How we voiced the temple

Sound zones, a fragment of the scheme

Temples, synagogues, churches, churches, mosques and other objects that are very specific in terms of acoustics are very nontrivial.

The big time of reverberation , the inability to change the architectural acoustics (and sometimes even install microphones), the need to take into account cultural and historical features, and learning how to use this system of personnel without technical education is not a complete list of the reasons for the difficulties of such a project.
')
The task of dubbing the cathedral was not quite normal for us - we had not been engaged in such projects before. The drama of the situation was given by the fact that the consecration of the cathedral was to take place in a few months.

When we arrived at the scene for the first time, the morning service was on. From the pulpit, the minister read the prayers. Within a radius of 4-5 meters, we dealt with the words, but as soon as we moved farther from the priest, all sounds merged into an even hum due to the endless transition process. It was clear that with a large crowd of people the speech would be completely indistinguishable.

Our task was to solve this problem. It was necessary to ensure sound amplification, speech transmission and at the same time minimize the influence of the gigantic volume of a concrete room on the forcing vibrations of acoustics. The main problem was that any sound from the point where the priest stood, repeatedly reflected from the walls, floor and ceiling and sounded for about 10 more seconds. That is, the ratio of signal power (speech) and generated noise due to the reflections was extremely low.

The task was complicated by the zoning of the temple square: “The children are baptized over there, seminars are held here, and your system should not disturb them,” the priests explained to us. “But when services come, everyone should hear what we read and on the street too . ”

While working on the altar dubbing solution, we had a long thought about how to place the microphones. “Stretching over the throne? Or can bring the cable from the bottom? Well, or at least on the silver candlestick ... Microphones cannot fly — there are no miracles, ” I voiced the options. "There are, my son ," snapped the priest.

Task

Why did the task arise at all? The fact is that the canons of church architecture and, consequently, acoustics have been observed for centuries. Long reverberation time was decorated with long notes of organ, church choir. Almost all prayers have a rhythm that is supported by sound reflections, and the long notes must also be decorated with reverb. However, it is worth noting that the monotonous (characteristic of the ROC) recitative in Old Slavonic from the point of view of acoustics is not very discriminating even in an anechoic chamber.

The sound of religious buildings has always differed from the design of warning systems for stations, airports and other similar objects. What matters is not so much the intelligibility of speech (STI index), but the uniform sound field over the entire area of ​​the parish.

Ideally, the architecture of the structure itself is such that, despite the acoustic unevenness of the surfaces, the temple is a connected rectangular and cylindrical resonators with their own modes of oscillation. Speaking in simpler language, the architecture and materials of the temple should “raise” the minister’s voice up and direct it to each listener below so that the sound goes “volumetric” as if “from heaven”. But at the time of the start of the project at this site, the architecture not only did not solve this problem, but, on the contrary, prevented it from parsing the voice.

The building of the temple is a total connected cylindrical and rectangular resonators with their own frequencies. Such resonators are described by solutions of the Helmholtz equation containing Bessel functions. One of the problems of the novelties is that in tribute to the traditions, architects provide for the reduction of the Q-factor of cylindrical resonators (built-in clay pots under the dome), removing the “unnecessary” Bessel fashions of the neck and the dome from the spectrum, but they forget to design heating ducts, from the local heating plant. This leads to the fact that the additional connection between the “upper” and “lower” resonators disappears and the minister’s voice does not sound volumetric, but from the point from which he reads psalms. Given the size of the temple, the service is more like a local whispering. In addition, not everyone has strong voices. We had to take into account all these nuances.

And, of course, everyone wanted everything to be automatic. To even the button no one did not press. Otherwise, the whole effect will disappear: what kind of service, if the clergyman first presses a button or climbs into the config?

Design

To begin with, we made measurements and began to build the full model in EASE 4.3. This allowed us to carry out the basic calculations, but it is obvious that not a single model, not a single set of parameters will give an accurate idea of ​​the sound picture of an object. In addition, statistical calculation methods and geometric acoustics methods are insensitive to obstacles and wavelengths. Nevertheless, even without the use of wave techniques, the data obtained as a result of the calculation made it possible to arrange the acoustic systems in an optimal way, select the initial power, choose the radiation direction in such a way as to minimize late reflections and obtain a fairly uniform structure of early reflections.


Detail of the speaker layout

This is what allowed us to minimally change the temple’s own sound and at the same time create a uniform sound field. Since the model had more than 2000 surfaces, the calculation was carried out in a cluster of “clouds” of KROK, in a virtual machine with a capacity of 32 cores and 32 GB of RAM. Glory to professors Anert and Feistel for the opportunity to parallel the processes in EASE. Each calculation module AURA took only 9 hours.


Fragment of the presentation of the results of the module AURA

The most difficult task was to create a microphone system, and more convenient for use. And suitable for working with reverb more than 7 seconds. There was only one way out - linear microphones and radio systems (the microphone park in the ready solution, by the way, is updated and expanded, since the matrix switching system provides for this).

The range of tasks for dubbing was quite large. First of all, it was necessary to ensure the multi-zone structure of the complex in essence in a single space. As already mentioned, here and services are held, and seminars, weddings, etc. Separated sound zones can be created only with directional controlled radiation, taking into account the fact that the standard reverberation time (RT60) was about 8 seconds according to the measurement data. Therefore, the first decision that was made - we will sound with linear arrays. Of course, this is not a know-how, now almost all Catholic cathedrals are voiced that way. Previously, special air ducts, resonators in the dome and other solutions were used, that is, the architecture of the temple already had an “integrated” sound reinforcement system - but our facility was a “replica”, therefore electronics was required.

The choice of vendors is quite wide: Intellevox, HKAudio, Bose, MeyerSound, JBL (for fairness, I must say that Meyer and Harman had not released their CAL and CBT by that time) and so on. But, recalling the restrictions on price, delivery speed and quality, the choice was Bose.

We also made a multimedia system for the conference hall of the lower church, but this is a completely different story.


Fragment of the switching circuit

The square was divided into zones: the central Pritartarnaya, 4 zones behind the columns, the solea, the porch, and the outer zones outside the temple. The central zone is voiced by the stack solution - two-meter arrays (24 four-inch speakers) Bose MA12. Of course, in the near zone one can feel the presence of side petals of DN, but already at a distance of 2–3 meters one can enjoy a cylindrical wave (1 kg). This decision is justified by a large voicing area. The power supplied to the speaker systems is designed in such a way that the signal from the main sound columns is masked in time by the signal of the speaker systems behind the casing zones.

The enclosed zones are voiced by “homemade” arrays - four vertically mounted DS16SE speakers. This approach provided a directional gain of about 2. The distance between the centers of the speakers in these arrays was made larger than in MA12, which provided greater control of the directivity in the low-frequency region.

Problems with setting on the spot did not arise, because when modeling were pretty accurate. There were difficulties with feedback (microphone effect, something that does not take into account EASE), however, due to automatic filters and compensation of basic resonant modes, the effect was reduced to a minimum. In general, after designing and installing equipment, all that remains is to set up all the amplification channels, detect self-excitation frequencies, record several presets for different service scenarios, and teach the staff how to use the system.

At the time of publication of this topic, the system has long been commissioned and continues to work flawlessly.