What went wrong, and how to fix it
Christopher Blair – full time acoustician, part-time conductor, 3rd time blogger. “For me the evening can’t end soon enough. I head back to my hotel with a splitting headache triggered by the blare of the orchestra and that spot in the Mahler where a percussionist strikes a rail with a sledgehammer….There’s enough blame to go around, of course, but by now I’ve become a convenient scapegoat. My dream of a great hall and my reputation as an acoustician both appear to be going up in smoke.” – Leo Beranek, in Riding the Waves, The MIT Press, 2008
Much has been said of the famous opening of Philharmonic Hall (now Avery Fisher Hall) at Lincoln Center and the acoustic brouhaha that ensued. But most out of the profession are unaware of the political pressures and fatally flawed design process that were the true parents of this disaster.
In December 1959, under the title “Final Design” the New York Times published a description of the approved design using architect Max Abramowitz’s sketch of a hall with no more than 2400 seats, a volume of 700,000 cubic feet, an adjustable canopy at the front of the hall, and side balconies that extended horizontally, parallel to the main floor. If only this room had been built.
Shortly after publication, several New York newspapers, particularly the Herald-Tribune, complained about the 2400-seat limit, arguing that Philharmonic Hall should contain at least as many as the 2760-seat Carnegie Hall, then scheduled for demolition. The Lincoln Center building committee, under such relentless public pressure, instructed the architect to increase the seat-count by whatever means necessary. Because of Abramovitz’ insistence that there be no direct contact between Bolt, Beranek, and Newman (BBN) and the building committee, the acoustician never had the opportunity to debate this change.
To accommodate the increased seating the downward slope of the side balconies were sharply increased by the architect so that the two ends where actually located at different floor levels. This change effectively rendered the underside of the side balconies useless as an essential reflector of early lateral energy to the main floor. This change was only discovered by accident at BBN, when co-consultant Russell Johnson viewed some drawings sent to the mechanical team. Surely, for such a severe architectural change, the drawings would have been sent to BBN for review of the room acoustics implications. They were not.
Abramowitz and his team worked for over a year on how to make the canopy and ceiling panels look good. In the end he finally decided that the ceiling panels should extend the full length of the hall. Beranek, with the experience of designing the successful canopy at Tanglewood behind him, felt this would not work, but reluctantly agreed to the change as long as the canopy elements could be pulled up to the ceiling if necessary. Abramowitz, however, fearing the ceiling panels might bang together in an earthquake, decided, without consulting the acoustician, that the ceiling panels should be welded together as in a huge raft, rendering them un-adjustable. In addition, the contractor misread the drawings and welded the first row of fixed ceiling panels six feet lower than specified in the drawings. The still-adjustable canopy over the orchestra consequently could not be raised to a higher setting as the orchestra requested (revisit last Monday’s Adaptistration discussion) for visual reasons.
Diffusion is an essential element in concert hall design, breaking up and scattering sound in many directions, to improve distribution and remove high frequency harshness. Abramowitz intended to include the surface irregularities designed by BBN, but was overruled by the building committee faced with a large cost overrun in construction. An interior designer was hired over the objections of both architect and acoustician to hide the deletion of these surfaces. Ultimately the “solution” was to paint the walls blue and illuminate them with blue floodlights!
BBN embarked on an extensive study of the halls deficiencies using model studies and developed several solutions that would have corrected many of the hall’s most glaring deficiencies. These solutions were approved by the Board. However, after an unfortunate lunchtime meeting with Maestro George Szell (a vituperative enemy of BBN),William Schuman, then president of Lincoln Center, set aside the Board’s approval of the BBN recommendations, and appointed an advisory committee to propose alternative solutions. These solutions were implemented…and made the hall worse than before!
There have been subsequent renovations of what is now known as Avery Fisher Hall, some of them making aspects of the room worse (lowering the hard ceiling over the stage platform), more of them making it a little better. But aside from the last major renovation with acoustical design by Cyril Harris, the approach has often been piecemeal touch-ups, not really challenging basic early assumptions of how this room should work.
Some of the impediments to progress may be political as well as financial. Unlike the Metropolitan Opera, which owns its building, Lincoln Center owns Avery Fisher Hall and the New York Philharmonic is only its principal tenant. (A decision by the Philharmonic made at the very beginning of Lincoln Center that some may regret today.)
With today’s improved knowledge, there is no physical reason why a shoebox concert hall of the basic interior dimensions and volume of Avery Fisher Hall shouldn’t possess magnificent acoustics. Some of the problem lies in the presence of the third level of the side balconies, which both reflects too much early energy to the floor and effectively destroys a potentially resonant “hard-cap” in the room. In 2002, working on the room acoustic design for the Nashville Symphony in a 1:20 physical scale model, I experimented with an addition of a third sidewall shelf. No real change in volume or absorption, just geometry. The measured reverberation time dropped by almost 30% at mid-frequencies.
But one doesn’t have to rely on models. Some of the difference can be heard at full scale. In the 1990’s the Kennedy Center Concert Hall, another Cyril Harris room similar in geometry to Avery Fisher, underwent a $12 million renovation in which the front half of the third level side tiers was removed. The increase in reverberation and envelopment experienced in the hall caused by simply eliminating a reflection, some audience absorption, and creating an effective hard-cap was pronounced.
Another problem in the design of Avery Fisher Hall is the limited volume of the stage enclosure surrounding the orchestra platform. Sound levels on this stage are much higher than they should be for the orchestra to hear itself and the room response well (“forward masking” again rearing its ugly head). Experiments have been made in recent years with Mostly Mozart concerts of moving the orchestra farther out into the room. The result creates less masking and better balance of early to late energy for both orchestra and audience. This approach should make even more of an auditory improvement with larger orchestras and is worthy of additional study.
There are additional improvements that might be considered, such narrowing the room at the floor level and upgrading the wall diffusion, but I sense eyes glazing over and will stop here.
12 thoughts on “Orchestral Acoustics 101: Avery Fisher Hall”
My eyes are not glazing over at all. This is fantastic, thank you. I’ll stop blaming BBN, given that the terrible process kept them from doing their job!
I have always been amazed at how inept high-powered architects can be. If anyone out has ever dealt with one, they’ll know that they can be harder to deal with than opera singers, Hollywood starlets, and politicians combined. The only solution is – as soon as they start getting difficult – fire them immediately and call a good carpenter.
Hall acoustics for classical music requires the use of psychoacoustics. Waves do not actually reach the brain at all but are converted into pulse trains by the inner ear. These pulse trains arrive at the brain through the nervous system and they enter a 30 millisecond delay line where they are examined for patterns. This means that the brain cannot perceive events happening faster than about 33 per second. Above this frequency, the brain finds only periodic sounds. You can give the brain a richer musical experience by presenting it with more pattern information derived from delayed arrivals of the original sound, but not delayed more than about 30 milliseconds. Longer than that and the delays become separate events: echoes. The key to hall design is to have many surfaces that will provide delayed sound under 30 ms. This means you must have surfaces that reflect sound that are no more than about 18 feet farther from each listener than the original sound source is. This is why absolute hall dimensions are so critical. When path lengths exceed this limit the sound becomes distant because the delays are heard as echoes.
An understanding of psychoacoustics is certainly important place in the study of architectural acoustics. (It was part of the curriculum I taught at MIT.) However, one must be careful not to place too much emphasis on one auditory mechanism. I have spent a significant portion of my professional career over the past couple of decades working with orchestras to correct the acoustics of rooms with “many surfaces” providing early energy as described above. A little early energy goes a long way…too much and problems arise (see first article in Orchestral Acoustics 101 series).
There are a number of experiments one can try with a problem Hall. The first would be to set up your chamber orchestra toward the back of the stage with no player more than 15 feet from the back wall. This way the back wall is able to contribute early reflections under 30 milliseconds. By delaying the direct sound, the time of arrival ratio of direct to reflected sound in improved and the side walls of the stage are able to project more early reflections, as well.
A second experiment is to set up your chamber orchestra on the front edge of the stage and surround it with Wenger choral shell modules with the angled overhanging panels. These will provide a surprising amount of early reflections.
Eventually, it may seem necessary to build a trapezoidal adjustable orchestral shell on the stage with a 10-foot extending roof. This is necessary because the ceiling of Avery Fisher hall is too high. The shell should be adjustable in order to accomodate different sizes of ensembles.
The problem of insufficient orchestral bass is usually solved by having an acoustical stage: a lightly built surface over a resonating chamber with grilled windows on the apron. The model for this design is Carnegie Hall!
It is my opinion that the ceiling elements at Philharmonic Hall and its later incarnation as Avery Fisher Hall are not too high, but, rather, too low.
Low ceilings increase the amount of early energy onstage, increasing loudness on the platform, and because of masking, making it more difficult for an orchestra to hear itself and the supporting return from the hall.
Two of the most esteemed halls in the world, the Musikverienssaal in Vienna, and the Concertgebouw in Amsterdam, both have ceilings which are more than 55 feet above the platform. Boston Symphony Hall’s ceiling is even higher, but the orchestra is located at the end of this room where there is a lower sloped ceiling approximately 40 feet above the players.
The ceiling at Avery Fisher Hall over the stage is much lower than this.
Modifications have been made to try to make hearing better for the musicians, with limited success. Recent seating experiments for Mostly Mozart concerts have moved the orchestra entirely out of the stage enclosure with a very open suspended canopy above them. The results are promising, and indicate a direction for part of the solution for this troublesome room.
A responsive platform surface is very important for support and projection of energy of the cellos and basses. The worst thing one can do on a stage is to apply wood directly to concrete (as I have been told was done at the Concert Hall of the recently opened National Grand Theater in Beijing).
There are instances where a resonance chamber under the platform has been tried (Boettcher Concert Hall in Denver, Sala Nezahuacoyotl in Mexico City) but to my knowledge this is not the case at Carnegie Hall. The grilles at the front of the stage there are for air return, not for acoustical reasons.
Years ago, Jaffe Associates designed Boettcher Concert Hall in Denver and they also designed a stage for us at UNC. They built an acoustical stage and the acoustician told me that the idea came from Carnegie Hall. More recently, after a redesign, reports circulated that the stage at Carnegie Hall had lost some of its bass reinforcing ability and I wondered if they had added concrete or something. Our stage really boosts basses and cellos but if you place a guitar amp on it, it boosts the bass from that, too.
Regarding ceiling height, I understand that a high ceiling is necessary in order to achieve a sufficient interior volume because, all other things remaining the same, reverberation time increases with interior volume. On the other hand, there need to be closer surfaces to produce early reflections. In a wide hall, the side walls don’t perform this task well, so it must be achieved from the ceiling. Thus, the standard solution is to install clouds or some other type of projecting panels much lower than the ceiling, just beyond the stage. Our Monfort Concert Hall in Greeley, CO. has a large hardwood construction that looks like a vast Venetian blind curving up from the upper edge of the shell to the ceiling many feet out. It seems to work well.
Leo Beranek talks about the need for early reflections (he says under 20 millisconds) in order to achieve what he calls “intimacy.”
It seems to me that these are standard, well-known solutions. Why they were not employed in Avery Fisher Hall remains a mystery to me.
Lack of early energy (a condition which would indeed result in lack of clarity and “intimacy”)is not the problem evident at Avery Fisher Hall. Too much early energy (contributing to harshness and deficient reverberation)is.
Adding a lower canopy in this room would make the problem worse, not better.
Sometimes “standard, well-known solutions” don’t fit the problem.
Acoustical musicians playing without amplification produce a finite and limited amount of acoustical power. A symphony orchestra playing FFF might produce about 100 watts.
If the sound level is too high in one place (on stage) it is probably too low in another place (the audience) and it needs to be distributed. This is generally done with reflecting surfaces – flat or convex panels located close to the stage. I would say that Avery Fisher Hall in great need of a judicious number of sound distributing and diffusing panels.