A model of design

• Features

• 27/11/2015

As computers keep getting more powerful, is there still a place for physical models when designing a avenue’s acoustics? Or has computer modeling relegated physical models to the dustbin of acoustical history? The answer depends on whom you ask.


Not surprisingly, the choice often comes down to cost, with large – say, 1/10-scale – physical models running €300,000 and up. That alone often tips the business case toward computer modeling. But software also has a significant up-front costs.

“The software is expensive,” says Darren Cheshier, CTS-D, an engineer at Conference Technologies, a U.S.-based integrator. “It is hard to convince system integrators/consultants to invest in this kind of software when an experienced acoustical engineer has designed successful spaces for years. After all, acoustical technology has not and will not change. New materials will come out, but the science behind it does not change.”

The cost of computer modeling also can vary by the venue’s complexity.

“We have three rates for acoustic modelling depending on the complexity and detail we deem necessary,” says Tony Stacey, senior acoustic consultant at London-based AMS Acoustics Ltd., which does only computer modeling. “For example, the cost of modelling a simple shoe-box-type space would be less than a complex atrium area.”

Regardless of whether a project uses computer or physical models – or, in some cases, both – there’s not much debate over the value of modeling for many applications.

“The need for modeling is determined by the uses for the model,” says Sam Berkow, founder of and principal consultant at SIA Acoustics, a New York-based design and consulting firm. “We strongly believe that modeling loudspeaker systems is a very effective way to predict performance and provide insight into critical elements of the design, such as angles of an array hang or time delays in a directional sub-woofer array. In our firm, Steve Sockey uses computer modeling extremely effectively to determine the best ways to steer low frequency arrays, which results in much better sounding stages.”

The case for computers

One big, obvious advantage of computer modeling is that it’s all virtual, so it’s easier to accommodate new variables as the project moves along. Other advantages include:

 Flexibility. “You can do as much or as little detail as required,” says Conference Technologies’ Cheshier. “If an architect needs a starting point, the computer modeling software can still calculate without the major details.”

 Predictability. Computer models provide can better predictability than physical models or calculations alone. “I have used EASE on some of the most challenging acoustical spaces, and it works very well,” Cheshier says.

 Impact. Seeing is believing, but so is hearing. “It is a great sales tool and justification tool as to why acoustics are important within a space,” Cheshier says. “You can have them listen to a before and after scenario before the work is done.”

 Portability. It’s easy to burn the model to a CD or upload it to an FTP site, even if the architect or client is half a world away. “I can produce a CD with a file than anyone can view,” Cheshier says. “So even if your client is out of town, they can see and hear the results.”

But computer modeling has its share of disadvantages – such as up-front costs – and caveats. One caveat is the role of experience, both in acoustics and with the software. “There is a big learning curve,” Cheshier says. “I went to a three-day class, and that was a very intensive course. Even after that, I am still learning how to use the software.”

Others agree that the software’s power often hinges on the person using it.

“Experience is most helpful in two areas: understanding what to include in a model – i.e., which architectural details are important – and how to interpret the data,” says SIA Acoustics’ Berkow. “In our firm, computer models are generated by draftsman, who work with our consultants to refine the models as required. Experience in critical listening and developing an understanding of how sound behaves in rooms is critical to understanding the results of modeled calculations.”

Experience also can help avoid over- or underestimating, or just plain overlooking, certain factors.

“Experience is still the be all and end all in any acoustic modeling,” says AMS Acoustics’ Stacey. “Simply plugging in absorption coefficients taken from tables is never enough. You need to consider how the material is mounted and what the diffusion may be. You also need to consider what’s been missed out, [such as] the area next to and coupled to the area you’re modelling. It’s not unusual for us to say ‘I don’t believe that answer!’ and rerun the prediction after changing some parameters.”

Room for improvement

Although there’s no shortage of modeling software available, some consultants have created their own because commercial packages lacked certain features and tools.

“We surveyed the software market, but unfortunately we couldn’t find any useful software,” says Yasu Toyota, chief acoustician at Tokyo-based Nagata Acoustics. “So we decided to develop it ourselves.”

Like any other type of software, modeling applications continue to evolve and improve, with experience again playing a key role – this time in terms of adding features that some users couldn’t find in the marketplace.

“I developed SIA-Smaart, an acoustic measurement and sound system optimization tool, in large part to be able to measure and compare results from prediction programs I had worked on,” Berkow says. “While we are thrilled with many of the tools currently available, there are a number of areas where today’s models can be improved: Acoustical models can be improved by using extended source and receiver directionality and locations. Sound system models can certainly be improved by more effectively modeling the interaction of the room’s acoustical performance with the performance of a loudspeaker system.”

Getting physical

Even as computer modeling becomes more sophisticated, physical modeling still has its place. The main reason, ironically enough, is money: Although physical models are expensive, there’s a business case for them if that money can be recouped in the form of, for example, more donations when a symphony hall’s benefactors are wowed by a model that they can walk around or even poke their heads inside.

Physical models also can have a greater wow factor than computer images when they appear in a newspaper or on TV. At least that’s been the experience of Toyota and others.

“This may not be popular to say, but in a non-technical sense, physical models are often very effective as marketing and sales tools, where computer images often have less impact,” says SIA Acoustics’ Berkow.

In the case of the Kauffman Center for the Performing Arts in the U.S. city of Kansas City, the physical model’s ability to wow benefactors was a side benefit rather than its raison d'être.

“The choice to have a physical model built was based on scientific testing for accuracy, in conjunction with the acoustical modeling software,” says Jane Chu, the Kauffman Center’s president and CEO. “There is an added benefit in showing the model to benefactors; however, that is not why we built it.”

Another money-oriented reason is that physical models can ferret out problems that would cost far more to fix once construction is complete than the cost of the model itself. That’s one of the reasons why the Kauffman Center and its acoustic designer, Nagata Acoustics, opted to do both physical and computer modeling.

“The acoustical model is extensive: several hundred thousand dollars,” Chu says. “However, there have been times when new performance halls have opened, only to find out that the acoustics were compromised. This is embarrassing and costs millions of dollars to fix. Creating a scientifically precise acoustical model is operating under the ‘measure twice, cut once’ approach. Preparation before construction at the cost of several hundred thousand dollars is much cheaper than having a celebrated opening that ultimately costs several million dollars to fix afterwards, not to mention the negative exposure.”

Some consultants say that echoes are one example of how physical models can complement computer modeling.

“The biggest merit of a physical model is that we can use actual sound,” says Toyota, whose firm has used a combination of physical and computer models on the Kauffman Center and several recent European projects. “This is very, very helpful for us to find out things such as echoes, which are not predicted by computer models.”

Long lead times

Even so, physical models have their share of drawbacks, some of which aren’t obvious. For example, just as it takes time – typically several months – to build a physical model, so do changes made as a result of the initial tests. But the ability to remodel the model can be at the mercy of the environment, adding to the lead time before the next round of tests.

That was the case at the Kauffman Center. The initial tests took place in January 2007 and found that a few wall angels needed to be changed and that some absorptive materials needed to be added to mitigate echoes in some areas. But making those changes in the model wasn’t something that could be done in a week or two.

“Our builders spent four months making adjustments to the model,” Omni Models said in a Kauffman Center press release. “Kansas City’s rain and sticky June weather really made it difficult to get varnishes dry before testing could begin again.”

On the plus side, the local market also can cut the cost of physical models.

“It totally varies, depending on the market,” Toyota says. “For instance, when we built a 1/10-scale model in China, it was very cheap. The cost in the United States and Europe would be [higher].”