1、OR-05-6-4 Relating Human Productivity and Annoyance to Indoor Noise Criteria Systems: A Low Frequency Analysis E.E. Bowden ABSTRACT A number of indoor noise criteria systems are used to quantifj, the background noise in a built environment, includ- ing Noise Criteria (NC), Balanced Noise Criteria (N
2、CB), Room Criteria (RC), Room Criteria Mark II (RC Mark Ir), A- weighted Equivalent Sound Pressure Level (LA,d, and others. An ongoing debate exists in the acoustical community over which criterion is the most appropriate to use in the variety of ambient noise situations encountered. In an effort to
3、 quanti- tatively support the use of an individual criterion, this project subjectively correlates these various criteria with human task performance and perception. Eleven subjects participated in a pilot study by completing typing and proofreading tasks, as well as subjective ratings ofloudness, a
4、nnoyance, andspectral quality. Results show that there were no signijcant diferences in productivity scores among the 12 noise exposures tested; howevel; signijcant relationships were found between indoor noise criteria predictions of level and subjective perception of loudness and annoyance. In thi
5、s study, RC and RC-Mark II were found to be the most correlated with level perception, although NC, NCB, and LAes were also strongly correlated. Additionally, interesting relationships were found between subjective perceptions of rumble or roar and criteria predic- tions ofsuch. The authors are in t
6、he process of extending the pilot study to more subjects, as well as examining the efects of tonal and jluctuating background noise spectra on criteria predictions. INTRODUCTION Indoor background noise can dramatically impact occu- pants by causing annoyance, affecting productivity, hindering speech
7、 communication, impacting sleep, and degrading over- L.M. Wang, PhD Member ASHRAE all occupant comfort and satisfaction. In extreme cases, exces- sive background noise can even result in hearing damage. Noise is a complex entity, and the effect on occupants can vary depending on factors such as leve
8、l or loudness, how the sound varies across frequency, and even how it varies across time. Acoustic specialists have used various criteria over the decades to quanti human perception of the background noise in a room. Most of the descriptors consist of single-number ratings that summarize the backgro
9、und noise level over a range of frequencies. Some provide additional descriptors of quality that evaluate the spectral characteristics of the back- ground noise. Noise Criteria (NC), Balanced Noise Criteria (NCB), Room Criteria (RC), Room Criteria Mark II (RC Mark II), and A-weighted Equivalent Soun
10、d Pressure Level (LAeq) are five criteria systems commonly used by mechanical engi- neers, architects, and acousticians in the United States. The criteria systems are popular tools in setting appropriate back- ground noise levels in built spaces based on type of occupancy. However, an ongoing debate
11、 exists in the acoustical commu- nity over which criterion is the most appropriate to use in the variety of background noise situations encountered. The pool of data linking the use of these various criteria to actual human reaction continues to grow. This study seeks to add to this database by exam
12、ining the correlations between indoor noise criteria systems and human productivity, loud- ness, annoyance, and spectral quality. Previous Research Many previous studies have sought to evaluate the effects of background noise on humans. Beranek (1956), Keighley (1966, 1970), Hay and Kemp (1972a, 197
13、2b), and Blazier (1 98 1) are among those who have developed criteria systems Erica E. Bowden is a doctoral student and Lily M. Wang is an assistant professor in the Architectural Engineering Program, University of Nebraska, Lincoln. 684 02005 ASHRAE. reflecting occupant response to office noise. Re
14、cent years have seen a resurgence of researchers linking subjective perception of ambient noise with measured sound spectra (Tang et al. 1996; Tang 1997; Tang and Wong 1998,2003; Ayr et al. 200 1 , 2003). Subjects of these studies were asked to rate their general perception of the background noise w
15、ith regard to several factors including annoyance, loudness, and satisfac- tion. Their responses were then related to background noise measurements and criteria systems. Tang and Ayr consistently found LAep to be highly correlated with subjective auditory sensation in office surveys. Persson Waye an
16、d Rylander (2001), on the other hand, found that LAeq was not a good predictor of annoyance to long-term noise exposure in resi- dences. This discrepancy indicates that the types of spaces analyzed and the measurement method can affect the perfor- mance of criteria predictions. The effect of low fre
17、quency noise in particular has been the focus of much research. In addition to subjective reaction to background noise, productivity was also evaluated in several studies. Kyriakides and Leventhall (1 977) investigated performance on central and peripheral vision tasks under three acoustic condition
18、s: audio frequency noise at 70 dBA, an infrasound noise band from 2 Hz to 15 Hz at 1 15 dBA, and an audio frequency noise band from 40 Hz to 16 kHz at 90 dBA. They found that the peripheral vision task was affected by noise, and the effect of infrasound increased over the 36 minutes spent on the tas
19、k. Landstrm et al. (1991) examined the effects of three different ventilation noise signals on occupant performance, wakefulness, and annoyance. The signals were broadband (40 dBA), 1 O0 Hz tonal broadband (40 dBA), and the same tonal noise masked by means of low frequency pink noise (41 dBA). Lengt
20、h of exposure to each noise signal was 50 minutes, during which subjects performed tasks for the first 40 minutes and rested for the final 10 minutes. Performance on figure identification tasks was found to be lower during the 100 Hz tonal signal than the masked tonal signal. Holmberg et al. (1993)
21、used five different ventilation noise exposures: gradually falling frequencyAeve1 spectral character (35 dBA and 40 dBA), 43 Hz raised filtered broad- band noise (40 dBA), 43 Hz tonal broadband noise (40 dBA), and naturally occurring background noise (20 dBA). Subjects were exposed to each noise for
22、 60 minutes, during which time they completed proofreading tasks. Although no significant differences between exposures were obtained on performance tests, the results did indicate that the frequency character should be considered when evaluating the effects of ventila- tion noise on annoyance sensa
23、tion and productivity. In 1997, Persson Waye et al. evaluated the effect on performance and work quality of two ventilation spectra, one of predominately mid-frequency character (NC 35) and the other of predominantly low frequency character (NC 35). Total time spent under each exposure was 60 minute
24、s. The study concluded that the low frequency noise interfered more strongly with performance on three cognitive tasks than the Figure 1 View of the test chamber, with subject (S), i-ceiling speaker (LS), and subwoofer (SUB) locations. mid-frequency noise. The difference between productivity scores
25、in this study indicates that NC curves do not fully assess the negative impact of low frequency noise on task perfor- mance. Furthermore, as in the Leventhall study, there was an indication that the effects of noise developed over time. Pers- son Waye et al. (2001) extended the study and found that
26、low frequency noise negatively impacts demanding verbal tasks, while the effects on more routine tasks were less clear. Addi- tionally, results indicated that low frequency noise may be more difficult to adapt to. This study aims to further the research on background noise and work performance with
27、12 new background noise exposures, all of differing loudness and spectral content. The ability ofindoor noise criteria systems to relate to productivity scores, as well as auditory perception of noise, is examined in detail. METHODOLOGY Subjects Eleven subjects (five male and six female) participate
28、d in the pilot study. Subjects ranged in age from 19 to 29 with a mean age of 24. All subjects were prescreened for typing abil- ity, auditory ability, and visual function. The subjects were all found to have a minimum typing ability of 20 words per minute using Skillcheck typing test software. Adeq
29、uate visual function was verified with the Keystone Ophthalmic Telebin- ocular, which provides a quick measure of phorias, fusion readiness, binocular visual efficiency at far and near, stereop- sis, visual acuity, and color vision. Finally, a GSI 17 audiom- eter was used to verify that all subjects
30、 had hearing thresholds below 25 dB hearing level (HL) from 125 Hz to 8 kHz. Test Chamber The experiment was performed in a 906 fi3 (25.7 m3) test chamber. A view of the floor plan is shown in Figure 1, with test subject and loudspeaker locations noted. The room is ASHRAE Transactions: Symposia 685
31、Table 1. Noise Exposure Design Matrix Neutral 30 dB at 1000 Hz 40 dB at 1000 Hz 50 dB at 1000 Hz LEVEL Rumble Roar +5 to 10 dB in 31.5 and 63 Hz octave bands +10 dB in 125,250, and 500 Hz octave bands I- High SPECTRAL QUALITY 16 31.5 63 125 250 500 1000 2000 4000 8000 Octave Band Center Frequency (H
32、z) Figure 2 Frequency character of test chamber background noise levels. furnished as a typical office with carpeting, gypsum board wail construction, and acoustical ceiling tiles, and it exhibits a reverberation time of 0.25 seconds at 500 Hz. The naturally occurring background noise level in the t
33、est chamber is rela- tively low, as shown in Figure 2. Additionally, the surrounding structure achieves STC 47 to minimize noise intrusion from adjacencies. The spaces immediately surrounding the struc- ture were unoccupied during testing, with the exception of the researcher sitting quietly in an a
34、djacent room. The room was maintained as a comfortable working environment at approx- imately 68F (20C), with overhead fluorescent lighting at an average illuminance of 7 1 foot-candies (764 lux) at the work plane. Experimental Procedure A flowchart of the experimental procedure is shown in Figure 3
35、. Ninety-second adaptation times to the background noise were used at the beginning of each new noise exposure to allow the subject to audiologically adjust to the change in background noise. Subjects were instructed to sit and relax during this period. Productivity tests and a subjective rating por
36、tion followed. Each noise exposure trial lasted approxi- mately 12 minutes. To reduce overall fatigue, testing took Hiss +IO dB in 2000,4000, and 8000 Hz octave bands Productivity Tasks 1-b time (min) O 1.5 10 12 Figure3 view of the experimental sequence for a single noise exposure. place over two s
37、essions on two separate days, with each subject scheduled at approximately the same time on both days. Each session lasted approximately one and a half hours, for a total testing time per subject of three hours. Noise Exposures Twelve different background noise exposures which simulate ventilation n
38、oises that might be encountered in real- world environments were used in this study. Each exposure was controlled to be nonvarying over time and nontonal. The exposures can be generally categorized as having three differ- ent levels (low, medium, and high) and four different spectral qualities (neut
39、ral, nimbly, roaring, and hissy). A matrix of the noise exposure design is given in Table 1. The neutral signals followed a slope of approximately -5 dB/octave band. Rumbly sounding signals were achieved by raising the levels of the 31.5 and 63 Hz octave bands by 5 to 10 decibels. Similarly, roaring
40、 and hissy sounding signals were achieved by raising the levels by approximately 10 decibels from 125 to 500 Hz and 2000 to 8000 Hz, respectively. Control over the 16 Hz octave band was limited due to subwoofer response and mixing capabilities. Octave band measurements of the mid- level signals are
41、presented in Figures 4 through 7. All 686 ASHRAE Transactions: Symposia 70 tin 7 70 60 -Jm En u= a O * 16 31.5 63 125 250 500 1000 2000 4000 8000 Octave Band Center Frequency (Hz) Figure 4 Frequency character of the mid-level neutral noise exposure. 70 1 I I I I_ I 16 31.5 63 125 250 500 1000 2000 4
42、000 8000 Octave Band Center Frequency (Hz) Figure 5 Frequency character of the mid-level rumbly noise exposure. 70 60 ” - - al-50 al 50 -lm -Jm $ %40 a 040 E ? 30 ? 30 am am -0- -0-0 c - 20 O O u) En $N :- 5 - 20 a 10 O rn 10 O 16 31.5 63 125 250 500 1000 2000 4000 8000 Octave Band Center Frequency
43、(Hz) 16 31.5 63 125 250 500 1000 2000 4000 8000 Octave Band Center Frequency (Hz) Figure 6 Frequency character of the mid-level roaring noise exposure. measurements were made at the test subjects location using a Larson Davis 824 sound level meter. Noise exposures were presented over two loudspeaker
44、s: an Armstrong i-ceilingTM loudspeaker and a JBL NorthridgeTM E250P subwoofer. The exposures were presented in random order, and no two subjects heard the same order of presentation. Mixing and amplification of the loud- speakers was achieved with an Armstrong i-ceilingTM D200 1 Digital Processor a
45、nd a D4 100 Amplifier. All test signals were generated by filtering white noise into the desired spectra with Cool Edit 2000 software. The i-ceilingTM loudspeakers are typically used in open office plans for masking systems and look like acoustical lay- in ceiling tiles. The subwoofer was covered in
46、 an acoustically transparent fabric to resemble an end table. In a post-study survey, it was found that the majority of subjects were unable Figure 7 Frequency character of the mid-level hissy noise exposure. to identi the source of the noise and merely commented that it seemed to be coming from abo
47、ve the ceiling somewhere. In this sense, the localization of the noise resembled typical ventilation installations. At the end of the study, subjects were also asked if the background noises reminded them of anything they had heard before. Responses included “air- conditioners,” “mechanical noise,”
48、“vents,” and “the noise in my office,” indicating that most of the exposures were gener- ally considered to be similar to office noise they heard before. A few subjects commented that the more hissy-sounding signals were less natural sounding. Productivity Tests Productivity was evaluated under each
49、 background spec- trum via two types of computer-based tests. The test and soft- ware were developed in conjunction with the National ASHRAE Transactions: Symposia 687 Rate the noise you are hearing according to the following qualities: Not Quiet Loud o o o o o o o very Rurnbly Rumbly Hissy Roaring Roaring Hissy Not o o o o o o o Very Figure8 Scale used for subjective ratings of loudness, annoyance, and spectral quality. Research Council of Canada (Scovil et al. 1995a, 1995b). A typing test required the subjects to retype paragraphs presented on the computer monitor. Paragraphs w
