Appropriate Characterization of Background Noise Levels in the Workplace

TL;DR: In this article, an extensive study is being completed at the University of Nebraska investigating the effects of various types of air-conditioning noise on occupant productivity and perception, and results from performance tests and questionnaires are used to evaluate a number of these noise criteria systems.

Abstract: Overall loudness is not the only quality of indoor background noise that affects occupants. The distribution of noise across frequency (pitch), whether or not the noise contains tones, and whether or not the noise changes over time must also be considered. There are several indoor noise criteria systems available to quantify the background noise in rooms, but many of them do not account for these factors. These systems are commonly used by architects and engineers, and often incorporated into manufacturer’s data, design guides, and standards. There is much debate over which of the criteria systems best reflect how occupants will respond to noise. To examine these issues, an extensive study is being completed at the University of Nebraska investigating the effects of various types of air-conditioning noise on occupant productivity and perception. Results from performance tests and questionnaires are used to evaluate a number of these noise criteria systems. Specific implications of this project on characterizing noise in the workplace will be discussed.

Summary

Background

• It can have psychological and physiological effects on people, but individual reactions to a particular noise are very subjective (Kryter 1985) .
• Other properties of noise can also impact one's perception of noise, such as how the noise is distributed across frequency .
• Motors and fans often exhibit low-frequency pure tones under 350 Hz.
• Unfortunately there is some debate over which criteria is the most appropriate to use.
• Few have tried to determine whether or not the indoor noise criteria represent this relationship.

General Methodology

• The spaces surrounding the lab remain unoccupied during testing, with the exception of a researcher sitting quietly in an adjacent room.
• The project is being run over three phases.
• In each phase, subjects are exposed to a wide range of background noise conditions which are typical of HVAC noise situations that might be encountered in offices.
• The noise conditions are presented over two loudspeakers: an overhead panel speaker and a subwoofer in the corner.
• Their scores are then related to criteria ratings of the noise.

Phase I: Broadband (non-tonal, non-fluctuating) noise

• Phase I examined the effects of 12 broadband, non-tonal, non-fluctuating background noise conditions.
• The conditions can be generally categorized as having three levels of loudness (low, medium, and high), and four different spectral qualities (neutral, rumbly, roaring, and hissy).
• Each subject came in for two testing sessions, with each session consisting of one "warm-up trial" followed by six of the 12 noise conditions.
• Every effort was made to schedule a subject's two test sessions at approximately the same time of day.
• In the proofreading test, subjects compared two columns of 10digit numbers and marked when the columns did not match.

Phase I Results

• Statistical relationships between the three main variables: noise criteria, perception, and performance were analyzed using linear mixed models and Pearson Product Moment Correlations (SPSS 2005, Field and Hole 2003) .
• Another aspect of the criteria systems evaluated in this phase was their ability to accurately predict spectral imbalance.
• Some of the criteria also provide an assessment of noise induced vibration (NIV), which was not evaluated in this study.
• For this study, RC spectral ratings of rumble agreed with perception (p<0.01), but not necessarily hiss, whereas RC Mark II ratings matched roar (p<0.01) and hiss, but not rumble (p<0.01).
• In looking at the performance test results, no significant relationships were observed between the test scores and criteria predictions of level.

Intermediate Study: Effects of Exposure Time and Types of Tests Used

• Phase I, a study was conducted to look more specifically at the types of tasks used in this research and the time length of exposure to the noise conditions.
• To highlight, in Phase I it was found that typing was significantly affected by perception of noise, but no significant results were found with the proofreading tasks.
• Three types of tests (typing, verbal reasoning, and math) were evaluated in this intermediate study.
• It is thought that this methodology will more accurately evaluate a subject's "true" feelings about any one noise condition, without influence from the other conditions included in a particular phase.

Phase II: Tonal noise

• Results from the intermediate study were used in a Phase II.
• This portion of the research focuses specifically on discrete tones in noise, how they affect productivity and perception, and how this relates to the indoor noise criteria methods.
• Six background noise conditions that are controlled to be non-fluctuating over time are being used in this phase: 1. Mid-level neutral condition: broadband noise, simulated with -5 dB/octave band slope, intersecting 40 dB at 1000 Hz.
• This noise condition was also used in Phase I. 2. Low frequency tonal condition: broadband noise with a low frequency tonal component, measured from an existing facility, with the overall level equalized to 49 dB L Aeq 3.
• Low frequency tonal condition: similar to condition 2, with the prominence ratio of the tone increased (American National Standards Institute 1995) 4. Mid-frequency tonal condition: broadband noise with a mid-frequency tonal component, measured from an existing facility, with the overall level equalized to 48 dB L Aeq .

5. Mid-frequency tonal condition:

• Broadband noise with a high frequency tonal component, measured from an existing facility, with the overall level equalized to 48 dB L Aeq, also known as High frequency tonal condition.
• Under each condition, subjects complete typing, verbal reasoning, and math tests.
• Following each testing session, the noise is turned off, and subjects are asked to what degree they prefer the current environment as compared to the environment they have been working under for the past hour.
• Results from this phase are still under analysis.
• The subjective performance and perception results are being compared to criteria ratings of the background noise conditions.

Conclusions

• A variety of indoor noise criteria methods, including NC, NCB, RC, RC Mark II, and L Aeq are used by architects, mechanical systems designers, and acousticians to quantify background noise in buildings.
• A series of studies are being conducted at the University of Nebraska to relate several of the more commonly used criteria systems to human performance and perception under HVAC-type noises which might be encountered in working environments.
• The noises span a range of sound levels, spectral qualities (neutral, rumbly, roaring, and hissy), and characteristics (broadband, with tones, and with time-fluctuating components).
• Research is currently underway to more specifically examine the impact of tones and fluctuations in background noise on perception and performance, and the relationship to indoor noise criteria.
• The impacts this research is having on the fields of acoustical, architectural, and mechanical engineering are very exciting.

Figures

Figure 3 – Phase I subjective loudness ratings versus subjective annoyance ratings averaged across all subjects. Standard error bars are shown. The loudness rating scale ranges from 1 = very quiet to 7 = very loud, and the annoyance rating scale ranges from 1 = not annoying to 7 = very annoying.

Table 1 – The spectral quality predictors available with the five criteria systems analyzed

Figure 2 – Phase I subjective loudness perception versus LAeq level. Standard error bars are shown.

Figure 1 – Three items investigated in the research

Full text

University of Nebraska - Lincoln University of Nebraska - Lincoln
DigitalCommons@University of Nebraska - Lincoln DigitalCommons@University of Nebraska - Lincoln
Architectural Engineering -- Faculty Publications
Architectural Engineering and Construction,
Durham School of
3-2006
Appropriate Characterization of Background Noise Levels in the Appropriate Characterization of Background Noise Levels in the
Workplace Workplace
Erica E. Bowden
University of Nebraska - Lincoln
Lily M. Wang
University of Nebraska - Lincoln
, lwang4@unl.edu
Follow this and additional works at: https://digitalcommons.unl.edu/archengfacpub
Part of the Architectural Engineering Commons
Bowden, Erica E. and Wang, Lily M., "Appropriate Characterization of Background Noise Levels in the
Workplace" (2006).
Architectural Engineering -- Faculty Publications
. 11.
https://digitalcommons.unl.edu/archengfacpub/11
This Article is brought to you for free and open access by the Architectural Engineering and Construction, Durham
School of at DigitalCommons@University of Nebraska - Lincoln. It has been accepted for inclusion in Architectural
Engineering -- Faculty Publications by an authorized administrator of DigitalCommons@University of Nebraska -
Lincoln.

1
Appropriate characterization of background noise levels in the workplace
Erica E. Bowden
1
, Lily M. Wang
Architectural Engineering Program, University of Nebraska – Lincoln, PKI, 1110 S.
67
th
St., Omaha, NE 68182, USA
1
corresponding author: PH (402)554-2074; email: ebowden@mail.unomaha.edu
Abstract
Overall loudness is not the only quality of indoor background noise that
affects occupants. The distribution of noise across frequency (pitch), whether or not
the noise contains tones, and whether or not the noise changes over time must also be
considered. There are several indoor noise criteria systems available to quantify the
background noise in rooms, but many of them do not account for these factors. These
systems are commonly used by architects and engineers, and often incorporated into
manufacturer’s data, design guides, and standards. There is much debate over which
of the criteria systems best reflect how occupants will respond to noise. To examine
these issues, an extensive study is being completed at the University of Nebraska
investigating the effects of various types of air-conditioning noise on occupant
productivity and perception. Results from performance tests and questionnaires are
used to evaluate a number of these noise criteria systems. Specific implications of
this project on characterizing noise in the workplace will be discussed.
Background
Noise is generally thought of as “unwanted sound.” It can have psychological
and physiological effects on people, but individual reactions to a particular noise are
very subjective (Kryter 1985). One individual may find a particular noise pleasing,
while another person may be quite disturbed by the same sound. Because of the
subjective nature of noise perception, quantifying background noise and determining
appropriate levels that will satisfy the majority of occupants is a difficult process.
To complicate the issue, overall loudness is not the only quality of
background noise that affects occupants. Other properties of noise can also impact
one’s perception of noise, such as how the noise is distributed across frequency
(pitch). For example, does the noise contain a lot of low frequency energy? If so,
people might perceive it as sounding “rumbly”, or it might even induce vibration in a
building. If the noise has a lot of high frequency energy, then it might be described
as sounding “hissy.” Another commonly used descriptor is “roaring,” which means a
noise has a lot of middle frequency energy.
A second property of noise that affects occupants is whether or not the noise
contains one or more tones, or discrete pitches. Tonal noise is quite commonly

2
produced by heating, ventilating, and air-conditioning (HVAC) systems in buildings.
For example, motors and fans often exhibit low-frequency pure tones under 350 Hz.
Chillers can produce higher frequency tones. There are many more examples of
sources that produce tones.
A third property to consider is how the noise behaves over time. Does the
noise remain relatively steady or does it fluctuate? If it fluctuates, the change may be
rapid (as in a rattling diffuser), or slower (as in a variable air volume (VAV) system
that adjusts to changing occupancy).
There are several indoor noise criteria systems available to quantify the
background noise in rooms, but many of them do not fully account for these factors.
The criteria systems are commonly used by architects and engineers, and often
incorporated into manufacturer’s data, design guides, and standards. Some of the
more commonly used criteria include Noise Criteria (NC; Beranek 1957), Balanced
Noise Criteria (NCB; Beranek 1989), Room Criteria (RC; Blazier 1981), Room
Criteria Mark II (RC Mark II; Blazier 1997), and A-weighted Equivalent Sound
Pressure Level (L
Aeq
; International Organization for Standardization 1987). These
criteria provide a single number rating which describes the overall level, or loudness
of the sound. Some also include descriptors, such as rumbly, roaring, and hissy,
which give an indication of how the background noise will sound to the occupant.
Unfortunately there is some debate over which criteria is the most appropriate
to use. Previous studies have shown that there can be large differences among indoor
noise criteria systems for the same spectrum (Goodfriend 1975, Tocci 2000, Bowden
and Wang 2003). There is no consensus that a particular criterion generally performs
better than the others, with preferences among professional societies and consultants
varying greatly.
In working environments, the ability of the criteria systems to relate to
performance is also of interest. There have been several studies examining the effects
of low frequency noise on performance (Landström et al 1991, Holmberg 1993,
Persson Waye et al 1997, Persson Waye et al 2001). Some of the conclusions from
these studies are that task performance can be affected by noise, the impact can
potentially change over time, and the frequency character should be considered.
Although this previous work has established that a relationship between noise and
performance exists, few have tried to determine whether or not the indoor noise
criteria represent this relationship.
To examine these issues, the authors are investigating the effects of various
types of air-conditioning noise on occupant productivity and perception. As
illustrated in Figure 1, results from performance tests and questionnaires are used to
evaluate criteria ratings of NC, NCB, RC, RC Mark II, and L
Aeq
, as well as compared
against each other. End results aid in understanding appropriate applications of the
various criteria methods.

3
Figure 1 – Three items investigated in the research
General Methodology
All of the testing is conducted in the “Indoor Environment Lab” at the
University of Nebraska. This 25.7 m
3
(906 ft
3
) lab resembles a typical office and was
specially constructed to be acoustically and thermally controlled. The spaces
surrounding the lab remain unoccupied during testing, with the exception of a
researcher sitting quietly in an adjacent room. The naturally occurring background
noise in the lab is low at 36 dB L
Aeq
, and the reverberation time is 0.25 seconds at 500
Hz.
The project is being run over three phases. In each phase, subjects are
exposed to a wide range of background noise conditions which are typical of HVAC
noise situations that might be encountered in offices. The noise conditions are
presented over two loudspeakers: an overhead panel speaker and a subwoofer in the
corner. The overhead loudspeaker resembles a ceiling tile, and the subwoofer is
disguised to resemble an end-table. Under the different noise conditions, subjects
complete performance tests and perception questionnaires. Their scores are then
related to criteria ratings of the noise.
Phase I: Broadband (non-tonal, non-fluctuating) noise
Phase I examined the effects of 12 broadband, non-tonal, non-fluctuating
background noise conditions. The conditions can be generally categorized as having
three levels of loudness (low, medium, and high), and four different spectral qualities
(neutral, rumbly, roaring, and hissy). The neutral signals follow a slope of
approximately -5 dB/octave band. The rumbly, roaring, and hissy signals were
obtained by raising the levels in the low, mid, and high frequency ranges by 5-10 dB.
Further details on the Phase I noise conditions are available in Bowden and Wang
(2005).
Twenty-eight subjects were exposed to each noise condition for
approximately 12 minutes. During that time, they spent 90 seconds adapting to the
noise, followed by approximately ten minutes of performance testing and then a short
questionnaire. Each subject came in for two testing sessions, with each session
consisting of one “warm-up trial” followed by six of the 12 noise conditions. Every
effort was made to schedule a subject’s two test sessions at approximately the same
time of day.
Two types of tests were used in this phase: a typing test and a proofreading
test. Both tests were administered through a computer program developed by the
National Research Council of Canada (Scovil et al 1995a, 1995b). In the typing test,
Indoor Noise
Criteria Ratings
PerceptionPerformance

4
subjects were scored for speed and accuracy as they re-typed text presented on a
computer monitor. In the proofreading test, subjects compared two columns of 10-
digit numbers and marked when the columns did not match.
The questionnaire consisted of five questions about the background noise.
Specifically, subjects had to rate their impressions of the loudness, rumble, roar, hiss,
and annoyance on seven point discrete scales.
Phase I Results
Statistical relationships between the three main variables: noise criteria,
perception, and performance were analyzed using linear mixed models and Pearson
Product Moment Correlations (SPSS 2005, Field and Hole 2003). Additional
statistical methods are being considered for future analysis.
Phase I results showed that noise conditions with higher criteria level ratings
were perceived as louder, more rumbly, more roaring, and more hissy by test subjects
(p<0.01). Figure 2 shows an example, where conditions with higher L
Aeq
levels were
perceived as louder. This is the expected trend. Perception of loudness and
annoyance were also found to be correlated, as demonstrated in Figure 3 (p<0.01).
30
35
40
45
50
55
60
65
70
1234567
Subjective Loudness Perception Rating
L
Aeq
Level of Noise Condition
very quiet very loud
Figure 2 – Phase I subjective loudness perception versus L
Aeq
level. Standard
error bars are shown.

Frequently asked questions

Q1. What have the authors contributed in "Appropriate characterization of background noise levels in the workplace" ?

Specific implications of this project on characterizing noise in the workplace will be discussed. 

Q2. What were the three predictors used by the criteria systems?

That is, the rumbly, roaring, and hissy predictors used by several of the criteria systems were compared to subjective perception. 

Q3. What are the main conclusions from these studies?

Some of the conclusions from these studies are that task performance can be affected by noise, the impact can potentially change over time, and the frequency character should be considered. 

Q4. What are the common indoor noise criteria methods?

A variety of indoor noise criteria methods, including NC, NCB, RC, RC Mark II, and LAeq are used by architects, mechanical systems designers, and acousticians to quantify background noise in buildings. 

Q5. How long did the subjects spend adapting to the noise?

During that time, they spent 90 seconds adapting to the noise, followed by approximately ten minutes of performance testing and then a short questionnaire. 

Q6. What are the other factors that are included in the tests?

Other questions about the temperature, lighting, and work station comfort are included as well, although the only factor manipulated in the test room is the noise. 

Q7. What was the effect of typing speed on the noise?

it was found that typing speed slowed down significantly as subjects perceived the noise as louder or more hissy (p<0.01). 

Q8. What is the purpose of this research?

Research is currently underway to more specifically examine the impact of tones and fluctuations in background noise on perception and performance, and the relationship to indoor noise criteria. 

Q9. What is the difference between the two studies?

An intermediate study further revealed that performance did not change significantly over time, but that shorter noise exposures (20 minutes) may allow for more back-to-back perception comparison between sequential noise conditions.