The use of computer simulations to sort out acoustic and electro acoustic problems has grown rapidly during the last 10 years. Today there are several suppliers of these programs, but our experience is based on the use of EASE, which is the most commonly used program for these purposes.

Below we will describe a simulation done with the 900 years old cathedral in the city of Lund in southern Sweden.

The church was built from sandstone and to be the site for the first archbishop in the Nordic countries. The job was carried out by stone cutters from Germany and Italy. Who was in charge of the construction is not known.

 

The first steps to be taken in a computer simulation are to get hold of useful drawings of the building as well as the interior’s surface materials.

We received some “drawings” of the church but after a short evaluation it was clear that these drawings – from 1963 – were not very useful, as several measures were missing and not very accurate.

In order to do a correct job we had to go to the site and use a laser measurer to find out the real size and shape of the church. This was a useful and important step as some measures were up to a few meters wrong. Acoustics is not a millimetre science but definitely have to have less error than meters.

This job meant 1,5 days of measurements in the church in order to get everything as accurate as needed for a simulation.

It can be pointed out that measurement errors in absolute figures are depending on the size of the building. The larger the buildings the more tolerant are the absolute measures, when it comes to the acoustic results.

During the measurements a model was built in Google Sketchup. Sketchup is a very intuitive program for building complex models. As a major part of an old church consists of many arches and vaults, these structures are very time consuming to build in the computer. It is also of importance to get their geometry right as these surfaces usually focus the sound into “hot-spots”, which can create severe problems with the acoustics of the building. In a church such as the cathedral in Lund you will find many vaults and arches to work with before the computer model is finished. All this is quite obvious as a cathedral usually is large building and the structure had to be built in a safe way to the visitors, without the tools of modern structure calculations, computer program etc.

Now with a model imported into EASE, all interior surfaces have to be evaluated and given an absorption coefficient, which is corresponding to the material used. This is made automatically in cooperation between Sketchup and EASE, using different ”colors” in Sketchup, and then giving them a unique material in EASE.

An old church usually has quite long reverberation time, generally giving a low STI (Speech Transmission Index) value.

When a church such as the cathedral in Lund was built, all sermons were held in Latin so the STI-values did not matter as few persons understood Latin. Today when the sermons are held in the native language the visitors expect to both hear and understand what the content is in the sermon.

The few and most important absorbents in the church are the visitors themselves as their clothes are the major absorbents. However these absorbents vary a lot due to the different amount of visitors coming to the church. Therefore it is valuable to do the simulations with different amount of visitors.

The first values of importance are the RT60 values in the church. Here it is of importance which formula to be used, Sabine or Eyring or Schröder. Here the judgement and experience has to be used.

After we have found out the RT60 values we have to find out whether these values are correct or not for the building. The usual value tends to be too long for the building indicating more absorbents are needed to get an optimal value. However due to the age of these types of church buildings they are protected against interior changes, meaning the acoustic changes to the better are very limited if any.

 

Next step is to find out the best-suited loudspeakers for the building and their correct placements. This can often be a tricky job as there are positions not allowed for loudspeaker placements and usually few good placements available. And for each position a simulation should be done to check the coverage, STI and sound level values.

Here again a long practical experience will be very helpful, as the following parameters have to be considered:

• Correct direction of the loudspeaker sound in comparison to the source. We always try to have the loudspeakers located so the priest is easy to locate in the church. This job we consider as important and we are confident that the priest share this opinion.
• Correct placement of the speakers to get a complete coverage of all visitors of the church.
• Correct placement of the speakers to get as high STI-value as possible for as many visitors as possible.
• The loudspeakers placement also has to be in accordance with what is expected from an aesthetic point of view.

The last point above is usually one of the hardest to meet, and can take longest time to convince the decision makers.

However with the new loudspeaker technology involving steerable column speakers many improvements have been done such as:

• More flexible placement of the loudspeaker due to vertical steering
• Often more than one beam available
• Vertical beam coverage is flexible
• Internal DSP-power makes the loudspeaker more intelligent
• Easier cabling and more redundant system due to several inputs

The new system in Lund’s cathedral is installed and the selected speakers are Renkus-Heinz Iconyx systems of various sizes due to the work they are expected to do.

The image below shows a top view of the model