Acoustic Design of PAGA and PAVA Systems

15 October 2018 by James Ward, Senior Consultant

Public Address General Alarm and Voice Alarm (PAGA / PAVA) systems may have to convey vital safety information, often in hazardous work areas where lack of clarity could have catastrophic consequences. So why are so many systems barely audible and what is the key to achieving good speech intelligibility?

PAGA system loudspeaker on offshore platform
Figure 1: A PAGA system loudspeaker on a new offshore platform with acoustic design undertaken by Spectrum Acoustic Consultants

The Problem

You may be familiar with the scenario of being in a railway station where you hear an announcement but cannot understand what is being said. There seems to be plenty of loudspeakers, and the volume may even be high enough to be clearly audible above all of the other noise, yet the words may as well be in another language because you are unable to understand them. This is bad enough if it means you miss your train’s platform change, but imagine if you were instead on an oil platform and the message was safety critical. Fortunately, acoustic modelling at the design stage can drastically improve the speech intelligibility of PAGA/PAVA loudspeaker systems and is therefore within the capability of all installers with correct acoustic design.

The Physics

When you yodel in Switzerland and the sound comes back to you 2 seconds later from a reflection off the neighbouring mountain, it’s called an echo. But when you’re in a cave and you clap your hands, rather than hear one distinct echo you might just hear a mush of noise that gradually dies down. This is called reverberant noise and is comprised of hundreds of echoes of your clap, all reflecting off the cave walls again and again; you just can’t hear them individually because they are so close together, due to the small distance between the cave walls as compared to the more distant neighbouring mountain with its long yodel echo.

Echo is fun clapping and reverberation is great for classical music, but add significant amounts of either to speech and you will find that it becomes progressively less intelligible. Fortunately there is an established measurement parameter for speech intelligibility; the Speech Transmission index or STI, and it can not only be measured but also calculated, which means it can be used to optimise the design of PAGA/PAVA systems for all applications where speech intelligibility is important (e.g. offshore platforms, petrochemical plants, railway stations, sports stadiums, Airports and Shopping Centres).

Mathematically, the measure of speech intelligibility (STI) is dependent upon a number of factors, including (but not limited to):-

  • Background noise level
  • Signal to noise ratio (s/n ratio)
  • Reverberant noise level

As such, controlling these factors as far as possible enables the system acoustic designer to optimise speech intelligibility.

As an example, think of a modern lecture theatre specifically designed for the speaker to be clearly audible. It is likely to have a low background noise level, being well insulated from traffic or other external noise. It will also have a loudspeaker system generating a clearly audible noise level at all locations within the hall, resulting in a good signal to noise ratio. Finally, it is also likely to have acoustic ceiling, and possibly wall, tiles to control the reverberant noise level.

It is the management of these factors which enables the optimisation of speech intelligibility for a given PAGA/PAVA system in a particular acoustic environment.

The Acoustic Design Process for Good Speech Intelligibility

Establish Background Noise Levels

For a good PAGA/PAVA system design it is important to know the typical noise level in the areas where the system must be audible and this will usually be established by either undertaking a noise survey of the area or, if the plant is new or not operational, by calculation using a computer based noise prediction model.

Measuring noise levels on an offshore platform
Figure 2: Measuring noise levels on an offshore platform

Create Acoustic Model of Required Coverage Area and Loudspeaker Layout

Installers typically have an initial proposed layout for PAGA / PAVA loudspeaker locations and these will be used for a first run computer model simulation of the PAGA / PAVA system. It is important to represent the physical properties of the area (acoustic and geometric) so that the simulation can take into account the effects of acoustic reflections, screening and absorption. The software then calculates the speech intelligibility of a simulated test signal fed through the system, taking into account the acoustic properties of the surroundings as well as the established background noise levels.

3D noise model of petrochemical plant showing PAGA system loudspeaker locations
Figure 3: 3D noise model of Petrochemical Plant showing PAGA system loudspeaker locations

Adjust Loudspeaker number, power tappings, locations and orientations to optimise Signal to Noise Ratio

Optimising the speech intelligibility of the system (measured as the Speech Transmission Index or STI) is not simply a case of turning the volume up, since this can simply result in a very loud but still unclear signal. Indeed, it may be necessary to increase the number of loudspeakers but to have them set at a lower power tapping so that you will be close enough to hear one loudspeaker clearly at any location within the required coverage area. Simulation software can demonstrate the improvements made, or otherwise, when alterations are made and can show with contours the values of STI achieved throughout the required coverage area.

Computer simulation plot of PAGA sound transmission index on an offshore oil platform
Figure 4: Computer simulation plot of PAGA/PAVA Sound Transmission Index (STI) on an offshore oil platform (higher values mean better speech intelligibility)

Control Reverberant Noise Level

In enclosed spaces with hard acoustically reflective surfaces there may be a need to reduce the strength of the reverberant sound field to further improve speech intelligibility. Think again of the cave, of large old enclosed railway stations or of underground tube train platforms and you will be familiar with the effects these hard surfaces have on noise and the audibility of loudspeaker announcements.

However, if you have passed through a modern railway station with perforated steel on the walls and/or ceiling areas then you may also have noticed that they were more acoustically ‘dead’, and hopefully announcements were more intelligible. If this was the case then it is likely that there was acoustically absorptive material behind those perforations, designed to reduce the level of reverberant noise and so improve speech intelligibility in the area. This type of treatment is incorporated in multiple ways inside many types of building, and may take the form of soft wall finishes, suspended absorbers or acoustic ceiling tiles. The only difference in industrial applications is that the soft acoustically absorbent material is usually protected by a steel outer layer, with perforations to allow the sound to pass through it. This type of acoustic wall panel is often incorporated into the acoustic design of internal plant rooms on offshore oil platforms, or onshore refineries, in order to reduce the reverberant noise level. For existing plants, absorption can be improved by installing proprietary industrial acoustic panels.

Industrial Profiled Steel Wall Panelling
Figure 5: Industrial profiled steel wall panelling showing perforated area which protects soft acoustically absorbent material enclosed behind it

For external areas, the degree of acoustic reflection is typically much less significant so that the acoustic treatment of reflective surfaces is seldom required.

Measure Installed System Speech Transmission Index (STI) during Commissioning

If required, the STI of an installed system can be measured using appropriate instrumentation. This is commonly undertaken following the STIPA method (STI for PA systems) by transmitting a specialist sound test signal through the loudspeaker system and measuring it at representative receiver locations with a noise meter equipped to decode and evaluate the received signal. This is then expressed as a measure of the speech intelligibility of the system (STI or STIPA).

Conclusions

Many PAGA/PAVA systems continue to be installed without making use of established engineering design methods for optimising their intelligibility.

This article has aimed to highlight the importance of good speech intelligibility and to show how good performance can be built-in using established design and measurement parameters along with standard acoustic engineering principles.

Key Standards relating to PAGA/VA system design and performance are listed below:

BS EN/IEC 60849 Sound systems for emergency purposes
BS 5839 Fire detection and fire alarm systems for buildings
BS EN 60268 Sound system equipment
NORSOK T-100

Acoustic Design of PAGA and PAVA Systems