ºÚ¶´ÉçÇø

Like almost everything else on this stunning building, the shape of the debating chamber is unique. Located behind the transparent riverside facade, the chamber's spectacular glass walls will soar dramatically over the entire height of the building. Coiled around – and combined with – these glass walls is the access ramp, currently under construction, that spirals up from the ground floor auditorium to the roof, nine storeys above. The ramp will allow members of the public to see into the auditorium on their circular climb to the roof. The chamber itself is shaped like a wine-flask. From the bulbous base that will enclose the auditorium, the chamber's walls taper gently inwards and upwards to form the curving neck of the flask, before coming to an abrupt halt at the roof.

Before construction could commence, the chamber's design had to be finalised. Despite the obvious need for excellent acoustics in the space, Patel was concerned that the design team would struggle to comprehend the acoustic problems that the chamber's shape would create. "The acoustician is the one person nobody understands," he says disconsolately. He attributes this to the designer's inability to comprehend the black art of sound: "The rest of the design team is more used to dealing with visual images."

If he was to communicate with the rest of the design team, Patel knew he would have to speak their visual language. As luck would have it, he had spent the past three years at Arup developing the world's first acoustic visualisation computer program. To ensure that designers comprehend a room's acoustics, the program is based on a standard architectural modelling programme – 3D Studio Max – which is more commonly used to create photo-realistic renderings of buildings and to generate animated walkthroughs. The development of the debating chamber was the first time Patel had used the program as a design tool. The big question was: would it help the architects understand the impact of their design?

From the outset, the visualisation proved its worth. Early in the design process, while the architects were still trying to establish the chamber's basic form, Patel used a simplified version of the program to model the chamber's acoustic behaviour. Just by scanning a vertical cross-section into the program, and identifying the position in the auditorium the speaker would occupy, Patel was able to generate simple animations to show how the speaker's voice would travel around the room.

In the program, a circular ripple of arrows emanating from the speaker represents the sound of the speaker's voice and spreading out to fill the chamber – in the same way sound would travel. When the arrows collide with a wall, they are reflected back into the chamber, just like sound waves do when they collide with a solid object. By studying the animation the architects can easily see how sound is directed and where problems occur (see "How the visualisation works", page 60). "It meant we could demonstrate early in the design process the acoustic properties of the space," says Richard Hyams, project director for Foster and Partners.

The acoustician's aim is to ensure that all sound, direct and reflected, reaches the listener in the magic time of 50 milliseconds. Any longer and the speaker's voice appears to echo; any shorter and the room sounds dead and lacks atmosphere. Unfortunately, the computer animation showed that the chamber's smooth, curved walls were reflecting a speaker's voice back to them in more than 80 milliseconds. "The chamber's geometry focused the sound energy at the speaker, which would create an echo," says Patel.

The problem was caused by the vertical lower part of the enclosure's walls. The visualisation program proved invaluable in resolving this problem by enabling the consequences of a series of design modifications to be viewed almost instantaneously. "We were able to do in two days what would usually take two weeks," says Patel.

To direct the sound out of the flask, the designers altered the angles of the enclosure's lower walls to reflect the sound upwards and out of the space. This reduced the problem significantly, but it did not remove it. Further modifications were tested, and more and more of the wall sections were tilted upwards, so that in the end a vertical slice through the chamber showed a saw-tooth cross-section (above).

By now the final form of the chamber was starting to evolve and Patel and the designers moved from 2D animations to more sophisticated 3D ones. Using a version of the 3D Studio Max software, Patel was able to import sophisticated AutoCAD drawings into the program to model the chamber in three dimensions. Again, by selecting the position where the speaker would stand, the designers were able to flood the 3D space with sound arrows and watch an animation of their behaviour.

As the pressure mounted to finalise the chamber's design, the 3D representation of the chamber was being updated on an almost daily basis. During the day, Foster's team would incorporate the latest design revisions into the 3D model. This was then sent to Arup's New York office for overnight analysis and returned to Foster's office the next morning as an animation, together with a message saying "we need to tweak this or to modify that".

As the form of the chamber was being finalised, the spiral access ramp that had previously been situated outside the chamber came to be integrated into the walls, where it was used to support the glazing. Patel used this ramp to improve the chamber's acoustic properties still further by applying acoustic insulation to its underside in the model. The designers also added carpet to the chamber's floor to help absorb sound. The model is able to simulate the effect of sound absorption by giving these surfaces "stickiness". As the animation is run, sound arrows that land on the absorbent surfaces stick to them and remain attached.

However, "absorption costs money", says Patel. By analysing the animation, the designers were able to see which absorbent surfaces were most effective, to so to decide where to target the funding for the optimum result. "The benefit of the system and its movie output is that you get a great presentation which you can than present to the client or user," says Hyams.

With the chamber's design now finalised, the acoustic engineers were able to check the chamber's performance by taking a seat in a virtual chamber. By running the visualisation program, the acousticians were able to see the route the sound took to each seat. However, for this animation, rather than use arrows, Patel simulated the sound wave with coloured balls. He says that to have arrows raining down on you "is a bit disturbing".

The debating chamber's glazed enclosure is scheduled for completion in September but even then it will be some time before the auditorium is fully fitted out. Only then will the team be able to test the room's acoustics to find out how successful the program has been. For Patel, visualising the acoustics has made the task of communication much easier. Hyams agrees: "The visualisation was absolutely key to developing a chamber that is perfectly tuned," he says.