
Cornell University Ergonomics Web
DEA3500: Ambient Environment: Acoustics and Speech Privacy
NOISE IN BUILDINGS
Acoustics and Speech Privacy
A major problem in the design of many modern office buildings concerns the provision of adequate levels of speech privacy.  This problem is particularly acute in open-plan offices, libraries etc.  Main noise sources in such settings are:
- telephones
 - typewriters
 - printers/photocopiers
 - environmental services e.g. air conditioning
 - conversations
 
What has tended to happen is that in many buildings the background noise level has fallen (c.f. factories) and the consequence of this is that the signal / noise ratio has increased.
The problem of speech privacy concerns the intelligibility of the encroaching sound.
Most intelligible speech occurs in the 2-4 KHz range (full speech range = 100-8000 Hz, most energy between 100-600 Hz), irrespective of gender, although females tend to have voices one octave higher than their male counterparts.  Basically there are only two ways of overcoming the privacy problem:
-  reduce the signal strength
 - increase the noise level
 
Ways of reducing signal strength
Room interiors provide surfaces that can either absorb or reflect sound.  For a material/ surface to be sound absorbing it needs to be soft, porous, fuzzy, and thick.  If the material is hard and solid, it will reflect sound and if it is also thin it may resonate and even amplify the sound.
Noise Reduction Coefficient
The sound-absorbing capability of a material can be represented by a single figure - the noise reduction coefficient (NRC).
NRC = (a250 + a500 + a1000 +
a2000)/4 
													
                                      
- where a = coefficient of absorption = ratio of reflected sound to incident sound at a specific frequency.
 - (250, 500, 1000, 2000 are octave band midpoints)
 - full range of octave band midpoints - 31.5, 63, 125, 350, 500, 1000, 2000, 4000, 8000 Hz
 
Major sources of absorption/reflection in a room are:
  - Ceiling - major sound surface in many rooms.  As room size increases
    so the ceiling increases in importance.  Ceilings are often constructed from or covered by some form of sound absorbing mineral tile (NRC  > .90).  However, ceiling tiles do not provide a uniform surface e.g. joints between tiles, and also there are light fittings either recessed or suspended.
    Flat lucite/perspex lenses over fluorescent tubes are the worst fittings for sound reflection; parabolic, deep cell diffusers are the best for sound absorption.  Sometimes suspended ceiling baffles in a checkerboard pattern are used.
Walls - These are usually the next most important surface.  Their importance increases as room size decreases.  Typically walls have very poor sound absorbing qualities and this is often made worse by putting sound reflectors against the walls e.g. filing cabinets.  Carpeting the walls will increase sound absorption, as does placing cork tiles/corkboard on the walls.  (However, the NRC for these surfaces is still .2-.3.)  Windows are also problematic here.
 
  - Floor - Carpeting the floor will increase the NRC, but only up to about 0.3.  Moving to thicker carpeting is often not a cost-effective solution because much of the floor area is covered by furniture with a worse NRC.  Carpeting will reduce impact noise though (dropping on table vs. carpet).
 
  - Furniture - Most furniture is designed to be functional or aesthetically pleasing rather than to be a good sound absorber.  In many types of offices, screening has been purchased in an effort to improve visual and acoustic privacy.  However, often the screening is only at a height of 4' (1.22
    meters) which effectively is 0 feet high for speech privacy because it approximates the height of a person's mouth and ears when seated.
 
To have any impact at all the screening needs to be at least 5 feet high and one needs solid mass in this screening to block sound energy.
Ways of increasing the noise level:
Acoustic masking/sound conditioning -
When steps have been taken to reduce the signal strength, then measures to increase noise can be adopted.  Background music or
Musak may not be an effective way of doing this because the frequency range is too wide.  Electronic sound conditioning (white noise) directed to the 2-4 KHz spectrum is the most effective way of achieving this.
Signal - wanted sound,
Noise - unwanted sound
example:  In library, people talking around you while you're trying to study.  To people talking, talk is signal; to studier, talk is noise.
Speech tends to be particularly distracting when understandable - one strains to ignore the sound.
Unexpected sounds and uncontrollable sounds tend to be especially disturbing.
In office settings many noise sources: typewriters/word processors, etc.
These intruding noises are especially severe in open office type designs.
Go to the next lecture