Skip to content
Upcoming courses

Voice alarm system

Tools
From Fire Safety Academy Wiki
Revision as of 18:14, 2 November 2026 by Lisa (talk | contribs) (Add Intermediate SVG figures with Fire Safety Academy watermark at planned locations; preserve article text)
(diff) ← Older revision | Latest revision (diff) | Newer revision → (diff)
Contents
  1. 1How a voice alarm system works
  2. 2Why spoken messages are more effective
  3. 3Effect on pre-movement time in design
  4. 4Phased evacuation and zoned messages
  5. 5Audibility and intelligibility
  6. 6Reliability and lessons from the World Trade Center
  7. 7See also
  8. 8Notes
  9. 9References

A voice alarm system is an evacuation alarm system that warns the occupants of a building with spoken messages through loudspeakers. The messages replace or accompany tones such as bells, horns or a slow whoop.[1][2] The messages can be prerecorded or spoken live by staff or the fire service from a control point.[3][4] Larger systems divide the building into speaker zones, so that different floors can receive different messages.[5][6] In the sources the same function is also called an emergency voice communication system, a public address system or an emergency warning and intercommunication system.[4][6]

Spoken messages address a known problem: people often hesitate when they hear a plain alarm signal.[7] Research summarised in the SFPE Handbook shows that occupants respond faster to clear spoken instructions than to bells or sounders, and fastest to informative live messages.[8][3] The Introduction to Fire Safety Engineering likewise states that a slow whoop is less effective than the spoken word.[1] Design methods reflect this: in New Zealand, C/VM2 (second edition, 2025) halves the pre-travel activity time for some occupancies when a voice alarm signal is used.[9] Voice systems also make phased evacuation possible, because occupants who stay in the building can be told what to do.[10] The sources stress that messages must be audible and intelligible, and that they lose credibility when they conflict with what occupants see.[3][8]

How a voice alarm system works

A voice alarm system is one part of a fire alarm system, whose basic purpose is to detect fires, notify occupants and summon emergency responders.[4] The fire alerting part consists of audible and visual signals activated by the fire detection system.[2] The audible devices are usually horns, bells, chimes or speakers, and the visual indicators are usually strobe lights.[2] In a voice alarm system the audible devices are loudspeakers that can carry both tones and speech.[4][6]

The Australian verification method data sheets describe a range of emergency warning systems.[6] At the simple end, a single-zone system uses one tone generator and an amplifier to give alert and evacuation tones only.[6] More complex systems add public address capability and visual alarms.[6] Multi-zoned evacuation systems are used where a building is divided into zones for phased evacuation.[6] Each zone is then controlled separately and has its own speakers for tones and public address, and warden intercom points connect each zone with the chief warden.[6][11]

The towers of the World Trade Center in New York had such a system. Its fire command station gave manual control of the emergency voice and alarm communication, by individual floor, by a group of floors or throughout the building.[4] The alarm tones were silenced while the loudspeakers were in use and restarted automatically afterwards.[4] Firefighters could make announcements over the same system, and floor wardens had a two-way telephone link to the command station.[4] The earlier public address system had two sound channels, so that simultaneous and independent announcements could be made to selected speaker zones.[5]

Phased evacuation systems commonly use two tones.[12] An alert tone tells occupants to prepare for evacuation and to await further instructions from wardens, security staff or speaker announcements.[12][13] An evacuation tone means that the building must be left immediately.[13] In a two-stage system, a building-wide alert can sound first while the evacuation tone sounds on the fire floor, and the system then cascades through the other floors.[14]

Section of a multi-storey building: the fire floor hears an evacuation message, other floors an alert tone, all controlled from a command station.
A zoned voice alarm system in a phased evacuation: evacuation signal and spoken instructions on the fire floor, alert tone on the other floors, controlled from the fire command station. Based on Australian Building Codes Board (2020a), pp. 97 and 243 to 254, and Keough and Grill (2005), p. 44.

Why spoken messages are more effective

How people respond to a fire is a central question of human behaviour in fire. According to the Introduction to Fire Safety Engineering, the reaction depends on a person's role, experience, culture, training and personality, on the perceived threat and on the actions of others.[1] It also depends on the signals perceived, such as the evacuation alarm, smoke and heat.[1] The means of alarm largely determines whether people take the signal seriously.[1]

Plain alarm signals have a weak record. A review of research in the United Kingdom, quoted by Bryan, found that the response to fire alarm bells and sounders tends to be less than optimal.[7] People are often unsure whether the noise is a fire alarm at all, or merely a test or a drill.[7] Proulx adds that occupants take longest to respond to the fire alarm signal if they have had no training or if the alarm often sounds for reasons other than fire.[3] She regards three or more false alarms within six months as a large number.[3]

Spoken messages give occupants information that a tone cannot give. The quality, pitch and volume of a voice, and the content of the message, may reinforce the perceived threat.[7] In evacuation drills in an underground station, passengers did not move at all after the fire alarm signal.[15] They started to move only 15 s after precise live messages that gave the type and location of the incident and told them what to do.[15] Proulx concludes that messages through a voice communication system or directly from staff are the signals that occupants take most seriously.[15]

The type of message also matters. Proulx considers informative live messages probably the best way to start a fast response, while prerecorded messages are rarely as effective and contain less information.[3] The Introduction to Fire Safety Engineering agrees that live announcements, whether through a public address system, in person or by telephone, are followed faster than automatic alarms.[1] A well-trained evacuation organisation in the building makes it many times more likely that an automatic alarm is taken seriously.[1]

A message must also agree with what people see. If a spoken instruction conflicts with other cues, such as the smell or sight of smoke, its credibility may be questioned and the information ignored.[8] In a fire in the South Tower of the World Trade Center on 17 April 1975, smoke spread to the 9th to 22nd floors.[8] At 9:10 a.m. a message advised occupants in the core lobbies to stay calm and return to their offices.[8] They stayed in the lobbies and grew more concerned, and an evacuation message followed at 9:16 a.m.[8]

Effect on pre-movement time in design

The main design benefit of a voice alarm is a shorter pre-movement time, the time between the warning and the start of purposeful movement to a place of safety.[1] In New Zealand, C/VM2 (second edition, 2025) adds this time to the detection time, the notification time and the travel time to give the required safe egress time.[16] For standard evacuation strategies the notification time is taken as 30 s.[17]

C/VM2 gives fixed pre-travel activity times that depend on the occupants and, for some building types, on the alarm signal.[9] For retail shops, exhibition spaces and restaurants, it gives 60 s in the enclosure of origin and 120 s elsewhere with a standard alarm signal.[9] With a voice alarm signal these become 30 s and 60 s.[9] For hotels and motels, the time for occupants remote from the fire falls from 600 s with a standard signal to 300 s with a voice signal.[9] For offices, where occupants are awake and familiar with the building, the same values apply to any alarm signal.[9] C/VM2 lists the type of warning system among the factors that these values are meant to cover.[18]

Grouped bar chart of pre-travel times for retail and hotel occupants with standard and voice alarm signals; voice bars are half as long except one.
Pre-travel activity times in C/VM2 (second edition, 2025) with a standard alarm signal and with a voice alarm signal. Based on Ministry of Business, Innovation and Employment (2025b), p. 30.

Proulx reproduces a similar table from the British Draft for Development DD240 (1997).[19] It sets three levels of warning system.[19] Level W1 is live directives through a voice communication system from a control room with closed-circuit television, or live directives with well-trained uniformed staff.[19] Level W2 is non-directive prerecorded voice messages or an informative visual display, with trained staff, and level W3 is a fire alarm signal with untrained staff.[19] For shops and other assembly buildings, the estimated delay before evacuation starts is less than 2 minutes with W1, 3 minutes with W2 and more than 6 minutes with W3.[19]

As a worked example, consider occupants of a retail shop who are remote from the enclosure of origin, under C/VM2.[9] With a standard alarm signal the pre-travel activity time is 120 s; with a voice alarm signal it is 60 s. If detection, notification and travel times are unchanged, the voice alarm therefore shortens the required safe egress time for these occupants by 60 s. For remote occupants of a hotel the same change saves 300 s, half of the 600 s that C/VM2 assigns with a standard signal.[9] The International Fire Engineering Guidelines (2005) note that there is no validated method for calculating pre-movement times and refer designers to the data in the SFPE Handbook.[20]

Two time bars of detection, notification, pre-travel and travel; the voice alarm bar has a pre-travel segment of 60 s instead of 120 s.
Required safe egress time for remote occupants of a retail shop with a standard alarm and with a voice alarm: only the pre-travel segment changes, from 120 s to 60 s. Based on Ministry of Business, Innovation and Employment (2025b), pp. 26, 29 and 30.

Phased evacuation and zoned messages

Voice communication systems were developed in response to phased evacuation.[10] In a phased evacuation, some occupants are relocated within the building or held where they are, and they need instructions until they are told to leave.[10] Requirements for voice systems first appeared in US national standards in the mid-1980s, for example in the 1985 edition of NFPA 72F.[10] In the same period New York City adopted Local Law 16 of 1984.[10]

The International Fire Engineering Guidelines describe the approach for tall buildings.[21] In high-rise buildings with suitable emergency warning and intercommunication systems, trained personnel may manage the evacuation.[21] Occupants on floors furthest from the fire are first placed on alert and evacuated progressively, only if the fire continues to develop.[21] The SFPE Handbook adds that stay-in-place and partial evacuation strategies depend on giving occupants information about the location and extent of the fire.[22]

Zoned messages need careful timing in the analysis. Where phased evacuation applies, the alarm reaches some areas later than others, and these delays must be included in the notification time.[12] Delays for alarm investigation and verification must also be included, because they can have a significant effect on the required safe egress time.[12]

Audibility and intelligibility

A voice message is useless if it is not heard and understood.[2] The SFPE Handbook gives sound levels that alarm signals need to reach. A level of 55 to 70 dBA is reported to wake a college-age person with normal hearing.[2] In the US, NFPA 72 requires signals 15 dBA above ambient in areas where people may sleep.[2] British standards require 65 dBA, or 5 dBA above ambient noise, where people are awake, and 75 dBA at the bedhead where people sleep.[2]

To estimate these levels, the handbook starts from the sound power level of a sounder. This is the acoustic power that the device radiates, expressed in decibels. It follows from the output that the manufacturer states at a given distance, using a relation of Butler, Bowyer and Kew:[23]

LW=L+20log10r+11dB

Symbol Description Unit
LW sound power level of the horn, bell, speaker or other sounder, referenced to 10−12 W dB
L output stated by the manufacturer at distance r dBA
r distance at which L is stated m

The handbook uses this relation for sounders in corridors, before corrections for the number of directions in which sound spreads, the wall and floor finishes, the distance and the sound insulation of walls and doors.[23] As a worked example, a typical compression-driver fire alarm speaker powered at 2 W gives 94 dBA at 3.05 m.[23] Its sound power level is then 94 + 20 × log10(3.05) + 11 ≈ 115 dB.[23] Doubling the power to 4 W raises the sound power level by only 3 dB, because decibels are logarithmic.[24]

Louder is not always better for speech. The handbook warns that, if the fire alarm system is also used for voice communication, a speaker tapped at 4 W in a small corridor might sound distorted and be unintelligible.[24] Proulx therefore advises assessing both audibility and intelligibility, so that occupants actually perceive the information.[3] At the World Trade Center, the Port Authority analysed speaker intelligibility, meaning the clarity of the message, and audibility, meaning whether it is loud enough to be heard.[25] The aim was a guideline for speaker locations that would ensure acceptable levels of both.[25] Speakers were to be placed so that they would be heard clearly above the ambient noise.[26]

Reliability and lessons from the World Trade Center

The Australian data sheets list a default reliability of 0.9 for voice alarms, taken from the work of Zhang and colleagues, the same value as for central alarms.[27] A local alarm has a value of 0.75 in that list.[27] In the World Trade Center towers, the loudspeaker system was required to remain 50% operable throughout the building if the other 50% failed.[26]

On 11 September 2001, the fire alarm systems of the towers worked, but not all functions performed as intended.[28] The alarm signal had to be activated manually, and in WTC 1 it was delayed by 12 minutes after the aircraft impact.[29] There was no confirmation that the emergency voice communication system worked on any floor of WTC 1 after the impact, although alarm tones were heard below the impact floors.[30] In WTC 2, voice messages and alarm tones were heard above the impact floors.[30] Even with multiple communication risers, the systems degraded, and in WTC 1 all notification and communication functions above the impact floors appear to have been lost.[29]

The content of the messages mattered as much as the hardware. About three minutes before WTC 2 was hit, a building-wide announcement told its occupants that the fire was in WTC 1, that WTC 2 was secure and that they should return to their offices.[31] Fifty-four percent of the occupants of WTC 2 reported receiving information from the public address system.[32] Some people followed the instruction and returned to their offices, while many others continued to evacuate.[32] A second announcement at 9:02 a.m., one minute before the impact, said that occupants could start an orderly evacuation if conditions on their floor warranted it.[32]

See also

Notes

  1. ↑ 1.0 1.1 1.2 1.3 1.4 1.5 1.6 1.7 Van Rossum (2027), §7.3.
  2. ↑ 2.0 2.1 2.2 2.3 2.4 2.5 2.6 Schifiliti et al. (2002), p. 4-32.
  3. ↑ 3.0 3.1 3.2 3.3 3.4 3.5 3.6 Proulx (2002), p. 3-350.
  4. ↑ 4.0 4.1 4.2 4.3 4.4 4.5 4.6 Keough and Grill (2005), p. 44.
  5. ↑ 5.0 5.1 Keough and Grill (2005), p. 27.
  6. ↑ 6.0 6.1 6.2 6.3 6.4 6.5 6.6 6.7 Australian Building Codes Board (2020a), p. 97.
  7. ↑ 7.0 7.1 7.2 7.3 Bryan (2002), p. 3-315.
  8. ↑ 8.0 8.1 8.2 8.3 8.4 8.5 Bryan (2002), p. 3-316.
  9. ↑ 9.0 9.1 9.2 9.3 9.4 9.5 9.6 9.7 Ministry of Business, Innovation and Employment (2025b), p. 30.
  10. ↑ 10.0 10.1 10.2 10.3 10.4 Lew et al. (2005), p. 150.
  11. ↑ Australian Building Codes Board (2020a), p. 98.
  12. ↑ 12.0 12.1 12.2 12.3 Australian Building Codes Board (2020a), p. 243.
  13. ↑ 13.0 13.1 Australian Building Codes Board (2020a), p. 244.
  14. ↑ Australian Building Codes Board (2020a), p. 254.
  15. ↑ 15.0 15.1 15.2 Proulx (2002), p. 3-349.
  16. ↑ Ministry of Business, Innovation and Employment (2025b), p. 26.
  17. ↑ Ministry of Business, Innovation and Employment (2025b), p. 29.
  18. ↑ Ministry of Business, Innovation and Employment (2025b), p. 31.
  19. ↑ 19.0 19.1 19.2 19.3 19.4 Proulx (2002), p. 3-351.
  20. ↑ National Research Council of Canada et al. (2005), p. 2.8-3.
  21. ↑ 21.0 21.1 21.2 National Research Council of Canada et al. (2005), p. 2.8-2.
  22. ↑ Schifiliti et al. (2002), p. 4-2.
  23. ↑ 23.0 23.1 23.2 23.3 Schifiliti et al. (2002), p. 4-34.
  24. ↑ 24.0 24.1 Schifiliti et al. (2002), p. 4-36.
  25. ↑ 25.0 25.1 Keough and Grill (2005), p. 49.
  26. ↑ 26.0 26.1 Keough and Grill (2005), p. 45.
  27. ↑ 27.0 27.1 Australian Building Codes Board (2020a), p. 117.
  28. ↑ Keough and Grill (2005), p. 103.
  29. ↑ 29.0 29.1 Keough and Grill (2005), p. 106.
  30. ↑ 30.0 30.1 Keough and Grill (2005), p. 104.
  31. ↑ Averill et al. (2005), p. 142.
  32. ↑ 32.0 32.1 32.2 Averill et al. (2005), p. 143.

References

  • Australian Building Codes Board. (2020a). Data sheets: Fire Safety Verification Method (Handbook annex, Print version 1.1).
  • Averill, J. D., Mileti, D. S., Peacock, R. D., Kuligowski, E. D., Groner, N., Proulx, G., Reneke, P. A., & Nelson, H. E. (2005). Occupant behavior, egress, and emergency communications (NIST NCSTAR 1-7). National Institute of Standards and Technology.
  • Bryan, J. L. (2002). Behavioral response to fire and smoke. In P. J. DiNenno, D. Drysdale, C. L. Beyler, W. D. Walton, R. L. P. Custer, J. R. Hall Jr., & J. M. Watts Jr. (Eds.), SFPE handbook of fire protection engineering (3rd ed., pp. 3-315–3-341). National Fire Protection Association; Society of Fire Protection Engineers.
  • Keough, R. J., & Grill, R. A. (2005). Fire alarm systems (NIST NCSTAR 1-4C). National Institute of Standards and Technology.
  • Lew, H. S., Bukowski, R. W., & Carino, N. J. (2005). Design, construction, and maintenance of structural and life safety systems (NIST NCSTAR 1-1). National Institute of Standards and Technology.
  • Ministry of Business, Innovation and Employment. (2025b). Verification Method C/VM2: Framework for fire safety design (2nd ed.).
  • National Research Council of Canada, International Code Council, Department of Building and Housing, & Australian Building Codes Board. (2005). International fire engineering guidelines (2005 ed.). Australian Building Codes Board.
  • Proulx, G. (2002). Movement of people: The evacuation timing. In P. J. DiNenno, D. Drysdale, C. L. Beyler, W. D. Walton, R. L. P. Custer, J. R. Hall Jr., & J. M. Watts Jr. (Eds.), SFPE handbook of fire protection engineering (3rd ed., pp. 3-342–3-366). National Fire Protection Association; Society of Fire Protection Engineers.
  • Schifiliti, R. P., Meacham, B. J., & Custer, R. L. P. (2002). Design of detection systems. In P. J. DiNenno, D. Drysdale, C. L. Beyler, W. D. Walton, R. L. P. Custer, J. R. Hall Jr., & J. M. Watts Jr. (Eds.), SFPE handbook of fire protection engineering (3rd ed., pp. 4-1–4-43). National Fire Protection Association; Society of Fire Protection Engineers.
  • Van Rossum, E. (2027). Introduction to Fire Safety Engineering. Fire Safety Academy.
Retrieved from "https://fire-safety.academy/w/index.php?title=Voice_alarm_system&oldid=2620"