Mass Notification Systems

The National Fire Protection Association (NFPA) has continued to quantify its position on the intelligibility of voice systems used for Mass Notification.

Documents Regarding Intelligibility and Mass Notification
Rt. Click,
Save Target As
Mass Notification Systems Fire Protection Engineering
- Fall 2005
Mass Notification Systems
Speech Intelligibility Fire Protection Engineering
- Fall 2002
Speech Intelligibility
Mass Notification Systems - Design and O&M Department of Defense
- October 2005
Mass Notification Systems - Design and O&M
Minimum Antiterrorism Standards for Buildings Department of Defense
- July 2002
Minimum Antiterrorism Standards for Buildings

Return to the Intelligibility Instrument Page.

The following is excerpted from the 2006 edition of NFPA 72.

3.3.208* Voice Intelligibility. Audible voice information that is distinguishable and understandable. (SIG-NAS)

A.3.3.208 Voice Intelligibility. As used in this Code, intelligibility and intelligible are both applied to the description of voice communications systems intended to reproduce human speech. When a human being can clearly distinguish and understand human speech reproduced by such a system, the system is said to be intelligible. Satisfactory intelligibility requires adequate audibility and adequate clarity. Clarity is defined as freedom from distortion of all kinds (IEC 60849, Sound systems for emergency purposes, Section 3.6). The following are three kinds of distortion responsible for the reduction of speech clarity in an electroacoustic system:
(1) Amplitude distortion, due to non-linearity in electronic equipment and transducers.
(2) Frequency distortion, due to non-uniform frequency response of transducers and selective absorption of various frequencies in acoustic transmission.
(3) Time domain distortion, due to reflections and reverberation in the acoustic domain.
Of these three kinds of distortion, frequency distortion is partially, and time domain distortion is totally, a function of the environment in which the system is installed (size, shape, and surface characteristics of walls, floors, and ceilings) and the character and placement of the loudspeakers (transducers).* The purpose of the voice/alarm signaling service shall be to provide an automatic response to the receipt of a signal indicative of a fire emergency. Subsequent manual control capability of the transmission and audible reproduction of evacuation tone signals, alert tone signals, and intelligible voice directions on a selective and all-call basis, as determined by the authority having jurisdiction, shall also be required from the fire command center.

A. It is not the intention that emergency voice/alarm communications service be limited to English-speaking populations. Emergency messages should be provided in the language of the predominant building population. If there is a possibility of isolated groups that do not speak the predominant language, multilingual messages should be provided. It is expected that small groups of transients unfamiliar with the predominant language will be picked up in the traffic flow in the event of an emergency and are not likely to be in an isolated situation. If a fire alarm system is a component of a life safety network and it communicates data to other systems providing life safety functions or it receives data from such systems, the following shall apply:
(1) The path used for communicating data shall be monitored for integrity. This shall include monitoring the physical communication media and the ability to maintain intelligible communications.* Where required, emergency voice/alarm communications systems shall be capable of the reproduction of prerecorded, synthesized, or live (e.g., microphone, telephone handset, and radio) messages with voice intelligibility.
A. Voice intelligibility should be measured in accordance with the guidelines in Annex A of IEC 60849, Second Edition: 1998, Sound systems for emergency purposes. When tested in accordance with Annex B, Clause B1, of IEC 60849, the system should exceed the equivalent of a common intelligibility scale (CIS) score of 0.70. Intelligibility is achieved when the quantity Iav-, as specified in B3 of IEC 60849, exceeds this value. Iav is the arithmetical average of the measured intelligibility values on the CIS and (sigma) is the standard deviation of the results.
Objective means of determining intelligibility are found in Part 16 of IEC 60268, The objective rating of speech intelligibility by speech transmission index. Subject-based techniques for measuring intelligibility are defined by ANSI S3.2, Method for Measuring the Intelligibility of Speech Over Communications Systems. ANSI S3.2 should be considered an acceptable alternative to ISO TR 4870, where referenced in IEC 60268, Part 16, Second Edition: 1998, The objective rating of speech intelligibility by speech transmission index.
The designer of an intelligible voice/alarm system should possess skills sufficient to properly design a voice/alarm system for the occupancy to be protected. System designs for many smaller occupancies can be accomplished satisfactorily, if not optimally, based upon experience with the performance of other systems in similar spaces. For existing construction, relatively simple acoustic measurements combined with knowledge of the chosen loudspeakers performance characteristics can frequently result in satisfactory performance using mathematical formulas developed for the purpose.
For occupancies that do not yet exist, the designer should have an understanding of the acoustic characteristics of the architectural design, as well as the acoustic performance properties of available loudspeakers. Architecturally, this includes the physical size and shape of the space, as well as the acoustic properties of the walls, floors, ceilings, and interior furnishings. A proper design analysis can sometimes reveal that an intelligible system is not achievable unless some features of the architectural design are changed. The designer should be prepared to defend such conclusions and, if necessary, refuse to certify the installation of such a system. While hand calculationsand experience work well for simpler installations, more complex designs are frequently better and more cost-effectively analyzed using one of a number of readily available computer-based design programs.
The designer and the authority having jurisdiction should both be aware that the acoustic performance parameters of the chosen loudspeakers, as well as their placement in the structure, play a major role in determining how many devices are necessary for adequate intelligibility. The numerical count of devices for a given design and protected space cannot, by itself, be used to determine the adequacy of the design. Sometimes, the acoustic problems of certain placement constraints can be satisfactorily overcome through the careful selection of loudspeakers with the requisite performance characteristics, rather than by increasing their number.
There might be applications where not all spaces will require intelligible voice signaling. For example, in a residential occupancy such as an apartment, the authority having jurisdiction and the designer might agree to a system that achieves the required audibility throughout, but does not result in intelligible voice signaling in the bedrooms. The system would be sufficient to awaken and alert. However, intelligibility might not be achieved in the bedrooms with the doors closed and the sounder in the adjacent hallway or room. In some cases this can require that messages repeat a sufficient number of times to ensure that occupants can reach a location where the system is sufficiently intelligible to be understood. Systems that use tone signaling in some areas and voice signaling in other areas would not require voice intelligibility in those areas only covered by the tone.

A. Audio levels are commonly measured using units of decibels, or 1/10 Bell, abbreviated dB. When measured using a sound level meter, the operator can select either an A-weighted, B-weighted, or C-weighted measurement. The C-weighted measurement is nominally flat from 70 Hz to 4000 Hz, and the B-weighted measurement is nominally flat from 300 Hz to 4000 Hz. The A-weighted measurement filters the input signal to reduce the measurement sensitivity for frequencies to which the human ear is less sensitive and is relatively flat from 600 Hz to 7000 Hz. This results in a measurement that is weighted to simulate the segment of the audio spectrum that provides the most significant intelligibility components heard by the human ear. The units used for measurement are still dB, but the shorthand for specifying use of the A-weighted filter is typically dBA. The difference between any two sound levels measured on the same scale is always expressed in units of dB, not dBA.

From Ch 10, Table
Methods of verification of voice intelligibility should include, but not be limited to, any one of the following methods:
(1) Standard subject-based test methods such as described in ANSI S3.2, Method for Measuring the Intelligibility of Speech Over Communications Systems.
(2) Methods and instruments that measure certain physical parameters and provide a common intelligibility scale score such as described in IEC 60849, Sound systems for emergency purposes.
The use of test methods that provide a common intelligibility scale score may be used for existing systems but should not be used to require revisions to systems that were designed prior to the 2002 edition of this Code. Also, refer to Section 1.4.

Return to the Intelligibility Instrument Page.

top Home Pro Audio Page   Sales Tech Support     Check Price Place Secure Order