National Academies Press: OpenBook
« Previous: Chapter 3 - Research Approach
Page 41
Suggested Citation:"Chapter 4 - Findings and Applications." National Academies of Sciences, Engineering, and Medicine. 2016. Evaluating Methods for Determining Interior Noise Levels Used in Airport Sound Insulation Programs. Washington, DC: The National Academies Press. doi: 10.17226/23473.
×
Page 41
Page 42
Suggested Citation:"Chapter 4 - Findings and Applications." National Academies of Sciences, Engineering, and Medicine. 2016. Evaluating Methods for Determining Interior Noise Levels Used in Airport Sound Insulation Programs. Washington, DC: The National Academies Press. doi: 10.17226/23473.
×
Page 42
Page 43
Suggested Citation:"Chapter 4 - Findings and Applications." National Academies of Sciences, Engineering, and Medicine. 2016. Evaluating Methods for Determining Interior Noise Levels Used in Airport Sound Insulation Programs. Washington, DC: The National Academies Press. doi: 10.17226/23473.
×
Page 43
Page 44
Suggested Citation:"Chapter 4 - Findings and Applications." National Academies of Sciences, Engineering, and Medicine. 2016. Evaluating Methods for Determining Interior Noise Levels Used in Airport Sound Insulation Programs. Washington, DC: The National Academies Press. doi: 10.17226/23473.
×
Page 44
Page 45
Suggested Citation:"Chapter 4 - Findings and Applications." National Academies of Sciences, Engineering, and Medicine. 2016. Evaluating Methods for Determining Interior Noise Levels Used in Airport Sound Insulation Programs. Washington, DC: The National Academies Press. doi: 10.17226/23473.
×
Page 45
Page 46
Suggested Citation:"Chapter 4 - Findings and Applications." National Academies of Sciences, Engineering, and Medicine. 2016. Evaluating Methods for Determining Interior Noise Levels Used in Airport Sound Insulation Programs. Washington, DC: The National Academies Press. doi: 10.17226/23473.
×
Page 46
Page 47
Suggested Citation:"Chapter 4 - Findings and Applications." National Academies of Sciences, Engineering, and Medicine. 2016. Evaluating Methods for Determining Interior Noise Levels Used in Airport Sound Insulation Programs. Washington, DC: The National Academies Press. doi: 10.17226/23473.
×
Page 47
Page 48
Suggested Citation:"Chapter 4 - Findings and Applications." National Academies of Sciences, Engineering, and Medicine. 2016. Evaluating Methods for Determining Interior Noise Levels Used in Airport Sound Insulation Programs. Washington, DC: The National Academies Press. doi: 10.17226/23473.
×
Page 48
Page 49
Suggested Citation:"Chapter 4 - Findings and Applications." National Academies of Sciences, Engineering, and Medicine. 2016. Evaluating Methods for Determining Interior Noise Levels Used in Airport Sound Insulation Programs. Washington, DC: The National Academies Press. doi: 10.17226/23473.
×
Page 49
Page 50
Suggested Citation:"Chapter 4 - Findings and Applications." National Academies of Sciences, Engineering, and Medicine. 2016. Evaluating Methods for Determining Interior Noise Levels Used in Airport Sound Insulation Programs. Washington, DC: The National Academies Press. doi: 10.17226/23473.
×
Page 50
Page 51
Suggested Citation:"Chapter 4 - Findings and Applications." National Academies of Sciences, Engineering, and Medicine. 2016. Evaluating Methods for Determining Interior Noise Levels Used in Airport Sound Insulation Programs. Washington, DC: The National Academies Press. doi: 10.17226/23473.
×
Page 51
Page 52
Suggested Citation:"Chapter 4 - Findings and Applications." National Academies of Sciences, Engineering, and Medicine. 2016. Evaluating Methods for Determining Interior Noise Levels Used in Airport Sound Insulation Programs. Washington, DC: The National Academies Press. doi: 10.17226/23473.
×
Page 52
Page 53
Suggested Citation:"Chapter 4 - Findings and Applications." National Academies of Sciences, Engineering, and Medicine. 2016. Evaluating Methods for Determining Interior Noise Levels Used in Airport Sound Insulation Programs. Washington, DC: The National Academies Press. doi: 10.17226/23473.
×
Page 53
Page 54
Suggested Citation:"Chapter 4 - Findings and Applications." National Academies of Sciences, Engineering, and Medicine. 2016. Evaluating Methods for Determining Interior Noise Levels Used in Airport Sound Insulation Programs. Washington, DC: The National Academies Press. doi: 10.17226/23473.
×
Page 54
Page 55
Suggested Citation:"Chapter 4 - Findings and Applications." National Academies of Sciences, Engineering, and Medicine. 2016. Evaluating Methods for Determining Interior Noise Levels Used in Airport Sound Insulation Programs. Washington, DC: The National Academies Press. doi: 10.17226/23473.
×
Page 55
Page 56
Suggested Citation:"Chapter 4 - Findings and Applications." National Academies of Sciences, Engineering, and Medicine. 2016. Evaluating Methods for Determining Interior Noise Levels Used in Airport Sound Insulation Programs. Washington, DC: The National Academies Press. doi: 10.17226/23473.
×
Page 56
Page 57
Suggested Citation:"Chapter 4 - Findings and Applications." National Academies of Sciences, Engineering, and Medicine. 2016. Evaluating Methods for Determining Interior Noise Levels Used in Airport Sound Insulation Programs. Washington, DC: The National Academies Press. doi: 10.17226/23473.
×
Page 57
Page 58
Suggested Citation:"Chapter 4 - Findings and Applications." National Academies of Sciences, Engineering, and Medicine. 2016. Evaluating Methods for Determining Interior Noise Levels Used in Airport Sound Insulation Programs. Washington, DC: The National Academies Press. doi: 10.17226/23473.
×
Page 58
Page 59
Suggested Citation:"Chapter 4 - Findings and Applications." National Academies of Sciences, Engineering, and Medicine. 2016. Evaluating Methods for Determining Interior Noise Levels Used in Airport Sound Insulation Programs. Washington, DC: The National Academies Press. doi: 10.17226/23473.
×
Page 59
Page 60
Suggested Citation:"Chapter 4 - Findings and Applications." National Academies of Sciences, Engineering, and Medicine. 2016. Evaluating Methods for Determining Interior Noise Levels Used in Airport Sound Insulation Programs. Washington, DC: The National Academies Press. doi: 10.17226/23473.
×
Page 60
Page 61
Suggested Citation:"Chapter 4 - Findings and Applications." National Academies of Sciences, Engineering, and Medicine. 2016. Evaluating Methods for Determining Interior Noise Levels Used in Airport Sound Insulation Programs. Washington, DC: The National Academies Press. doi: 10.17226/23473.
×
Page 61
Page 62
Suggested Citation:"Chapter 4 - Findings and Applications." National Academies of Sciences, Engineering, and Medicine. 2016. Evaluating Methods for Determining Interior Noise Levels Used in Airport Sound Insulation Programs. Washington, DC: The National Academies Press. doi: 10.17226/23473.
×
Page 62
Page 63
Suggested Citation:"Chapter 4 - Findings and Applications." National Academies of Sciences, Engineering, and Medicine. 2016. Evaluating Methods for Determining Interior Noise Levels Used in Airport Sound Insulation Programs. Washington, DC: The National Academies Press. doi: 10.17226/23473.
×
Page 63
Page 64
Suggested Citation:"Chapter 4 - Findings and Applications." National Academies of Sciences, Engineering, and Medicine. 2016. Evaluating Methods for Determining Interior Noise Levels Used in Airport Sound Insulation Programs. Washington, DC: The National Academies Press. doi: 10.17226/23473.
×
Page 64
Page 65
Suggested Citation:"Chapter 4 - Findings and Applications." National Academies of Sciences, Engineering, and Medicine. 2016. Evaluating Methods for Determining Interior Noise Levels Used in Airport Sound Insulation Programs. Washington, DC: The National Academies Press. doi: 10.17226/23473.
×
Page 65
Page 66
Suggested Citation:"Chapter 4 - Findings and Applications." National Academies of Sciences, Engineering, and Medicine. 2016. Evaluating Methods for Determining Interior Noise Levels Used in Airport Sound Insulation Programs. Washington, DC: The National Academies Press. doi: 10.17226/23473.
×
Page 66
Page 67
Suggested Citation:"Chapter 4 - Findings and Applications." National Academies of Sciences, Engineering, and Medicine. 2016. Evaluating Methods for Determining Interior Noise Levels Used in Airport Sound Insulation Programs. Washington, DC: The National Academies Press. doi: 10.17226/23473.
×
Page 67
Page 68
Suggested Citation:"Chapter 4 - Findings and Applications." National Academies of Sciences, Engineering, and Medicine. 2016. Evaluating Methods for Determining Interior Noise Levels Used in Airport Sound Insulation Programs. Washington, DC: The National Academies Press. doi: 10.17226/23473.
×
Page 68
Page 69
Suggested Citation:"Chapter 4 - Findings and Applications." National Academies of Sciences, Engineering, and Medicine. 2016. Evaluating Methods for Determining Interior Noise Levels Used in Airport Sound Insulation Programs. Washington, DC: The National Academies Press. doi: 10.17226/23473.
×
Page 69
Page 70
Suggested Citation:"Chapter 4 - Findings and Applications." National Academies of Sciences, Engineering, and Medicine. 2016. Evaluating Methods for Determining Interior Noise Levels Used in Airport Sound Insulation Programs. Washington, DC: The National Academies Press. doi: 10.17226/23473.
×
Page 70
Page 71
Suggested Citation:"Chapter 4 - Findings and Applications." National Academies of Sciences, Engineering, and Medicine. 2016. Evaluating Methods for Determining Interior Noise Levels Used in Airport Sound Insulation Programs. Washington, DC: The National Academies Press. doi: 10.17226/23473.
×
Page 71
Page 72
Suggested Citation:"Chapter 4 - Findings and Applications." National Academies of Sciences, Engineering, and Medicine. 2016. Evaluating Methods for Determining Interior Noise Levels Used in Airport Sound Insulation Programs. Washington, DC: The National Academies Press. doi: 10.17226/23473.
×
Page 72
Page 73
Suggested Citation:"Chapter 4 - Findings and Applications." National Academies of Sciences, Engineering, and Medicine. 2016. Evaluating Methods for Determining Interior Noise Levels Used in Airport Sound Insulation Programs. Washington, DC: The National Academies Press. doi: 10.17226/23473.
×
Page 73
Page 74
Suggested Citation:"Chapter 4 - Findings and Applications." National Academies of Sciences, Engineering, and Medicine. 2016. Evaluating Methods for Determining Interior Noise Levels Used in Airport Sound Insulation Programs. Washington, DC: The National Academies Press. doi: 10.17226/23473.
×
Page 74
Page 75
Suggested Citation:"Chapter 4 - Findings and Applications." National Academies of Sciences, Engineering, and Medicine. 2016. Evaluating Methods for Determining Interior Noise Levels Used in Airport Sound Insulation Programs. Washington, DC: The National Academies Press. doi: 10.17226/23473.
×
Page 75
Page 76
Suggested Citation:"Chapter 4 - Findings and Applications." National Academies of Sciences, Engineering, and Medicine. 2016. Evaluating Methods for Determining Interior Noise Levels Used in Airport Sound Insulation Programs. Washington, DC: The National Academies Press. doi: 10.17226/23473.
×
Page 76
Page 77
Suggested Citation:"Chapter 4 - Findings and Applications." National Academies of Sciences, Engineering, and Medicine. 2016. Evaluating Methods for Determining Interior Noise Levels Used in Airport Sound Insulation Programs. Washington, DC: The National Academies Press. doi: 10.17226/23473.
×
Page 77
Page 78
Suggested Citation:"Chapter 4 - Findings and Applications." National Academies of Sciences, Engineering, and Medicine. 2016. Evaluating Methods for Determining Interior Noise Levels Used in Airport Sound Insulation Programs. Washington, DC: The National Academies Press. doi: 10.17226/23473.
×
Page 78
Page 79
Suggested Citation:"Chapter 4 - Findings and Applications." National Academies of Sciences, Engineering, and Medicine. 2016. Evaluating Methods for Determining Interior Noise Levels Used in Airport Sound Insulation Programs. Washington, DC: The National Academies Press. doi: 10.17226/23473.
×
Page 79
Page 80
Suggested Citation:"Chapter 4 - Findings and Applications." National Academies of Sciences, Engineering, and Medicine. 2016. Evaluating Methods for Determining Interior Noise Levels Used in Airport Sound Insulation Programs. Washington, DC: The National Academies Press. doi: 10.17226/23473.
×
Page 80
Page 81
Suggested Citation:"Chapter 4 - Findings and Applications." National Academies of Sciences, Engineering, and Medicine. 2016. Evaluating Methods for Determining Interior Noise Levels Used in Airport Sound Insulation Programs. Washington, DC: The National Academies Press. doi: 10.17226/23473.
×
Page 81
Page 82
Suggested Citation:"Chapter 4 - Findings and Applications." National Academies of Sciences, Engineering, and Medicine. 2016. Evaluating Methods for Determining Interior Noise Levels Used in Airport Sound Insulation Programs. Washington, DC: The National Academies Press. doi: 10.17226/23473.
×
Page 82
Page 83
Suggested Citation:"Chapter 4 - Findings and Applications." National Academies of Sciences, Engineering, and Medicine. 2016. Evaluating Methods for Determining Interior Noise Levels Used in Airport Sound Insulation Programs. Washington, DC: The National Academies Press. doi: 10.17226/23473.
×
Page 83
Page 84
Suggested Citation:"Chapter 4 - Findings and Applications." National Academies of Sciences, Engineering, and Medicine. 2016. Evaluating Methods for Determining Interior Noise Levels Used in Airport Sound Insulation Programs. Washington, DC: The National Academies Press. doi: 10.17226/23473.
×
Page 84
Page 85
Suggested Citation:"Chapter 4 - Findings and Applications." National Academies of Sciences, Engineering, and Medicine. 2016. Evaluating Methods for Determining Interior Noise Levels Used in Airport Sound Insulation Programs. Washington, DC: The National Academies Press. doi: 10.17226/23473.
×
Page 85
Page 86
Suggested Citation:"Chapter 4 - Findings and Applications." National Academies of Sciences, Engineering, and Medicine. 2016. Evaluating Methods for Determining Interior Noise Levels Used in Airport Sound Insulation Programs. Washington, DC: The National Academies Press. doi: 10.17226/23473.
×
Page 86
Page 87
Suggested Citation:"Chapter 4 - Findings and Applications." National Academies of Sciences, Engineering, and Medicine. 2016. Evaluating Methods for Determining Interior Noise Levels Used in Airport Sound Insulation Programs. Washington, DC: The National Academies Press. doi: 10.17226/23473.
×
Page 87
Page 88
Suggested Citation:"Chapter 4 - Findings and Applications." National Academies of Sciences, Engineering, and Medicine. 2016. Evaluating Methods for Determining Interior Noise Levels Used in Airport Sound Insulation Programs. Washington, DC: The National Academies Press. doi: 10.17226/23473.
×
Page 88
Page 89
Suggested Citation:"Chapter 4 - Findings and Applications." National Academies of Sciences, Engineering, and Medicine. 2016. Evaluating Methods for Determining Interior Noise Levels Used in Airport Sound Insulation Programs. Washington, DC: The National Academies Press. doi: 10.17226/23473.
×
Page 89
Page 90
Suggested Citation:"Chapter 4 - Findings and Applications." National Academies of Sciences, Engineering, and Medicine. 2016. Evaluating Methods for Determining Interior Noise Levels Used in Airport Sound Insulation Programs. Washington, DC: The National Academies Press. doi: 10.17226/23473.
×
Page 90
Page 91
Suggested Citation:"Chapter 4 - Findings and Applications." National Academies of Sciences, Engineering, and Medicine. 2016. Evaluating Methods for Determining Interior Noise Levels Used in Airport Sound Insulation Programs. Washington, DC: The National Academies Press. doi: 10.17226/23473.
×
Page 91
Page 92
Suggested Citation:"Chapter 4 - Findings and Applications." National Academies of Sciences, Engineering, and Medicine. 2016. Evaluating Methods for Determining Interior Noise Levels Used in Airport Sound Insulation Programs. Washington, DC: The National Academies Press. doi: 10.17226/23473.
×
Page 92
Page 93
Suggested Citation:"Chapter 4 - Findings and Applications." National Academies of Sciences, Engineering, and Medicine. 2016. Evaluating Methods for Determining Interior Noise Levels Used in Airport Sound Insulation Programs. Washington, DC: The National Academies Press. doi: 10.17226/23473.
×
Page 93
Page 94
Suggested Citation:"Chapter 4 - Findings and Applications." National Academies of Sciences, Engineering, and Medicine. 2016. Evaluating Methods for Determining Interior Noise Levels Used in Airport Sound Insulation Programs. Washington, DC: The National Academies Press. doi: 10.17226/23473.
×
Page 94
Page 95
Suggested Citation:"Chapter 4 - Findings and Applications." National Academies of Sciences, Engineering, and Medicine. 2016. Evaluating Methods for Determining Interior Noise Levels Used in Airport Sound Insulation Programs. Washington, DC: The National Academies Press. doi: 10.17226/23473.
×
Page 95
Page 96
Suggested Citation:"Chapter 4 - Findings and Applications." National Academies of Sciences, Engineering, and Medicine. 2016. Evaluating Methods for Determining Interior Noise Levels Used in Airport Sound Insulation Programs. Washington, DC: The National Academies Press. doi: 10.17226/23473.
×
Page 96
Page 97
Suggested Citation:"Chapter 4 - Findings and Applications." National Academies of Sciences, Engineering, and Medicine. 2016. Evaluating Methods for Determining Interior Noise Levels Used in Airport Sound Insulation Programs. Washington, DC: The National Academies Press. doi: 10.17226/23473.
×
Page 97
Page 98
Suggested Citation:"Chapter 4 - Findings and Applications." National Academies of Sciences, Engineering, and Medicine. 2016. Evaluating Methods for Determining Interior Noise Levels Used in Airport Sound Insulation Programs. Washington, DC: The National Academies Press. doi: 10.17226/23473.
×
Page 98
Page 99
Suggested Citation:"Chapter 4 - Findings and Applications." National Academies of Sciences, Engineering, and Medicine. 2016. Evaluating Methods for Determining Interior Noise Levels Used in Airport Sound Insulation Programs. Washington, DC: The National Academies Press. doi: 10.17226/23473.
×
Page 99
Page 100
Suggested Citation:"Chapter 4 - Findings and Applications." National Academies of Sciences, Engineering, and Medicine. 2016. Evaluating Methods for Determining Interior Noise Levels Used in Airport Sound Insulation Programs. Washington, DC: The National Academies Press. doi: 10.17226/23473.
×
Page 100
Page 101
Suggested Citation:"Chapter 4 - Findings and Applications." National Academies of Sciences, Engineering, and Medicine. 2016. Evaluating Methods for Determining Interior Noise Levels Used in Airport Sound Insulation Programs. Washington, DC: The National Academies Press. doi: 10.17226/23473.
×
Page 101
Page 102
Suggested Citation:"Chapter 4 - Findings and Applications." National Academies of Sciences, Engineering, and Medicine. 2016. Evaluating Methods for Determining Interior Noise Levels Used in Airport Sound Insulation Programs. Washington, DC: The National Academies Press. doi: 10.17226/23473.
×
Page 102
Page 103
Suggested Citation:"Chapter 4 - Findings and Applications." National Academies of Sciences, Engineering, and Medicine. 2016. Evaluating Methods for Determining Interior Noise Levels Used in Airport Sound Insulation Programs. Washington, DC: The National Academies Press. doi: 10.17226/23473.
×
Page 103
Page 104
Suggested Citation:"Chapter 4 - Findings and Applications." National Academies of Sciences, Engineering, and Medicine. 2016. Evaluating Methods for Determining Interior Noise Levels Used in Airport Sound Insulation Programs. Washington, DC: The National Academies Press. doi: 10.17226/23473.
×
Page 104
Page 105
Suggested Citation:"Chapter 4 - Findings and Applications." National Academies of Sciences, Engineering, and Medicine. 2016. Evaluating Methods for Determining Interior Noise Levels Used in Airport Sound Insulation Programs. Washington, DC: The National Academies Press. doi: 10.17226/23473.
×
Page 105
Page 106
Suggested Citation:"Chapter 4 - Findings and Applications." National Academies of Sciences, Engineering, and Medicine. 2016. Evaluating Methods for Determining Interior Noise Levels Used in Airport Sound Insulation Programs. Washington, DC: The National Academies Press. doi: 10.17226/23473.
×
Page 106
Page 107
Suggested Citation:"Chapter 4 - Findings and Applications." National Academies of Sciences, Engineering, and Medicine. 2016. Evaluating Methods for Determining Interior Noise Levels Used in Airport Sound Insulation Programs. Washington, DC: The National Academies Press. doi: 10.17226/23473.
×
Page 107
Page 108
Suggested Citation:"Chapter 4 - Findings and Applications." National Academies of Sciences, Engineering, and Medicine. 2016. Evaluating Methods for Determining Interior Noise Levels Used in Airport Sound Insulation Programs. Washington, DC: The National Academies Press. doi: 10.17226/23473.
×
Page 108
Page 109
Suggested Citation:"Chapter 4 - Findings and Applications." National Academies of Sciences, Engineering, and Medicine. 2016. Evaluating Methods for Determining Interior Noise Levels Used in Airport Sound Insulation Programs. Washington, DC: The National Academies Press. doi: 10.17226/23473.
×
Page 109
Page 110
Suggested Citation:"Chapter 4 - Findings and Applications." National Academies of Sciences, Engineering, and Medicine. 2016. Evaluating Methods for Determining Interior Noise Levels Used in Airport Sound Insulation Programs. Washington, DC: The National Academies Press. doi: 10.17226/23473.
×
Page 110
Page 111
Suggested Citation:"Chapter 4 - Findings and Applications." National Academies of Sciences, Engineering, and Medicine. 2016. Evaluating Methods for Determining Interior Noise Levels Used in Airport Sound Insulation Programs. Washington, DC: The National Academies Press. doi: 10.17226/23473.
×
Page 111
Page 112
Suggested Citation:"Chapter 4 - Findings and Applications." National Academies of Sciences, Engineering, and Medicine. 2016. Evaluating Methods for Determining Interior Noise Levels Used in Airport Sound Insulation Programs. Washington, DC: The National Academies Press. doi: 10.17226/23473.
×
Page 112

Below is the uncorrected machine-read text of this chapter, intended to provide our own search engines and external engines with highly rich, chapter-representative searchable text of each book. Because it is UNCORRECTED material, please consider the following text as a useful but insufficient proxy for the authoritative book pages.

41 C H A P T E R 4 4.1 Aircraft Flyovers The aircraft flyover test is used in a number of sound insulation programs. This method simultaneously measures the exterior free-field incident sound of flyovers and the diffuse sound field in the test room within the structure. The difference in the two A-weighted SEL values is subtracted to yield the NLR of the room. In practice, synchronized digital programmable sound level meters (SLMs) are positioned in the free field outside the home and in the room to simul- taneously record the SEL of each flyover event. The SLMs record multiple SEL events, allowing for computation of the NLR for each event and statistics for a series of flyover events. Typically, multiple interior rooms are measured simulta- neously. These measurements generally follow a national standard for field NLR measurement: ASTM E966. The flyover method is assumed to provide a reasonable approximation of the NLR in each room, but does have limitations and sources of error, as indicated in detail later in this section. Summary: The research team conducted aircraft flyover measurements at ten homes near the San Diego International Airport (SAN) and at four homes in Boston. The conclusions are as follows: • Measurements need to be conducted in vacant homes, as occupant contamination easily occurs. • Outdoor microphones should be set in the free field or flush mounted to the ground or build- ing façade (Figure 4-1). Near field measurement [1 m to 2 m from (3.3 ft. to 6.6 ft.) façade] is not suggested. • Measurements should be made in one-third-octave bands from 50 Hz through 5 kHz or octave bands from 63 Hz through 5 kHz, and the A-weighted value would be computed rather than using direct A-weighted measurements. • Measurement sample time should not be faster than every 0.5 sec (500 ms). • In general, noise reduction is higher with the flyover measurement method than the loud- speaker measurement method; a correction of 2 to 4 dB is suggested to compensate for ground reflection and/or reflected noise off the façade under test. • To determine the NLR from sequential measurement and computation of single events the research team suggests (1) sorting NLR values by standard deviation from the mean from highest to lowest, (2) computing the standard deviation and confidence interval for the initial list, then (3) sequentially deleting the top value in the list until the desired standard deviation and/or confidence intervals are obtained. • The research team finds the use of a properly placed single microphone in interior rooms to be adequate for most NLR measurements. Findings and Applications

42 Evaluating Methods for Determining Interior Noise Levels Used in Airport Sound Insulation Programs 4.1.1 Measurement Procedures Aircraft flyover measurements are made using a pair of synchronized digital programmable SLMs. One is located outside in the free field away from reflecting surfaces and extraneous noise sources; the other is located in the room to be measured for NLR, locating it away from the imping- ing façade and away from locations amplifying room acoustic effects. Both meters run continu- ously, recording aircraft flyover noise and all noise between aircraft flyover events. The meters are interrogated by computer software that matches events by time, computes the SEL value of the simultaneous events, and subtracts the interior SEL from the exterior SEL to yield the NLR. The preferred microphone locations are either free field, away from reflecting surfaces or flush mounted on either the ground or against the façade (whichever provides the best normal incidence to the flyovers). Free-field measurements require that 2 to 4 dB be subtracted from measured values to eliminate the influence of ground reflections and reflections from the façade under test. The reasoning for the correction is as follows: ASTM E966 and ISO 1996-2 include a correction for reflected noise when the sound source (e.g., loudspeaker) is located a horizon- tal distance away from the façade. The flyover measurement is similar, except that it is in the vertical plane and the reflection comes from the ground rather than the façade. There are also secondary reflections that occur with both the flyover and exterior loudspeaker methods. For the loudspeaker measurement, there are reflections from the ground and neighboring buildings. For the flyover measurement, there are reflections from neighboring buildings and the façade of the home being tested. In general, these secondary reflections are minor. Other factors that affect the correction are ground surface type (e.g., soft soil versus hard concrete), microphone location (height above ground or distance from reflecting objects), and aircraft angle. Also, as verification, the research team conducted both measurements and computer model- ing. First, the research team conducted simultaneous exterior measurements of flyovers with Figure 4-1. Exterior microphone for flyover measurement in Boston.

Findings and Applications 43 one microphone flush with the ground and one microphone 2 m (6 ft.) above the ground. The team measured a 2 dB difference between the flush and elevated positions, thus sup- porting the 2 dB correction used in the analysis. Second, the research team modeled various outdoor conditions in acoustical modeling software for flyover measurements (e.g., con- crete vs. grass, the effect of neighboring building reflections) and found that exterior noise levels can increase by 1 to 2 dB with reflective ground surfaces and numerous surrounding buildings. Surface-mounted microphone measurements require that 5 dB (per ASTM E966) be sub- tracted for the reflected near field. Figure 4-2 shows the flush microphone configuration. Modern programmable digital SLMs allow a variety of sampling rates, often as rapidly as every few milliseconds. They also enable recording of overall A-weighted levels, octave band, or one-third octave band values. Care must be taken in the setting of SLMs. It may seem advisable to use very rapid sampling rates in order to maximize the sound level data, but this incurs a problem. Digital SLMs have a limitation of sampling rate with bandwidth; very fast sampling at wide bandwidths (such as A-weighting) produces significant errors. The research team proposes a sampling rate of a half second (i.e., 500 ms). As discussed in Section 4.9.2, a situation arises with moderate flyover sound levels and sub- stantial ambient interior sound levels where the interior level matches or exceeds the exterior level. This only occurs in the high frequencies where flyover noise is greatly attenuated. In order to avoid this situation, the research team proposes that all interior and exterior sound level data be recorded in one-third octave bands from 50 Hz through 5 kHz or octave bands from 63 Hz through 4 kHz, but at no higher frequencies; A-weighted sound levels are then computed in post processing. Therefore, overall A-weighted monitoring is not suggested. Figure 4-2. Flush microphone mounting.

44 Evaluating Methods for Determining Interior Noise Levels Used in Airport Sound Insulation Programs 4.1.2 Data Analyses The example results of the flyover measurement procedure above are shown in Table 4-1. This is from a measurement for the living room of test subject San Diego #1. This event was selected because it recorded the many flyover events. There are several obvious “outliers” or events where the NLR value is obviously incorrect (see events at 10:05:52 and 10:12:13). These occur from corrupted or missing data. San Diego #1: Living Room Flight # Aircra SELout SELin NLR ANOMS Event Time 9:27:42 66.1 9:30:35 61.2 9:34:02 SWA4602 B737 90.30 62.7 27.6 9:35:27 SWA758 B737 86.40 61.6 24.8 9:37:08 SWA1582 B737 82.00 53.7 28.3 9:40:34 AWE2040 A320 88.40 59.7 28.7 9:42:33 SWA4105 B733 89.50 60.7 28.8 9:44:04 SKW2599 CRJ2 85.90 57.6 28.3 9:45:37 SWA736 B737 87.20 59.6 27.6 9:47:35 ASA718 B738 87.30 59.9 27.4 9:49:32 SKW6323 E120 84.90 57.9 27.0 9:50:59 SWA4657 B737 86.60 58.9 27.7 9:53:11 CPZ5743 E170 85.90 57.9 28.0 9:57:14 N301KR LJ45 84.40 55.3 29.1 9:59:29 SWA1582 B737 90.30 62.5 27.8 10:01:24 UAL289 A319 86.60 58.8 27.8 10:02:59 ASA240 B739 88.70 61.8 26.9 10:05:52 JBU619 A320 63.40 59.5 3.9 10:09:01 G/A JET 82.40 54.4 28.0 10:12:13 RGY710 BE40 81.80 10:15:08 UAL1563 B738 88.50 60.8 27.7 10:19:13 UAL229 B752 90.00 60.8 29.2 10:21:26 DAL833 B739 88.80 61.1 27.7 10:23:47 EJA669 C56X 82.60 54.8 27.8 10:26:22 AWE581 A320 87.10 58.8 28.3 10:28:05 SWA4791 B738 82.30 56.7 25.6 10:30:18 UAL709 A320 84.50 58.2 26.3 10:31:46 NKS470 A319 86.20 58.3 27.9 10:36:17 SWA2468 B738 88.80 61.8 27.0 10:38:24 DAL1687 MD90 85.40 58.3 27.1 10:40:35 SKW171Z CRJ9 85.20 58.9 26.3 10:43:21 SWA777 B737 89.00 62.8 26.2 10:45:18 SWA4791 B738 90.20 62.2 28.0 10:47:29 SWA633 B733 90.20 60.3 29.9 10:49:34 DJR829 C550 79.30 50.4 28.9 10:51:37 SWA2397 B738 89.00 61.6 27.4 10:53:02 AAL1565 B738 89.50 61.9 27.6 10:54:54 SKW2621 CRJ2 85.80 56.5 29.3 10:56:55 AAL1228 B752 91.40 63.6 27.8 10:59:33 SWA4679 B737 88.70 66.0 22.7 11:01:08 SWA1532 B737 87.20 59.3 27.9 11:03:17 SWA238 B737 88.50 60.0 28.5 11:05:07 JBU189 A320 81.00 52.6 28.4 Table 4-1. Flyover data analysis examples.

Findings and Applications 45 A general rule of thumb is to measure about 25 flyover events to obtain a valid assessment of the NLR to within 0.5 dB. This varies with each individual measurement site. A site with quieter flyovers, noisy interior, and/or background (exterior) noise will require more measurements to converge to a steady average value. It is typical to compute a running average of the NLR and the standard deviation with suc- cessive flyover events. As one progresses down the running average, the change in average and standard deviation value will become less and eventually stabilize to within a small range, such as 0.5 dB. This convergence may be increased by first discarding outliers. However, there is no established rule for identifying outliers, and there is no standard by how much a value deviates from the mean before it is discarded. Some consultants suggest two standard deviations. Even with outliers, the rate of convergence depends upon the order in which values 11:38:30 53.5 11:40:36 65.3 11:43:03 JAL66 B788 90.00 61.5 28.5 11:45:49 N818SE C650 80.70 49.9 30.8 11:50:42 63.4 11:55:52 61.6 11:59:52 57.9 12:01:50 SWA4479 B737 87.80 59.8 28.0 12:11:07 ASA238 B738 88.90 60.6 28.3 12:13:08 DAL2378 A320 88.80 60.5 28.3 12:15:45 DAL2267 B739 88.50 60.7 27.8 12:17:58 UAL284 A320 86.90 59.1 27.8 12:20:27 WJA1435 B738 88.20 56.3 31.9 12:23:34 SKW2611 CRJ2 84.00 56.0 28.0 12:25:08 44.2 12:27:33 SWA4317 B737 86.30 57.3 29.0 12:31:27 UAL356 A320 87.20 59.4 27.8 12:44:40 SWA3538 B737 89.80 62.1 27.7 12:47:52 SWA2076 B737 85.40 57.4 28.0 12:49:29 UAL1155 B739 88.50 60.5 28.0 12:55:16 AAL2382 B738 89.20 61.3 27.9 12:57:08 CPZ5749 E170 86.30 12:59:47 DAL513 MD90 86.00 57.3 28.7 13:03:07 SWA2280 B737 80.20 13:05:23 SWA2052 B733 91.50 Note: The 2 dB reflecon correcon has not been applied to the above data. A blank indicates either no event recorded at interior or no event noted by the airport monitoring system. 11:07:50 SWA2029 B737 85.70 57.3 28.4 11:09:37 SWA665 B737 87.70 61.2 26.5 11:17:50 JBU189 A320 82.40 56.4 26.0 11:20:02 DAL2506 B738 89.80 62.4 27.4 11:22:13 SWA2144 B737 86.90 59.1 27.8 11:24:41 SKW6325 E120 85.00 11:28:17 SWA4761 B737 86.90 11:30:33 VRD956 A319 87.10 11:32:36 JBU189 A320 87.90 59.5 28.4 11:34:31 SWA3339 B737 87.20 58.8 28.4 11:36:08 FFT551 A320 87.70 59.9 27.8 San Diego #1: Living Room Flight # Aircra SELout SELin NLR Table 4-1. (Continued).

46 Evaluating Methods for Determining Interior Noise Levels Used in Airport Sound Insulation Programs of various variances from the mean are encountered. That is, the chronology of the events affects the analysis. To avoid this problem, the research team recommends that all flyover event values, including even the most obvious outliers, first be ordered by their deviation from the mean from highest to lowest. Then events may be discarded sequentially down the list with a new mean and standard deviation computed for the new list. This allows for much more rapid convergence to a small standard deviation. Also, it may be useful to sequentially compute the convergence interval for each new list. One standard that may be considered in selecting a mean and standard deviation is a confidence interval (CI) of ±0.5 dB at the 95% probability level. That is, 95% of all events in the sample fall within ±0.5 dB of the mean. The confidence interval is: CI X Z n = ± σ  Where X = mean value s = standard deviation n = number of events Z = random variable related to probability level (Z = 1.96 for 95%: Z = 1.645 for 90%; and Z = 2.576 for 99%) Note that, unlike the standard deviation, the CI does not continue to drop with fewer values. This is because the CI is proportional to the standard deviation and inversely proportional to the square root of the number of values. Table 4-2 is a table of this method for the data presented in Table 4-1. San Diego #1—Sorted by Standard Deviaon n NLR |NLR NLRAvg| NLR AVG Std Dev 95% CI 65 80.7 52.4 80.7 64 3.9 24.4 28.3 7.35 ± 1.80 63 22.7 5.6 27.5 3.25 ± 0.80 62 31.9 3.6 27.8 1.27 ± 0.32 61 24.8 3.5 27.9 1.10 ± 0.28 60 25.6 2.7 27.9 0.98 ± 0.25 59 30.8 2.5 27.9 0.90 ± 0.23 58 26.0 2.3 28.0 0.85 ± 0.22 57 26.2 2.1 27.9 0.77 ± 0.20 56 26.3 2.0 27.9 0.74 ± 0.19 55 26.3 2.0 28.0 0.70 ± 0.19 54 26.5 1.8 28.0 0.67 ± 0.18 53 29.9 1.6 28.0 0.64 ± 0.17 52 26.9 1.4 28.1 0.61 ± 0.16 51 27.0 1.3 28.0 0.55 ± 0.15 50 27.0 1.3 28.1 0.54 ± 0.15 49 27.1 1.2 28.1 0.52 ± 0.15 48 29.3 1.0 28.1 0.50 ± 0.14 47 29.2 0.9 28.1 0.48 ± 0.14 46 27.4 0.9 28.1 0.46 ± 0.13 45 27.4 0.9 28.1 0.43 ± 0.13 Table 4-2. Sorting of flyovers to find outliers.

Findings and Applications 47 44 27.4 0.9 28.1 0.42 ± 0.13 43 29.1 0.8 28.1 0.42 ± 0.12 42 29.0 0.7 28.1 0.41 ± 0.12 41 27.6 0.7 28.1 0.38 ± 0.12 40 27.6 0.7 28.1 0.36 ± 0.11 39 27.6 0.7 28.1 0.35 ± 0.11 38 28.9 0.6 28.1 0.35 ± 0.11 37 27.7 0.6 28.1 0.34 ± 0.11 36 27.7 0.6 28.1 0.32 ± 0.10 35 27.7 0.6 28.1 0.32 ± 0.11 34 27.7 0.6 28.1 0.32 ± 0.11 33 28.8 0.5 28.1 0.31 ± 0.11 32 27.8 0.5 28.1 0.31 ± 0.11 31 27.8 0.5 28.1 0.29 ± 0.10 30 27.8 0.5 28.1 0.29 ± 0.10 29 27.8 0.5 28.1 0.29 ± 0.10 28 27.8 0.5 28.1 0.29 ± 0.11 27 27.8 0.5 28.2 0.28 ± 0.11 26 27.8 0.5 28.2 0.28 ± 0.11 25 27.8 0.5 28.2 0.28 ± 0.11 24 27.8 0.5 28.2 0.27 ± 0.11 23 28.7 0.4 28.2 0.26 ± 0.11 22 28.7 0.4 28.2 0.25 ± 0.11 21 27.9 0.4 28.2 0.24 ± 0.10 20 27.9 0.4 28.2 0.22 ± 0.09 19 27.9 0.4 28.2 0.21 ± 0.10 18 28.0 0.3 28.2 0.21 ± 0.09 17 28.0 0.3 28.2 0.20 ± 0.09 16 28.0 0.3 28.2 0.19 ± 0.09 15 28.0 0.3 28.3 0.19 ± 0.10 14 28.0 0.3 28.3 0.18 ± 0.10 13 28.0 0.3 28.3 0.17 ± 0.09 12 28.0 0.3 28.3 0.16 ± 0.09 11 28.5 0.2 28.3 0.13 ± 0.08 San Diego #1—Sorted by Standard Deviaon n NLR |NLR NLRAvg| NLR AVG Std Dev 95% CI 10 28.5 0.2 28.4 0.08 ± 0.05 9 28.4 0.1 28.4 0.07 ± 0.05 8 28.4 0.1 28.3 0.05 ± 0.04 7 28.4 0.1 28.3 0.05 ± 0.04 6 28.4 0.1 28.3 0.05 ± 0.04 5 28.3 0.0 28.3 0.04 ± 0.04 4 28.3 0.0 28.3 0.00 ± 0.00 3 28.3 0.0 28.3 0.00 ± 0.00 2 28.3 0.0 28.3 0.00 ± 0.00 1 28.3 0.0 28.3 0.00 ± 0.00 Note: The 2 dB reflecon correcon has not been applied to the above data. A blank cell indicates either no event recorded at interior or no event noted by the airport monitoring system. Table 4-2. (Continued).

48 Evaluating Methods for Determining Interior Noise Levels Used in Airport Sound Insulation Programs 4.1.3 Sources of Error: Overview Reflections: Up to 4 dB increase in sound level from ground reflections and reflections from the façade under test (if the outdoor microphone is located near the building façade). This requires a -2 to -4 dB adjustment to all exterior measurements. Non-NEM fleet mix: The sound spectra of the aircraft flyover samples should reflect the energy average for all aircraft used in the NEM from the FAR Part 150 study for the airport. Average annual fleet mix may not be measured on a single day. Typical flyover measurements record aircraft during a single operation type (i.e., all landings or all takeoffs) and cannot reflect that of the annual mix. Extraneous noise sources: Non-flyover noise, both on the exterior and interior, is recorded and is included with the aircraft noise measurements. These sources include occupants, local vehicles, construction, recreation, and other neighborhood activities. Non-reverberant sound field: A single stationary microphone in a room does not give a good measure of the diffuse sound field. Laboratory tests use large rooms of special dimensions, often with moving microphones and/or vanes, and nonparallel walls to minimize standing wave effects. Microphones in small rooms are significantly influenced by location, particularly with pre-retrofit testing where the location relative to a poor sound-attenuating window may have considerable effect. Room absorption: The total NLR is primarily from the CTL characteristics of the structure but also from the room acoustics controlled by the size and absorptive properties of the receiv- ing room (NLR = TL ± room absorption). Therefore, a considerable change in room furnishing between the pre- and post-retrofit testing causes a significant change in room absorption and will affect the measured NLR. Varying sound spectra: Different aircraft under different operating conditions are recorded for the pre-construction and post-construction acoustical measurements. These different operating conditions result in different spectra and incidence, therefore posing another source of error. 4.1.3.1 Reflections The NLR results for all rooms in all homes were compared with those from the ground-level and elevated loudspeaker measurements. A strong trend was found that the NLR values from the flyovers exceeded those measured by the loudspeaker methods by approximately 2 dB. Assum- ing that the main noise transmission paths are the same for both the flyover and the loudspeaker methods, this indicated that the exterior flyover SEL values were biased 2 dB high, since the inte- rior measurements for both methods were the same. This 2 dB increase is discussed in acoustical standards and may arise from ground reflections in the neighborhoods measured. There was no discernible trend between hardscape or softscape ground surfaces. Based on the above, all flyover data in this report was corrected by 2 dB. It should be noted that this correction is not commonly used by acoustic consultants in the existing airport sound insulation programs, and more research is needed to further understand the correction for ground reflection. It also appears, based on measurements and modeling, that some reflections may occur if the outdoor microphone is placed near the home under test. The combined ground and façade reflection correction may approach 4 dB, but more research and analysis is needed. 4.1.3.2 Non-NEM Fleet Mix The ideal procedure for measuring the NLR at a home would be to use an exterior sound spectrum representing the energy average of all aircraft operations from the NEM for the design year. This would entail computing the spectral contribution from all aircraft types, climb profiles,

Findings and Applications 49 power settings, volume of daytime and nighttime operations, landings, etc. used to prepare the NEM. This would be a practically impossible task. However, the noisiest aircraft operations dominate the noise contribution at any location. For instance, it would take 10, 80-dB-SEL events to provide the same NEM contribution as a single 90-dB-SEL event. When recording the NLR results from a series of flyovers, consultants typically compute the arithmetic average of valid NLR results. But, this may not provide the best estimate of the NLR because: • The loudest events dominate the exterior noise exposure value from the NEM, as explained above, and • The loudest events are the least affected by extraneous noise sources, because they have a greater capability to mask, or drown out, the extraneous noise. For this study, the research team compared the arithmetic average results with energy average results. Energy average results strongly bias the loudest events. The equations for the two averages are: Arithmetic average: 1NLR NLRi n i n∑ = = Energy average: NLR SEL SELout in = ↑    − ↑    10 10 10 10 10 log       == ∑∑ i n i n n 11 Table 4-3 shows the differences between arithmetic average and energy average for the flyover tests of 10 homes near San Diego International Airport (SAN). The results in Table 4-3 show no clear trend in the difference between the arithmetic and energy average NLR values. The mean difference between the two is an almost negligible 0.1 dB. Biasing the NLR values from the loudest events may not always be advisable since extraneous noise in a single loud event tends to override the average of other uncontaminated events. 4.1.3.3 Extraneous Noise Sources The potential for contamination from extraneous noise sources exists with every NLR mea- surement technique. An advantage of the flyover test method is the array of results allowing for statistical assessment and identification of outlying, or far-off, NLR results not close to the other values. Two methods of identifying outlying results for elimination are: • Flagging and eliminating those NLR values that are two standard deviations from the mean and • Using the median rather than the mean results. Flagging those events two standard deviations from the mean is a standard statistical tech- nique. However, it is arbitrary in that it provides no means for identifying the source of the deviation; that particular event may in fact be valid, but for reasons not understood. Similar arbitrary data screening techniques are also available such as a more lenient three standard devi- ations from the mean, or simply some static value for the mean, such as ±3 dB from the mean. Using the median in lieu of the mean is a simpler way of screening outliers. The mean, or aver- age, sums all values and divides by the number of values to obtain the result. The median, on the other hand, is the L50 percentile or that value where half of the values exceed it and the other half are below it. Table 4-4 compares the average and median values for the SAN flyover NLR events. It is evident from Table 4-4 that the mean and median values are entirely similar for “well- behaved” statistical results; the average difference between the mean and median values in the table is 0 dB. Therefore, median values seem to provide a good statistical sample.

50 Evaluating Methods for Determining Interior Noise Levels Used in Airport Sound Insulation Programs 4.1.3.4 Non-Reverberant Sound Field Acoustical consultants have long recognized the lack of an ideal reverberant sound field in measuring rooms within homes. This problem exists with all measurement methods, but is ame- liorated to some degree with the loudspeaker method where the measurement within the room is attended by manually moving the microphone to achieve a better spatial average of the sound field. The long duration of measurements with the flyover method makes attended measure- ments infeasible, so a single stationary microphone is typically used. This microphone is located away from the incident building elements to avoid bias from the particular TL properties of nearby building elements. Additionally, microphone locations also avoid corner areas and sites midway between parallel wall surfaces in order to minimize standing wave effects. This project studies the effects of a single stationary microphone in two ways, both taken from the laboratory standard for TL testing: • Use of multiple microphones in the receiving room to view differences in interior noise levels and to average their results and • Use of a moving microphone as is the standard practice for laboratory TL testing. Table 4-5 shows the test results for the various rooms tested with moving microphone and/or with multiple microphones. Residence Room Arithmec Energy Difference Average (dB) Average (dB) (Energy Arith), dB San Diego #1 Living Room 26.3 26.4 0.1 Dining Room 28.0 28.1 0.1 San Diego #2 Living Room 24.3 24.3 0.0 Bedroom 4 20.4 20.4 0.0 San Diego #3 Living Room 23.7 23.8 0.1 Master BR 24.7 24.8 0.1 San Diego #4 Living Room 23.2 23.2 0.0 Bedroom 1 26.3 26.4 0.1 San Diego #5 Living Room 24.0 24.3 0.3 Bedroom 1 30.0 30.1 0.1 San Diego #6 Living Room 24.9 25.0 0.1 Dining Room 27.4 27.5 0.1 San Diego #7 Living Room 25.1 25.2 0.1 Family Room 26.5 26.7 0.2 San Diego #8 Dining Room 19.4 19.6 0.2 Master BR 22.2 22.5 0.3 San Diego #9 Living Room 21.8 22.0 0.2 Bedroom 1 31.7 31.9 0.2 San Diego #10 Bedroom 1 17.5 17.6 0.1 Bedroom 2 25.0 25.2 0.2 Average 0.1 Table 4-3. Flyover measurements, comparison of arithmetic average and energy average NLR values.

Findings and Applications 51 Residence Room Arithmec Average Median Difference (Arith Median) San Diego #1 Living Room 26.3 26.3 0.0 Dining Room 28.0 28.2 -0.2 San Diego #2 Living Room 24.3 24.2 0.1Bedroom 4 20.4 20.4 0.0 San Diego #3 Living Room 23.7 24.0 -0.3Master BR 24.7 24.4 0.3 San Diego #4 Living Room 23.2 23.5 -0.3Bedroom 1 26.3 26.3 0.0 San Diego #5 Living Room 24.0 24.3 -0.3Bedroom 1 30.0 30.2 -0.2 San Diego #6 Living Room 24.9 24.9 0.0 Dining Room 27.4 27.5 -0.1 San Diego #7 Living Room 25.1 25.1 0.0 Family Room 26.5 26.8 -0.3 San Diego #8 Master BR 22.2 21.7 0.5 Dining Room 19.4 19.2 0.2 San Diego #9 Living Room 21.8 21.9 -0.1Bedroom 1 31.7 31.6 0.1 San Diego #10 Bedroom 1 17.5 17.5 0.0Bedroom 2 25.0 25.3 -0.3 Average 0.0 Table 4-4. Flyover measurements, comparison of average and median NLR events. Residence Room Mic 1 Mic 2 Rotang Average Std. Dev. 95%CI Mic Mic1 : Mic2 All Values San Diego #1 Living Room 25.9 27.4 25.6 26.7 0.79 ±0.89 San Diego #2 Living Room 24.4 23.9 24.5 24.2 0.26 ±0.30 San Diego #3 Living Room 24.0 23.1 24.0 23.6 0.45 ±0.62 San Diego #4 Bedroom 1 26.9 25.8 26.9 1.14 ±1.58 San Diego #5 Living Room 24.2 23.8 24.2 0.39 ±0.54 San Diego #6 Dining Room 26.9 27.8 27.4 27.4 0.37 ±0.42 San Diego #7 Family Room 26.2 26.9 26.5 26.6 0.29 ±0.32 San Diego #8 Dining Room 19.1 19.9 19.1 19.5 0.38 ±0.43 San Diego #9 Living Room 20.9 22.4 22.0 21.7 0.63 ±0.72 San Diego #10 Bedroom 1 18.2 17.0 17.2 17.6 0.52 ±0.59 Note: The configuraon of microphones in the San Diego # and San Diego #5 homes allowed for only one fixed microphone in the same room as the rotang microphone. Table 4-5. Comparison of averaged microphone alternative NLR results.

52 Evaluating Methods for Determining Interior Noise Levels Used in Airport Sound Insulation Programs The long boom length of the rotating microphone required that large rooms be measured; most rooms were living rooms. The two stationary microphones were generally placed in oppo- site room quadrants to measure the noise environment in distinct areas. The measurement results in Table 4-5 show: • Little distinction among the independent values, average values, and values from the rotating microphone and • No trend among the measurement techniques; rotating microphone NLR values do not trend either higher or lower than those from the stationary microphones. The rotating microphone selected was the same model used in laboratory testing. However, it never worked properly under power and was therefore rotated manually for all measurements. This complicated the measurements because it required the receiving room to be occupied and added to the potential for noise contamination. 4.1.3.5 Room Absorption The issue with acoustical absorption in rooms exists with all NLR measurement techniques. Typically, an empty room of moderate size without carpeting or curtains may be 3 dB louder than the same room fully furnished. That is, the overall noise reduction is controlled primarily by the sound TL characteristics of the building elements and also by the room absorption. The basic relationship is (DOT-FAA-AEQ-77-9, 1977): 10 610NLR TL log S Ap ( )= − − where S = surface area of the assembly exposed to the noise source and A = the total absorption in the room at the source frequency The absorption varies with frequency as does the TL, so this computation must be done in octave bands or one-third octave bands. Typically the 6 dB term for perfect acoustic reflection is reduced to 5 dB to reflect actual less-than-perfect reflection. The total room absorption may be determined by measuring the room dimensions and the reverberation time (symbol RT60), or time for an impulsive sound to decay by 60 dB. Rever- beration time is measured by emitting a high level broadband impulsive sound, often by a large balloon, and recording the decay rate. Many contemporary SLMs have a feature built in for this measurement. The general relationship between reverberation times is given by the Sabine equa- tion (Bies and Hansen, 2003): 55.25 60RT V S c = ∝ where RT60 = reverberation time (60 dB decay) in seconds V = room volume (m3) c = speed of sound (m/s) ∝ = the acoustical absorption coefficient at the source frequency Those homeowners desiring to qualify for the sound insulation program under the PGL/5100.38D guidelines may increase their chances of qualifying by removing furnishings from rooms to be tested prior to the qualification NLR measurements, thereby increasing measured receiving room sound levels and reducing the reported interior DNL values. 4.1.3.6 Varying Sound Spectra Flyover events of varying sound spectra with the same A-weighted SEL value produce differ- ent NLR results. This is related to the issue of a non-NEM fleet mix (see 4.1.3.2) and it exists with

Findings and Applications 53 other measurement techniques as well. For example, an 80-dB flyover with a concentration of low frequency energy will produce a lower NLR than another 80-dB flyover with more energy concentrated in the higher frequencies. This is due to the TL property of all building assemblies to attenuate sound more effectively in the higher frequencies. The NLR effects of any particular case may be examined by viewing the SEL spectrum of the flyover, the CTL properties of the room, and the room absorption characteristics. To examine this more thoroughly, the research team (a) analyzed the differences in NLR from specific aircraft types from the SAN flyovers measurements and (b) computed the CTL for various typical aircraft in the INM database. Table 4-6 shows the results from the homes for which the research team recorded the most valid flyovers. These are the living room in the San Diego #3 residence for aircraft departures and the San Diego #6 dining room for aircraft arrivals. From Table 4-6, there is no clear NLR trend for the specific aircraft currently operating from SAN. However, certain classes of aircraft do not operate from SAN, so an independent study was subsequently conducted. The effect of spectral changes from other aircraft was computed by selecting the standard spectra for certain classes of aircraft from the database for the FAA standard INM computer pro- gram used to develop the noise contours for the Part 150 studies (DTS-34-FA065-LR1, 1999), and an ideal “mass law” TL curve for a residence. The mass law curve at STC 39 (OITC 31) slopes upward at 8 dB per octave from 24 dB at 125 Hz to 54 dB at 4 kHz. This curve is a good average for the CTL, or the TL from all incident sound on all exposed surfaces of non-retrofit homes. The aircraft spectra are taken from the FAA noise certification tests (under 14CFR36) for the specific aircraft in a class. Aircra San Diego #3 Departures San Diego #6 Arrivals Type Count Average Std Dev Count Average Std Dev All MD 3 23.7 0.2 A319 5 23.5 1.8 4 27.7 0.6 A320 6 24.9 0.4 5 27.3 0.9 B733 5 27.1 0.6 B737 19 23.8 1.1 12 27.4 1.0 B738 13 23.2 1.0 9 27.0 1.2 B739 2 27.3 0.1 B752 3 24.4 0.7 3 27.6 0.6 All CRJ# 3 28.4 1.3 All E## 1 24.2 3 26.6 0.7 Average 24.0 Average 27.4 Std Dev 0.58 Std Dev 0.50 Legend: All MD All MD 80 models A319 Airbus A319 A320 Airbus A320 B733 Boeing 737 300 B737 Boeing 737 700 B738 Boeing 737 800 B739 Boeing 737 900 B752 Boeing 757 200 All CRJ# All Canadair Regional Jet All E## All Embraer turboprops Table 4-6. Comparison of NLR values for several SAN aircraft.

54 Evaluating Methods for Determining Interior Noise Levels Used in Airport Sound Insulation Programs Table 4-7 compares the computed exterior and interior SEL (an integrated measure of the total sound energy of a noise event), values for the aircraft classes, and computes the attendant NLR values. Table 4-7 shows that there is generally little difference among the medium, heavy, and regional jet aircraft classes for the turbofan aircraft. However, the NLR values for the turboprop aircraft show a significant reduction in NLR of about 5 dB. This is due to the strong low frequency com- ponents from prop blade pass frequencies that attenuate less than the more broadband jet noise from turbofan jets. No airport operation is composed exclusively of a single class of aircraft, so the NLR differ- ences in Table 4-7 are significantly moderated in the mix of aircraft for any particular airport. The SEL values and the varying sound spectra both play a significant role in determining the proper noise spectrum to use for evaluating the NLR of a residence in terms of the NEM spec- trum at a particular location. NLR results from turboprop approaches and departures were compared with those from other aircraft. Little change was found, contradicting the theoretical analysis using the FAA INM spectral classes. But the INM data is 16 years old and modern commuter turboprops now typically employ five or six-bladed propellers which are often in the scimitar configuration. These significantly diminish the tones emitted by earlier turboprop versions. Therefore it can be concluded that turboprop spectral changes are generally not a significant factor in NLR assessment. 4.2 Ground-Level Exterior Loudspeaker The ground-level exterior loudspeaker measurement method follows the measurement pro- cedure outlined ASTM E966, and this method is commonly employed by acoustical consultants working on sound insulation projects (Figure 4-3). The method involves locating a loudspeaker approximately 6.1 m to 12.2 m (20 to 40 feet) from the façade of the room under test, at a height of 6 to 8 feet above grade. The only difference between the elevated exterior loudspeaker method (discussed in the following section) and the ground-level method is the height of the loudspeaker above grade. Summary: The research team conducted ground-level loudspeaker measurements at 10 homes in San Diego. In Boston, the research team conducted ground-level loudspeaker measurements in nine homes, but the aircraft flyover source spectrum was only available in four of the homes. The conclusions are as follows: • On average, calculated noise reduction using the OITC spectrum was slightly lower (0.6 dB) than that of the flyover spectrum. This slight difference was similar to the results of the elevated loudspeaker measurements. Aircra Types Departures Approaches SELout NLR SELin SELout NLR SELin B737 300, 3B2, 400, 500 73.5 30.5 43.0 75.4 31.3 44.1 B757 & B767; A300, 310, 320 75.7 31.0 44.7 77.5 33.4 44.1 MD81, 82, 83 72.8 29.9 42.9 72.6 28.7 43.9 B747 10Q, 200, 720A, 420B, 400 77.0 32.4 44.6 77.5 33.8 43.7 2 engine turboprop, DHC6 77.7 24.8 52.9 69.8 27.3 42.5 4 engine turboprop, DHC7, DHC8 77.8 22.8 55.0 69.8 27.3 42.5 Table 4-7. Comparison of SEL values and NLR values for aircraft classes.

Findings and Applications 55 • In general, the results varied little from measurement to measurement; however, there were a few outliers which warranted closer examination. • Similar to the elevated loudspeaker, the NLR decreased by approximately 1 dB when the roof and walls were measured at the exterior (rather than just the walls). • On average, the calculated noise reduction using the OITC spectrum was slightly lower (0.9 dB) than that calculated using the flyover spectrum. • Noise reduction varied by less than 1 dB when measurements were repeated. • Noise reduction increased when the wall and roof were measured, which is the opposite of what happened when the loudspeaker was located inside of the building. 4.2.1 Measurement Procedure The ground-level exterior loudspeaker noise reduction measurement was conducted as follows: 1. A loudspeaker capable of generating 90 dB to 100 dB was mounted at a height of 1.8 to 3.0 meters (6 to 10 feet) above grade, at a distance of, on average, 9.1 m (30 feet) from the building façade. The loudspeaker was pointed in the direction of the room under test, at a horizontal angle of 45° from the façade. 2. Ambient (background) noise measurements were conducted outside and inside of the room under test. The ambient measurement allowed the technician to verify that the noise gener- ated by the loudspeaker was sufficiently above the ambient noise level; corrections to the measurement were made if the ambient noise level approached the level of the loudspeaker. This ensured that noise reduction was accurately measured and quantified. 3. Pink noise was generated by the loudspeaker. A measurement of the diffuse sound field just outside of the room under test [e.g., 1 to 2 meters (3 to 6 feet) from the façade] was made. This measurement consisted of a spatial average of the noise levels at the façade (and roof, where applicable). 4. A measurement of the loudspeaker-generated pink noise was then made inside of the room. Again, a spatial average was conducted with the technician maintaining a minimum distance of 0.3 to 0.6 m (1 to 2 feet) between the microphone and walls, ceilings, floors, etc. 5. The one-third octave measurement data (50 Hz to 5 kHz) were then analyzed in a spreadsheet, where the interior pink noise level was subtracted from the exterior pink noise level. Three cor- rections were then applied: (a) the subtraction of ambient noise, (b) a 2-dB correction (based Figure 4-3. Ground-level loudspeaker measurement in San Diego.

56 Evaluating Methods for Determining Interior Noise Levels Used in Airport Sound Insulation Programs on ASTM E966-10) to account for the reverberant noise build-up at the façade, and (c) a correction to account for the different frequency spectrum of an aircraft as compared to pink noise. These corrections are discussed in more detail in the following section. 4.2.2 Data Analysis and Correction Factors To calculate the amount of noise reduction provided by the building envelope, the data must be analyzed and corrected after completion of the measurement. There are three primary cor- rections that are made to the measured data; each is summarized below: 4.2.2.1 Ambient Noise Correction Ambient noise can affect the accurate determination of noise reduction. For the exterior, ground-level loudspeaker measurement, ambient noise is typically only an issue inside of the room under test. This is because exterior ambient noise is significantly below the level of the loudspeaker (e.g., 50-dB ambient as compared to 90-dB pink noise from the loudspeaker); mea- surements are not conducted during an aircraft overflight, as overflights could approach the noise level of the loudspeaker. Inside the room, ambient noise is generated by household items such as refrigerators, HVAC systems, and other domestic equipment. Typically, the technician will turn off noisy devices; however it is not always possible to do so. If the ambient noise level is within 10 dB of the pink noise level inside the room, then the ambient noise level is logarithmically subtracted from the pink noise level. This ensures the cal- culated noise reduction is only based upon pink noise intrusion through the building envelope, and not contaminated by interior noise sources. 4.2.2.2 Reverberant Noise Build-Up The second correction applied is to account for reverberant (reflected) noise build-up at the façade. This correction is outlined in ASTM E966 and ISO 1996-2. When noise from the loud- speaker impinges on a building façade, some of the noise transmits through the façade into the residence, while some of the noise reflects back away from the façade. At a measurement distance of 1 to 2 meters (3 to 6 feet) from the façade, this reflected noise yields a noise level 2 to 3 dB higher than what would be measured without the presence of the building façade. Since the measure of concern is noise that transmits through the building envelope, it is important to eliminate the reverberant noise from the measurement data. ASTM E966 (2004) stipulates a 3-dB correction, while E966 (2010) stipulates a 2-dB correction. Effectively, correcting for the reverberant field reduces the measured noise reduction by 2 to 3 dB (depending on which cor- rection factor is used). The correction was lowered in the 2010 version of the standard, as purportedly the 2-dB cor- rection more closely aligns to actual field experience (per ASTM), while the 3-dB correction is a laboratory/theoretical value. Based on analysis of the measurement data, the 2-dB correction factor more closely aligns with the true noise reduction of a façade. 4.2.2.3 Spectral Correction As discussed in Section 4.2.1, pink noise is generated by the loudspeaker. Pink noise is used because it is a reference sound source with equal energy across all frequency bands, and it is not logistically feasible to accurately generate the same frequency spectrum as generated by typical aircraft. However, the FAA eligibility standards are based upon aircraft noise levels, and not pink noise levels. As such, it is necessary to “convert” pink noise reduction into aircraft noise reduc- tion. In order to convert pink noise reduction to aircraft noise reduction, the pink noise reduction is subtracted from a reference aircraft spectrum to yield a theoretical aircraft noise level inside of the tested room.

Findings and Applications 57 While the conversion utilizes standard acoustical calculations, there is no standard guidance on how to determine the reference aircraft spectrum. Many consultants involved with sound insulation programs measure multiple arrivals and departures at a given airport and then aver- age these events to calculate the reference spectrum. However, there are many questions raised by this practice: (a) how many aircraft flyovers need to be measured to produce a statistically valid sample, (b) how does one account for homes located at varying distances from the runway ends (i.e., the aircraft spectrum at a home far from the runway will be different from that of a home near a runway), and (c) how does one account for future aircraft fleet mix changes. For this report, the research team has converted pink noise reduction into aircraft noise reduc- tion using two reference spectra: the OITC reference spectrum and the average aircraft spectrum measured at each test home over a period of 3 to 5 hours (measured as a part of the flyover noise reduction measurements). The research team has analyzed the difference in measured noise reduction using the OITC and the average aircraft spectrum measured at each test home. Table 4-8 shows the noise reduction calculated using both of these frequency spectra. For the San Diego measurements, ground-level loudspeaker measurements were conducted in 10 homes. At Boston, ground-level loudspeaker measurements were conducted in nine homes, but aircraft flyover noise spectrum was only available in four out of the nine homes. Conclusion: On average, the calculated noise reduction using the OITC spectrum was slightly lower (0.6 dB) than that calculated using the flyover spectrum, and this difference was similar to that for the elevated loudspeaker. Residence Room Reference Spectrum NLR (dB) San Diego #1 Living SAN 24.5 OITC 22.2 Difference 2.3 Dining SAN 22.2 OITC 20.1 Difference 2.1 San Diego #2 Living SAN 25.2 OITC 25.8 Difference -0.6 Office SAN 18.4 OITC 17.5 Difference 0.9 San Diego #3 Living SAN 21.9 OITC 22.0 Difference -0.1 San Diego #4 Living SAN 21.6 OITC 19.9 Difference 1.7 Bedroom 1 SAN 25.0 OITC 23.2 Difference 1.8 Table 4-8. Exterior ground-level loudspeaker, NLR comparison of source spectra. (continued on next page)

58 Evaluating Methods for Determining Interior Noise Levels Used in Airport Sound Insulation Programs Residence Room Reference Spectrum NLR (dB) San Diego #5 Living SAN 21.3 OITC 20.7 Difference 0.6 Bedroom 1 SAN 26.3 OITC 25.0 Difference 1.3 San Diego #6 Living SAN 20.9 OITC 20.2 Difference 0.7 Dining SAN 26.4 OITC 24.9 Difference 1.5 San Diego #7 Family SAN 25.7 OITC 26.4 Difference -0.7 Difference 0.2 Living SAN 24.6 OITC 24.4 San Diego #9 Living SAN 19.4 OITC 18.8 Difference 0.6 Bedroom 1 SAN 25.6 OITC 25.0 Difference 0.6 San Diego #10 Bedroom 1 SAN 17.6 OITC 17.7 Difference -0.1 Bedroom 2 SAN 25.1 OITC 25.9 Difference -0.8 Boston #1 (storm windows closed) Living BOS 37.0 OITC 37.4 Difference -0.4 Bedroom 2 BOS 30.5 OITC 29.7 Difference 0.8 Boston #3 (storm windows closed) Living BOS 25.3 OITC 25.4 Difference -0.1 Bedroom 1 BOS 26.4 OITC 26.3 Difference 0.1 San Diego #8 Dining SAN 19.6 OITC 19.2 Difference 0.4 Master Bedroom SAN 28.3 OITC 27.1 Difference 1.2 Table 4-8. (Continued).

Findings and Applications 59 4.2.3 Repeatability At one of the test homes, the research team repeated the ground-level loudspeaker measure- ment with no change in loudspeaker position to determine whether the results changed from test to test. The goal was to determine whether the measurement engineer could induce significant variation in the test results. Table 4-9 summarizes the findings. Conclusion: In general, the results varied little from measurement to measurement; however, there were a few outliers which warrant closer examination. 4.2.4 Measurement of Exterior Wall and Roof vs. Exterior Wall Only Similar to the elevated exterior loudspeaker method, the research team conducted measure- ments of noise reduction using two methods: (1) making an exterior spatial measurement of the wall and roof-ceiling assembly and (2) making a spatial measurement of just the exterior wall. The resultant difference in noise reduction is presented in Table 4-10. Conclusion: Similar to the elevated loudspeaker, the NLR decreased by approximately 1 dB when the roof and walls were measured at the exterior (rather than just the walls). 4.3 Elevated Exterior Loudspeaker The elevated exterior loudspeaker measurement method generally follows the measure- ment procedure outlined in ASTM E966; this method has been employed by various acousti- cal consultants working on sound insulation projects. The method involves the suspension of a loudspeaker approximately 6.1 m to 12.2 m (20 feet to 40 feet) above grade, set back 6.1 m to 12.2 m (20 feet to 40 feet) from the façade of the room being measured. In theory, the elevated position is used because it more closely approximates the origin of the elevated aircraft noise (i.e., up in the sky). The loudspeaker generates a diffuse sound field at the exterior building windows open) Difference 0.6 Study BOS 25.7 OITC 26.5 Difference -0.8 Average Difference (Flyover – OITC) 0.5 Standard Deviaon 0.9 Note: SAN and BOS reference spectrum refers to the average flyover spectrum measured at the home. Boston #6 (storm windows open) Dining BOS 25.0 OITC 25.4 Difference -0.4 Bedroom 2 BOS 24.9 OITC 24.9 Difference 0.0 Boston #8 (storm Living BOS 24.8 OITC 24.2 Residence Room Reference Spectrum NLR (dB) Table 4-8. (Continued).

60 Evaluating Methods for Determining Interior Noise Levels Used in Airport Sound Insulation Programs Measured Noise Reducon (dB) Residence Room Ref. Spectrum Descripon Msmt. 1 Msmt. 2 Difference Standard Deviaon San Diego #1 Living OITC Exterior Wall Only 22.1 22.7 0.6 0.4 SAN Exterior Wall Only 24.3 25.0 0.7 0.5 Dining OITC Exterior Wall Only 20.1 20.0 0.1 0.1 SAN Exterior Wall Only 22.3 22.0 0.3 0.2 San Diego #2 Living OITC Exterior WallOnly 25.7 26.9 1.2 0.8 SAN Exterior WallOnly 24.7 25.6 0.9 0.6 San Diego #3 Living OITC Exterior WallOnly 21.7 22.3 0.6 0.4 SAN Exterior WallOnly 21.5 22.2 0.7 0.5 San Diego #4 Living OITC Exterior Wall Only 19.7 20.0 0.3 0.2 SAN Exterior Wall Only 21.2 21.9 0.7 0.5 Bedroom 1 OITC Exterior Wall Only 23.0 23.3 0.3 0.2 SAN Exterior Wall Only 24.8 25.1 0.3 0.2 Average 0.6 0.4 Note: SAN reference spectrum refers to the average flyover spectrum measured at the home. Table 4-9. Exterior ground-level loudspeaker, repeatability of measurement. Residence Room Ref. Spectrum Descrip on Noise Reduc on (dB) San Diego #6 Living OITC Exterior Wall and Roof 19.8 Exterior Wall Only 20.6 Difference -0.8 SAN Exterior Wall and Roof 20.4 Exterior Wall Only 21.4 Difference -1.0 Dining OITC Exterior Wall and Roof 24.4 Exterior Wall Only 25.4 Difference -1.0 SAN Exterior Wall and Roof 25.6 Exterior Wall Only 27.1 Difference -1.5 Table 4-10. Exterior ground-level loudspeaker, comparison of wall/roof and wall-only measurement.

San Diego #5 Living OITC Exterior Wall and Roof 20.0 Exterior Wall Only 21.3 Difference -1.3 SAN Exterior Wall and Roof 20.6 Exterior Wall Only 22.0 Difference -1.4 San Diego #7 Family OITC Exterior Wall and Roof 24.1 Exterior Wall Only 24.7 Difference -0.6 SAN Exterior Wall and Roof 24.4 Exterior Wall Only 25.0 Difference -0.6 San Diego #8 Dining OITC Exterior Wall and Roof 18.7 Exterior Wall Only 19.7 Difference -1.0 SAN Exterior Wall and Roof 18.9 Exterior Wall Only 20.2 Difference -1.3 Master Bedroom OITC Exterior Wall and Roof 26.2 Exterior Wall Only 27.9 Difference -1.7 SAN Exterior Wall and Roof 27.5 Exterior Wall Only 29.1 Difference -1.6 Residence Room Ref. Spectrum Descrip on Noise Reduc on (dB) San Diego #10 Bedroom 1 OITC Exterior Wall and Roof 17.3 Exterior Wall Only 18.0 Difference -0.7 SAN Exterior Wall and Roof 17.4 Exterior Wall Only 17.8 Difference -0.4 Bedroom 2 OITC Exterior Wall and Roof 25.9 Exterior Wall Only 25.9 Difference 0.0 SAN Exterior Wall and Roof 25.1 Exterior Wall Only 25.1 Difference 0.0 Average Difference OITC -0.9 SAN -1.0 Overall -0.9 Standard Deviaon OITC 0.5 SAN 0.6 Overall 0.5 Note: SAN reference spectrum refers to the average flyover spectrum measured at the home. Table 4-10. (Continued).

62 Evaluating Methods for Determining Interior Noise Levels Used in Airport Sound Insulation Programs façade; measurements of this diffuse field are taken along with measurements of the reverberant sound field in the room under test. Section 4.3.1 describes the measurement procedure in detail. (Figure 4-4.) Summary: The research team conducted elevated loudspeaker measurements in five homes in San Diego. In Boston, the research team also conducted elevated loudspeaker measurements in five homes, but aircraft flyover noise spectrum was only available in two out of the five homes. The conclusions are as follows: • On average, the calculated noise reduction using the OITC spectrum was slightly lower (0.3 dB) than that of the flyover spectrum. • When measurements were repeated, the noise reduction did not significantly change. • Noise reduction decreased by 0.8 dB when the exterior measurement included the roof. 4.3.1 Measurement Procedure The elevated exterior loudspeaker noise reduction measurement was conducted as follows (in a manner similar to the ground-level loudspeaker): 1. A loudspeaker capable of generating 90 dB to 100 dB at a distance of 9.1 m (30 feet) was elevated above the roof plane of the home under test. Typically, the height was 6.1 m to 12.2 m Figure 4-4. Exterior elevated loudspeaker in San Diego.

Findings and Applications 63 (20 feet to 40 feet) above grade; a bucket/crane truck was required to accomplish this. The loudspeaker was pointed in the direction of the room under test, with the goal of a 45 degree horizontal angle from the façade. 2. Ambient (background) noise measurements were conducted outside and inside of the room under test. The ambient measurement allows the technician to verify that the noise gener- ated by the loudspeaker was sufficiently above the ambient noise level; corrections to the measurement were made if the ambient noise level approached the level of the loudspeaker. This ensured that noise reduction is accurately measured and quantified. 3. Pink noise was generated by the loudspeaker. A measurement of the diffuse sound field just outside of the room under test [e.g., 1 to 2 meters (3 to 6 feet) from the façade] was made. This measurement consisted of a spatial average of the noise levels at the façade (and roof, where applicable). 4. A measurement of the loudspeaker-generated pink noise was then made inside of the room. Again, a spatial average was conducted with the technician maintaining a minimum distance of 0.3 to 0.6 meters (1 to 2 feet) between the microphone and walls, ceilings, floors, etc. 5. The one-third octave measurement data (50 Hz to 5 kHz) was then analyzed in a standard spreadsheet, where the interior pink noise level was subtracted from the exterior pink noise level. Three corrections were then applied: (1) the subtraction of ambient noise, (2) a 2 dB correction (based on ASTM E966-10) to account for the reverberant noise build-up at the façade and, (3) a correction to account for the different frequency spectrum of an aircraft as compared to pink noise. These corrections are discussed in more detail in the following section. 4.3.2 Data Analysis and Correction Factors To calculate the amount of noise reduction provided by the building envelope, the data must be analyzed and corrected after completion of the measurement. There are three primary cor- rections that are made to the measured data, as previously discussed in Section 4.2.2. For the spectral correction, the research team converted pink noise reduction into aircraft noise reduction using two reference spectra: the OITC reference spectrum and the average aircraft spec- trum measured at each test home over a period of 3 to 5 hours (measured as a part of the flyover noise reduction measurements). Table 4-11 shows the noise reduction calculated using both of these frequency spectra. For the San Diego measurements, elevated loudspeaker measurements were conducted in five homes. At Boston, elevated loudspeaker measurements were also conducted in five homes, but aircraft flyover noise spectrum was only available in two out of the five homes. Conclusion: On average, the calculated noise reduction using the OITC spectrum was slightly lower (0.3 dB) than that calculated using the flyover spectrum. 4.3.3 Repeatability At one of the test homes, the research team repeated the elevated loudspeaker measurement with no change in loudspeaker position to determine whether results changed from test to test. The goal was to determine whether the measurement engineer could induce significant variation in the test results. Table 4-12 summarizes the findings. Conclusion: When measurements were repeated, the noise reduction did not significantly change.

64 Evaluating Methods for Determining Interior Noise Levels Used in Airport Sound Insulation Programs Study OITC 27.1 Difference -0.5 Average Difference (Flyover – OITC) 0.3 Standard Deviaon 0.6 Note: SAN and BOS reference spectrum refers to the average flyover spectrum measured at the home. Residence Room Reference Spectrum NLR (dB) San Diego #6 Living SAN 21.0 OITC 20.1 Difference 0.9 Dining SAN 23.7 OITC 22.4 Difference 1.3 San Diego #7 Family SAN 25.6 OITC 25.3 Difference 0.3 Living SAN 23.8 OITC 23.4 Difference 0.4 San Diego #8 Dining SAN 19.9 OITC 19.4 Difference 0.5 Master Bedroom SAN 28.6 OITC 27.8 Difference 0.8 San Diego #9 Living SAN 19.8 OITC 19.0 Difference 0.8 Bedroom 1 SAN 26.1 OITC 25.6 Difference 0.5 San Diego #10 Bedroom 1 SAN 18.9 OITC 19.4 Difference -0.5 Bedroom 2 SAN 25.3 OITC 25.9 Difference -0.6 Boston #6 (storm windows open) Dining BOS 25.0 OITC 25.7 Difference -0.7 Bedroom 2 BOS 25.1 OITC 25.1 Difference 0.0 Boston #8 (storm windows open) Living BOS 25.0 OITC 24.3 Difference 0.7 BOS 26.6 Table 4-11. Elevated loudspeaker, comparison of source noise spectra.

Findings and Applications 65 4.3.4 Measurement of Exterior Wall and Roof vs. Exterior Wall Only Aircraft noise enters into a residence via the building envelope. This includes the exterior wall and roof-ceiling assembly. While windows are typically the main path for noise intrusion, the roof-ceiling assembly contributes to noise intrusion and attic insulation or other roof-ceiling treatments are included as a part of the sound insulation treatment package. Measurements of noise reduction were made following two methods: (1) making an exte- rior spatial measurement of the wall and roof-ceiling assembly and (2) making a spatial mea- surement of just the exterior wall. The resultant difference in noise reduction is presented in Table 4-13. Conclusion: The noise reduction decreased by 0.8 dB when the exterior measurement included the roof. Measured NLR (dB) Residence Room Ref. Spectrum Descripon Msmt. 1 Msmt. 2 Difference Standard Deviaon San Diego #6 Living OITC Exterior Wall Only 20.6 20.5 0.1 0.1 SAN Exterior Wall Only 21.6 21.5 0.1 0.1 San Diego #6 Living OITC Exterior Wall and Roof 19.5 19.8 0.3 0.2 SAN Exterior Wall and Roof 20.2 20.7 0.5 0.4 Average 0.3 0.2 Note: SAN reference spectrum refers to the average flyover spectrum measured at the home. Table 4-12. Exterior elevated loudspeaker, repeatability of measurement. Residence Room Reference Spectrum Descripon NLR (dB) San Diego #6 Living OITC Exterior Wall and Roof 19.7 Exterior Wall Only 20.6 Difference -0.9 SAN Exterior Wall and Roof 20.5 Exterior Wall Only 21.6 Difference -1.1 Dining OITC Exterior Wall and Roof 22.0 Exterior Wall Only 22.7 Difference -0.7 SAN Exterior Wall and Roof 23.3 Exterior Wall Only 24.0 Difference -0.7 Table 4-13. Exterior elevated loudspeaker, comparison of wall/roof and wall only measurement. (continued on next page)

66 Evaluating Methods for Determining Interior Noise Levels Used in Airport Sound Insulation Programs San Diego #7 Family OITC Exterior Wall and Roof 25.2 Exterior Wall Only 25.3 Difference -0.1 SAN Exterior Wall and Roof 25.4 Exterior Wall Only 25.7 Difference -0.3 Living OITC Exterior Wall and Roof 23.1 Exterior Wall Only 23.7 Difference -0.6 SAN Exterior Wall and Roof 23.4 Exterior Wall Only 24.2 Difference -0.8 San Diego #8 Dining OITC Exterior Wall and Roof 18.7 Exterior Wall Only 20.1 Difference -1.4 SAN Exterior Wall and Roof 19.2 Exterior Wall Only 20.6 Difference -1.4 Master Bedroom OITC Exterior Wall and Roof 27.1 Exterior Wall Only 28.4 Difference -1.3 SAN Exterior Wall and Roof 27.8 Exterior Wall Only 29.3 Difference -1.5 Residence Room Reference Spectrum Descripon NLR (dB) San Diego #10 Bedroom 1 OITC Exterior Wall and Roof 19.2 Exterior Wall Only 19.6 Difference -0.4 SAN Exterior Wall and Roof 18.6 Exterior Wall Only 19.1 Difference -0.5 Bedroom 2 OITC Exterior Wall and Roof 25.5 Exterior Wall Only 26.3 Difference -0.8 SAN Exterior Wall and Roof 24.9 Exterior Wall Only 25.6 Difference -0.7 Average Difference OITC -0.8 SAN -0.9 Overall -0.8 Standard Deviaon OITC 0.4 SAN 0.4 Overall 0.4 Note: SAN reference spectrum refers to the average flyover spectrum measured at the home. Table 4-13. (Continued).

Findings and Applications 67 4.4 Interior Loudspeaker The interior loudspeaker measurement method was created by one acoustical consultant work- ing on airport sound insulation programs. This method has been employed at a few airport sound insulation programs, and the research team understands that the local FAA representatives over- seeing those airport sound insulation programs approved the measurement method. However, this method is not addressed in ASTM E966 or other national or inter national standards. The advantage to the interior loudspeaker method is twofold: (1) there are no loudspeaker location logistical issues as there can be with the exterior loudspeaker methods and (2) noise generated by the loudspeaker does not disturb adjacent residents (except when testing multi-family units). The interior loudspeaker method involves locating a loudspeaker inside of the room under test and conducting measurements of the reverberant sound field inside of the room and at the exterior of the room’s façade. Section 4.4.1 describes the measurement procedure in detail. There is a significant flaw in the interior loudspeaker method. If a façade were comprised of a single building assembly (e.g., wall system with no doors or windows) exterior sound measured from this uniformly radiating surface would decrease minimally close to the façade and then more rapidly with increased distance. The nonlinear rate of sound reduction with distance from a uniformly radiating rectangular plane depends upon the dimensions of the plane (Rathe, 1969; Bies and Hansen, 2003). The measurement issue becomes more complex with a multi-element façade with multiple radiating planes of varying dimensions. Summary: The research team conducted interior loudspeaker measurements at 10 homes in San Diego. In Boston, the research team conducted interior loudspeaker measurements in nine homes, but aircraft flyover noise spectrum was only available in four out of the nine homes. The conclusions are as follows: • On average, the calculated noise reduction using the OITC spectrum was slightly lower (0.9 dB) than that calculated using the flyover spectrum. • The noise reduction varied by less than 1 dB when measurements were repeated. • The noise reduction increased when the wall and roof were measured. This is opposite of what happened when the loudspeaker was located outside of the building. 4.4.1 Measurement Procedure The interior loudspeaker noise reduction measurement was conducted as follows: 1. A loudspeaker capable of generating 90 to 100 dB at a distance of 9.1 m (30 feet) was placed inside of the room under test. The loudspeaker was pointed toward an interior room corner (i.e., not towards the façade) so as to generate a diffuse sound field inside of the room. 2. Ambient (background) noise measurements were conducted inside and outside of the room under test. The ambient measurement allows the technician to verify that the noise gener- ated by the loudspeaker was sufficiently above the ambient noise level; corrections to the measurement were made if the ambient noise level approached the level of the loudspeaker. This ensured that noise reduction was accurately measured and quantified. 3. Pink noise was generated by the loudspeaker. A measurement of the reverberant sound field inside the room under test was made. This measurement consisted of a spatial average of the noise levels in the room. The engineer maintained a minimum distance of 0.3 to 0.6 meters (1 to 2 feet) between the microphone and walls, ceilings, floors, and loudspeaker. 4. A measurement of the loudspeaker-generated pink noise was then made at the exterior of the façade of the room under test. The microphone was held a distance of 1 to 2 meters (3 to 6 feet) off the façade and a spatial average was conducted. 5. The one-third octave measurement data (50 Hz to 5 kHz) was then analyzed in a spreadsheet, where the exterior pink noise level was subtracted from the interior pink noise level. Three

68 Evaluating Methods for Determining Interior Noise Levels Used in Airport Sound Insulation Programs corrections were then applied: (1) the subtraction of ambient noise, (2) a 5 dB correction to account for reverberant build-up in the source room, and (3) a correction to account for the different frequency spectrum of an aircraft as compared to pink noise. These corrections are discussed in more detail in the following section. 4.4.2 Data Analysis and Correction Factors The interior loudspeaker noise reduction measurement employs the same ambient correction and data analysis procedure as the exterior loudspeaker methods. These procedures are outlined in Section 4.2.2. 4.4.2.1 Spectral Correction Similar to the exterior loudspeaker methods, the research team has analyzed the difference in measured noise reduction using the OITC and the average aircraft spectrum measured at each test home. Table 4-14 shows the noise reduction calculated using both of these frequency spectra. Residence Room Reference Spectrum NLR (dB) San Diego #1 Living SAN 24.3 OITC 22.4 Difference 1.9 Dining SAN 25.2 OITC 22.4 Difference 2.8 San Diego #2 Living SAN 23.2 OITC 22.0 Difference 1.2 Office SAN 22.8 OITC 21.7 Difference 1.1 San Diego #3 Living SAN 23.9 OITC 23.2 Difference 0.7 Master Bedroom SAN 25.0 OITC 23.9 Difference 1.1 San Diego #4 Living SAN 22.6 OITC 20.5 Difference 2.1 Bedroom 1 SAN 26.9 OITC 24.8 Difference 2.1 San Diego #5 Living SAN 23.6 OITC 22.9 Difference 0.7 Bedroom 1 SAN 35.5 OITC 32.6 Difference 2.9 San Diego #6 Living SAN 23.1 OITC 22.2 Difference 0.9 Table 4-14. Interior loudspeaker, comparison of exterior noise spectra.

Residence Room Reference Spectrum NLR (dB) San Diego #9 Living SAN 21.8 OITC 20.7 Difference 1.1 Bedroom 1 SAN 27.6 OITC 26.1 Difference 1.5 San Diego #10 Bedroom 1 SAN 23.7 OITC 22.7 Difference 1.0 Bedroom 2 SAN 27.0 OITC 26.6 Difference 0.4 Boston #1 (storm windows closed) Living BOS 27.7 OITC 27.4 Difference 0.3 Bedroom 2 BOS 27.5 OITC 27.8 Difference -0.3 Boston #3 (storm windows closed) Living BOS 23.2 OITC 23.2 Difference 0.0 Bedroom 1 BOS 28.3 OITC 26.8 Difference 1.5 Boston #6 (storm windows open) Dining BOS 25.0 OITC 27.1 Difference -2.1 Bedroom 2 BOS 26.9 OITC 28.3 Difference -1.4 San Diego #6 Dining SAN 26.1 OITC 24.2 Difference 1.9 San Diego #7 Family SAN 22.8 OITC 22.5 Difference 0.3 Living SAN 18.5 OITC 18.5 Difference 0.0 San Diego #8 Dining SAN 20.9 OITC 20.3 Difference 0.6 Master Bedroom SAN 27.3 OITC 27.0 Difference 0.3 Boston #8 (storm windows open) Living BOS 27.7 OITC 25.9 Difference 1.8 Study BOS 23.7 OITC 23.4 Difference 0.3 Average Difference (OITC – Flyover) 0.9 Standard Deviaon 1.1 Note: SAN and BOS reference spectrum refers to the average flyover spectrum measured at the home. Table 4-14. (Continued).

70 Evaluating Methods for Determining Interior Noise Levels Used in Airport Sound Insulation Programs For the San Diego measurements, interior loudspeaker measurements were conducted in ten homes. In Boston, interior loudspeaker measurements were conducted in nine homes, but aircraft flyover noise spectrum was only available in four out of the nine homes. Conclusion: On average, the calculated noise reduction using the OITC spectrum was slightly lower (0.9 dB) than that calculated using the flyover spectrum. 4.4.3 Repeatability At one of the test homes, the research team repeated the ground-level loudspeaker measure- ment with no change in loudspeaker position to determine whether the results changed from test to test. The goal was to determine whether the measurement engineer could induce significant variation in the test results. Table 4-15 summarizes the findings. Conclusion: The noise reduction varied by less than 1 dB when measurements were repeated. 4.4.4 Measurement of Exterior Wall and Roof vs. Exterior Wall Only Similar to the exterior loudspeaker methods, the research team conducted measurements of noise reduction using two methods: (1) making an exterior spatial measurement of the wall and roof-ceiling assembly, and (2) making a spatial measurement of just the exterior wall. The resultant difference in noise reduction is presented in Table 4-16. Conclusion: The noise reduction increased when the wall and roof were measured. This is opposite of what happened when the loudspeaker was located outside of the building. 4.5 Architectural Survey and NLR Computation It is not always feasible to measure the noise reduction of each habitable room in a residence, nor to acoustically test every residence potentially eligible for sound insulation. However, there are acoustical calculation methods available to estimate the NLR of a room/home. These calculation Measured NLR (dB) Residence Room Ref. Spectrum Descripon Msmt. 1 Msmt. 2 Difference Standard Deviaon San Diego #1 Living Room OITC Exterior Wall Only 22.5 22.2 0.3 0.2 SAN Exterior Wall Only 24.5 24.0 0.5 0.4 San Diego #6 Living Room OITC Wall and Roof 22.3 22.9 0.6 0.4 SAN Wall and Roof 23.7 23.3 0.4 0.3 Dining Room OITC Wall and Roof 24.3 25.3 1.0 0.7 SAN Wall and Roof 26.9 26.7 0.2 0.1 San Diego #4 Living Room OITC Exterior Wall Only 20.8 20.2 0.6 0.4 SAN Exterior Wall Only 22.8 22.3 0.5 0.4 Bedroom 1 OITC Exterior Wall Only 24.6 24.9 0.3 0.2 SAN Exterior Wall Only 26.7 27.1 0.4 0.3 Average 0.5 0.3 Note: SAN reference spectrum refers to the average flyover spectrum measured at the home. Table 4-15. Interior loudspeaker, repeatability of measurement.

Findings and Applications 71 Table 4-16. Interior loudspeaker, comparison of wall/roof and wall only measurement. NLR (dB) Residence Room Ref. Spectrum Descripon San Diego #6 Living OITC Exterior Wall and Roof 22.6 Exterior Wall Only 20.2 Difference 2.4 SAN Exterior Wall and Roof 23.5 Exterior Wall Only 20.8 Difference 2.7 Dining OITC Exterior Wall and Roof 24.8 Exterior Wall Only 22.3 Difference 2.5 SAN Exterior Wall and Roof 26.8 Exterior Wall Only 23.7 Difference 3.1 San Diego #7 Living OITC Exterior Wall and Roof 20.1 Exterior Wall Only 16.8 Difference 3.3 SAN Exterior Wall and Roof 20.0 Exterior Wall Only 16.9 Difference 3.1 San Diego #8 Dining OITC Exterior Wall and Roof 21.3 Exterior Wall Only 19.3 Difference 2.0 SAN Exterior Wall and Roof 21.9 Exterior Wall Only 19.9 Difference 2.0 Master Bedroom OITC Exterior Wall and Roof 31.8 Exterior Wall Only 30.2 Difference 1.6 SAN Exterior Wall and Roof 32.6 Exterior Wall Only 31.0 Difference 1.6 San Diego #10 Bedroom 1 OITC Exterior Wall and Roof 23.0 Exterior Wall Only 22.4 Difference 0.6 SAN Exterior Wall and Roof 24.0 Exterior Wall Only 23.3 Difference 0.7 Average Difference OITC 2.1 SAN 2.2 Overall 2.1 Standard OITC 0.9 Deviaon SAN 1.0 Overall 0.9 Note: SAN reference spectrum refers to the average flyover spectrum measured at the home.

72 Evaluating Methods for Determining Interior Noise Levels Used in Airport Sound Insulation Programs methods are used extensively during the design process of new buildings, and the results are used to inform the architect of the required TL performance of the building envelope components (e.g., windows, doors). In general, the acoustical calculation methods utilize a library of TL data for typical building components. This performance data comes from a variety of sources, as summarized below: • Exterior walls: National Research Council of Canada, California Office of Noise Control. • Windows: National Research Council of Canada, window vendors, glazing manufacturers (glass only). • Doors: National Research Council of Canada, door vendors. • Roof-Ceiling Assemblies: National Research Council of Canada, roofing manufacturers (limited). In order to accurately calculate noise reduction, an architectural survey must be conducted. Section 4.5.1 summarizes the components of an architectural survey. 4.5.1 Architectural Survey The goal of the architectural survey is to catalog the pertinent details of the building envelope to inform the calculation procedure. The façade and room elements are inspected and pertinent details are logged. The following summarizes the types of data gathered: • Exterior Wall – Wall type: brick, stud, concrete masonry unit. – Stud thickness and presence/type of wall insulation. – Exterior sheathing type: vinyl siding, wood siding, stucco, brick, etc. – Interior sheathing type and thickness: gypsum board, tongue-and-groove, etc. • Windows – Dimensions. – Frame material: aluminum, vinyl, wood, etc. – Glazing configuration: thickness of panes, air space between panes, laminated or float. – Condition of window: leaky, average, good. – Operation type: double-hung, horizontal sliding, casement, fixed, etc. • Doors (including storm doors) – Dimensions. – Material: wood, fiberglass, metal, etc. – Glazing: size, type, thickness of lites, airspace between lites. – Gasketing: type, material, condition, quality of seal (credit card test). – Door bottom: type, material, condition, quality of seal (credit card test). • Roof-Ceiling Assembly – Type: flat, pitched (e.g., gable, hip), etc. – Roof material: asphalt shingle, tile, tar-and-gravel, etc. – Attic insulation: thickness, type. – Roof vents: type, quantity. • Room Details – Dimensions. – Floor type: carpet, hardwood, tile, etc. – Wall type: gypsum board, etc. – Ceiling: type (flat or vaulted), material. – Furnishings: sofas, beds, bookshelves, entertainment center, etc. The data from the architectural survey is then input into software or a spreadsheet to calculate the noise reduction of the room.

Findings and Applications 73 4.5.2 Acoustical Calculation Procedure In order to calculate the noise reduction of a room, there are three primary steps: (1) deter- mine the TL of each element, (2) calculate the CTL of all elements, and 3) convert TL to noise reduction. For the first step, the engineer must first determine the transmission of each build- ing element. This is accomplished by selecting the laboratory TL of the building element most similar to that of the room being calculated. This requires information from the architectural survey and the use of engineering judgment, as often the building element encountered in the field does not perfectly match the available laboratory TL data. In addition, for elements that are degraded (e.g., leaky windows, poor door weather stripping), a correction must be applied to account for this less-than-ideal condition (as laboratory TL data is based upon perfect conditions). After the laboratory data has been selected for each building element, the CTL of the room must be calculated. This is accomplished using a logarithmic formula (shown below), which takes into account the surface area of each building element and its TL performance. In general, the building element with the most surface area will control the CTL; however, since the calcula- tion (and the decibel scale) is logarithmic, the CTL can be significantly degraded by a building element with low TL. Transmission coefficient i avg 1 1 Si Si i n i n ∑ ∑τ = τ = =  where 10 10 TL τ = −     then composite transmission loss CTL 10 log 1 avg 10= τ     The final step in the acoustical calculation process is to convert TL into noise reduction. This consists of making a correction for room factor, which is the amount of acoustically absorptive material in a room. Section 4.1.3 of this report provides more information on the effect of absorptive materials on noise reduction. In general, the more acoustically absorptive material in a room, the higher the NLR will be. The research team performed reverberation time (RT60) measurements in every tested room. The research team measured reverberation time data to convert TL data to noise reduction using the following formulas: 10 logNR CTL S A = −     where NR = Noise reduction (dB), usually A-weighted CTL = Composite Transmission Loss (dB) S = area of the transmitting surface(s) (m2 or ft2) A = absorption, k room volume RT ×    60 per 1⁄3 octave band from 80 Hz to 5 kHz (Sabine), where k = 0.049 in Imperial units, and k = 0.161 in International System (SI) units. For every room acoustically tested, an architectural survey was conducted and noise reduc- tion calculations were performed using two models: IBANA-Calc and a spreadsheet using the reverberation formula and the CTL formulas shown above.

74 Evaluating Methods for Determining Interior Noise Levels Used in Airport Sound Insulation Programs 4.5.2.1 IBANA Calculation Model The first model employed was the IBANA (Insulating Buildings Against Noise from Aircraft) software, which was created by the National Research Council Canada. This model employs a graphical user interface and requires the user to input the following: • Source spectrum (in one-third octave bands) and exterior noise level (see Figure 4-5); default spectra include: – Standard aircraft. – Canadian Mortgage Housing Corporation (CMHC) reference source. – OITC reference source. – Chapter (Stage) 2 jets. – Chapter (Stage) 3 jets. – Helicopters. – Custom spectrum can be added by the user. • Floor area (m2 or ft2). • Acoustical absorption as a percentage of floor area. • Surface area and TL of building elements (see Figure 4-6); the software includes transmission loss data for: – 2 × 4 stud exterior walls. – 2 × 6 stud exterior walls (with and without resilient channels). – Staggered stud exterior walls (with and without resilient channels). – Doors. – Glazing and windows. – Wood joist roofs. – Wood truss roofs. – Raised heel wood truss roofs (with and without vents). – Steel deck roofs. – Custom building elements can be added by the user. Figure 4-5. IBANA source spectrum selection screen.

Findings and Applications 75 • Optional correction factors (not used): – Air absorption. – Vertical angle of incidence. – Horizontal angle of incidence. – Horizontal angle of view. – Ground reflection. For each set of inputs, a scenario is created. The user is able to name the scenario to allow for tracking of multiple scenarios. The composite noise reduction and resultant interior noise level is calculated by IBANA for each scenario (Figure 4-7). 4.5.2.2 Spreadsheet Calculation Model The research team created a spreadsheet incorporating the CTL formula shown in Section 4.5.2. For each tested room, the research team input the surface area of each building element and used the same TL values as used in the IBANA modeling. This yielded a composite (or average) TL for each tested room. The research team then used the reverberation time formula contained in Section 4.5.2 to convert the CTL into composite noise reduction. Figure 4-8 shows a sample calculation spreadsheet. 4.5.3 NLR Calculation Results The research team calculated the noise reduction of each room that was acoustically tested. Table 4-17 summarizes the results of the calculations and compares them to the loudspeaker and flyover measurement results. Specifically, the average NLR was computed for each NLR testing method, and variation in results from that average was computed for each method. The research team found the following from this comparison: • The NLR calculations generally agreed with the measurement findings. • There were some outliers in the NLR calculations, which increased the standard deviation. Figure 4-6. IBANA building element input screen.

76 Evaluating Methods for Determining Interior Noise Levels Used in Airport Sound Insulation Programs • It is important that the field survey be very detailed so the calculations are accurate; even then, it is possible to miss flanking paths (noise leaks) that would overstate the calculated NLR. • The IBANA computation method was preferable to the spreadsheet method because of the database and other factors incorporated. 4.6 Air Infiltration and Noise Reduction As a part of the field measurements in San Diego and Boston, the research team conducted air infiltration (blower door) measurements at most of the homes. The purpose of this testing was to determine whether there was a correlation between air infiltration and façade noise reduction as, in theory, higher infiltration values should correspond to lower noise reduction (i.e., the leakier the façade, the more paths for noise to enter the residence). Summary: • A comparison of air infiltration to measured noise reduction was made for homes where flyover and loudspeaker noise reduction was measured. • Based on published data, the research team would expect lower noise reduction when air infiltration was high (i.e., leakier homes allowed more noise intrusion). • The research team found that higher measured air infiltration with the blower door test did not correspond to lower NLR. 4.6.1 Measurement Procedure At ten homes in San Diego and five homes in Boston, air infiltration tests were conducted in rooms where acoustical testing took place. The air infiltration tests were conducted per ASTM E-779-10; a brief description of the measurement procedure follows: 1. A blower door fan was set up in the doorway of the room to be tested. If there were doorways or openings to other rooms (e.g., a bathroom), this door/opening was sealed airtight. See Figure 4-9. Figure 4-7. IBANA calculation results screen.

Figure 4-8. TL calculation sample spreadsheet.

Average NLR (no int. spkr.) Diff. from Avg NLR Elevated Loudspeaker Diff. from Avg NLR Ground Loudspeaker Diff. from Avg NLR Interior Loudspeaker Diff. from Avg NLR Diff. from Avg NLR Diff. from Avg NLR Residence Room Flyover Spreadsheet IBANA San Diego #1 Living 26.2 26.3 -0.1 24.5 1.7 24.3 2.0 28.1 -1.9 26 0.2 Dining 26.4 28.0 -1.6 22.2 4.3 25.2 1.2 28.5 -2.1 27 -0.6 San Diego #2 Living 26.1 24.3 1.9 25.2 1.0 23.2 2.9 28.1 -2.0 27 -0.9 Office 23.3 20.4 2.9 18.4 4.9 22.8 0.5 27.3 -4.0 27 -3.7 San Diego #4 Living 23.4 23.3 0.1 21.6 1.8 22.6 0.8 24.7 -1.3 24 -0.6 Bedroom 1 26.8 26.4 0.4 25.0 1.8 26.9 -0.1 29.7 -3.0 26 0.8 San Diego #5 Living 25.0 24.0 1.0 21.3 3.7 23.6 1.4 29.8 -4.8 25 0.0 Bedroom 1 29.0 30.0 -1.1 26.3 2.7 35.5 -6.6 30.5 -1.6 29 -0.1 San Diego #6 Living 21.9 24.9 -3.0 21.0 0.9 20.9 1.0 23.1 -1.2 17.8 4.1 25 -3.1 Dining 27.5 27.4 0.1 23.7 3.8 26.4 1.1 26.1 1.4 31.0 -3.5 29 -1.5 San Diego #7 Family 26.4 26.5 -0.1 25.6 0.8 25.7 0.7 22.8 3.6 27.2 -0.8 27 -0.6 Living 25.1 25.1 0.0 23.8 1.3 24.6 0.5 18.5 6.7 27.5 -2.4 24 1.1 San Diego #8 Dining 21.2 19.4 1.9 19.9 1.3 19.6 1.7 20.9 0.3 23.3 -2.1 24 -2.8 Master Bed 26.7 22.2 4.5 28.6 -1.9 28.3 -1.6 27.3 -0.6 27.3 -0.6 27 -0.3 San Diego #9 Living 21.3 21.8 -0.5 19.8 1.5 19.4 1.9 21.8 -0.5 22.3 -1.0 23 -1.7 Bedroom 1 28.3 31.7 -3.4 26.1 2.2 25.6 2.7 27.6 0.7 30.3 -2.0 28 0.3 San Diego #10 Bedroom 1 20.1 17.5 2.6 18.9 1.2 17.6 2.5 23.7 -3.6 23.5 -3.4 23 -2.9 Bedroom 2 25.9 25.0 0.9 25.3 0.7 25.1 0.8 27.0 -1.1 28.3 -2.4 26 -0.1 Boston #6 Dining 23.1 20.2 2.9 25.0 -1.9 25.0 -1.9 25.0 -1.9 21.3 1.8 24 -0.9 Bedroom 2 22.3 22.0 0.3 25.1 -2.8 24.9 -2.6 26.9 -4.6 19.3 3.0 20 2.3 Boston #8 Living 24.8 24.2 0.6 25.0 -0.2 24.8 0.0 27.7 -2.9 25.0 -0.2 25 -0.2 Study 27.0 29.5 -2.5 26.6 0.4 25.7 1.3 23.7 3.3 26.2 0.8 27 0.0 Average Difference 0.4 0.5 1.4 0.1 -1.3 -0.7 Standard Deviaon 1.9 1.7 1.8 2.9 2.1 1.4 Notes: 1) Correcons applied—Flyover: 2 dBA, exterior loudspeaker: 2 dBA, interior loudspeaker: 5 dBA. 2) Average NLR difference calculated by first averaging all of the NLR across all measurement methods (except interior loudspeaker and sound intensity), and then subtracng the NLR from one method (e.g., flyover) from the average NLR. Interior loudspeaker not used in average as this method does not follow naonal standards and the 5-dB correcon applied to the data is based on limited field measurements (i.e., not fully veˆed). 3) All differences are calculated by subtracng from the average noise reducon. 4) Blank cells indicate no elevated loudspeaker test was performed at the corresponding residence. Table 4-17. Comparison of NLR calculations to acoustical measurements.

Findings and Applications 79 2. The fan was then turned on and either created a positive or negative pressure differential between the room under test and the adjacent spaces (including the outside). Typically, a negative pressure differential was created. The fan was ramped up until there was a pressure difference of 50 Pascals between the room under test and the adjacent spaces. 3. The amount of airflow required to maintain the 50 Pascals was then measured. For this proj- ect, multi-point measurements were made, meaning airflow was measured at various pres- sure differentials to provide for more accurate air infiltration values. Blower door testing was conducted in most of the rooms acoustically tested; however, there were some instances where it was not appropriate to conduct the blower door measurement. For example, when a living room was acoustically tested, this living room was often open to a dining room and/or kitchen. When the research team acoustically tests this condition, the team is primarily measuring noise entering the living room from the front door and living room win- dows. However, the air infiltration test is quantifying infiltration from the living room plus the kitchen and/or dining room. Thus, in this case, it would not be reasonable to compare measured air infiltration to noise reduction. Figure 4-10 shows a sample air infiltration graph. 4.6.2 Comparison of Air Infiltration to Noise Reduction Figures 4-11 and 4-12 compare the measured air infiltration [in terms of cubic feet per minute (cfm)] to measured façade noise reduction. The research team has included the Figure 4-9. Blower door test.

80 Evaluating Methods for Determining Interior Noise Levels Used in Airport Sound Insulation Programs measured flyover noise reduction and the noise reduction measured using an exterior loud speaker. As can also be seen in Figures 4-11 and 4-12, NLR is not correlated to the measured air infiltration. The lack of correlation may be due to the following: • Air infiltration from interior walls, ceilings, and floors: When an air infiltration test is performed, air can leak into a room via the building façade and interior walls, floors, or ceilings. This would lead to higher infiltration values than infiltration from the façade alone. Unfortunately, it is not feasible with the current testing protocol to only measure infiltration from the façade. • Some infiltration paths may not be noise paths: leaky doors and windows reduce noise reduc- tion (Sabine et al. 1975); however, infiltration via vents, flues, and other openings may not be a significant path of noise intrusion. For example, a fireplace may allow for significant air infiltration, but not be a significant path of noise when there is a damper and/or solid fire- place doors. 4.7 Sound Intensity 4.7.1 Introduction Sound intensity is an attractive concept for airport sound insulation programs because it does not require measurement of the reverberant field in the receiving room. Ideally, this would elim- inate the errors and anomalies associated with non-uniform reverberant fields and the standing wave effects of reflections from parallel surfaces, as reflections create resonances. Intensity mea- sures both the pressure (a scalar or non-directional parameter) and the velocity (a vector, having both magnitude and direction) of the sound. Therefore, it is theoretically possible to directly measure the intensity or power flow into the receiving room (the room under test). Moreover, in contrast to pressure measurements, intensity measurements made close to room surfaces and with a high enough resolution should reveal hot spots, sound leaks, and other problems, and not just total power flow into the test room. Figure 4-10. Typical air infiltration graph.

Findings and Applications 81 San Diego #6 Living San Diego #6 Dining San Diego #1 Living San Diego #8 Dining San Diego #5 Flyover N LR Ext. Spkr N LR A ir Infiltraon (cfm ) A ir Infiltraon trendline Living San Diego #4 Bedroom 1 San Diego #2 Living San Diego #9 Bedroom 1 San Diego #10 Bedroom 1 San Diego #10 Bedroom 2 Air Infiltraon (cfm) 1350 1050 750 450 150 Noise Level Reducon (dBA) 15 20 25 30 35 San Diego #5 Bedroom 1 San Diego #7 Living San Diego #2 Office San Diego #4 Living San Diego #8 Master Bed Figure 4-11. San Diego comparison of air infiltration and noise reduction.

82 Evaluating Methods for Determining Interior Noise Levels Used in Airport Sound Insulation Programs Summary: The research team conducted sound intensity measurements in five homes within the vicinity of the San Diego International Airport. Additional sound intensity measurements were conducted at a research team member’s residence in Champaign, IL, for comparison pur- poses. The research team took measurements from outdoors to indoors, as well as from indoors to outdoors. The conclusions are as follows: • Classical acoustic theory presumes that TL is the same in both directions (i.e., a wall performs equally whether the source of noise is inside of a home or outside of a home). Through this study, the research team found this generally to be true. • In these measurements of the two directions, the research team used the pressure incident on the surface of the source side of the wall and windows rather than positions 0.3 m to 2 m (1 to 6 feet) from the wall surface, and it is believed that the generally equal performance is due to the more equal treatment of the source side measurements. • The best strategy when conducting sound intensity measurements is to measure as close to the exterior wall as feasible. • Based on the research team’s experience, the technology and time required preclude the use of sound intensity for airport sound insulation programs at this time. • There are a number of enhancements suggested to provide for better sound intensity results and possibly allow for the use of sound intensity for airport sound insulation programs in the future. • Sound intensity holds the promise of someday being an extremely effective method because the results are virtually independent of weather effects, the measurements are independent of resonance effects, there are no neighbor noise problems, and nearby reflectors are not a problem. 4.7.2 Purpose The purpose of the intensity measurements is to see if sound power (i.e., the acoustic inten- sity over the building element measured) can be used to determine a building envelope’s TL and, if so, to gauge the accuracy, complexity, and costs of using sound intensity with respect to Bedroom 2 Boston #2 Bedroom 2 Boston #4 Living Room Boston #5 Bedroom 2 Boston #4 Bedroom 1 Boston #3 N oi se L ev el R ed uc o n (d BA ) 26 27 28 30 29 A ir In fil tr a on (c fm ) 1500 1200 Speaker Airflow at 50 Pa Air Infiltraon trendline 900 300 600 Figure 4-12. Boston comparison of air infiltration to noise reduction.

Findings and Applications 83 conventional sound pressure measurements. One of the most important questions posed about sound intensity measurements is as follows: Does the flow of sound power into or out of a room and its measurement ameliorate the resonance issues that are prevalent with measurements of sound pressure (i.e., the loudspeaker or flyover method). When one uses pressure, measure- ments from outdoor sources to indoor microphones do not agree well with measurements from indoor sound sources to outdoor microphones; reciprocity does not seem to apply as it should with acoustical theory. This study is concerned with the accurate measurement of the aircraft noise TL from outside to inside residences, and sound intensity measurements were conducted to determine whether they could provide the most accurate TL measurement. Aircraft noise usually impinges on a residence at an angle, e.g., 45°. With three-dimensional sound intensity, one can measure the angle at which the power is flowing as it travels from out- doors through the building façade into a room. 4.7.3 Background Acoustic velocity is difficult to measure directly, so a pair of opposing microphones is typically used to approximate the pressure gradient, which is the change in pressure (Dp) with a change in distance (Dx). The original intensity meters, developed in the early 1980’s, used a single pair of phase-matched microphones. For middle frequencies (e.g., 100 to 4000 Hz), a 2-cm to 3-cm (0.8 in. to 1.2 in.) space was established between the microphone pair as the distance (Dx). At 100 Hz, a 2.5-cm (1.0 in.) spacing is 1/120 of a wavelength, so the expected change in phase is (1/120) p 360 or 3°. Accurate measurement with a resolution of 3° requires that the phase matching be 10 times better than that being measured; in this case, phase matching should be 0.3° or less. A three-dimensional sound intensity system with visualization was used in this study, and the phase error for this assembly is specified as 0.8 dB at 60 Hz and is negligible above 60 Hz. The configuration of the two microphones in the classical intensity meter form what is called an acoustic dipole (two close acoustic receivers of opposing phase). However, the simple dipole that is formed by the two microphones is rather insensitive to direction. More recently, three pairs of microphones have been used to measure intensity in three dimensions. Nagata et al. (2005) reports on their tests using their three-axis, six-microphone array. Among other things they report and show that they find peaks in a spectrum. This is correct, but they do not find the entire spectrum with sharp resolution; their frequency range is 200 to 2000 Hz. The latest intensity probes are based on a tetrahedron that uses four phase-matched micro- phones. In this configuration, there are six unique microphone pairs, with two pairs for each axis. This configuration yields significantly better angular resolution than is given by dipoles. The system’s manufacturer reports a resolution of 3 cm to 5 cm (1.2 in. to 1.9 in.) when the camera is positioned at 1.5 m (4.9 ft.) or less. This implies a resolution of at least 5 cm at 1.5 m (1.2 in. at 4.9 ft.), and this indicates an angular resolution of 2°. They also report what is termed “orientation errors” and state that they are less than 10° for the frequency range from 100 to 4000 Hz. Thus, these orientation errors appear to be the limiting factor for angular resolution. The microphone goes one octave higher because with the tetrahedron array, there is a symme- try that permits correction for an error that cannot be done when using six microphones. They achieve the added octave at low frequencies by enhanced phase matching. The basics of this project’s test plan can be summed up in three simple steps: 1. Find and rent equipment, 2. Make measurements, and 3. Analyze the data.

84 Evaluating Methods for Determining Interior Noise Levels Used in Airport Sound Insulation Programs While this plan appears to be straightforward, to obtain the 2.5° resolution specified for this microphone assembly, one has to know the position of the microphone assembly with respect to the object or room under test, and the pitch and yaw of the microphone assembly itself. Since the goal is to gather directional data on sound power flow, it is not sufficient just to know the x, y, and z coordinates of the microphone assembly; one must also know the orientation of the assem- bly with respect to the source. In this case, the research team needed to know if the assembly was directly facing the wall, pointed up or down, or to the left or right. Pitch is a measure of up or down and yaw is a measure of left or right. This positioning and orientation capability should be at least as precise as the measurements that are being attempted, and preferably by 2 times or more. With the older dipole sensors, this location and orientation were not very significant because they had little directivity. Their resolution capability was about 45° at mid frequencies, poorer at lower frequencies, and better at higher frequencies. One may ask what dipole intensity meters were good for. It turns out that intensity meters of the type described were used to measure the total sound power radiated by different items of equipment and machinery. This made the measurement of sound power feasible at locations that would otherwise not be suitable, and it obviated the need for a reverberant room or an anechoic chamber to test machinery. The earlier meters did a good job of measuring the total radiated sound, but they were not good at identifying individual sources of radiated sound energy. 4.7.4 Sound Intensity Measurements 4.7.4.1 Instrumentation The basic instrumentation used for the intensity measurements was the three-dimensional sound intensity system with visualization, a standard Type I sound level meter, and a noise source. The instrument exists mainly as software and runs from a portable computer. The microphone, a wand assembly, and a continuously running video camera were connected to the portable computer via USB. The camera was used to track the position of the microphone wand assembly. Specifically, the camera and computer tracked a light on the wand to determine the position of the wand as a function of time. The angular position of the wand was determined from the angular location of the light bulb in the picture, and the distance of the wand from the camera is determined by the size of the light bulb in the picture (see Figure 4-13). The wand primarily consists of the microphone assembly and two gyroscopes mounted perpendicularly from one another that measured the pitch and yaw. Figure 4-13. Sound intensity system (instrument synchronization).

Findings and Applications 85 Figure 4-13 shows the wand being synched to a known coordinate in the room; the wall to be measured is shown in the background. The red, green and blue colors showing the three coordinates are not actually physically in the room, but are superimposed showing the reference for the coordinate system. The wand’s light must be easily visible or the system does not work. 4.7.4.2 San Diego Measurements Measurements were conducted in five homes within the vicinity of the San Diego Interna- tional Airport. Traditional aircraft sound TL measurements were conducted at these five houses, using a loudspeaker both inside and outside, and aircraft flyover measurements were also made. The intent was to be able to compare the intensity measurements to the traditional measure- ments, and to a small extent the research team was able to do this. However, even with the manufacturer of the three-dimensional sound intensity system helping (after spending a day doing preliminary measurements in San Francisco), the research team had to overcome the fact that the sound intensity system had been designed to use on a shop floor. The light on the wand works well indoors in a commercial setting, but did not work in sunlit rooms that are typically found in the residences. In an indoor setting, the room is darker and a nearly white bulb sticks out against the dark background. In the outdoor setting, the human eye can’t see if the light is on or off. Two or more different bulb colors are needed for different lighting situations. This appli- cation was unknown to the instrument manufacturer until the research team talked to them, but the instrument is capable of having different light colors, the software just needs to be created to implement it. As such, the research team had to find ways to make the instrument work in the existing lighting. This involved using shade and dark clothes to block the light coming through windows until the research team achieved enough contrast to make the instrument functional. There were several other smaller problems, all of which took time to understand and correct. However, the biggest limitation was that the traditional loudspeaker TL measurements were being conducted much more quickly than the intensity measurements could be made. Extra time was required because (1) the research team was not adept in the operation of the sound intensity instrument (a second round of measurements were conducted in Champaign, IL, where more time could be spent on the measurements) and (2) the research team did not (and in retrospect, still does not) know the required measurement time. So being conservative, the research team attempted to sweep broad areas, typically using measurement durations of 1 to 3 minutes (the limit is about 5 minutes). If great resolution is not required, it could be that 2 to 4 seconds of measurements spread over two to three camera positions would suffice. The following section describes the measurements made in Champaign and is followed by all the data that turned out to be useful from measurements gathered in San Diego. 4.7.4.3 Champaign Measurements For sound intensity to be useful, the research team needed to be able to measure the net sound energy flowing into the room via all pathways. The starting point was to measure the sound flowing into a room through an outdoor-facing wall having one or more windows. These mea- surements were conducted at a research team member’s Champaign residence. Figure 4-14 shows a partial first floor layout, indicating the dining room as the room where the measure- ments were conducted and the primary wall on which they were conducted. The dining room has two walls that face the outside, one with a pair of double-hung windows, the other with no openings. The construction is standard 2 × 4 stud wall (actual dimensions 38 mm × 89 mm) construction on 406-mm (16-inch) centers, 122 cm (4 feet) of plywood as corner bracing, and insulation board for remainder of the external facing surface on the 2 × 4 stud wall.1 The 1 In this section and for other house construction details, English units are retained, not in small part, because some of these are nominal and not true dimensions (e.g., a 2 x 4 is neither 2 inches on one side nor is it 4 inches on the other).

86 Evaluating Methods for Determining Interior Noise Levels Used in Airport Sound Insulation Programs Figure 4-14. Champaign home floor plan and measurement location.

Findings and Applications 87 cavities in the wall were fully filled with batt insulation, and the wall was finished with a brick veneer. The windows are 43-year-old narrow line, double pane models. The space between the panes is fairly small, less than 1 cm (2.5 inch). With this construction, the research team expected the vast majority of the sound flow through the window, and that is what the results appear to show. Since the overall purpose is to measure the transmission loss from outdoors to indoors, the research team concentrated first on measuring the sound power flowing through the wall into the room from an outdoor loudspeaker. The signal source was pink noise (i.e., noise with equal energy per octave). The loudspeaker was mounted 1.8 m (5.9 ft.) in the air and 9.1 m (23.8 ft.) from the face of the house in the normal direction 90° from the wall. An outdoor microphone was always located 1.8 m (5.9 ft.) from the wall. According to ISO and ASTM, with distances that are 1 m to 2 m (3.3 ft. to 6.6 ft.) from a wall, it is normal for the sound level (A-weighted) to be increased by 3 dB over what would occur without the reflecting wall. At the surface of the wall, this increase becomes 6 dB. Therefore, the difference between measurements at the wall and at the microphone is 3 dB minus the change for dis- tance spreading: the change from 7.3 m to 9.1 m (23.9 ft. to 29.8 ft.). For comparison purposes, measurements were made with a second microphone at about 20 positions on the wall with the microphone windscreen just touching the surface of the wall. The discussion later with Table 4-18 shows that the values measured were consistent with the 3 and 6 dB predictions given above. The three-dimensional sound intensity instrument is specified as being able to measure a 2.0 m by 2.5 m (6.6 ft. by 8.2 ft.) area at a distance of 2.5 m (8.2 ft.) from the camera. However, the mea- surement accuracy is 3 cm to 5 cm (7.6 in. to 12.7 in.), but only up to a camera distance of 1.5 m (4.9 ft.) beyond which it degrades. At a camera distance of 1.5 m (4.9 ft.), the measurement area is limited to 1.2 m by 1.5 m (3.9 ft. by 4.9 ft.). The measurements were made by sweeping the microphone array all throughout the desig- nated measurement area. To measure a whole wall, the wall had to be divided into several seg- ments, each measured separately. The sound intensity instrument does this easily and efficiently. The results are a field of vectors that together indicate the sound power flow. However, it takes time to learn and understand fully how the meter is processing the data. The software reports the total sound power flowing in a normal direction through a designated plane surface, in this case, the dining room wall. In actuality, the total sound power is really a summation of the sound power over all the vectors. This topic, understanding and analyzing the results, is developed in Section 4.7.5 4.7.5 Sound Intensity Results 4.7.5.1 Champaign Although there is a fair amount of data, the most interesting and elucidating are the com- posite data for the sound flowing from outdoors to indoors, as well as from indoors to out- doors, and the comparison of these two. Figure 4-15A shows the composite sound flow out of the room. It shows the window space filled with vectors largely normal to the surface. These vectors begin to spread out and diverge as they move away from the window in a very smooth and regular fashion, with the lowest levels being where they have turned the most. In contrast, Figure 4-15B shows the sound going through the window into the room. The sound diverges towards the wall but soon meets the corner where it becomes lower in level and the flow (apparently) becomes turbulent. In some places the direction changes by 90° from one vector to the next. An example of this is circled in red in Figure 4-15B.

88 Evaluating Methods for Determining Interior Noise Levels Used in Airport Sound Insulation Programs Figure 4-16 shows the composite of the wall stitched together, in this case from five indi- vidual measurement episodes, one from each quadrant around the window, and one from the center of the window. To make the comparison of the results for the two flow directions easier to view, Figure 4-17 contains a side view and again shows the frontal view of the composites for each of the two directions of flow. Again, one can see only straight lines of the inside to outside composite and the considerable curling turbulent structure of sound flow from outside to inside the room. Appendix A contains the spectra for each of the five measurement episodes shown in Figure 4-16. The sound intensity instrument does not provide composite spectra; spectra are only available for individual measurement episodes. Figure 4-18 shows the frontal view for each of the three measurement episodes that comprise the in to out composite. The sound intensity instrument also outputs a one-third octave spec- trum covering the full frequency range from 100 to 4000 Hz; however, it is not quite clear how the spectrum can and should be utilized. Spectra for all measurements are placed in Appendix A and referred to in the text. 4.7.5.2 San Diego Data were collected at five houses in San Diego, CA. The first house wasn’t successfully mea- sured, because the research team was still familiarizing itself with the equipment. For the second house, San Diego #1, the research team was able to collect satisfactory data on one window and wall in the living room. The spectra for these two measurements are somewhat similar, especially when compared with other spectral data that were measured for other building elements. The vector data of Figure 4-19 for San Diego #1 are consistent with the research team’s expecta- tions. In this figure (top view), the reader will note the difference in the depth at which these two mea- surement episodes were taken: the first episode is at a depth of 2.5 cm to 42 cm (1 in. to 15-18 in.), the second is at a depth of 30 cm to 60 cm (~12 in. to ~24 in.). This difference is used later in the analysis. On the second day, the research team was only able to collect usable data at the San Diego #1 residence due to a neighbor’s complaint at the San Diego #3 residence. The San Diego #1 home provided the research team with two measurement locations of the same living room window, one focusing on the wall left of the window and including part of the window, and the other focusing only on the window. A discussion of the frequency spectra can be found in Appendix A. The difference in spectra for these measurements seems to suggest that spectra may not provide very useful data. The spectrum for one of the two left of the window measurement episodes actu- ally looks more similar to the spectrum taken of the window only condition (see Appendix A). A. Inside to Outside B. Outside to Inside Figure 4-15. Whole wall composite sound flow.

Figure 4-16. The “stitching” together of a composite from individual measurements (the “A” in parentheses signifies that the measure is A-weighted).

90 Evaluating Methods for Determining Interior Noise Levels Used in Airport Sound Insulation Programs In contrast, the vector plots of power flow that constitute Figure 4-20 show that the two mea- surement episodes taken left of the window look similar to one another, and different from the measurement episode that was only through the window. The vectors indicate flow in from the window, as expected, but then they appear to turn and split into at least two streams. One stream appears to be going down at an angle into the party wall between the bedroom and the living room, and the other appears to be going down tangentially towards the floor. The measurement episode that only includes flow through the window appears to be different from the measure- ments taken of the wall. Also, it is clear from the bottom right image that the sound entering through this window is almost perpendicular to it. The research team took measurements at San Diego #5, focusing on the living room window and wall (Figure 4-21). Two measurement episodes were conducted with the speaker at 45° from the room corner and two more with the speaker at 90° from the room center. These measure- ments of the living room wall included the front door to the house and a buttress in between the door and window at 90°. The door had large air leaks, which imply sound leaks. Figure 4-21 shows the combined pair of measurement episodes conducted with the loud- speaker at 45°, and the combined pair with the loudspeaker at 90°. It is not certain if, or how, these structural components affected the data, but it is clear the results were not what were A. Inside to Outside B. Outside to Inside Figure 4-17. Comparative front (left) and side (right) views for the whole wall composite sound flow.

Findings and Applications 91 Figure 4-18. Front view for each of the three measurements shown in Figure 4-17.

92 Evaluating Methods for Determining Interior Noise Levels Used in Airport Sound Insulation Programs expected. The data are very complicated, neither straight through nor at a common angle; thus far the research has not been able to interpret the results. 4.7.5.3 Analysis 4.7.5.3.1 Relative Analysis The research team made use of the results collected from the intensity measurements to examine the sound power flow from out to in versus in to out. Classical acoustic theory is that the TL is the same in both directions. For each comparison, there are SPL measurements on the sound source side of the room wall, and sound intensity measurements on the side opposite of the source. These are the composite images presented, where the in composite refers to the source outdoors and the out composite refers to the source indoors. The sound level for the source outdoors is 95.1 dB at the face of the wall, and the level is 108.3 dB at the face of the windows when the source is indoors. Figure 4-14 shows the placement of the source with respect to the dining room wall. Note: One can see the distance that the second measurement was from the wall compared to the first which was up against the wall. Figure 4-19. San Diego #1 vectors. Note: From the top of the page: top, side, and front views of the measurement near the wall (left side) and the measurement that begins about 30 cm (11.8 in) from the wall (right side), roughly twice the distance of the first measurement.

Findings and Applications 93 As an indication of the reasonableness of the 95.1 dB at the face of the wall, the measurements taken at 1.8 m (5.9 ft.) from the wall were consistently 93.9 dB (the loudspeaker at full volume is a very steady source). Table 4-18 shows a theoretical calculation of the level at the face of the wall based upon the measurement of 93.9 dB at a distance of 1.8 m (5.9 ft.) from the face of the wall. In com- parison, the level measured at the face of the wall is 95.1 dB compared to the prediction of 95.0 dB. For the TL measurements, the source side is given in terms of pressure in decibels and the receiver side is given in terms of power flow, both into or out of the dining room, as appropriate. So this TL result is not the traditional TL. However, the research team also calculates the traditional TL for some of the measurements for purposes of comparison. Above, the source side pressure levels are given as 95.1 dB and 108.3 dB for source outside and inside, respectively. The following discusses and develops the power flows into the receiver region that correspond to the two SPLs given above. The composite power is calculated in the following fashion. For the on-screen display2, the intensity meter calculates the total sound power collected over the duration of each of the separate Figure 4-20. San Diego #1 vectors. Note: From the top: top, side, and front views of the two measurements. 2 Inexplicably, the sound intensity instrument displays power in watts on the computer screen, and, as noted, this power is the total of all the vector powers. However, when the sound intensity instrument outputs the vectors to a file, it outputs vectors representing energy in Joules, which are the power multiplied by the duration of the episode.

94 Evaluating Methods for Determining Interior Noise Levels Used in Airport Sound Insulation Programs episodes that make up the composite. In calculating this power, the sound intensity instrument uses only those vectors going through the area of the wall designated by the user. Each vector gathered during a measurement episode is a separate estimate of the power being measured. The sound intensity instrument sums the power normal to the designated surface over all the vectors that constitute the episode. The best estimate of the total is the mean of the vectors, which is the reported total energy divided by the number of vectors. The results of this calculation are shown in Table 4-19 for the outdoor and indoor data separately, given in decibels. Figure 4-21. San Diego #5 vectors. Note: From the top, side, and front views of the respective pairs of combined measurement episodes with the loudspeaker at 45° (left side) and at 90° (right side).

Findings and Applications 95 For example, for the first line in Table 4-19, the total power calculated by the sound intensity instrument and displayed on the meter is a watt level of 72.1 dB, and it is gathered with 1242 vec- tors. In this case, subtract 10*log(1242) which is equal to 30.9 dB from 72.1 dB to obtain the mean estimated power in watts of 41.2 dB (to divide by a number when using decibels it is more convenient to subtract a dB-like transformation of the number n to 10*log(n). In this example, one needs to divide the total energy (joules in dB) by the number of vectors in dB, which is equal to 30.9 dB from 72.1 dB to obtain the mean estimated power in watts of 41.2 dB. There is one more major issue before the sound power can be estimated. The two composites developed for the Champaign house (in to out and out to in) each have the problem that the separate measurements going into the composite are not independent. Rather, there is overlap to a greater or lesser degree from one measurement to another. So, these five measurements that make up the out to in composite must be viewed as five not-so-independent estimates of the power flow. Posion measured Time data collected Number of vectors Sum of vector power estimates (was, dB) 10*log (number of vectors) dB Average power (was, dB) 4:24 p.m. 1,242 72.1 30.9 41.2 4:58 p.m. 1,896 71.7 32.8 38.9 Inside 5:08 p.m. 1,966 63.9 32.9 31.0 5:22 p.m. 1,437 67.2 31.6 35.6 5:52 p.m. 1,540 73.0 31.9 41.1 7:11 p.m. 1,687 82.9 32.3 50.6 Outside 7:20 p.m. 1,034 85.8 30.1 55.7 7:27 p.m. 1,642 83.7 32.2 51.5 Table 4-19. Calculation of composite power. Assume measured total at wall equals free field at wall plus 6 dB, 1.8 m (5.9 .) from wall; 7.3 m (23.9 .) from speaker, where the speaker is 9.1 m (29.8 .) from wall: 93.9 dB Assume wall is +6 dB for pressure doubling at ISO +3 dB "posi†on" when 1.8 m (5.9 ft.) from wall (ISO +3 dB range is within 1 to 2 m [3.3 to 6.6 .] from wall): 3 dB Difference for distance (speaker to mic = 7.3 m versus speaker to wall = 9.1 m) is 20*log(7.3/9.1): -1.9 dB Result: (Predicted, 93.9 + 3 – 1.9 = 95 dB) 95.0 dB For comparison, measured at wall: 95.1 dB Note: All levels in the table are A-weighted. Table 4-18. Validation of pressure measurements for out to in situations.

96 Evaluating Methods for Determining Interior Noise Levels Used in Airport Sound Insulation Programs As an example, consider the following two limit situations. In situation one, two measurements exactly replicate one another. In this situation, there are two totally dependent measurements and neither one alone is the correct result. In situation two, the two measurements do not overlap anywhere; they each measure different parts of the source. In this situation, the correct process is to sum the two results. In general, if one measurement is dominant, the solution is to take this highest estimate as the best estimate. Consider Table 4-20, which contains example pairs of decibel numbers and their resulting sum. In Table 4-20 the “sum” of two decibel levels that are more than 6 dB apart (these are in rows 7 and higher) have a resultant level that is not much higher than the higher of the two levels that are being “added” together. For all situations herein where two levels are being added together and they differ by 6 dB or more, the higher of the two is termed the “dominant” level, and in these situations the “sum” of the two levels is taken to be the higher level itself because the error is 1 dB or less. When the two decibel levels are less than 6 dB apart, they are usually “added” together, but the result is still relatively small (70 dB + 70 dB = 73 dB). So 3 dB is the largest number that can be added to the higher of the two levels being “summed,” and it occurs when the two levels to be combined are equal. The specific concern is combining the data from more than one measurement of the same general area (e.g., a window or door). As noted above, to get the correct answers, one needs to combine the different measurements on the same element such that data that overlap are only counted once. For combining two sets of measurements, note the following: if the two are totally independent then their sum ranges from 0 to 3 dB above the higher of the two measurements. It is +3 when the two measurements are equal and less than 1 dB when the two measurements differ by more than 6 dB. With respect to Table 4-20, the added level is effectively zero when the difference between the two measurements is greater than 20 dB. The second endpoint is when the two measurements are totally dependent. In this case, one of the two levels being “summed” is redundant and effectively discarded. The “sum” is just the higher level of the pair. So even if both levels are, for example, 70 dB, the “sum” is 70 dB. That is, one could get a result of 73 dB when the correct result was 70, or one could a result of 70 dB when the correct result was 73. So in both these endpoint cases, the largest an error can be is 3 dB. Row number A B Sum 1 70 70 73.0 2 70 69 72.5 3 70 68 72.1 4 70 67 71.8 5 70 66 71.5 6 70 65 71.2 7 70 64 71.0 8 70 63 70.8 9 70 62 70.6 10 70 61 70.5 11 70 60 70.4 12 70 50 70.0 Table 4-20. The sums of pairs of decibel numbers.

Findings and Applications 97 In the results herein, there is overlap among the 5 out to in measurement episodes, in particu- lar, episode 1 and 5 are almost equal, 41.2 and 41.1 respectively. In Figure 4.16 one can see that about 60% of episode 1 is overlapped by episode 5, and about 40% of episode 5 is overlapped by episode 1. A 40% overlap is just under 1.5 dB and a 60% overlap is almost 1.7 dB, so 1.6 dB is added to the highest single value. That is, 1.6 dB is added to 41.2 for a power estimate of 42.8 dB. For the in to out situations, episode 2 is dominant, so the 55.7 dB level of episode 2 is taken as the total power estimate. 4.7.5.3.2 Absolute Analysis In terms of comparing the TL from out to in versus in to out, A-weighted and flat-weighted source side pressures were measured using a precision sound level meter during the intensity measurements. The sound power levels in Table 4-20 are taken from Table 4-19 and show a reasonably good agreement between out to in and in to out, especially for the A-weighted values. The intensity meter display was flat-weighted so there was no way of get- ting A-weighted power levels directly from the meter. However, the sound intensity instrument also outputs a spreadsheet that includes one-third octave bands from 100 to 4000 Hz, and it is possible therefore to calculate the flat-weighted and A-weighted levels in one-third octave bands for both pressure and power. Appendix B provides further discussion on the spreadsheet output and contains a table of differences between flat-weighting and A-weighting. If two spectra differ only in amplitude, then the difference between A-weighting and flat- weighting for each of these two spectra is a constant. This suggests that for what is being mea- sured here, the difference between the flat-weighted power and the A-weighted power should be about equal for both indoor and outdoor measurements, since both are measurements of the same source with the same spectrum but in two different physical configurations: (1) the out- door measurement when the loudspeaker is outside and 9.1 m from the wall and (2) the indoor measurement when the loudspeaker is inside and 3.3 m from the wall. The data in Appendix B confirms this constant difference, which is shown in the appendix to be about 3 dB. The differences are all on the order of 3 dB between flat-weighted and A-weighted values. What this says is that the difference between A-weighted pressure and flat-weighted power is 3 dB compared to what one would get with A-weighted pressure and A-weighted power. That also suggests that to the first approximation, A-weighted pressure minus A-weighted power equals flat-weighted pressure minus flat-weighted power. And indeed in Table 4-21, the differ- ence between pressure (A-weighted) minus power (flat-weighted) and pressure (flat-weighted) minus power (flat-weighted) is 3.8 dB. Traditional TL A second comparison made is to calculate the traditional TL. This uses the data from the intensity meter, which includes sound pressure and sound velocity in addition to intensity for each vector. Table 4-22 lists the total A-weighted Leq for the entire duration of A Weighted Flat Weighted A & Flat Mixed Flat Weighted Flat Weighted Flat Weighted Pressure (dB) Power (dB) Difference (dB) Pressure (dB) Power (dB) Difference (dB) In to Out 107.4 55.7 51.7 111.2 55.7 55.5 Out to In 95.1 42.8 52.3 100.4 42.8 57.6 Differences: In to Out minus Out to In 0.6 In to Out minus Out to In 2.1 Table 4-21. Comparison of the measurement of transmission loss (out to in vs. in to out).

98 Evaluating Methods for Determining Interior Noise Levels Used in Airport Sound Insulation Programs each of the five measurements from out to in. This table shows a rather large spread of pressure levels, just under 3 dB. However, comparison of the SPL results with the sound power levels of Table 4-22 reveals a strong correlation between pressure and intensity. Because the estimate of sound power was based on only the two higher sound power levels, the estimate of the receiving room SPL is based on the two pressure levels that correspond to these two higher sound power levels. One SPL is 67.2 dB and the other is 67.4 dB, so the clear choice for the estimated sound pressure received in the dining room is 67.3 dB. The corresponding source side pressure as given above is 95.1 dB, so the indicated loss is 27.8 dB, which appears to be a reasonable value for the construction described earlier. Assessing the reasonableness of the estimate of power flow into the dining room In this section, the reasonableness of the estimate of power flow into the dining room is tested. As expressed above, the intensity meter simultaneously measures and records pressure and velocity as well as intensity. So for the out to in situation, the sound power is flowing into the dining room where a moderately reverberant field is established. With reference to Figure 4-14, the west end of the dining room should be the “reverberant” end of the dining room. This was somewhat verified by a walk around the din- ing room using a hand-held, Type 1 SLM with the loudspeaker as the source positioned indoors at about 30 cm (11.8 in.) before the middle of the entryway to the dining room from the living room. This walk around revealed a constant sound field for about the first 1.5 m (4.9 ft.) to about 1.8 m (5.9 ft.) from the window-wall (west wall). Since the intensity measurements herein extend to at most 0.6 m (2 ft.) or so from the window-wall, the instrument probe should be in a reverberant field. From the basic theory of room acoustics, 4 W2 op c a( )= ρ  where rco is the characteristic impedance of an acoustic wave in air, 415 Rayl, a is the room absorption in metric Sabine, P is the reverberant pressure in the room, and W is the sound power flowing into the room. This equation is used to find the total absorption, a, in the room and to compare this total with the calculated absorption based on the room’s furnishings. In decibels with W = 42.8 dB and P = 66.1 dB (the decibel levels measured herein), one gets: P 10 log 4 c 10 log , so, log W P 10 log 4 c 4.28 6.61 3.22 0.89 and 10^0.89 7.8 metric Sabine o o W a a a ( ) ( ) ( ) ( ) ( ) = + ρ − = − + ρ = − + = = =     Measurement Number Average A weighted Leq Sound Power Levels from Table 4 19 1 67.2 41.2 2 65.8 38.9 3 64.5 31.0 4 64.6 35.6 5 67.4 41.1 Energy Average: 66.1 38.9 Table 4-22. A-weighted Leq and sound power for the five measurement episodes that form the outdoor to indoor composite.

Findings and Applications 99 To assess the reasonableness of the estimate of power flow into the dining room, the total room absorption calculated using the room acoustics formula above is compared with the total room absorption based on surface sizes and their finishes, furnishings, and people in the room. The two openings to the dining room are the doorway to the kitchen and the large opening to the living room. These spaces are treated as having an absorption coefficient of 0.9. With reference to Fig- ure 4-14, one can see that the dining room is open to the kitchen and very open to the living room. The kitchen is completely open to the family room, which spans the north side of the house and has a 0.97 m (38 in.) by 1.72 m (68 in.) opening to a front hall that is very open to the living room, which extends west to an open doorway to the kitchen. Clearly, all the spaces are very open to one another. Thus, little sound is expected to be flowing back into the dining room and is estimated as equivalent to lowering the absorption coefficient from 1.0 to 0.9. With these caveats, the total absorption results for the dining room are calculated in Table 4-23 to be 8.3 metric Sabine (88.8 Imperial Sabine). This value of 8.3 metric Sabine, calculated from the room furnishings, compares favorably with the 7.8 metric Sabine (84 Imperial Sabine) calculated above using room acoustics, especially given the limits on the assumption of a reverberant space. 4.7.5.4 Enhancements What Constitutes TL When Measuring Intensity? The measurement of TL using intensity, of necessity, represents a departure from current airport sound insulation program practice to at least some degree. In theory, the power flow on either side of a wall should differ by the losses only due to the wall itself, whereas the current TL is calculated by the pressures impinging on the outside wall compared to the reverberant energy internal to the room. The research team has shown that, with intensity, the measurement from outdoors to in and indoors to out is the same, where on one side of the wall pressure is measured and the other side of the wall, power is measured. Thus, reciprocity has been demonstrated with the use of intensity. Dining Room Details Length (in.) Height (in.) Area (m2) Absorpon Coefficient Absorpon (a) Walls Whole window (west) wall 144 96 8.9 na na Window area 72 54 2.5 0.10 0.3 Wall area minus window area na na 6.4 0.05 0.3 South wall 140 96 8.7 0.05 0.4 Living room (east) wall (2*42") 84 96 5.2 0.05 0.3 Living room wall connued 57.5 16 0.6 0.05 0.0 North wall (24" + 42") 66 96 4.1 0.05 0.2 North wall connued 28.5 16 0.3 0.05 0.0 Floor 144 140 13.0 0.10 1.3 Ceiling 144 140 13.0 0.10 1.3 Openings Kitchen door 28.5 78 1.4 0.9 1.3 Living room entryway 57.5 78 2.9 0.9 2.6 Furnishings Three chairs, cushioned seats 18 18 0.2 0.6 0.1 People One person 0.2 Total Absorpon (metric Sabine) 8.3 Table 4-23. Calculation of total absorption in the dining room.

100 Evaluating Methods for Determining Interior Noise Levels Used in Airport Sound Insulation Programs Since there is reciprocity for both directions, the measurement from indoors to outdoors is the more promising direction. In this scenario, one creates a reverberant sound field in the room under test and determines the power flow from inside to outside by measuring the power flow out through the wall surface. It has been shown that this is feasible in the measurements at Champaign, and the difference between power flowing through the wall and power flowing through the window has been clearly measured. In theory one could have extended that and measured the intensity through the south wall that had no window, or through the roof which has a second story between it and the dining room. It would appear that one can measure the main wall and any hot spots, and beyond that it is not clear that measurements are warranted in any case. Measurements taken from indoors to outdoors have a number of advantages. By locating the sound source indoors, one can generate a very loud indoor level and easily measure the power flow from inside to outside by measuring intensity on the outside. With this method there are no resonance problems, so it appears that this methodology offers a means to make more repeatable measurements. From the research team’s measurements, it is clear that this new TL measurement would, of necessity, be numerically different than the old TL because the old TL was on the order of 25 dB and was the pressure difference on either side of the wall. For this new intensity measure- ment, one can get an idea of the difference between pressure and power as in Table 4-21 where the A-weighted difference between the energy average pressure and power is about 52 dB. This makes sense because it is known that the TL is about 25 dB, and (from Table 4-21) the difference between pressure and power for an intensity measurement is about 27 dB, for a total difference on the order of 52 dB. So, with this new measurement, the numerical values will be on the order of minus ~50 dB, rather than minus ~25 dB. Basic Power or Intensity Measurements In the scanning of the surface, the research team did essentially a three-dimensional scan that included the wall surface and about 30 to 60 cm (~1 to 2 feet) out from the wall surface, which is a three-dimensional volume being 30 to 60 cm (~1 to 2 feet) times the dimensional surface. When one stops to think about the process, it becomes clear that using more than the minimum depth required to scan the surface is not the right way to measure power through a surface. What one wants is the narrowest surface depth feasible and as close as possible to the wall through which the power flow is being measured. If one measures close to the wall with a depth of 30 cm [i.e., 0 to 30 cm (0 to ~1 feet)] from the wall], and one measures further from the wall such as a depth of 30 to 60 cm (~1 to 2 feet), then to the first order (except for dispersion), the power flow should be the same through either of these surfaces. The research team did this very test at the San Diego #1 site. As shown in Figure 4-19, the power flow from out to in looks very similar when measured within 30 cm (~1 foot) of the window, and within 30 to 60 cm (~1 to 2 feet) of the window. The only difference is that there is a little more dispersion at 60 cm (~2 feet) versus 30 cm (~1 foot). A similar test was done in Champaign for one of the five measurements for the indoor-measured cluster. The sound intensity was scanned both close to, and further from the wall/window for about the same time. To analyze this, the research team found the total energy within 30 cm (~1 foot) of the wall/window and within 60 cm (~2 feet) of the wall/window. As expected, when the distance (depth) from the wall is doubled, double the energy is measured. But, as noted ear- lier, the power should be energy per unit time and for purposes of this discussion the research team considers the time to flow 1 foot as the unit of time. So, flowing from 0 to 30 cm (0 to ~1 foot) yielded half the energy realized by flowing from 0 to 60 cm (0 to ~2 feet). But, in both cases, to calculate the power the vector sum is divided by time, where the time to go from 0 to 60 cm (~0 to 2 feet) is double the time required to go from 0 to 30 cm (0 to ~1 foot); therefore, the totals get divided by a given time increment for going from 0 to 30 cm (0 to ~1 foot) and double that time increment to go from 0 to 60 cm (~0 to 2 feet). So the power flow does not change and it is clear that the power flow should have been measured with just one layer from

Findings and Applications 101 0 to 30 cm (0 to ~1 foot) from the wall. Any greater depth, or surface removed from the wall, is either unnecessary, less precise, or both. The best strategy is to measure as close to the wall as possible, with the shortest depth possible. For most of the situations in this study, this ideally would have been 0 to 10 cm (~0 to 4 inches) from the wall or perhaps as much as 0 to 15 cm (~0 to 6 inches) from the wall. The goal should be to measure within this narrow distance from the wall, uniformly across the entire surface under test. A great assistance to the user would be software that divided the wall into sectors designated by the user and based on such things as hot spots and the resolution desired. For example, consider the wall in the Champaign dining room discussed in Section 4.7.4.3. As shown in Figure 4-22, the window is designated as a hot spot and divided into four elements for measurement. The adjacent walls are subdivided in a similar but not equal fashion. As shown, the dividing lines for the window are normally maintained to the extremities of the wall, giv- ing the same spacing in one direction off the window. Beneath the window, two wall elements are shown that have the same width as elements of the window, but the vertical height of these wall elements is not generally equal to the vertical spacing of the window elements. Four more elements are shown to the sides of the window, two on each side, that follow the vertical spac- ing of the window, but for these four elements the horizontal spacing will not generally equal that of the window. In the two lower corners, neither direction will normally equal that of the window. In the top two corners, the vertical spacing of the windows is not used because of the small distance from the top of the window to the ceiling, about 20 cm (8 inches). Rather, these two corner elements are set as shown. Finally, directly above the window it again departs from following the lines of the window because of the short height. In a practical way this might be accomplished by the computer drawing the standard witness lines outside the hot spot and the user designating to the computer which elements to merge. As a special case, if one had a uni- form wall without windows, then the entire wall would be treated as a “hot spot,” with no part of the wall outside the hot spot. More complicated situations would need to be accomplished by dividing the wall into sub- walls. For example, if there was a room with two separate windows of different size, one could Figure 4-22. Test wall divided into hot spot (window) and the remainder (hard wall), with each divided into sub-elements.

102 Evaluating Methods for Determining Interior Noise Levels Used in Airport Sound Insulation Programs take the point midway between the two windows, or something close to that, and use the vertical line that goes through that point as the line dividing the wall into two sub-walls, each of which would be treated separately as above. In designating elements, the smallest permitted element dimension would be (for the current sound intensity microphone assembly) 5 cm (2 in.), and as a practical matter the smallest ele- ment dimension should be closer to 10 cm (4 in.). In any event, no element dimension should be shorter than the distance that sound intensity instrument can resolve based on the setup in use. As a practical aid to the user, a screen should display the division of the wall into elements, where each element would initially be portrayed in some color such as red. This color would change to a contrasting color, such as green, when the requisite number of vectors had been gathered in that element. In this way, the operator can see what has been filled in adequately and what still needs to be filled in. Also, the software should produce a big (very prominent) warn- ing in real-time when the instrument is outside of the specified box (e.g., too far from the wall). These capabilities become increasingly important as the elements become smaller. Given that the largest surface area the sound intensity instrument can address at one time is about 2 × 2.5 m (6.5 to 8.2 ft.), and that the camera must be 2.5 meters back from the wall, the resolution is certainly no more than 10 cm (3.9 in.), and perhaps even larger. If the maximum resolution was 10 cm (3.9 in.), then one could divide the 2 × 2.5 meter (6.5 × 8.2 ft.) area into 500 10 × 10 cm (3.9 in.) elements, 25 across the 2.5 meter (8.2 ft.) width and 20 spanning the 2 meter (6.5 ft.) height. With 5 vectors per element, this would result in 2,500 vectors, which is probably the maximum that could be envisioned. The overarching requirement for making unambiguous measurements of power flow through a surface is that there be a single defining vector for each element that represents the average of all the vectors through that element, that all the elements be independent and not overlapping, and that the sum of the elements equals the area of the surface under test. The single defining vector for each element would represent the power flowing through that element. The component of the real power normal to that element would be the real power flowing through that element’s por- tion of the wall under test. The report should provide the intensity for the surface of each element, which is the power through that element divided by the area of that element. The data included in the report should be the area of the element, the unique element number given to that element by the software or the user, the position of that element in the wall surface, and the intensity at the surface of the element, or alternatively, the power normal to the surface of the element, or both. (Although not relevant to sound through walls, similar concepts, although more complicated geometrically, could be applied to conformal surfaces that were used to surround a test object.) The most important point is that all overlap must be avoided. One cannot tolerate a situa- tion where there is overlap between two or more elements, and the sum of elements must cover the entire wall under test. With the current version of the sound intensity instrument software, it appears to be virtually impossible to precisely measure the sound power flow through a wall. Additional Processing Problem to Avoid in Addition to Overlap As indicated above, when the sound intensity instrument calculates a power as displayed on the attached computer screen, this power is the sum of the individual vectors that go through the user-selected rectangular area. In this application, the vectors apparently have units of watts. When one prints out the vectors in a spreadsheet, the vectors, apparently, have units of joules (watts multiplied by time). Appendix B has a detailed discussion of the screen displayed vectors and the outputted vectors, and how they differ. Advanced Analysis Possibilities Potentially, software can be developed to more or less auto- mate the total measurement process on the basis of just a few user-specified parameters. In general,

Findings and Applications 103 the user is interested in mapping the power flow through a wall or through a conformal surface that surrounds some machine. The most general analysis would divide the surfaces into logical rela- tively large elements. Logical means that the window might be broken into certain-sized elements and the wall would be broken into similar but not necessarily identical elements. Steps to the automate process would be: 1. The user would designate to the sound intensity instrument the number of vectors through each element desired. 2. The user would then proceed to measure for as long as it took for all the elements under measurement to turn from “red to green” as described above. 3. The user would specify the largest standard deviation acceptable. 4. The software would calculate the mean and standard deviation of the data measured for each element. It is expected that a minimum of about 25 vectors will be required to meet standard deviations on the order of a couple of dB. Note: with a scan size of 2 by 2.5 m (6.5 to 8.2 ft.), this 25 vector per element requirement suggests that the smallest element size be about 20 by 25 cm (8 × 10 inches) in order to keep the total number of vectors in a reasonable range. 5. The measurements would be complete for elements having standard deviations within the specified limit and the power flow would be given by the average of the vectors measured for that element. 6. If the standard deviation for any element exceeded the standard deviation criterion, this would indicate that that element had some hot spots in part of the area of the element compared to the rest. In that case, the software should divide that element into four sub-elements and the measurement for just that element should be again repeated until the user-specified number of vectors is measured for each sub-element. The standard deviation should then be calculated by the sound intensity instrument for each sub-element to see if it now meets the standard deviation criterion. 7. This process should continue until the element size is smaller than the maximum resolution that the system is capable of, about 5 cm (1.8 in.) in a 1 m × 1.5 m (3.2 ft. × 4.9 ft.) area or 10 cm (3.6 in.) in a 2 m × 2.5 m (6.5 ft. × 8.2 ft.) area. The only things the user would need to specify would be the number of vectors per element (generally 25 vectors should be a sufficient quantity), and the standard deviation desired (it is expected that the standard deviation will be in the range from 2.5 to 5 dB). Of course, the user would need to wave the sound intensity instrument over the surface area in question until all areas were covered to the degree required. This may entail several data collections necessitated for resetting the gyroscopes. Measuring more than the minimum number of vectors in any element is not a problem, as long as all the vectors in that element or sub-element are averaged together. But again, ultimately, there must be one average vector for each element or sub-element, and the power flowing normal to the surface of that element or sub-element must be the real part of the single average vector through that element or sub-element, and the intensity must be the power represented by this vector divided by the area of the element or sub-element. In addition to the above there are several minor suggestions as follows: 1. As it is now, the cradle of the sound intensity instrument must be placed on a level surface such as a table, some ledge, or a chair or the floor. The sound intensity instrument cradle could be supported by a camera tripod if there were a 1/4-20 camera tripod mount situated at the balance point for the sound intensity instrument cradle with wand. This would provide much greater flexibility in positioning the cradle. 2. As it is now, the calibration tool must be placed on a flat surface and must lean against some vertical surface. The checkerboard, like the cradle, needs greater flexibility in positioning and supporting it. The checkerboard should include holes in the two upper corners with a small cord affixed to these two holes for hanging the checkerboard in various places. This could

104 Evaluating Methods for Determining Interior Noise Levels Used in Airport Sound Insulation Programs include a pair of magnets with a hook on one of the magnets so that magnets could be on either side of drapery, or a shade, and would then support the checkerboard cord on the hook. Or, the magnet with the hook could be affixed to metal surfaces like a filing cabinet, and again supporting the checkerboard cord by the hook on the magnet. Also, the cord on the checker- board could be attached to a hook that had a 1/4-20 camera tripod mount, and supported by another tripod. There needs to be more ways to conveniently support the instrument cradle and the checkerboard. 3. There needs to be several colors for the light on the sound intensity instrument so that good contrast is obtained with various lighting conditions and various wall colors. The current whitish light works well indoors with little sunlight in the room, but does not work well out- doors or in a sunlit room, except at dusk or night. 4.7.5.5 Conclusions 1. From the measurements, the research team has been able to conclude that intensity can form the basis for portraying sound flow through a wall into a room, but that the enhancements described above are necessary for these measurements to be feasible and obtainable in accept- able time duration. 2. Measuring intensity minimizes, or eliminates, the problem generated by reflections off of surfaces and other problems generated by complex (i.e., real and imaginary) pressure waves, because only the real power through the wall is measured. 3. The research team has shown that measuring from indoors to outdoors, and using reciproc- ity is the clearest way to use intensity for TL measurements. Under this scenario, one creates a loud reverberant field in the indoor space and measures the power flow out of that space to the precision and the extent required. This likely amounts to a redefinition of TL, but, as discussed in the report, measurements made this way hold the promise of being much more repeatable because the resonant effects of using pressure are eliminated. However, the defini- tion and numerical value for measured TL change and the numerical values will be on the order of minus ~50 to 55 dB, rather than minus ~25 dB. 4. Currently, resonance effects for a microphone outdoors and close to the house under test are subject to ground reflection and wall reflection issues that can create large errors. The most obvious means to gain regularity is to position the microphone where it measures the free field impinging on the house. On the other hand, creating a reverberant field inside a room and measuring intensity outside makes measurements feasible on any day that the weather is not too poor outside to make acoustical measurements, e.g., too much wind or precipitation. These measurements will be repeatable and will have good signal-to-noise ratio. 4.8 Comparison of Results Across All Methods Table 4-24 provides a comparison of all results across measurement and calculation methods. The research team found the following: 1. There is decent agreement between all methods, once corrections have been applied. 2. The flyover and elevated loudspeaker methods had the lowest NLR difference and lowest standard deviation. 3. The ground-level loudspeaker had a relatively high standard NLR difference, meaning it may be under predicting NLR performance. This may be because of the difference in angle of incidence between a ground-level source and an elevated source such as an aircraft/elevated loudspeaker. 4. The interior loudspeaker, with the 5-dB correction, had a low NLR difference but high stan- dard deviation. This puts into doubt its validity at this time. 5. The spreadsheet and IBANA calculation methods tended to over predict NLR.

Findings and Applications 105 4.8.1 Outliers There were some rooms measured where the noise reduction results fell outside of the expected range of noise reduction and/or the flyover noise reduction varied significantly from the exterior loudspeaker noise reduction. A deeper analysis was performed for each of these rooms to deter- mine why the results were atypical. The following summarizes the findings: • San Diego #1 – Dining Room: The difference between the flyover and exterior loudspeaker noise reduction was 9 dB, which is significantly more than the expected 3 to 4 dB. Why: Based on the review of the airport’s typical arrival flight tracks (viewed via the air- port’s tracking software), a majority of the arrival flight paths are just south of the home. The dining room is at the north end of the home, and the dining room windows would be Average NLR (no int. spkr.) Diff. from Avg NLR Elevated Loudspeaker Diff. from Avg NLR Ground Loudspeaker Diff. from Avg NLR Interior Loudspeaker Diff. from Avg NLR Diff. from Avg NLR Diff. from Avg NLR Residence Room Flyover Spreadsheet IBANA San Diego #1 Living 26.2 26.3 -0.1 24.5 1.7 24.3 2.0 28.1 -1.9 26 0.2 Dining 26.4 28.0 -1.6 22.2 4.3 25.2 1.2 28.5 -2.1 27 -0.6 San Diego #2 Living 26.1 24.3 1.9 25.2 1.0 23.2 2.9 28.1 -2.0 27 -0.9 Office 23.3 20.4 2.9 18.4 4.9 22.8 0.5 27.3 -4.0 27 -3.7 San Diego #4 Living 23.4 23.3 0.1 21.6 1.8 22.6 0.8 24.7 -1.3 24 -0.6 Bedroom 1 26.8 26.4 0.4 25.0 1.8 26.9 -0.1 29.7 -3.0 26 0.8 San Diego #5 Living 25.0 24.0 1.0 21.3 3.7 23.6 1.4 29.8 -4.8 25 0.0 Bedroom 1 29.0 30.0 -1.1 26.3 2.7 35.5 -6.6 30.5 -1.6 29 -0.1 San Diego #6 Living 21.9 24.9 -3.0 21.0 0.9 20.9 1.0 23.1 -1.2 17.8 4.1 25 -3.1 Dining 27.5 27.4 0.1 23.7 3.8 26.4 1.1 26.1 1.4 31.0 -3.5 29 -1.5 San Diego #7 Family 26.4 26.5 -0.1 25.6 0.8 25.7 0.7 22.8 3.6 27.2 -0.8 27 -0.6 Living 25.1 25.1 0.0 23.8 1.3 24.6 0.5 18.5 6.7 27.5 -2.4 24 1.1 San Diego #8 Dining 21.2 19.4 1.9 19.9 1.3 19.6 1.7 20.9 0.3 23.3 -2.1 24 -2.8 Master Bed 26.7 22.2 4.5 28.6 -1.9 28.3 -1.6 27.3 -0.6 27.3 -0.6 27 -0.3 San Diego #9 Living 21.3 21.8 -0.5 19.8 1.5 19.4 1.9 21.8 -0.5 22.3 -1.0 23 -1.7 Bedroom 1 28.3 31.7 -3.4 26.1 2.2 25.6 2.7 27.6 0.7 30.3 -2.0 28 0.3 San Diego #10 Bedroom 1 20.1 17.5 2.6 18.9 1.2 17.6 2.5 23.7 -3.6 23.5 -3.4 23 -2.9 Bedroom 2 25.9 25.0 0.9 25.3 0.7 25.1 0.8 27.0 -1.1 28.3 -2.4 26 -0.1 Boston #6 Dining 23.1 20.2 2.9 25.0 -1.9 25.0 -1.9 25.0 -1.9 21.3 1.8 24 -0.9 Bedroom 2 22.3 22.0 0.3 25.1 -2.8 24.9 -2.6 26.9 -4.6 19.3 3.0 20 2.3 Boston #8 Living 24.8 24.2 0.6 25.0 -0.2 24.8 0.0 27.7 -2.9 25.0 -0.2 25 -0.2 Study 27.0 29.5 -2.5 26.6 0.4 25.7 1.3 23.7 3.3 26.2 0.8 27 0.0 Average Difference 0.4 0.5 1.4 0.1 -1.3 -0.7 Standard Deviaon 1.9 1.7 1.8 2.9 2.1 1.4 Notes: 1) Correcons applied—Flyover: 2 dBA, Exterior loudspeaker: 2 dBA, Interior loudspeaker: 5 dBA. 2) Average NLR difference calculated by first averaging all of the NLR across all measurement methods (except interior loudspeaker and sound intensity), and then subtracng the NLR from one method (e.g., flyover) from the average NLR. Interior loudspeaker not used in average as this method does not follow naonal standards and the 5 dB correcon applied to the data is based on limited field measurements (i.e., not fully veˆed). 3) All differences are calculated by subtracng from the average noise reducon. 4) Blank cells indicate no elevated loudspeaker test was performed at the corresponding residence. Table 4-24. Comparison of NLR calculations to acoustical measurements.

106 Evaluating Methods for Determining Interior Noise Levels Used in Airport Sound Insulation Programs shielded from aircraft noise during the flyover measurements from the large overhang at the east façade and the building itself for the north façade. The living room windows are not shielded (i.e., they have full line-of-sight to the arriving aircraft), and the room did not have the large flyover vs. loudspeaker noise reduction difference. • San Diego #6 – Living Room: The difference between the flyover and exterior loudspeaker noise reduction was 7 dB, which is significantly more than the expected 3 to 4 dB. Why: Based on the field notes, the large difference between the flyover and loudspeaker measurements is due to the presence of a pass-through air conditioning (PTAC) unit in the living room window. PTAC units provide little noise reduction and serve as a significant path of noise intrusion into a unit. The exterior loudspeaker test resulted in much lower noise reduction than the flyover test, as the loudspeaker is pointed directly at the PTAC, whereas an aircraft flyover is above the PTAC. During an aircraft flyover, there is shielding provided by the PTAC sheet metal enclosure and, possibly, the angle of incidence from the flyover results in less aircraft noise intrusion via the PTAC. The measurement results show significantly more mid- to high-frequency noise intrusion. – Dining Room: The difference between the flyover and elevated loudspeaker measurement was 6.8 dB, whereas the difference between the flyover and ground-level loudspeaker was 4.1 dB. Why: The reason for this difference is the same as the living room explanation above, as the dining room is open to the living room. • San Diego #8 – Master Bedroom: In this case, the flyover noise reduction was lower than the loudspeaker noise reduction by approximately 3 dB; typically, flyover NLR are 3 dB higher than loudspeaker NLR. Why: The loudspeaker noise reduction was higher than the flyover noise reduction because the research team was only able to generate loudspeaker noise at the side wall of the bed- room. The rear wall of the bedroom contained additional windows and a door (which was acoustically weak), but the research team was not able to direct the loudspeaker noise to the rear façade due to the detached garage located just outside of the bedroom. The research team would have expected much lower noise reduction from the loudspeaker test if it were able to generate noise at both the side and rear bedroom walls. • San Diego #9 – Bedroom 1: The difference between the flyover and exterior loudspeaker was approximately 9 dB versus the expected 3 to 4 dB. Why: There is a solid overhang at the bedroom window. Using noise modeling software, the research team modeled the noise level from an aircraft flyover with and without the overhang. The noise level at the bedroom window was 7 dB lower with the overhang. The overhang does not shield noise generated by the loudspeaker, thus it makes sense that the measured flyover noise reduction was significantly higher than the loudspeaker noise reduction (the overhang serves to increase the noise reduction of the windows). • Boston #6 – Dining Room: The flyover noise reduction was lower than the loudspeaker noise reduction by approximately 2 dB; typically, flyover NLR are higher than loudspeaker NLR. Why: The dining room is open to the living room and kitchen. During the flyover measure- ments, noise enters the home via all of these rooms and is measured by the meter in the dining room. For the loudspeaker measurement, noise is only generated toward the dining room and the meter only picks up noise intrusion via the dining room façade. There are significant paths of noise in the connected living room and kitchen (e.g., exterior doors, PTAC units), which lowers the measured noise reduction during the flyover test.

Findings and Applications 107 4.9 Comparison of Measurement Results from Loudspeaker and Flyover Testing It is reasonable to assume that if both the loudspeaker and the flyover measurement methods are valid and properly conducted, the NLR results from the two should closely agree. Unfortu- nately, this is not always the case, as shown by measurements for this study and previously those for the ATAC study, “Study of Noise Level Reduction (NLR) Variation” (Landrum and Brown, 2013). The ATAC study concludes, Figure 39 and Figure 40 summarize the measured variations as a result from various measurement methods, parameter changes, and absorption changes. The total variation of NLR measurements is com- prised of many causes, each introducing their own variations to the total. This study includes a subset of a number of possible causes that contribute to the total NLR variation. Section 6.1 listed various aspects that contribute to the NLR variation. To quantify the total NLR variation, the variation of individual components that contribute to the total NLR needs to be quantified separately. This ACRP Project 02-51 study carefully conducted loudspeaker and flyover measurements at SAN on the same rooms, identically furnished, generally on the same day. Thus, certain param- eters such as changes in acoustical absorption or architectural modifications were eliminated. Nonetheless, notable differences in NLR measurement results were encountered from the raw data results. However, when various adjustments were made in accordance with national and international standards for microphone position, the various methods agreed closely. To the knowledge of the research team, the ATAC study and this study are the first time that careful comparative acoustical measurements had been made to assess the differences in mea- surement results from the flyover and loudspeaker methods. Measurements made in the course of the RSIP’s do not make redundant measurements in the interest of time, efficiency, and cost. Acoustical theory and experience leads to two likely alternative possible causes for the discrepancies: • Angle of incidence: Acoustical theory and measurement show that under some circum- stances, sound impinging on a building element at a grazing angle may transmit sound more effectively (providing less sound attenuation) than that normal to the element or at moderate angles. • Insufficient flyover high-frequency sound energy: Most of the high-frequency sound energy from aircraft flyovers is absorbed by the atmosphere before it reaches a home, and the sound insulation of the building envelope further reduces this energy to levels that may be at or below the ambient high-frequency sound level in the residence. 4.9.1 Angle of Incidence All U.S. and international standards for sound insulation measurement and reporting specify “random incidence” testing whereby sound impinges the test specimen equally at all angles. This is achieved by creating a diffuse and random sound field in the sound source room while record- ing a spatial average of the sound field in the receiving room. Early laboratory sound TL tests for various window glass (circa 1960’s) often report STC values two to three points above those from later tests. This is because the early tests failed to create a random sound field; sound at grazing incidence was not effectively achieved. However, the effect of grazing incidence is complex and varies considerably with the type of material being acoustically tested. Figure 4-23 conceptually shows the primary TL effects in various frequency regions for a composite building element such as a wall section composed of wood paneling or stucco on one side, batt insulation in the interstitial space, and gypsum board on the other face. At random

108 Evaluating Methods for Determining Interior Noise Levels Used in Airport Sound Insulation Programs incidence, materials are typically acoustically isotropic in that they produce the same TL when tested from either side of the building element. The actual frequencies for the first panel resonance, other resonances, and coincidence areas may be computed from the mass and stiffness properties of the various materials. However, incidence effects are more prominent in monolithic materials, such as glass, than in composite building elements like walls and roof/ceiling systems. Additionally, incidence effects are greatest in the high frequencies, above the coincidence region, and generally minimal at low frequencies. Under ideal conditions, a difference in TL between random and optimum grazing angle (90°) of 5 dB is possible for monolithic materials, and up to 3 dB for composite materials. The effects of grazing incidence on residential sound insulation from aircraft have been studied by the National Research Council, Canada (Bradley et al. 2002, Bradley 2002). Incidence effects were found to be entirely negligible at loudspeaker angles down to 30°. However, at full grazing incidence a nominal 3 dB correction is found. The effective (corrected) incident level Ls″(f) is calculated: Ls f Ls f 10 log 180 D f dBi( ) ( ) ( )′′ = + φ     Where f is the horizontal angular view of the fly-by and 0 ≤ f ≤ 180. (Note that ASTM E966 and this report Figure 4-24 use ‘f’ for the vertical angle and ‘q’ for the horizontal angle.) The optional correction simply relates the incident sound energy to the portion of the aircraft flyover that is visible at the façade with a small empirical correction for diffraction. That is, the incident sound energy is reduced when the aircraft flyover is not completely visible at the façade. However, the application of angle of incidence is more complicated in practice for the fol- lowing reasons: • Building facades are composed of several various sized elements, monolithic and composite, each with different angle of incidence TL properties. • While the vertical angle, f, from the flyover may be fairly constant, the horizontal angle, q, varies throughout the event. • While one façade is receiving grazing incidence, the perpendicular side is receiving normal incidence. In corner rooms both occur at once. • As shown in Figure 4-24, part of the home is shadowing another portion of the home during part of the flyover. Figure 4-23. Theoretical transmission loss (Bies and Hansen, 2003).

Findings and Applications 109 • Façades are composed of different elements, some more susceptible to angle of incidence than are others. • The TL is not isotropic but orthotropic, meaning that the TL varies with direction. The actual model for orthotropic TL is more complicated than that outlined in Figure 4-23. It is therefore highly impractical to establish a protocol to account for angle of incidence effects. These effects also make it difficult to distinguish the effects of angle of incidence from insufficient flyover high-frequency sound energy in quantifying the sources of the loudspeaker and flyover NLR result discrepancies. The incidence effects discussed here are different than the coincidence dip effects of dual- glazed windows discussed in Appendix C. 4.9.2 Insufficient Flyover High-Frequency Sound Energy Turbine aircraft noise is generally broadband near the source; that is, it has fairly equal acous- tic energy per bandwidth. However, the effects of spherical radiation and atmospheric absorp- tion substantially attenuate the noise at RSIP residences, particularly in the higher frequencies. Figure 4-25 shows the attenuation of a Boeing 737 (300 thru 500 series) departure using the FAA standard spectra at 305 meters (1,000 feet) (John A. Volpe National Transportation Systems Figure 4-24. Aircraft flyover diagram.

110 Evaluating Methods for Determining Interior Noise Levels Used in Airport Sound Insulation Programs Center, 1999). The attenuated spectra were computed from the FAA spectrum at 1,000 ft., tak- ing into account spherical radiation and atmospheric absorption from the ANSI S1.26 standard. If insufficient flyover high-frequency sound energy were a significant factor, it would be expected that those residences with higher ambient interior noise would be the most influenced; that is, the difference between the loudspeaker and the flyover NLR would be the greatest for the higher ambient noise rooms measured. However, when this hypoth- esis was examined, it was determined that the low signal-to-noise ratio of interior flyover noise in the higher frequencies is a minor issue and generally does not create a significant discrepancy. However, several homes did encounter alteration of NLR results when the interior high- frequency ambient noise levels came close to the interior flyover noise levels. For this reason, the research team recommends that all exterior and interior measurements be band limited to an upper frequency limit of 5 kHz. 4.10 Suggested Research With this project, the research team has clarified several sources of systematic error and iden- tified improved ways to do these tasks. But not all of the improvements have been accomplished, since some of these improvements require further research. Figure 4-25. B737 departure noise level by distance.

Findings and Applications 111 4.10.1 External Sound Spectra A significant factor that affects the measured noise reduction is the external sound spectra (i.e., the noise “signature” of aircraft overflights at a given airport). The external DNL frequency spec- trum to be modeled (from the FAA Part 150 program Noise Exposure Map) is, in itself, unknown since it is the annual energy average of all aircraft over a particular location, with the 10 dB night- time penalty (a single nighttime flyover is equal to 10 daytime flyovers of the same level), under all annual meteorological conditions; it can only be estimated. The spectra of louder aircraft should be biased on an energy basis; for instance, a single flyover at 90 dB should be averaged equally with 10 flyovers at 80 dB. Many commercial airports with significant incompatible residential land use (i.e., above 65 DNL) have noise monitoring systems. These currently only measure A-weighted SEL values of individual flyover events, and compute daily DNL values. However, these monitors may be modified to collect spectral information in terms of SELs. This would allow for computation of the daily DNL aircraft spectrum at that location. Future programs could use this information to design the exterior sound spectrum to be used in testing and evaluation of NLR values. Since the sound spectra could not be addressed in the field measurements conducted for this report, the issue of external sound spectra is beyond the scope of this study. Additional research and investigation is suggested. 4.10.2 Ground and Façade Reflection for Flyover Measurements As noted in the study, the research team applied a 2-dB correction to the exterior flyover measurements to account for ground reflection. The research team believes that up to a 4 dB correction is needed if there is ground reflection in combination with reflections off the façade being tested (based upon modeling and calculations made). This reflection phenomenon, with respect to exterior flyover measurements, should be investigated in more detail so the correc- tions can be standardized and codified. 4.10.3 Interior Loudspeaker The research team found that the interior loudspeaker measurements resulted in systematically high NLR values. Based on additional measurements conducted after the initial round of field measurements, the research team applied a 5-dB correction to the interior loudspeaker NLRs. This correction is based upon reverberant noise build-up measured inside of the room. However, this correction is not codified in any standards, nor are any other aspects of the interior loud- speaker measurements (e.g., position of the receive microphone on the outside of the building). Further research should be conducted to standardize the interior loudspeaker measurement method so that results can be comparable to the exterior loudspeaker and flyover measurement methods. 4.10.4 Sound Intensity Prominent on this list of future research is the use of an indoor loudspeaker with measure- ments made outdoors, since this offers so many clear advantages, such as there being no problem with neighbors, no problem with microphone placement (at least for intensity), good signal-to- noise ratio, and the smallest uncertainty. Some of the research questions would be: 1. Where and how should the indoor sound in the source room be measured? 2. Where and how should the outdoor sound be measured? a. Should a wall be divided into its elements, i.e., the windows, doors, and regular wall con- sidered all separate, or as the total combined partition?

112 Evaluating Methods for Determining Interior Noise Levels Used in Airport Sound Insulation Programs [NOTE: The above task 2.a is the same for both sound intensity and the interior loud- speaker described just above. These two methods depart in terms of the outdoor measure- ments. The intensity will be measured on or just beyond the surface of the wall to measure the power flowing from the reverberant room. For the interior loudspeaker measurements using pressure (microphones), one needs one or more microphones at some “free field” or pressure-doubling positions to measure just the integrated power flowing from the test room.] b. Develop methods and procedures to measure the power flowing through the wall surface. c. Improvements to the intensity meter are detailed in Section 4.7.5.4. These include: a. One mandatory software requirement for making unambiguous measurements of power flow through a surface is that there be a single defining vector for each element that represents the average of all the vectors through that element, that all the elements be independent and not overlapping, and that the sum of the elements equals the area of the surface under test. b. The second mandatory software requirement is that all overlap must be avoided. One cannot tolerate a situation where there is overlap between two or more elements, and the sum of elements must cover the entire wall under test. c. Several small hardware changes. d. Several major enhancements that would be “nice to have.” 4.10.5 Field Measurement Uncertainty Another research need is data that can better quantify measured field uncertainties. For exam- ple, the research team suggests herein that the tolerance on positioning a repeat measurement to the original measurements is unknown. Data are needed to answer this multifaceted question. The term multifaceted is used because the comparison can be the same technicians and equip- ment doing the same measurements in the same room twice, or it could be as varied as different people from a different company with different equipment and no knowledge of the placement of the equipment by the previous company. The research team proposed that the emphasis would be on the simplest situation since most often, the same company and people make the before and after measurements in the same house. The goal would be to understand the variation in results based primarily on the tolerance of the equipment placement and perhaps some second-order factors. This testing would provide a better understanding of the tolerance of repeatability under the most careful methods and conditions.

Next: Bibliography »
Evaluating Methods for Determining Interior Noise Levels Used in Airport Sound Insulation Programs Get This Book
×
 Evaluating Methods for Determining Interior Noise Levels Used in Airport Sound Insulation Programs
MyNAP members save 10% online.
Login or Register to save!
Download Free PDF

TRB's Airport Cooperative Research Program (ACRP) Report 152: Evaluating Methods for Determining Interior Noise Levels Used in Airport Sound Insulation Programs provides guidance for selecting and implementing methods for measuring noise level reduction in dwellings associated with airport noise insulation programs. The report complements the results of ACRP Report 89: Guidelines for Airport Sound Insulation Programs.

READ FREE ONLINE

  1. ×

    Welcome to OpenBook!

    You're looking at OpenBook, NAP.edu's online reading room since 1999. Based on feedback from you, our users, we've made some improvements that make it easier than ever to read thousands of publications on our website.

    Do you want to take a quick tour of the OpenBook's features?

    No Thanks Take a Tour »
  2. ×

    Show this book's table of contents, where you can jump to any chapter by name.

    « Back Next »
  3. ×

    ...or use these buttons to go back to the previous chapter or skip to the next one.

    « Back Next »
  4. ×

    Jump up to the previous page or down to the next one. Also, you can type in a page number and press Enter to go directly to that page in the book.

    « Back Next »
  5. ×

    To search the entire text of this book, type in your search term here and press Enter.

    « Back Next »
  6. ×

    Share a link to this book page on your preferred social network or via email.

    « Back Next »
  7. ×

    View our suggested citation for this chapter.

    « Back Next »
  8. ×

    Ready to take your reading offline? Click here to buy this book in print or download it as a free PDF, if available.

    « Back Next »
Stay Connected!