
choir, loud choir and organ The latter sets the RT60 at 2.5 to 3 seconds. The room works so well
that two groups have used it to make CDs.! Steve makes the point that members of in the
congregation are accustomed to hearing good sound in other venues - symphony, professional
stage, opera - often without realizing that what they are hearing has been electronically tailored.
They just have a pleasant aesthetic experience.
Four microphones are used, two over the choir and two at the front of the congregation over the
chancel step. In addition, the pulpit microphone feeds into the system and the LARES system
becomes a time delayed sound reinforcement system during the sermon (similar to our current
FP sound system). The LARES system was installed in 1999 and cost $135,000. Over half of
that cost was in construction to place the fifty loudspeakers and provide space for the electronics
racks.
8E.3: Piedmont Community Church - LARES System (Rick Moore)
I had a conversation with Steve Mains, music director of the Piedmont Church. This
congregation has installed a LARES system for the listed benefits. Steve is quite happy with it
(and a natural salesman) but he's not affiliated with the company. I reviewed the material covered
by David Kelch and got a complete confirmation of that material (from memory). Then I
pointedly delved into the "congregational singing" question. Steve responded that the LARES
system in their sanctuary creates the sense of presence needed to support congregational singing.
He stated that he knew "just what I was getting at" and related that he had worried over the same
concerns I had regarding how a system can feedback an individual’s local acoustic neighborhood
without local mic’ing of each individual. His system uses 6 microphones for the whole
congregation. The key points Steve made were as follows:
Appendix 8 – page 151. The ‘presence’ effect is supported by hearing other congregational individuals, even if
they're not the ones singing 3 seats away – “we don't really discriminate the minor-delay
neighborhood field all that well, except to note it generally.”
2. The fed-back sound is delayed enough so that it's fidelity isn't so important ... speakers do
this just fine.
The level of enthusiasm from this Oberlin-trained organist and choir director warrants our further
investigation.
8E.4 Allin Church – a special case with possibilities for FP (Patrick Everett)
This falls under "OPTION 3. Electronic Input" of Appendix 8A
The Allin Church in Dedham has serendipitously demonstrated the power of electronic
reinforcement for the organ using electronic input. When faced with the need to rework their
organ console, the music director and organist, Martin Steinmetz discovered they could buy a
high-quality electronic organ with a console that could drive their pipe organ at less cost than
reworking their existing console. They took this opportunity to include this electronic organ at
negative extra cost! The total renovation cost was less than $100,000. This console rework would
have included a digital renovation as Martin considered that would have been necessary.
Being a musical and engineering perfectionist, Martin took great pains to have the electronic
organ voiced to match the pipe organ, as well as to make sure that the console would have a high
quality finish, and operational feel to be compatible with its function in that church. We need to
thank Al Armenti for introducing us to this interesting discovery. Our Task Force visited the
Allin church for a demonstration by Martin (discussed elsewhere in this report); we were very
impressed by the results.
Our task Force reported that it was difficult to tell the difference between a passage played
sequentially on the two instruments, when each organ was playing alone. Moreover; when the
passage was played simultaneously on both, with the electronic organ a little quieter than the
pipe organ, the combination sounded better than either alone. It is not clear just why this is so.
One possible reason is that the electronic organ has loudspeakers distributed around the
sanctuary, and the "surround effect" helps. Another is that, at least in most positions in the
sanctuary, the arrival time for the sounds must be different because of the multiple sources.
These differences would be measured in milliseconds (i.e. thousandths of a second); thus adding
to the presence, but having no significant effect on the longer-term decay of the sound.
In this case of the Allin Church, the console needed reworking anyway, and they chose the less
expensive, but not low quality, way of achieving this, which happily also added an electronic
organ. The accidental enhancement was a surprise and came at no extra cost. For any church
within several years of needing a rework of its console, this would now be a natural path to
consider. However, its success would depend on very careful voicing of the electronic organ to
the pipe organ. If this approach were to be considered then engaging Martin Steinmetz as a
consultant might be well advised.
It is important to note that the Allin Church approach demonstrates electronic augmentation of
the pipe organ by an electronic organ, and the system is immune to audio feedback. This is very
significant. It is an example of "OPTION 3 in Appendix 8A.
Appendix 8 – page 16It should be noted that both the OSC and OAC discussed the possibility of adding an electronic
organ, but had not considered this in the light of its being an electronic enhancement in quite this
way.
This approach should certainly be considered if we are anticipating reworking the console within
the next several years. Adding the necessary digital organ-controls is now considered the natural
thing to do anyway when reworking an old console. The whole electronic organ, except for the
speakers, is enclosed within its console. The console is no larger than the typical pipe-organ
console. However, as with any electronic enhancement, loudspeakers are required. In the case of
the Allin Church they total about twelve. They are not noticeable, since most of them are in the
balcony. Their balcony is not used for seating, since the congregation is small.
8E.4.1 What did the Allin Church achieve for less than $100,000?
They gained a new organ console. Steinmetz told us they were quoted $100,000 for reworking
their old console, including wiring it into their existing pipe organ, without achieving the
following benefits:
Better Sound: Martin Steinmetz believes, and our visiting Task Force agreed, that the
combination of the two organs playing together, with the electronic organ at a lower level,
sounded better than either organ alone. This in spite of either organ playing alone still sounding
pretty good
Backup: A backup organ when the pipe organ needs repair or maintenance. The Allin Church
has already experienced this advantage when the pipe organ was down for a few months for
extensive maintenance. Also an occasional cipher is of much less concern, since the electronic
organ can temporarily substitute.
Added pipes: The Allin pipe organ has 3000 pipes. Their Rodgers electronic organ duplicated
all of these, plus allowing some extra pipe sounds, including extending the low-frequency
range. With our 2200 pipes, there should be many more pipe sounds available for us. This
could be an attractive alternative to adding our missing pipe sounds individually, at probably
greater expense.
Synthesizer included: This can synthesize instruments and sounds not normally available from
a pipe organ. Not necessary, but a nice added feature.
"Playback" capability: All the organist's key and stop, etc, commands can be recorded and then
played back later. This allows the organist to walk around and effectively hear himself playing
at any point in the sanctuary. The potential of importance of this capability should not be
underestimated. The organist could also use this feature to play a duet between the pipe and
electronic organs.
8E.4.2 Issues and questions
The following issues may arise with respect to an Allin-type installation:
Obsolescence: The OSC and the OAC commented on the likely obsolescence of an
electronic organ. Note however that the reason computers quickly become obsolete is
because their capabilities are rapidly improving. If one does not need or want, the new
capabilities, then there is generally no need to replace the computer. Likewise with electronic
organs. When finally the time comes that replacement of obsolete parts becomes a problem,
Appendix 8 – page 17You can also read
