Document zd1Lr4XX5MEMrNQLJbZMj86Xz

346 75 120 -Hao < CD O cc V CM o o o d CHAPTER 25 75 150 500 150 500 600 r-2ao . -250 - -200 200 600 1200 _ -250 ^00 -ISO riso 1200 2400 -- 70 -- 60 ^-ioo -70 -60 -50 -40 -30 2400 4800 4800 9600 1960 Guide Ln` N PHONS- SONES --250 --200 ' -70 ~ -60 - - so "-40 ' -30 __ --70 _ - 60 --to -- 15 .-4 .~ 3 "4 --3 __ S -4 z < CD - ,s -- .5 u - .4 o -- .4 - .3 30 - 75 ISO 300 600 1200 2400 4800 PHONS-SONES 150 300 600 1200 2400 4800 9600 Ln-N OCTAVE PASS BAND IN CYCLES PER SECOND odnn in cooet con bn converted to lottdneu level In phont by rseom of (bn nomogram of the right. (from Reference 3. Uted by pernwcsfonj Rg. 2------Nomograms Relating 1, in Octave Bands to Loudness in Sones, N of the apparatus room or from a duct plenum. Where such a room or plenum is near occupied rooms, the noise transmitted through the walls must be determined. The combined sound pressure levels resulting at a listener's position should be com pared with the criteria as done in Step 4 for noise produced by grilles. CRITERIA FOR NOISE CONTROL Criteria for noise control in building spaces have evolved in the United States from several large scale studies and from the experiences of engineers who have recorded favorable and unfavorable reactions to known background noises in many types of rooms.* The acoustical engineer should carefully question the user of the building to make certain of the hind of uses that he has in mind. For example, a coliseum used for sports only, need not be as quiet as a coliseum also used for concerts. As another example, a hall for band concerts need not be as quiet as one for chamber music. In fact, some rooms are better if they are not too quiet. It has been found that libraries, drafting rooms, and study rooms that are too quiet are not so desirable as Sound Control 349 they would be with moderate noise levels. In very quiet rooms of these types, sounds such as whispering, turning of pages, creaking of the floor, coughing, closing of doors, etc., act to distract the occupants and to main* them complain of noise. Introduction of a continuous spectrum noise properly dis* tributed as a function of frequency produces a quieter condi tion even though the background noise is actually higher. Rumbles, rattles, audible single frequency components, beats between components, and intermittency in a noise often mako it objectionable even though the noise as measured seems to meet the criterion. Noise Specification There are at least four possible ways of specifying the noise. The first way is simply to give the octave band spectrum of the noise. This method provides the information that is essential to tiie design engineer and no substitute can suffice. The other three of these methods of specifying noise are single-number - numerical quantities that indicate something about the in fluence o! the noise on the average human being. These three methods are useful as summaries of the octave band spectrum, but, taken separately, they are of limited value to the design engineer. The definitions of each of four commonly used ways of specifying noise are as follows; 1. Octave Band Spectrum. The octave band spectrum is com prised of the readings of sound pressure level in the eight octave frequency bands (see the section on Acoustical Terminology). The readings are taken with the sound level meter weighting networks set on fiat (C-scale) and the units are decibels (db). The specification of a complete octave band noise spectrum is necessary for satisfactory acoustic design in all cases. 2. Speech Interference Level L9r. The speech interference level is a single number found by taking the arithmetic average of the sound pressure levels in the three octave frequency bands 600-1200, 1200-2400, and 2400-4800 cpa. The unit is the decibel (db). The speech interference level indicates the ease of speech communication in the presence o! the noise so rated. _ 3. Loudness Level Lh . The loudness level is a single number computed from octave band levels according to the procedure following. The unit is the pbon. To obtain the loudness level in pbons, obtain the sound pres sure levels in the eight octave bands and read from Fig. 2 the loudness in sones for the respective bands. The sum of the loud nesses will be; 2JV = AT, + N, + N, + N< + N, + N. + N, + N, The loudness in sones of the total noise is given by (10) N = Nm + 0.3 (Stf< - Nm) where Nm = loudness of the loudest band, sones- (11) = sum of the loudnesses of all eight bands, sones. To convert from loudness in sones to loudness level in phons, use Equation 9 or the nomogram at the right of Fig. 2. The loudness level in phons indicates the subjective loudness but does not indicate tne ease of speech communication. 4. Sound Level. The sound level is a single reading taken on the sound level meter with the meter operated as follows :* 4 Sound-LavaJ Rang* Below 55 db. 55-85 db_____ .Above 85 db. A B C (Flat) Sound level measured with the A-ecale can be correlatedmost closely with speech interference level but the two are not nu merically the same. The C-scale is always used with the octave band analyser. If the weighting scale is not stated with a meas urement, it may be assumed that the C scale was used. Criteria Noise criteria have been developed from detailed studies and engineering experience in a number of architectural spaces.* The acceptability of background noise for various types of activities normally performed in rooms and enclo sures can be related to two principal characteristics of the noise: (1) Speech interference level, Lai, and (2) Loudness level, Ln . Definitions of these two quantities are given in the preceding section on Noise Specification. The speech inter ference level is governed by the amount of high-frequency noise (between 600 and 4S00 cps) and the loudness level is governed by both the high- nnA low-frequency noise. For most types of spaces the loudness level in phons is 22 to 25 units higher than the speech interference level in deci bels. However, for factories, sports coliseums, restaurants, and similar spaces where cost of acoustic treatment dictates a maximum amount of compromise, the loudness level may rise as much as 30 unite above the speech interference level. Acceptable background noise criteria have been established for these two conditions and are presented in Figs. 3 and 4. Each curve is designated by a number indicating the speech interference level obtained. The numbers in the first column of Table 1 and Table 2 identify tire curve which defines recom mended noise levels for the various types of spaces listed. With regard to the choice between Figs. 3 and 4, Fig. 3 should be used whenever possible, particularly where there is a possibility that the sound may be fluctuating, such as often occurs with turbulence in a ventilating system. Fig. 4, or numbers in the range between Figs. 3 and 4, may be used when economic factors dictate a compromise and when there is a definite assurance that the noise will not be of a fluctuat ing type. In engineering design, all eight octave band levels selected from the curves of Fig. 3 or 4 should be used. In construction specifications, reference may be made to The Guide and the selected NC or NCA curve may be referred to by number, e.g., ATC-35. In case pure tones ore present, better judgment of the sub jective effect of tiie sound is obtained by using a octave band analysis.* Figs. 3 and 4 remain the same except that the numerical value of the ordinate is reduced 5 db. If the sound level in any band is more than 3 db above adjacent bands, add 5 db to this level before utilising the NC curves. Where octave band analyses of the noise cannot be made and a rough approximation of accepted criteria must be ob tained by use of a sound level meter only, the A-ecale values which would be read for the noise spectra specified in Column 1 are given in Column 2 for comparison purposes. It should be understood, however, that an A-scale reading gives a measure of the overall noise using a weighted network; it does not specify the spectrum. For example, the noise spectrum shown as NC-25 in Fig. 3 would give an overall A-scale meas urement of 35 db, but a measured value of 35 db does not mean that the actual noise measured has the same spectral shape or is as satisfactory as the specified NC-25 spectrum. The lack of octave band analysis can be partly offset by a subjective observation; if the noise sounds unnatural because of excessive amounts of high pitched or low pitched noise or has particularly noticeable amounts of pure tones, the spec trum will probably depart widely from the specified one. In