Fundamentals of Architectural Acoustics PDF

C H A P T E R 17 Fundamentals of Architectural Acoustics...

C H A P T E R 17 Fundamentals of Architectural Acoustics ACOUSTICS THE ACOUSTIC ENVIRONMENT PLAYS AN problems. Population density in office spaces has important role in supporting (or disturbing) an over- steadily increased, thus raising the amount of noise all sense of comfort in many of the spaces we occupy generated and decreasing the distance between on a daily basis—including both residential and occupants. Worse yet—from the acoustics point commercial/institutional spaces. For several rea- of view—many offices today are designed as open sons, many design solutions seem to shortchange the areas with, at best, only thin partial-height dividers acoustical environment. This is partly due to the per- (cubicles) separating workers. Forty percent or more ceived complexity of architectural acoustics, partly of a building budget may be allocated for mechani- due to lack of coverage of the topic in many archi- cal systems—most of which generate noise. Outside tecture programs. Good acoustics is not required by noise sources, such as cars, trucks, trains, and air- most building codes and is not a key element in the planes, can also present problems and require isola- majority of green building rating systems. Neverthe- tion of interior spaces from exterior sounds. This is less, providing acceptable acoustical conditions is a particularly problematic where natural ventilation fundamental part of good design practice. is used. Building owners and tenants are aware that quality acoustic environments are required for high 17.1 ARCHITECTURAL ACOUSTICS productivity and comfort in buildings and hence competitive rental or purchase values. The archi- Architectural acoustics may be defined as the tect is expected to provide such acoustic quality. A design of spaces, structures, and mechanical/elec- clear understanding of the principles explained in trical systems to meet hearing needs. With proper this and the following chapters will assist the archi- design efforts, wanted sounds can be heard properly tect in preparing straightforward designs alone and unwanted sounds (noise) can be attenuated and, in more complex instances, by knowledgeable or masked to the point where they do not cause collaboration with an acoustic consultant. Proper annoyance. Achieving good acoustics, however, acoustic design responses early in the design pro- has become increasingly difficult for a variety of cess are critically important, as after-the-fact acous- reasons. To lower construction costs, the weight of tic “repair” is often difficult (and, therefore, costly) various materials used in many of today’s buildings and sometimes impossible without substantial has been reduced from those prevalent 50–100 structural alterations (which are very costly). years ago. Since light structures generally transmit All acoustical situations have three common sound more readily than heavy ones, lightweight elements—a sound source, a sound transmission buildings have the potential for major acoustical path or paths, and a receiver of the sound. Through 739  740 CHAPTER 17 FUNDAMENTALS OF ARCHITECTURAL ACOUSTICS design, a source can be made louder or quieter and TABLE 17.1 Speed of Sound Propagation in a path can be made to transmit more or less sound. Various Media Working with sources and paths throughout the Speed various design phases (and into construction and Medium Meters per Second Feet per Second often occupancy) is the bread and butter of archi- tectural acoustics. The listener’s perception of Air 344 1130 Water 1410 4625 sound also may be influenced, although this is not Wood 3300 10,825 normally an “architectural” solution. This chapter Brick 3600 11,800 Concrete 3700 12,100 presents the fundamental bases of architectural Steel 4900 16,000 acoustics to assist a designer in defining appropriate Glass 5000 16,400 acoustic intents and criteria. Moreover, it describes Aluminum 5800 19,000 basic methods for reaching such intents through Note: These figures are approximate, since the listed materials vary the design process. in density. An average frequency is assumed. standpoint, the speed of sound in air is slow enough ACOUSTICS 17.2 SOUND that the travel time of a sound signal can be a key design issue. Sound can be defined in a number of different ways, depending upon the aspect of most interest or con- cern. Thus, sound can be described as a physical (b) Wavelength wave, or as a mechanical vibration, or simply as a The wavelength of a sound is defined as the distance series of pressure variations in an elastic medium. between similar points (peaks or troughs) on suc- For airborne sounds, the medium is air. For struc- cessive waves, which is the distance a sound travels ture-borne sounds, the medium may be concrete, in one cycle. The relationship among wavelength, steel, wood, glass, or combinations of these materi- frequency, and speed of sound is expressed as als. A much more limited definition of sound, more c appropriate to architectural acoustics, is that it is λ= (17.1) f simply an audible pressure variation. This estab- where lishes that architectural acoustics is concerned with the building occupant. It also suggests that there λ = wavelength, ft (m) may be inaudible pressure variations that cannot c = velocity of sound, fps (m/s) be heard. This is the case with vibration, which is a pressure variation that can be felt but not heard. f = frequency of sound, Hz Low-frequency sounds are characterized by long (a) Speed of Sound wavelengths, and high-frequency sounds by short wavelengths. Sounds with wavelengths ranging Sound travels at different speeds, depending upon from 0.5 in. to 50 ft (12 mm to 15 m) can be heard the medium. In air, at sea level, sound velocity by human beings. A simple nomograph (see Fig. is 1130 fps (344 m/s). This is 770 miles per hour 18.18) permits rapid determination of wavelength, (1239 kilometers per hour)—slow indeed when given sound frequency, and vice versa. compared to light, which has a speed of 186,000 miles per second (299,338 km per second). Since (c) Frequency sound travels not only in air but also through parts of a structure, it is of interest to know the speed of The number of times that a cycle of compres- sound in other media (Table 17.1). For architectural sion and rarefaction of air occurs in a given unit design purposes, speed variations due to changes in of time is described as the frequency of a sound. temperature and altitude (atmospheric pressure) For example, if there are 1000 such cycles in 1 may be ignored and, for most calculations, 1130 fps second, the frequency of the sound is 1000 cps— (344 m/s) may be used as the speed of sound in 1000 hertz (Hz) in standard nomenclature. Thus, air (usually within 3% error). From a practical in Fig. 17.1, higher frequencies would be shown  SOUND 741 ACOUSTICS Fig. 17.1 Sound pressure waves. (a) The continuous vibration from the bell causes a series of compressions and rarefactions of the air to travel outward longitudinally from the source. Amplitude information is carried by pressure; that is, greater amplitude means greater compression and greater rarefaction. Compression and rarefaction are shown diagrammatically as line density, although they are actually molecular phenomena, as shown in the upper drawing. (b) Two single impulses of different magnitude (amplitude) travel- ing away from the source. Note how amplitude information is carried by the difference in pressure. by compressions and rarefactions that are closer the upper frequencies, whereas most of the acoustic together, and lower frequencies by those that are energy exists in the lower frequencies. The critical farther apart. In architectural acoustics, frequency frequency range for speech communication is 300 is sometimes referred to using a term borrowed from to 4000 Hz. Overtones outside these core frequen- music—pitch. The higher a sound’s frequency, the cies, however, give the voice its characteristic sound higher its pitch, and vice versa. and specific identity. Telephone and radio commu- Sound frequency is integrally linked to speech nication are accomplished using a considerably and hearing. The approximate frequency range of a narrower frequency band by sacrificing some voice healthy young person’s hearing is 20 to 20,000 Hz. quality and intelligibility (see Fig. 17.2). The human speaking voice has a range of approxi- A sound composed of only one frequency is mately 100 to 600 Hz in fundamental frequencies, called a pure tone. Except for the sound generated but harmonics (overtones) reach to approximately by a tuning fork, few sounds are truly pure. Musi- 7500 Hz. Most speech information is carried in cal sounds (tones) are composed of a fundamental  742 CHAPTER 17 FUNDAMENTALS OF ARCHITECTURAL ACOUSTICS ACOUSTICS Fig. 17.2 The positions of speech and wide-range music in the aural field of the human ear are illustrated. Speech is in the nominally linear response area of the ear, as is most music. Beyond these frequencies, the ear effectively attenuates the incoming signal. frequency and integral multiples of the fundamen- tal frequency (harmonics). Most common sounds Fig. 17.3 Schematic representations of a pure tone, a musical are complex combinations of frequencies. Figure note, and a more complex sound (such as speech, music, or 17.3 shows examples of pure tones, musical notes, noise), showing the variation of sound pressure with time and and common sounds, while Fig. 17.4 shows the frequency. frequency ranges of some common devices and phenomena. a way of capturing frequency-specific information about sounds without becoming buried in detail. (d) Octave Bands (e) The Concept of Sound Magnitude The frequencies in the scale of Fig. 17.4 all stand in the ratio of 2:1 to each other—that is, The magnitude of a sound signal is a relatively 16:32:63:125:250, and so on. Borrowing again complex concept because there are a number of dif- from musical terminology, they are one octave ferent metrics (and associated terminology) in com- apart. These particular frequencies are also accepted mon use and because of the great range of values internationally as the center (reference) frequen- involved in day-to-day acoustic situations. Sound cies of octave bands used for the purpose of sound magnitude is often equated with loudness, which specification. For technical reasons, a geometric is a subjective, receiver-oriented response not lin- mean is used. Thus, 250 is the center frequency early related to the power of a sound (in watts). The of an octave band ranging from 250/ 2 to physical magnitude of sound is variously described 250 2 , with that particular octave being known as sound power, sound power level (PWL), sound pres- as the 250-Hz octave. If a finer division is required sure, sound pressure level (SPL), sound intensity, and for analysis purposes (unusual in architectural sound intensity level (IL). Each of these metrics has acoustics), ½-octave and 1 3-octave bands are used. a place, and each differs from the others and from Because frequency is so important to architectural subjective loudness. To clearly understand these acoustics, and because there are 19,980 discrete important concepts a comprehension of how we whole frequencies in the normal range of hearing, hear and how sound is propagated in free space is octave bands are used repeatedly during design as necessary.  HEARING 743 ACOUSTICS Fig. 17.4 Frequency ranges of common instruments. Wavelength (λ) is calculated on an assumed propagation velocity of 344 m/s (1130 fps). (f) Sound Propagation sounds that are inaudible to most adults can be a source of extreme annoyance to young students. For For simplicity, it is probably best to view sound as example, dentists report that high-speed drills and a series of pressure variations. In air, these pres- tooth-cleaning devices cause extreme auditory dis- sure variations take the form of periodic compres- comfort in many young patients. These devices pro- sions and rarefactions. The bell in Fig. 17.1 radiates duce sounds in the 15- to 20-kilohertz (kHz) range. a tone in all directions equally—that is, it creates a circular wavefront. As the material of the bell vibrates, it sets up vibrations of the same frequency (a) The Ear in the air, which can best be visualized in a sectional view. Notice that the pressure changes containing Referring to Fig. 17.6, the outer ear is funnel-shaped the sound information travel in the same direction and serves as a sound-gathering input device for as the wavefront—longitudinally. Sound is there- the auditory system. Sound energy travels through fore a longitudinal wave motion. This is unlike (for the auditory canal (outer ear) and sets in motion example) an AM radio signal, in which the wave the components of the middle ear, comprising the travels longitudinally but the information—that is, eardrum, hammer, anvil, and stirrup. The stirrup the modulation—is transverse. Sound is a mechan- acts as a piston to transmit vibrations into the fluid ical wave, whereas light and electricity are electro- of the inner ear. The motion of this fluid causes magnetic waves. movement of hair cells in the cochlea, which, in turn, stimulates nerves at the bases of the hairs. The nerves, in turn, transmit electrical impulses 17.3 HEARING along the eighth cranial nerve to the brain. These impulses we understand as sound. As noted in Section 17.2(c), the approximate fre- It is often assumed that the ear ignores phase quency response of a healthy young person’s hear- differences and combines frequencies. This may not ing is 20 to 20,000 Hz. The upper limit decreases always be the case, however, particularly when with age as a result of a process called presbycusis the frequencies are very far apart. For this reason, (Fig. 17.5). The loss is more pronounced in men a single-number representation of a complex sound than women. Recognition of this phenomenon can (the type of information found in Table 17.4) can be be of importance in schools, since very high-pitched misleading and must be used with caution.  744 CHAPTER 17 FUNDAMENTALS OF ARCHITECTURAL ACOUSTICS ACOUSTICS Fig. 17.5 These curves illustrate the average hearing threshold shift for men and women with increasing age, relative to a 25-year-old reference group. Note that at age 60 the male threshold shift at upper speech frequencies (4000 Hz) is 30 dB, compared to 20 dB for women, a difference of 10 dB or one-half of the subjective loudness (see Table 17.2). For men older than 50 and women older than 60, frequencies above 10 kHz are effectively inaudible, and even at normal speech frequencies of 1000 to 2000 Hz, subjective loudness is reduced to one-half of that of a 25-year-old listener. (Reproduced with permission from F. A. White. 1975. Our Acoustic Environment. Wiley, New York.) At the threshold of hearing (approximately of approximately 0.25 mm (0.01 in.)—an astonish- 0 dB), the displacement of air molecules imping- ing range indeed. ing on the eardrum, and the eardrum excursion, A number of “measures” of sound magnitude are approximately one angstrom unit (1 Å = 10–8 are encountered in architectural acoustics. Move- cm), which is approximately the diameter of an ment of the eardrum (and thus hearing) is caused atom. Were the ear an order of magnitude more directly by air pressure variations. Therefore, the sensitive, it would hear thermal noise. The human physical magnitude generally of most interest to ear is thus operating close to the practical limit of architectural acoustics is sound pressure (force sensitivity. At the other end of the magnitude spec- density), and its derivative—sound pressure level— trum, the threshold of pain corresponds to a sound which is the ratio of a given sound pressure to a pressure level of 130 dB and to an eardrum motion base level, expressed in decibels. (b) Equal Loudness Contours TABLE 17.2 Intensity Level Changes and Corresponding Subjective Loudness Changes The human ear is not uniformly sensitive over its entire frequency range of 20 Hz to 20 kHz. The Change in Intensity 120- to 130-dB upper limit (pain threshold) occurs Level (dB) Subjective Change in Loudness at all frequencies. At the lower limit, however, the 3 Barely perceptible 0-dB threshold occurs only at 1000 Hz. The ear is 6a Perceptible 7 Clearly perceptible in fact most sensitive at 3000 to 4000 Hz, at which 10 Twice (or half) as loud frequencies the threshold is about −5 dB (relatively 20 Four times (or one-quarter) as loud speaking). This type of nonlinear response exists aSix decibels corresponds to the change encountered when the dis- throughout the ear’s hearing range. To determine tance to the source in a free field is doubled (or halved). the nature of this nonlinearity, a large number of  HEARING 745 ACOUSTICS Fig. 17.6 Drawing of the human ear showing the principal parts and its function as a sound receptor. (Drawing reproduced with per- mission from F. A. White. 1975. Our Acoustic Environment. Wiley, New York. Notes are reproduced from Quieting: A Practical Guide to Noise Control. 1976.) tests were conducted with pure tones of different 1. All points on a single contour have the same frequencies, in which listeners were asked to equate subjective sensation of loudness. the subjective loudness of signals. These test results 2. The loudness level in phons (the number shown produced a family of curves called equal loudness in the center of each curve in Fig. 17.7) of the level contours (sometimes called Fletcher–Munson entire contour is defined by the decibel level of equal loudness contours after two of the principal that contour at 1000 Hz. researchers). These curves (Fig. 17.7) are inter- nationally recognized and standardized, and are The phon scale was constructed as a way used as the reference for normal hearing response. of defining perceived sound (subjective loud- They are also used to “weight” measuring devices, ness impressions) in terms of the ear’s nonlinear as explained later in this chapter. Note that by response. Assigning a single number—in a unit definition: called the phon—to an equal loudness contour  746 CHAPTER 17 FUNDAMENTALS OF ARCHITECTURAL ACOUSTICS ACOUSTICS Fig. 17.7 Standard equal-loudness contours. These curves are accurate for a listener with normal binaural hearing situated in the near field of a source producing pure tones directly ahead of the listener. The subjective (perceived) loudness of each contour is quantified in a unit called the phon. The phon value of each curve is shown in the center of the figure. Note that a sound pressure level of 60 dB corresponds to 30 phons at 50 Hz, 50 phons at 100 Hz, 63 phons at 500 Hz, 60 phons at 1 kHz, 68 phons at 4 kHz, and 60 phons at 6 kHz. This indicates the relative flatness of the ear’s response in the central frequency range and the sharp drop at low frequencies. makes it possible to compare subjective loudness 2. Maximum sensitivity occurs between 3 and impressions of two sounds, regardless of frequency 4 kHz—precisely the frequencies that convey or actual sound pressure level. Thus, in Fig. 17.8, the most information in human speech (see the subjective loudness of a whisper in the ear (20 Fig. 17.2). phons) is the same as that of distant pounding surf, 3. In a normal listening range of 45 to 85 dB, and despite the fact that the sound pressure level (SPL) in the most often used frequency range of 150 of the former is 20 dB and that of the latter is 80 dB. Hz to 6 kHz, the contour is substantially flat; The 60-dB differential is due to the ear’s sharp drop that is, the ear’s response is effectively linear in in sensitivity at low frequencies. However, because this zone. It is only at extremes of sound level the phon scale is nonlinear in terms of perceived and frequency that nonlinearity occurs. loudness (a given phon differential does not cor- The ear “averages” sounds over some mini- respond to the same perceived loudness change mum time increment, and sound impulses of throughout the phon scale) and because phons shorter duration will be perceived as quieter than cannot be combined arithmetically to give a resul- they would if received as a steady-state sound. tant subjective loudness (60 phons plus 50 phons is This minimum sound pulse length (i.e., the time not 110 phons), the scale has found few uses (one of required for the sound to achieve full loudness) which is to rate the noisiness of bathroom fans). varies within the range of 50 and 200 milliseconds The Fletcher–Munson contours demonstrate (ms) (1 20 to 1 5 s) and depends to an extent upon some interesting phenomena: the frequency content of the sound. The figure of 1. Sensitivity drops off sharply at low frequencies, 70 ms (1 14 s) is frequently found in the literature. particularly at low dB levels. (For this reason, This threshold time becomes important when con- most stereo amplifiers provide automatic bass sidering the effect of echoes (see Sections 18.11 to boost at low volume levels.) 18.14).  HEARING 747 ACOUSTICS Fig. 17.8 Common sound sources plotted at their dominant frequencies and levels as typically heard by the observer. The equal- loudness contours (see Fig. 17.7) show why certain sounds seem louder than others, despite the pressure levels that would indicate the contrary (Reprinted with modification from Quieting: A Practical Guide to Noise Control, 1976.) (c) Masking since it depends in part upon how “hard” a listener is listening. Masking is an extremely important and When two separate sound sources are perceived useful technique for noise control, wherein back- simultaneously, the perception of each is made more ground sound levels are deliberately manipulated difficult by the presence of the other. This effect is to mask other unwanted sounds (see Chapter 19 known as masking, which is defined technically as for details). The background sounds used for this the number of decibels by which the threshold of purpose are of a broadband continuous nature, audibility of one sound is raised by the presence of such as “pink” noise (see the Glossary in Chapter another sound. The masking effect is greatest when 19, Section 19.37), and are non-information- two sounds are close in frequency or frequency con- bearing. They serve to depress the intelligibility of tent, since the ear has greater difficulty separating lower-magnitude, information-bearing sounds like frequencies. Also, a low frequency will mask that would otherwise cause annoyance, and are a high frequency more effectively than the reverse particularly useful in open (landscaped) office for the same decibel levels. With broad-frequency plans where few noise control alternatives exist. sounds, the masking effect is difficult to predict, Frequently, mechanical equipment sound (such as  748 CHAPTER 17 FUNDAMENTALS OF ARCHITECTURAL ACOUSTICS that from air diffusers and fans) can be utilized to ates the unwanted sound signals. Normally, how- good advantage as masking sound. ever, the ear does precisely the opposite. It combines sounds that are clearly distinct from each other in frequency and phase. The three tones in a chord (d) Directivity struck on a piano are different in frequency and, The exact mechanism by which the binaural if played as a very rapid triplet, out of phase. Yet aspect of hearing detects direction is not entirely the ear combines them and hears a single sound, understood. It is clear, however, that the brain despite the fact that the maxima of the three tones can sense the direction of a sound source in many do not occur simultaneously. environments. This is a useful trait in both social and dangerous situations. In enclosed spaces, echoes from surface reflections (reverberation) will 17.4 SOUND SOURCES blur most directivity, and any “stereo” information will be almost completely dependent upon high fre- Building occupants encounter a wide range of quencies in the near field (the sound field that is sounds in the course of a day. Some of these sounds ACOUSTICS close to the source and therefore minimally affected are produced by human beings, others by mechani- by reverberation). Since high frequencies (with short cal equipment or natural phenomena. Some of wavelengths) travel in a relatively straight line, they the sounds are information-bearing, while others will reach a listener before their reflections and in so convey nothing intelligible. Speech and music are doing will cue the hearer as to their origin. Low fre- sounds of particular interest to building designers, quencies (and thus long wavelengths) are difficult to as are those sounds deemed to be noise. localize, however, because the wavelengths are large compared to the dimension between our ears. Low (a) Speech frequencies, even in the near field, are readily mixed with their reflections, thus confusing the location of As can be seen from Fig. 17.2, the ear’s sensitiv- their origin. ity is highest in the speech frequency and the nor- mal energy range. Individual speech sounds vary in duration between 30 and 300 ms, and the ear (e) Discrimination normally perceives these individually and clearly. Frequency recognition is accomplished in the Speech is composed of phonemes, which are individ- cochlea by the hair cells embedded in the basilar ual and distinctive sounds that, to an extent, vary membrane. The human auditory field spans the from language to language. Certain phonemes exist frequencies between 20 Hz and 20 kHz, although in one language and not in another. Since some the threshold of sensitivity is different at each fre- phonemes carry more information than others, it quency. The ear can hear and recognize distinct is these that good architectural acoustics must be frequencies, yet the hearing mechanism has the particularly careful to preserve and support in order ability, apparently as directed by the brain, either to to maintain intelligibility. In English, consonants hear individual frequencies or to combine them into carry much more information than vowels, as can a single more complex sound. Thus, when we hear readily be demonstrated by writing a sentence (1), a string quartet we can, generally at will, hear either then rewriting it without consonants (2), and then the entire quartet or each instrument individually. without vowels (3): With concentration (vision helps in this regard), a 1. Most speech energy is concentrated in the 100 trained ear can pick out a single instrument in an to 600-hertz range. orchestra of 120 pieces, even if there is more than 2. o ee eey i oeae i e 100 o 600 e ae. one such instrument in the same section. Conduc- 3. Mst spch nrg s cncntrtd n th 100 t 600 Hrtz tors do this regularly. Similarly, the ear can perform rng. the discrimination feat known as the cocktail party effect, that is, pick out one voice among background The male voice centers its energy at around 500 sounds that may be substantially louder than the Hz, the female voice at around 900 Hz. It is, how- wanted signal. In effect, the ear in this case attenu- ever, in the high frequencies that consonants have  EXPRESSING SOUND MAGNITUDE 749 most of their energy. Phonemes such as s and sh, baseline reference value. Use of the qualitative met- for example, have most of their energy above 2 kHz, ric of loudness is also common. and both are particularly important in conveying intelligible content. (a) Sound Power Normal speech averages between 55 and 65 dBA sound pressure level at 3 to 4 ft (0.9 to 1.2 m) Sound power is an independent property of a sound from the source, with a dynamic range from about source that quantifies the source’s acoustical out- 30 dBA for soft speech to somewhat above 65 dBA put. Sound power is constant for any given source for loud speech (at the same distance). Extremes of operating under defined conditions (a certain level speech are 10 dBA for a soft whisper and 80 dBA for of speech effort or a rotating speed on a motor) and a shout, but in both of these instances intelligibil- is not influenced by the nature of the surroundings ity is sharply reduced because of lack of consonant into which a source is placed. Thus, the sound power power. Indeed, in shouting, emphasis is actually on (output) of a chiller or orchestra is not changed by vowels, so that it is generally accepted that a 70- the distance to a receiver or the characteristics of a dBA sound pressure level is about the upper limit mechanical room or auditorium. ACOUSTICS of fully intelligible human speech. Singers who fre- Sound power is expressed in watts (of acousti- quently exceed 90 dBA do so at great loss of intel- cal power) and varies widely from source to source. ligibility. Some selected sound power values are: a jet engine, Another result of the high-frequency content 100,000 W; a symphony orchestra, 10 W; a loud of consonants, and therefore intelligibility, is its radio, 0.1 W; normal speech, 0.000010 W. Note directiveness. The higher the frequency, the greater the wide range of values just in this sample: from a sound’s directivity and the less its diffraction (abil- 105 to 10–6 W. Sound power values for natural ity to be heard beyond a partial barrier). There- phenomena need to be obtained from an appropri- fore, intelligibility of speech is greatest directly in ate information source. Sound power data for man- front of a speaker and least behind him/her. High- ufactured equipment and devices can be obtained frequency tones are most easily absorbed and least from the manufacturer. Manufacturers can often diffracted. provide information on the sound power of a product under specified conditions. Manufacturers cannot provide accurate information on sound (b) Other Sounds pressure levels, as these are influenced by environ- Instrumental music is much broader in dynamic mental variables beyond the manufacturer’s con- range and more complex in frequency than speech. trol or knowledge. It has no direct parallel to intelligibility. A person’s “reception” of music is a combination of physiologi- (b) Sound Pressure cal and psychological phenomena. As such, it is an experience beyond the scope of this book to cover in Sound pressure is the deviation from ambient air depth, but it is briefly examined in the subsequent pressure that is caused by sound waves. Sound discussion of room acoustics, auditoriums, and pressure is instigated by the acoustic power output music halls. Noises are dealt with in Section 17.6. of a sound source, but is modified by the environ- ment between the source and the receiver. Sound power is a characteristic of a source; sound pres- 17.5 EXPRESSING SOUND MAGNITUDE sure is the effect of a source as experienced at some specific location. Sound pressure must be referenced Six different quantitative measures of sound mag- to a particular point in a space, as pressure will nitude are commonly encountered in architectural usually vary from location to location in a room. acoustics. Three of these—sound power, sound Sound pressure is expressed in pascals (Pa) (or pressure, and sound intensity—are absolute mea- microbars in the I-P system). It is common practice sures; the other three—sound power level, sound to use SI units for all architectural acoustics met- pressure level, and sound intensity level—are ratio rics—even in the United States. Sound pressures for values that compare an absolute measure to a some common situations include: near a jet plane,  750 CHAPTER 17 FUNDAMENTALS OF ARCHITECTURAL ACOUSTICS TABLE 17.3 Comparison of Decimal, Exponential, and Logarithmic Statements of Various Acoustic Intensities Intensity (W/cm2) Intensity Level, Decimal Notation Exponential Notation Logarithmic Notation (dB) Examples 0.001 10−3 130 Painful 0.0001 10−4 120 0.00001 10−5 110 75-piece orchestra 0.000001 10−6 100 0.0000001 10−7 90 Shouting at 5 ft (1.5 m) 0.000000001 10−9 70 Speech at 3 ft (0.9 m) 0.00000000001 10−11 50 Average office 0.0000000000001 10−13 30 Quiet, unoccupied office 0.00000000000001 10−14 20 Rural ambient 0.000000000000001 10−15 10 0.0000000000000001 10−16 0 Threshold of hearing ACOUSTICS 200 Pa; the threshold of pain, 20 Pa; a loud night- (10,000,000,000,000:1). Table 17.3 gives an idea club, 2 Pa; next to a highway, 0.2 Pa; and normal of the physical significance of these numbers. Note speech, 0.02 Pa. As with sound power, note the sub- that the maximum (painful) acoustic intensity in stantial range of values in these sample situations. Table 17.3 is 0.001 W/cm2 or 10 W/m2. Sound intensity diminishes over distance. A point sound source of constant power radiating (c) Sound Intensity in free space—that is, at a location far from the The threshold of hearing—that is, the minimum effects of any reflecting surface—is represented sound intensity (I) that a normal ear can detect—is in the drawing of Fig. 17.9. The sound inten- 10−16 W/cm2. (The ear actually responds directly sity at any (defined) distance from the source is to pressure variations, but such pressures involve expressed as various energy densities.) The maximum sound P (17.2) I= intensity that the ear can accept without dam- A age is approximately 10−3 W/cm2. This gives where an intensity range of 1013, or 10 trillion to 1 I = sound (power) intensity, W/cm2 Fig. 17.9 The same total energy passes through areas A1 and A2. Since A is proportional to the square of r, the energy density or intensity is inversely proportional to r2.  EXPRESSING SOUND MAGNITUDE 751 P = acoustic power, W the wavelength of the sound produced. This type of A = area, cm2* source produces spherical waves. Line sources, such as strings, produce cylindrical waves. Large vibrat- Since the sound radiates freely in all ing surfaces, such as walls, produce plane waves. directions. The importance of these distinctions will become P clear in the discussion of sound barriers and diffrac- I= W/cm2 (17.3) 4πr 2 tion in Chapter 19. where r is the radius of an imaginary sphere enclos- ing the sound source. This is an implementation of (d) The Decibel the classic inverse square law, stating that intensity is Two problems arise immediately when dealing inversely proportional to the square of the distance from with quantities of the type encountered with sound the source. (In I-P units, this is power, pressure, and intensity. The numbers them- P selves are very small or very large. Furthermore, I= W/ft 2 (17.4) 930 × 4πr 2 the human ear responds logarithmically, not arith- metically, to sound pressure (or intensity); that is, ACOUSTICS since 1 ft2 = 930 cm2.) Using Equation 17.3 to doubling the intensity of a sound does not double determine the intensities I1 and I2 at distances r1 its loudness—such a change is barely perceptible. and r2 from point source P, we find that the intensi- To address these problems, it would be much more ties at distance r1 and r2 from the source stand in convenient if there were a scale that: the ratio I1 r2 2 1. Started at zero for the minimum sound (inten- = (17.5) sity or pressure) that can be heard. I2 r1 2 2. Used whole numbers rather than powers of 10. Figure 17.10 shows graphically how a sound pulse 3. Had some fixed relationships between an is attenuated in strength (but not in waveform) as arithmetic difference and a loudness change, it travels outward from a source by the action of say, 10 units equals a doubling (or halving) of distance. loudness. Thus, on such a scale, the difference The preceding derivation is based upon a point between 20 and 30, between 60 and 70, would source—that is, a source that is small relative to always be a doubling of loudness. Such a scale exists. It is the decibel scale. *It is traditional in architectural acoustics to express area The word level when appended to power, pres- in square centimeters, although the SI system requires sure, or intensity indicates a quantity expressed area to be stated in square meters. Conversion data for relative to a base quantity—in decibels. Intensity units are given in Table 19.16. See also Table 19.17 for a level is thus the ratio between a given intensity and listing of symbols and abbreviations. a base intensity. If intensity level is expressed as I IL = 10 log (17.6) I0 where IL = intensity level, dB I = intensity, W/cm2 I0 = base intensity (i.e., 10−16 W/cm2, the thresh- Fig. 17.10 Attenuation of a sound signal in air as it travels away old of hearing) from the source. The shape (information) remains constant when traveling in a nondispersive medium such as air. This is not the log = logarithm to base 10 case with travel in solids, where different frequencies travel at different velocities, causing a wave-shape change with time and then a scale is established that satisfies the three distance. Since velocity of propagation is constant, time and conditions established above. The quantity IL is distance are linearly related and can therefore be plotted on the same axis. (Drawing by Jonathan Meendering; ©2004 by Walter dimensionless, since it indicates simply a ratio Grondzik; all rights reserved.) between two numbers. It is expressed in decibels,  752 CHAPTER 17 FUNDAMENTALS OF ARCHITECTURAL ACOUSTICS however, to clarify its status as a ratio quantity. This I1 = (10−16) (106) = 10−10 W/cm2 proves a convenient way to express the large range By similar calculation, of sound magnitudes encountered. Table 17.3 shows the great convenience of using the loga- I2 = 10−11 W/cm2 rithmic decibel scale compared to either decimal 2. Add the intensities arithmetically. notation or exponential notation. Table 17.2 gives a short listing of subjective loudness changes I1 + I2 = 10−10 + 10−11 expressed in decibels. Note that 10 dB indicates a = (10 × 10−11) + 10−11 doubling of loudness, and 20 dB is loudness doubled twice—that is, a situation four times as loud. The Itot = 11 × 10−11 W/cm2 difference (Δ) between any two intensity levels can 3. Reconvert to decibels. To find the intensity level be expressed as (IL) corresponding to the combined or total I2 I intensity I1 + I2, simply apply Equation 17.6: ΔIL = IL2 − IL1 = 10 log − 10 log 1 I0 I0 I ILtot = 10 log tot V0 ACOUSTICS Therefore, ΔIL = 10 log I2 dB 11× 10−11 (17.7) = 10 log I1 10−16 A few examples using decibel notation and logarith- = 10 (log 11 + log 105) mic calculations should help establish this useful = 10 (1.04 + 5) system. By the way, the bel in decibel is in honor of = 60.4 dB Alexander Graham Bell (thus the capital B in dB). which is only a fraction larger than the original 60 dB of the stronger of the two sounds. As dem- onstrated in this example, decibels cannot be added EXAMPLE 17.1 Two sound sources (I1 and I2) pro- arithmetically. duce intensity levels of 60 and 50 dB, respectively, at a point. When these sources are operating simul- taneously, what is the total sound intensity level? EXAMPLE 17.2 Assume two sounds of 60 dB each. (Assume identical frequency content and random What is the combined sound intensity level in deci- phasing—that is, the phase relationship between the bels? two sources changes in a random manner.) SOLUTION SOLUTION Note that this example deals with intensity level, not One method would be to calculate levels as in intensity, since intensity itself has little significance Example 17.1. A shorter method is to find the dif- for architectural acoustics. The technique involved in ference between the sum and either of the (equal) adding two sound intensity levels has three steps: signals and add it to either individual signal. Using Equation 17.7: 1. Convert both to actual intensity. I I ΔIL = ILcomb − IL1 = 10 log comb IL = 10 log I1 I0 2I1 so = 10 log I1 I 60 = 10 log −116 10 = 10 log 2 = 10 (0.30) or I1 = 3 dB 6.0 = log 10−16 This answer (which is independent of any partic- ular sound level) yields the extremely important and Then, using the definition of a base 10 logarithm: useful fact that doubling a signal’s intensity raises I the intensity level by 3 dB. (In this case, the com- 106 = −116 10 bined intensity level would be 60 dB + 3 dB, or 63  EXPRESSING SOUND MAGNITUDE 753 dB.) Similarly, quadrupling a signal’s intensity raises 20 Isingle the resulting level by 6 dB. This is because = 10 log Isingle 4I ΔIL = 10 log = 10 log 20 I = 10 log 4 = 10 (1.3) = 13 dB = 10 (0.60) Therefore, the total sound intensity level will be = 6 dB ILtot = 80 dB + 13 dB = 93 dB Therefore, quadrupling 60 dB gives 66 dB (or, A chart for combining the decibel levels of two alternatively, 60 dB + 60 dB = 63 dB and 63 dB + sources is given in Fig. 17.11 that eliminates the 63 dB = 66 dB). This technique is very useful when somewhat lengthy procedure detailed in Example combining a large number of identical sound levels, 17.1. Referring to Table 17.2, note that the human as in the following example. ear is not responsive to fractional decibel changes; indeed, even a 3-dB change is barely perceptible. This being so, it is recommended that the detailed ACOUSTICS calculations, and even the chart, be reserved for EXAMPLE 17.3 A factory will contain 20 identical situations where a high degree of accuracy is machines, each of which generates a sound inten- required. sity level of 80 dB. What will be the combined sound For everyday calculations, the following ap- intensity level? (Ignore issues of frequency content, proximations may be used to combine the decibel phase, and sound fields.) levels of two sources: When the difference between two sources is SOLUTION 1 dB or less, add 3 dB to the higher decibel level to obtain the total. ΔIL = ILtot − ILsingle When the difference is 2 to 3 dB, add 2 dB. Itot = 10 log When the difference is 4 to 8 dB, add 1 dB. Isingle When the difference is 9 dB or more, ignore the lower-level source (add 0 to the higher). Fig. 17.11 Chart for adding two uncorrelated sound pressure levels when both are expressed in decibels. (Reprinted by permission from E. B. Magrab. 1975. Environmental Noise Control. Wiley, New York.)  754 CHAPTER 17 FUNDAMENTALS OF ARCHITECTURAL ACOUSTICS TABLE 17.4 Addition of Uncorrelated Sound are based results from a source in an open (unen- Pressure Levels closed), obstruction-free space (e.g., outdoors). Also, intensity measurement I1 must be taken at a dB Levels Sum sufficient distance d1 from the source that a free Lower Higher Approximatea Accurateb field has developed. See the discussion in Section 60 60 63 63.0 18.7 for an explanation of sound fields. 60 62 64 64.4 60 64 65 65.5 60 66 67 67.0 SOLUTION 60 68 69 68.7 From Equation 17.7 it is known that 60 70 70 70.5 I ΔIL = 2 a I1 See text for approximation rules. b and from Equation 17.5 that In architectural acoustics, decimal values of decibels are not warranted, and these values would be rounded off to a whole I1 d 22 number. = I2 d12 Therefore, since d2 = 2d1 ACOUSTICS A comparison in Table 17.4 of addition using these I2 (d1)2 1 = = rules and a more accurate method shows that at I1 (2d1)2 4 usual levels, the error resulting from simplification Substituting in Equation 17.7, we have is always less than 1%. I ΔIL = 10 log 2 Returning to the inverse square law expressed I1 in Equation 17.5, it is now possible to determine the = 10 log ¼ effect on sound intensity level of moving away from = 10 (−0.6) a sound source. = −6 dB which tells us that sound intensity level (not pres- sure) is reduced by 6 dB. Similarly, if the distance is EXAMPLE 17.4 Given a sound source that produces quadrupled, it is reduced by 12 dB. an intensity level IL at a distance d1 from a source (substitute any numbers desired or follow the prob- To summarize, the intensity level changes by 3 lem with symbols), what is the intensity level at twice dB with every doubling or halving of power and changes the distance? At four times the distance? by 6 dB with every doubling or halving of the distance Note: A sound intensity distribution that obeys from a point source. Figures 17.12 and 17.13 illus- the inverse square law on which these calculations trate the latter relationship. Fig. 17.12 Decibel level increase as a function of power (intensity) increase.  EXPRESSING SOUND MAGNITUDE 755 Fig. 17.13 Sound energy levels at varying distances from a source. Each doubling of distance reduces the intensity level by 6 dB. These relationships hold true only in a free field. (e) Sound Power Level and the decibel values of the two can be used interchange- ably. The actual intensity and the actual pressure ACOUSTICS Sound power levels may be derived from sound corresponding to a particular decibel level, how- power values using the ratio process described in Sec- ever, are different—completely different—in mag- tion 17.5(d). The reference sound power is 10−12 W. nitude and units. For instance, 70 dB may equal Sound power level is expressed in decibels. 10−9 W/cm2 intensity or 0.063 Pa pressure. The important fact, though, is that 70 dB corresponds (f) Sound Pressure Level approximately to a particular sound magnitude. It is necessary to say “approximately” because assign- Sound pressure levels are derived from sound pres- ing a single-number decibel level to a sound pres- sures using the ratio approach described in Section ents two difficulties: 17.5(d). The usual reference sound pressure cor- responds to the threshold of hearing and is taken 1. Sound pressure level varies with time, except to be 20 μPa or 2 × 10−5 Pa (2 × 10−4 microbars for a pure steady tone. [μbar]). See Table 17.5 for representative values of 2. The different components of most common sound pressure levels and corresponding subjec- (complex) sounds vary in pressure level. tive responses. As with intensity, this sound pres- Two techniques are used to overcome this sure reference is established as 0 dB for the purpose problem. If a sound has a dominant frequency, of calculating sound pressure level. Since the ear that frequency’s level can be used (Fig. 17.8). responds logarithmically to intensity and since This would be the case for a relatively constant pressure varies as the square root of intensity, we sound such as that of a motor, fan, or pump. Other can write the expression sounds that vary widely in constituent level and p2 SPL = 10 log frequency can be plotted on an octave-band chart p 02 using maximum level for minimum percentage or of time (Fig. 17.14). Where the position of the lis- p SPL = 20 log (17.8) tener is not specified in the table, it is assumed to p0 be at normal close distances: that is, 10 to 20 ft (3 where to 6 m) from a train, 3 to 5 ft (0.9 to 1.5 m) from a SPL = sound pressure level, dB radio, and the like. p = pressure, Pa or μbar As suggested previously, the combined effect of two sounds depends upon their frequency content. p0 = reference base pressure, 20 μPa or 2 × In the foregoing examples, we assumed signals 10−4 μbar either of identical frequency and random phase or Since the 0-dB base corresponds to the hearing of a very-wide-frequency spectrum—so wide that threshold for both sound intensity level and sound phase phenomena are not significant. In architec- pressure level, the decibel scales for sound pressure tural acoustics work, such an assumption is gener- level and sound intensity level have been equalized ally valid.  756 CHAPTER 17 FUNDAMENTALS OF ARCHITECTURAL ACOUSTICS Fig. 17.14 Sound pressure level curves of common noise sources plotted according to octave band frequency content. Values shown ACOUSTICS are averages of multiple measurements. (Reprinted from A Guide to Airborne, Impact, and Structure Borne Noise–Control in Multifam- ily Dwellings, 1968.) (g) Measuring Sound with subjective loudness impressions, most such instruments that provide a single-number output The need for a means of measuring sound levels in are furnished with weighting networks, the char- built projects to confirm that design criteria have acteristics of which are given in Fig. 17.16. The been met should be obvious. One very useful instru- A network corresponds to an inverted 40-phon ment is the integrating sound-level meter illus- contour and discriminates against low frequencies trated in Fig. 17.15. To correlate meter readings (see Fig. 17.7), as does the human ear. The B and C TABLE 17.5 Common Sound Pressure Levels Sound Pressure Level (dBA) Typical Sound Subjective Impression 150 (Short exposure can cause hearing loss) 140 Jet plane takeoff 130 Artillery fire, riveting, machine gun (Threshold of pain) 120 Siren at 100 ft (30 m), jet plane (passenger ramp), Deafening thunder, sonic boom 110 Woodworking shop, hard-rock band, accelerating Sound can be felt (threshold of discomfort) motorcycle 100 Subway (steel wheels), loud street noise, power lawnmower, outboard motor 90 Noisy factory, unmuffled truck, train whistle, machine Very loud, conversation difficult; ear protection shop, kitchen blender, pneumatic jackhammer required for sustained occupancy 80 Printing press, subway (rubber wheels), noisy office, (Intolerable for phone use) supermarket, average factory 70 Average street noise, quiet typewriter, freight train at Loud, noisy; voice must be raised to be 100 ft (30 m), average radio, department store understood 60 Noisy home, hotel lobby, average office, restaurant, normal conversation 50 General office, hospital, quiet radio, average home, Usual background; normal conversation easily bank, quiet street understood 40 Private office, quiet home 30 Quiet conversation, broadcast studio Noticeably quiet 20 Empty auditorium, whisper 10 Rustling leaves, soundproof room, human breathing Very quiet 0 Intolerably quiet Threshold of audibility  NOISE 757 ACOUSTICS Fig. 17.15 Modern general-purpose (type 2) integrating, data- logging sound level meter. This unit will measure and record isolated, nonrepetitive noise events in addition to continuous, fluctuating, and impulsive sounds, and will give both instanta- neous and equivalent (Leq) sound levels. The latter are used when an equivalent noise level of varying sound conditions over a selected measurement time is required. The meter will measure dBA, dBC, and linear with fast, slow, peak, or impulse response. When equipped with a filter set, the meter can be used for octave or 1 3 -octave analysis. (Photo courtesy of Quest Technologies, Inc.) Fig. 17.16 Internationally standardized A, B, and C weighting curves.  758 CHAPTER 17 FUNDAMENTALS OF ARCHITECTURAL ACOUSTICS networks correspond to the 70-phon and 100-phon of an overall view of hearing and sound sources. contours, respectively. In addition, a completely There are two basic approaches to the negative linear response is usually available. The reasoning effects of noise; a psychological-practical one and behind the use of these weighting networks is evi- a purely physiological one. The latter is concerned dent when they are compared to the equal loudness with the physical impact of noise on the body, curves of Fig. 17.7. The original intention was to including hearing loss and other deleterious condi- use the A network at levels of up to 55 phons, the B tions. The former is concerned with noise levels that network to 85 phons, and the C network at higher cause annoyance and disturbance to daily activi- levels. In practice, however, it was found that only ties, including work, relaxation, and rest. the A network corresponded fairly well to subjec- tive loudness reports. As a result, the B network (a) Annoyance has fallen into disuse, and the A network is used today for all measurements, regardless of loudness. Research has developed accurate data on percep- The discrepancy at high loudness levels is apparent tions of loudness. The concept of annoyance, how- when the 40 phon–A weighting network curve is ever, being primarily subjective and psychological, ACOUSTICS compared to the equal loudness curves above 80 is much more elusive. Tests have shown that in gen- phon. All measurements should be identified with eral, annoyance as a result of noise is: the weighting network used, such as 50 dBA or 1. Proportional to the loudness of the noise 100 dBC. 2. Greater for high-frequency than low-frequency More accurate measurements of complex noise sounds than are possible with a standard sound 3. Greater for intermittent than continuous noise level meter are made with sophisticated instru- 4. Greater for pure-tone than for broadband ments that measure intensity in octave bands and noise also often plot the results, as per the graphs in Fig. 5. Greater for moving or unlocatable (reverber- 17.14. Such measurements are necessary for accu- ant) noise than for fixed-location noise rate application of sound absorption and attenua- 6. Much greater for information-bearing noise tion materials whose characteristics are nonlinear than for nonsense noise over the frequency spectrum. Single-number dBA readings are known as overall levels and are useful as preliminary data and (b) Noise Criteria for broad-spectrum design. Table 17.5 lists com- To establish criteria for acceptable background noise mon sound levels as measured by the dBA scale. (i.e., noise whose extent of annoyance is consid- Such single-value numbers are useful to establish a ered acceptable), certain of these effects must be mental-aural comparison base and for use in maxi- neglected at this point for the sake of simplicity. mum noise exposure calculations, as discussed in (They can be and are considered in design and the following section. in establishing levels of masking noise [see Chap- ter 19].) Thus, ignore for the time being: 17.6 NOISE Factor 3, assuming, instead, continuous sounds Factor 4, as broadband noise is assumed Noise is variously defined as unwanted sound, Factor 5, as the noise source is assumed to be fixed sound with no intelligible content, and/or broad- in location band sound, depending upon the listener and the Factor 6, assuming that we can only consider gen- situation. Each definition is appropriate for various eral noise level, not actual content times and situations. It should generally be assumed that any sound can at some point be referred to as Thus, the particular and special characteristics of noise by someone. noises such as a barking dog (3), a whistle (4), a sin- Although noise effects and their control are gle passing vehicle (5), and intelligible sounds (6) the specific subject of Chapter 19, noise criteria are not considered when establishing conventional and their development are discussed here as part noise criteria.  NOISE 759 In order to quantify the concept of acceptable ratio of correctly identified syllables to total syllables background noise, it was necessary to remove it read is the Articulation Index. An AI in excess of from the purely psychological arena and relate it to 0.5 was considered indicative (Beranek, 1988a) of a physical phenomenon. This was done by studying a condition in which acceptable intelligibility could the effect of the two remaining annoyance factors be expected for male voices. The Speech Interference (1 and 2) on speech communication. This study Level, a simplified version of the AI, was devised by resulted in two design concepts: the Articulation Beranek. It consists simply of the arithmetic average Index (AI) and the Speech Interference Level (SIL). in decibels of the background sound pressure levels These are both determined by reading a carefully in the four octave bands centered on 500, 1000, selected set of phonetically balanced nonsense syl- 2000, and 4000 Hz, for which acceptable intelligi- lables to a test audience in the presence of different bility could be expected, for a given voice effort, at a levels and compositions of background noise. The given distance between the speaker and listener. ACOUSTICS Fig. 17.17 Application of NC curves. The typical spectrum plotted would be rated NC-36, as it exceeds NC-35 at 500 Hz by 1 dB. (See Table 19.7 for specific NC recommendations for interior spaces.)  760 CHAPTER 17 FUNDAMENTALS OF ARCHITECTURAL ACOUSTICS (c) Noise Criteria Curves (d) Room Criteria Curves Beranek also developed the well-known and widely Due to the recognized shortcomings of the criteria accepted and used noise criteria (NC) curves shown inherent in the NC curves, specifically that they are in Fig. 17.17. The NC curves are based on the con- undefined in the very low frequencies (16 and 31.5 sideration that most people prefer to speak at a level octave bands) and are not sufficiently stringent at no greater than 22 dB above the background noise frequencies above 2 kHz, a similar approach to set- level. The NC curves are derived by combining the ting criteria was developed—called room criteria SILs in decibels with this behavioral phenomenon; (RC) curves. These curves, shown in Fig. 17.18, they represent a loudness level 22 dB higher than were adopted by ASHRAE (the American Society of the SIL. The contours then represent the maximum Heating, Refrigerating and Air-Conditioning Engi- continuous background noise likely to be consid- neers) as the suggested noise limitation benchmark ered acceptable in a specified environment, and in preference to NC curves. The curves were used to generally correspond to background sound pres- evaluate the acceptability of background mechani- sure (noise) levels found reasonably acceptable in cal system noise for typical space types. RC curves selected commercial/institutional building envi- differ from NC curves in a number of aspects: ACOUSTICS ronments. A similar set of curves called noise rating (NR) curves find considerable application outside They are straight lines. the United States. They are less stringent than NC Their slope is constant at −5 dB per octave (deter- curves in low frequencies but more stringent in mined from extensive tests, mostly in the range high frequencies. of 40 to 50 dB). To apply the NC curves, the spectrum (from Regions labeled A and B as in Fig. 17.18 address 63 to 8000 Hz) of a specific noise being studied is the problem of very low frequencies and high overlaid on a graph of the NC curves. The lowest NC sound pressure levels, an issue that is ignored curve that is not exceeded by any portion of the plot in the NC criteria. This addition then deals with becomes the NC rating of the particular noise. Thus, rumble and vibration that can cause extreme specifying a maximum noise level of NC-30 for a annoyance for many occupants. space means that no portion of the sound pressure Referring to Fig. 17.18, the procedure for deter- level curve of any continuous background noise in mining the RC value of a specific item of equipment the space may cross the NC-30 contour. A piece of whose noise spectrum is known is as follows: equipment rated NC-35 has an octave-band spec- trum completely below NC-35. A fan rated NC-53 1. Calculate the arithmetic average of the sound indicates that at some point in its frequency spec- pressure levels (SPL) in the 500-, 1000-, and trum the fan exceeded NC-50 by 3 dB. 2000-Hz octave bands. This number is the RC The NC rating of a noise usually falls between value of that noise spectrum. 5 and 10 dB below the measured dBA for the noise, 2. Draw a straight line at a −5 dB slope through depending upon the shape of the noise spectrum. this value of RC at 1000 Hz. The virtue of the NC curves is that they provide a 3. Plot the SPL values for the octave band center single-number specification for sound across a wide frequencies on the RC curve sheet and compare frequency range—without losing all sense of the the plot to the RC line drawn in Step 2. frequency distribution. Their disadvantage is that 4. Classify the equipment sound quality from this they were derived for, and are most accurate with, comparison as follows: speech conversation conducted against a backdrop a. Neutral. If the octave band data plotted in of continuous, broadband noise. This is not the situ- Step 3 do not exceed the RC line drawn in ation that is typically most troublesome in an office; Step 2 by more than 5 dB at or below 500 indeed, continuous equipment noise may be help- Hz and more than 3 dB at or above 1000 Hz, ful in masking unwanted speech. Nevertheless, NC then the sound is considered neutral (bland, curves remain the most commonly used criteria for uncharacteristic), and the designator letter N establishing acceptable continuous background is placed after the RC level. A piece of mechan- noise levels. ical equipment with an N designation is then  NOISE 761 ACOUSTICS Fig. 17.18 RC curves. These curves were adopted by ASHRAE in lieu of NC curve criteria. See text for an explanation of their use. (Reprinted with permission from ASHRAE Handbook—Applications, 1995.) understood to have a neutral tone quality bands. In the A area, vibration will likely be that most people would classify as unobtru- felt in light construction and furniture, and sive and lacking specific character. ASHRAE rattling may occur in loosely constructed design guidelines for HVAC system noise lev- devices, cabinets, glassware, and the like. In els are listed as RC (N). the lower-energy B area, rattling will be less b. Rumble. If the octave band plot does exceed likely but vibration may still be felt. the RC line by more than 5 dB at 500 Hz or below, the spectrum is classified as “rumbly,” The result of this classification is to give the designer and the descriptor letter R is appended to its not only an average SPL for an item of equipment, RC level number. but also a sense of subjective sound quality that c. Hiss. If the octave band plot does exceed the should be of considerable assistance in determining RC line by more than 3 dB at or above 1000 the noise abatement measures to be taken. Hz it is classified as “hissy,” and the descrip- ASHRAE has updated the RC curve concept tor letter H is appended to its RC level. to what is now called the RC Mark II method. This d. Vibration. The shaded areas in Fig. 17.18 updated method is a bit more complex than the labeled A and B represent high sound pres- original RC approach. For further information, see sure levels in the 31.5- and 63-Hz octave the ASHRAE Handbook—HVAC Applications (2007).  762 CHAPTER 17 FUNDAMENTALS OF ARCHITECTURAL ACOUSTICS The original NC curves have also been modified ment noise conforming to the NC criteria. These (by Beranek) and issued as balanced noise criteria adjustments have yielded a “balanced” neutral (NCB) curves. They amended a failing of the NC sound, which reflects the chart’s name. Applica- approach by adding very-low-frequency coverage tion of these curves to mechanical equipment noise and the NCB curves were straightened, so that they is similar to that for RC curves, except that letter resemble RC curves above 125 Hz, except that the descriptors are not used. For a full discussion of the slope angle is −3.33 dB compared to −5 dB for the construction and application of BNC curves, see RC curves. The slope at the higher frequencies has Beranek (1988b). Given these competing options also eliminated the hiss that characterizes equip- for criteria, selection of NC, RC, RC Mark II, or NCB ACOUSTICS Fig. 17.19 The standard for exposure to noise in the workplace. (From OSHA, July 1988; current as of late 2009.)  NOISE 763 criteria for background noise is a function of design Since continuous noise exposure is most severe in intent. The approach that best meets the needs (and industry, regulatory legislation in the United States budget) of the owner should be chosen. has been directed at this area. In 1969 the Walsh–Healy Public Contracts Act was passed, and thereafter its provision for (e) High Noise Levels and Hearing maximum permissible exposure to noise levels Protection was incorporated into the Occupational Safety and It has long been recognized that continuous expo- Health Act. Both the act and the associated regu- sure to high noise levels causes a degree of tempo- latory agency, the Occupational Safety and Health rary deafness in most people and that long periods Administration, are known as OSHA. The relevant of such exposure, even on an intermittent 8-hour provisions of this act are reproduced in Fig. 17.19. workday basis, can produce permanent hearing To avoid overly complex regulations, limitations on impairment. Most experts place the safe 8-hour exposure are given as single-number dBA values. upper limit at 85 dBA. In addition, studies have Since workers rarely remain in a single acoustic indicated that continual exposure to noise levels environment for 8 hours, their total daily exposure, ACOUSTICS as low as 75 to 85 dBA can produce or contribute or time-weighted average (TWA) exposure, can be to numerous physical and psychological ailments, calculated from timed dBA measurements using including headache, digestive problems, tachy- formulas and tables given by OSHA and then com- cardia, high blood pressure, anxiety, and nervous- pared to permissible levels. Alternatively, a dosim- ness—an extensive catalog of human illnesses. eter (Fig. 17.20) can be used, which automatically integrates the noise to which a person is exposed over a given time period and reads out the permis- sible TWA exposure directly.