in: News & Features

January 11, 2012

Calderwood Hall at ISGM: An Acoustician’s Report


This report is my preview of the eagerly awaited new music hall at the Isabella Stewart Gardner Museum. The new Calderwood Hall is replacing the Tapestry Gallery as the site for concerts. My observations are based partly on a guided tour a few weeks ago during which we heard no music, and partly on my observations from today’s official press opening which included a rehearsal by A Far Cry chamber orchestra.

Calderwood Hall, designed by Renzo Piano and Yasuhisa Toyota, is built into a cube 44 feet on a side. Two rows of audience surround the musicians on the floor. The rest of the seating is in three tiers of four-sided balconies – each only one row deep. Seating capacity is approximately 300, similar to the 330 seats of the Shalin Liu Hall in Rockport, but the designs of these two halls, and their sounds, could not be more different.

BMInt staff photo

Current fashion for concert venues of all types tends toward a shoebox shape – long, narrow, relatively high ceilinged, and with the performers in some kind of a stage house. The goal in such spaces is usually a reverberant sound, often at the expense of clarity. But at the Gardner we are being treated to an entirely different concept. The musicians are on the floor, and the audience surrounds them, as close as is practically possible.

The design of the Calderwood is unusual for our time, but it is not historically unprecedented. Most chamber music was written for performance in small spaces – holding at most a few hundred people, and richly supplied with sound absorbing furniture and fabric. The idea that it should be performed in a reverberant space would have been very foreign. Even churches, particularly Protestant churches with their emphasis on vernacular text, were hung with tapestries and banners that reduced reverberation to low levels. The Thomaskirche in Leipzig, where Bach performed his Cantatas and Passions, is estimated to have had a reverberation time of only 1.6 seconds, a very low value for such a large space. It would have been much easier to hear all the words in that space than in the recent passion performance in Boston Symphony Hall.

There was a need for such spaces – and their existence drove the evolution of music. In the Baroque period music became more highly ornamented, both in chamber settings and in opera. Opera houses were typically U shaped, with audiences in tiers. According to Marshall Long in his excellent book Architectural Acoustics:

‘Baroque music was performed in salons, drawing rooms, and ballrooms, as well as in churches. In general the former were not specifically constructed for music, and tended to be small. The orchestras were also on the smallish side, around twenty-five musicians, much like chamber orchestras today. … The problem of distributing sound evenly to the listener was soon recognized, but there were few useful guidelines. In England Thomas Mace published (1676) suggestions for the designer in his Musick’s Monument or a Rememberancer of the Best Practical Musick..He recommended a square room with galleries on all sides surrounding the musicians, much like a theater in the round.’

The Calderwood Hall takes Mace’s suggestion literally. The advantages are clear, the disadvantages equally clear. Ensembles typically arrange themselves facing a particular direction – often with a conductor. Their instruments are directional, with much of the sound directed forward or upward. Audiences behind them certainly get a different sonic and visual picture. The violins will be softer, the violas louder (sometimes not a bad thing…) Those audience members will see the face of the conductor, the backs of the musicians. With the top off a piano radiates mostly upward – a great benefit to the audience in the balconies, a slight detriment to those on the floor.

I was fortunate today to have several conversations with Yasuhisa Toyota, the acoustician from Nagata Acoustics in Los Angeles who was responsible for the acoustics, and Scott Nickrenz, the artistic director of the music at the Gardner. Nickrenz, an accomplished violist, had always been dissatisfied with the sound of concerts in the Tapestry Gallery. Although I am not personally familiar with concerts there, I have heard many reports of very dry sound, poor sight lines, and a great difference in sound between the front of the hall and the rear.

I asked Toyota and Nickrenz how the idea of the new space arose. Why put musicians on the floor, with audience in three balconies? Nickrenz, perhaps anticipating the question, pulled from a folder a wonderful picture of an early Italian opera house. “This is what I said I wanted”, he said. I was in complete agreement, as I have felt for some time that putting the audience in tiers around the musicians is the best way to give a large number of people the kind of clarity that chamber music demands. At least two halls in this area give credence to the idea: Jordan Hall and Sanders Theater. “But why”, I asked “did you decide to put the audience on four sides of the musicians, and not just three?” Nickrenz explained (with Toyota nodding) that this idea arose at the first meeting between Toyota and Piano at Piano’s home in Paris. Piano had always been fascinated with cubic forms. Nickrenz said some good ideas take a long time to develop. This good idea took only 15 minutes. “Good,” they both decided, “let’s go for it”.

A view from on high. (BMInt staff photo

In the Calderwood the goal was to make the sound for each audience member as uniform as possible, giving each both a sonic and visual unobstructed view of the performance. When BMInt visited in December the reverberation time was quite low – about half a second. It is not obvious why the room was so absorptive. The visual walls are made of decorative plywood cut with linear slits. Eighteen inches behind the slits there is a structural wall. There were absorptive curtains in the space between the visual and the structural wall. These were mostly withdrawn into the corners of the structure during our visit, exposing the walls behind. During our December visit the sound in the balconies from people speaking on the floor was clear, highly intelligible, and not reverberant.

In contrast, Shalin Liu is a shoebox, with the musicians occupying a stage area in front of a large glass window. The shoebox has become a standard shape for a music performance space, possibly because of the success of our own Boston Symphony Hall. But Boston Symphony was designed for orchestral music, and has a capacity of 2700 seats. It is not clear that a shoebox is the ideal shape for a chamber music hall with a much smaller capacity. In a shoebox some the sound up close is louder and clearer than the sound in the rear. A shoebox hall also typically has large surface areas that are sound reflecting. These surfaces reflect sound between each other and eventually into the audience, increasing the loudness and adding – in large halls – a warm enveloping reverberation. In small halls these reflections are a mixed blessing, as they come sooner and stronger than they do in a large hall. When there are too many early reflections the sound is muddled, and notes are difficult to separate one from another.

In the front of a large shoebox orchestral hall the sound in is dominated by the direct sound – sound that travels directly from the instruments to the listener. There are typically a great many prompt reflections – both from surfaces of the stage house, and from the ceiling and walls of the hall. These reflections increase the loudness of the sound somewhat as you move back – but they do so at the expense of clarity. The direct sound decreases in pressure by 6dB for every doubling of the distance, so in the rear of the hall the direct sound is weak. Early reflections – those that bounce off only one or two surfaces – are similarly weakened as they travel to the rear. But there are a great many later reflections that have bounced off many surfaces before they reach the listener, and these combine to make the loudness more uniform. For listeners in the middle and the front of the hall it is this late reverberation that gives orchestral sound its warmth and envelopment. But in a small hall we do not need the extra loudness if the audience is close to the instruments. The direct sound is strong enough.

Coffered ceiling and side walls make corner reflector (BMInt staff photo)

So why do we need reflections in a chamber music hall? Were we happy with the dry sound in the Tapestry Gallery? The answer depends in part on what experience an individual audience member would like. When direct sound is un-muddled by early reflections it is perceived as close to the listener. It demands our attention. Clarity is very high, and with careful listening one can hear every note, and identify which instrument played it. Listeners intent on hearing every note will be delighted – those wanting a less demanding “well blended” sound may be disappointed. Assuming no-one in our distinguished audience is in the latter category, what other purposes do reflections serve, and does the Calderwood deliver them?

Musicians do not like playing in a reflection-free space. They rely on sound from the hall to judge the balance between their own instrument and those of their colleagues. A solo string player needs sound from the hall to judge the effects of bow speed and position on the timbre and tone of their playing. It is also much easier to hear one’s intonation when the sound in the hall can be heard. But reflections that arrive earlier than 50 milliseconds are masked by the sound of their own instruments. All the beneficial reflections arrive later than that – ideally more than 100ms later. Like the perception of warmth and envelopment to an audience member, musicians depend on late reverberation. How can you have late reverberation in a small, absorptive hall?

Turns out – you can. Reports from several musicians that have played in the new hall are quite favorable. A glance at the geometry of the hall explains why. The underside of each balcony makes a right-angle with the wall behind it, forming a corner reflector. Sound striking the underside of the balcony will bounce against the wall behind, and return back to where it came from. If we imagine a musician near the center of the floor, his sound will strike the underside of three balconies in four different directions. Reflected energy will return to him at twelve different times from twelve different directions. If the curtains are removed the corners of the side walls the balcony undersides add an additional twelve reflections. The dimensions of the space are such that the first set of reflections arrives before 50ms, but the later ones arrive well after. There is sufficient strength and time delay in the combination of reflections to give the musicians a satisfactory sense of the hall. But there is some danger: If there are too many of these reflections, clarity will suffer. The audience also benefits, although to a lesser degree. Reflections directed to the floor are not absorbed – at least if the floor is not packed with musicians. The sound will bounce off the floor and back up to the balcony, creating at least some late reverberation.

Detail of sidewall (BMInt staff photo)

Audience members also appreciate reflected energy. In a reflection-free space each member of an ensemble is perceived as a soloist, and the group never seems to blend together into an ensemble. This problem is well-known to recording engineers. Closely-miked instruments never sound together unless some early reflections are added. My initial impression in the Calderwood during the first visit was that that there would be sufficient reflections in the time range of 80ms to 100ms to provide a sense of ensemble. This would certainly be true on the floor, where the parallel wall surfaces will provide them. But once again, if there are too many reflections, clarity will suffer.

So – how does the new hall sound? Short answer: it sounds fantastic. Nickrenz is in heaven. I listened to the rehearsal of A Far Cry, the resident chamber orchestra of the Gardner, in many of the balcony seats. (We were shooed away from seats on the floor.) The sound in every seat, according to both Toyota and Nickrenz, was “clear and warm.” (I think they had rehearsed their speech…). I am delighted to concur. One can quibble, and maybe this was my charge in being assigned to this article by BMInt. So here goes.

David Griesinger deliberates. (BMInt staff photo)

When we visited in December the curtains behind the perforated walls were mostly withdrawn. The reverberation time was on the order of half a second – although I was not able to measure it. Yesterday the reverberation time was immediately perceived as longer – my guess is about 0.8 seconds, and quite uniform with frequency. (This means the bass response is rich.) The change in reverberation time between the visits is significant. I inquired as to what had changed. The answer from Toyota was that the curtains had been completely withdrawn into enclosed spaces. I worried a bit about this increase, as I had looked forward to unusual clarity in the hall – much like an early opera house. (The Staatsoper unter den Linden in Berlin (currently under renovation) had 1500 seats and a reverberation time of 0.9 seconds). Small spaces lose clarity quickly as the reverberation time increases. But I was relieved when I heard A Far Cry play. Yes, the hall is more reverberant than before, and yes, the clarity has gone down a bit. But it is still good in every seat, and very good in most.

It is worth explaining what I mean by clarity, as I use the term somewhat differently than current acoustic texts. I have found that humans can instantly detect when a sound is close to them or far away. The detection is usually subconscious, but it can be very important. We pay attention immediately to sounds that we perceive as close. Those far away can be attended to more leisurely.

The quality of sound that allows us to perceive “close” is the same quality that allows the brain to separate simultaneous sounds from several instruments, determine from what direction each came from, and attend to one of several conversations at the same time. To me, sound is “clear” when all these perceptions can be made. These perceptions are mostly subconscious, but when you listen carefully you can hear them. A Far Cry was rehearsing a Haydn’s C Major cello concerto and Britten’s Variations on a theme of Frank Bridge. They were arranged conventionally in a semicircle, with the violins on the left, the violas on the right, and the cello in the middle. When I was sitting in the balconies in front of the group the clarity was very high. Even in the third tier I could clearly localize the cello with my eyes closed. The violins were equally clear, the violas mostly inaudible, as is normal. When I moved to the rear of the orchestra in the third tier the balance was still very good, the music terrific, but with my eyes closed the cello was not sharply localized, just somewhere in front. (With eyes open you always hear it just where you see it.) The violas were suddenly audible and localizable. The sound was good – but not as gripping as in front. The cello was more localizable in the rear part of the second tier, and the cello localized sharply in the rear of the first tier. The localization and gripping clarity was good at the sides of the orchestra in all the tiers – but you have to look over the rail. If you settle back in your seat behind the glass balcony front the sound is “well blended”. It is good but not gripping.

In short – no problems, and an amazing success story. Nickrenz and others have plenty of time to play with the curtains in the walls. Just a bit more absorption might bring the clarity up a bit, at the expense of what is currently audible reverberation.

Calderwood represents a bold break with current fashion in chamber music hall design. It portends an acoustic much closer to the kind expected by the great composers of the Baroque and Classical periods – strong, balanced, and exceptionally clear. I look forward to hearing many different types of concert there.

Griesinger’s first ISGM concert review, “Acoustics vs. Performance: Claremont Trio at ISGM,” is here.

David Griesinger is a Harvard-trained physicist who is eminent in the field of sound and music. His website is here.


  1. “I asked Toyota and Nickrenz how the idea of the new space arose. Why put musicians on the floor, with audience in three balconies? Nickrenz, perhaps anticipating the question, pulled from a folder a wonderful picture of an early Italian opera house. “This is what I said I wanted”, he said.”
    Thanks,  David,  for this excellent essay. And “Amen” to all that from this corner. My colleagues and I had the privilege of touring in central Italy a few seasons ago, and the main thing I learned,  as we moved from one  18th/19th century theater to the next,  is that the “théatre italien” is one of the major contributions of the Italian peninsula to world civilization.  In that concept, no one is seated very far from the source of sound, the performers can sense the audience and its moods, the public can hear clearly,  and singers and players are not obliged to strive for volume at the expense of musicality.
    Concerning the question of reverberation versus “mat” sound,  I have a subjective memory of a pleasant “cushion” of air for the human voice in the mid-sized theaters we visited.
    May there be many new halls along these time-tested lines!

    Comment by Joel Cohen — January 12, 2012 at 11:58 am

  2. *** Yesterday the reverberation time was immediately perceived as longer – my guess is about 0.8 seconds, and quite uniform with frequency

    David, thanks for the article. Of course there are some elements in your article that I do not agree (who could believe?) and there are some that are just inaccurate. Reverberation time you mentioned can’t be uniform with frequency by default.  Higher frequencies decay much faster. The standard 60dB decay usually is measured around 50Hz, some measure lower – at 27.5Hz – the lowest key in piano’s sub-contra-octave. Still, no one talks about the uniform reverberation across any more or less significant bandwidth.  The ratio with which the HF decay relative other ranges is a very complicated subject, has zillion “depends” but it for sure has nothing to do with uniformity.

    Comment by Romy The Cat — January 12, 2012 at 9:06 pm

  3. >> Reverberation time you mentioned can’t be uniform with frequency by default.  

    Sure it can, over wide bands, to an extent
    (depends on the surfaces within the space! Sabine in Sanders, etc.)

    >>  no one talks about the uniform reverberation across any more or less significant bandwidth.  

    Sure they do. How are you understanding ‘uniform’?

    >> The ratio with which the HF decay relative other ranges is a very complicated subject

    not really

    >>  for sure has nothing to do with uniformity.

    Gross overstatement. To first order it does. 

    While your LF points are taken, why would you ever spout off so contradictorily and confidently about matters easily checked? (I mean, look at figures 3.2-6 and -7 in the NAB Engineering Handbook.) 

    Much less to someone (for sure not me) who’s an actual expert?

    Comment by david moran — January 13, 2012 at 4:57 pm

  4. David, thanks for this, though I confess most of it is way over my head.  Having been privileged to attend the A Far Cry concert on Friday night I can attest to the fact that this hall is vastly superior to the Tapestry Room.  I, too am looking forward to many concerts in this new hall but looking down on the tops of the performers’ heads (I was in the 2nd balcony) will take some getting used to.  Especially in the case of A Far Cry, I value being able to see the performers’ faces.  As a fellow audience member remarked on Friday, their faces radiate pure joy.  (But I could see all their hands, and from a new and different angle, which as an amateur violinist I do appreciate!) I’m so grateful to Scott Nickrenz and the rest of the Gardner leadership for having the vision and the courage to embark on this project.  It enhances what is already one of Boston’s greatest treasures.

    Comment by Mary Lincoln — January 16, 2012 at 12:47 pm

  5. To David Moran:

    David, I do not claim to be an “expert” but I do claim that if one applies a rudimentary common sense to the “matters easily checked” then it will be self-evident that the uniformity of reverberation decay across any more or less significant bandwidth is imposable. You do not need to refer me to handbooks; this is what the nerdy “experts” do. We, the people of common sense, operate by tangible facts. Take RT-60 meter and do your own measurements. Remind you that reverberation time is derivative of volume, surface area and absorption/reflection of materials at given frequency.  So, first:  the radiating frequency is a factor, second: any absorption materials have frequency-dependant coefficiency.

    I would not even mention that air itself absorbs higher frequencies at much higher rate than lower frequencies. You do not need to send me to a book to verify it. Take RTA analyzer and measure anechoicly sound at different ranges. I do not think you need to actually do it as it is very much obvious.

    I would not also mention that in real life the rate of decaying of lower frequencies might be much extended not only by volume but many other factors: like standing waves in the given location, forming cylindrical waves that decay twice slower than single-source wave and many others factors….

    So, when your “acoustician” comes to a listening space and measures reverberation time as the decay of white noise then s/he measure in fact the decay of the lowers note can handled in the given space. If you got let say 1.5 seconds decay than it was at the LOWEST NOTE. If you inject into the room not the noise with equal power spectral density but let say you high-passed it at 10 kHz then you will have a small fraction of that 1.5 second. So, how we can talk about uniformity of reverberation time across wide bandwidth is absolutely beyond me. Uniformity is impossible.

    BTW, I do insist that the ratio with which the HF decay relative other ranges is a very complicated subject. By shaping reflective surfaces (panels) at specific distances it is possible to moderate how slow your HF will be decaying and for some music and some specific recording environments it might produce very interesting result.  I would refer you to the ways how the Westminster Records recorded pianos in 50s and to many other examples. Still, it is a divisive and very complex subject… unless if it was done right…

    Comment by Romy The Cat — January 17, 2012 at 7:32 am

RSS feed for comments on this post.

Sorry, this comment forum is now closed.