Even for seasoned pros, designing a voice evacuation system to meet current intelligibility requirements can be challenging due to a wide variety of independent factors that can influence the results. However, National Fire Protection Association (NFPA) requirements have been created to limit the complexity of these systems by minimizing the potential for over-design.
NFPA 72-2010 defines intelligibility as the quality or condition of being intelligible (3.3.124). It defines intelligible as capable of being understood; comprehensible, clear (3.3.126). Those familiar with the code will recognize that this is a slight change from NFPA 72-2007, which defined intelligibility as audible voice information that is distinguishable and understandable (3.3.211).
These definitions can be ambiguous. Therefore, it is important to know how to properly design for and predict intelligibility for each installation based on objective factors. Perhaps an easier way to look at intelligibility is as the measure of the effectiveness of speech, or the percentage of a message that is understood correctly.
Taming the Terminology
Before going any further, there are several key terms to be aware of in order to properly understand and apply intelligibility requirements.
Acoustically Distinguishable Space (ADS) is a term added to NFPA 72-2010. Establishing ADSs is foundational to planning an intelligible system. Speech Transmission Index (STI or STIPA) is the most common quantitative methodology for measuring intelligibility using a test meter. It is a weighted average of the response to fluctuating modulation frequencies. The Common Intelligibility Scale (CIS) was created to map all quantitative intelligibility measurement methods to the same scale so that all different results can be compared.
Emergency Communication System (ECS) is an NFPA term that refers to large, site-wide notification systems. Mass Notification System (MNS) is a military term used for the same types of systems. Intelligibility would relate to the voice evacuation system portion of a fire or ECS system.
Key Intelligibility Factors
While the properties of the speaker have some impact on the intelligibility of a system, most factors have to do with the occupancy itself.
Signal-to-Noise Ratio — SNR is a comparison of the sound level produced by the speaker to the ambient or background noise in the room. In order to help achieve the needed intelligibility, it is important to ensure the speaker sound output is 10 to 15dB over ambient noise. NPFA 72-2010, Chapter 18, calls for an average 15dB over ambient. Going any higher than this results in diminishing returns in terms of improving intelligibility. Therefore, if a better intelligibility score is needed, more speakers should be used at lower tap settings as opposed to increasing the sound output on the existing speakers.
As Figure 1 shows, when installing speakers, each time the power output or number of installed speakers is doubled, the sound output increases by 3dB. Each time the distance between the listener and sound source is doubled, there is a 6dB loss in loudness.
Frequency Response — For voice evacuation, speakers should ideally have a frequency range between 150 and 11,000Hz because this is the frequency range that an adult voice produces. UL requires and tests for a narrower frequency range, between 400 and 4,000Hz, because this is closer to the average range in which humans can hear sound efficiently.
When measuring frequency response, it is important for frequency to be as flat as possible to produce the most intelligible sound. The flatter the response over the frequency range, the better the speaker intelligibility will be.
Harmonic Distortion — The average person can detect as little as 2-percent distortion when listening to sound output. Once the sound output reaches 15-percent distortion, it is considered nonintelligible. UL allows up to 20-percent distortion over the range 710-3,550Hz. However, this would make for a poor sounding speaker. Therefore, the least amount of measurable distortion is desired.
There are many factors that affect harmonic distortion. These include tolerance of the message generator and amplifier, loading of the audio amplifiers (load vs. available power), or mechanical factors like wires touching the cone of the speaker, excessive voltage drop in the speaker line, vibration caused by poor installation or damaged speakers. In addition, all manufactured equipment has distortion built into it. All of these different factors build on each other and have a cumulative effect on intelligibility.