Document
37. JUDGMENT TESTS OF AIRCRAFT NOISE By Karl D. Kryter, Paul J. Johnson, and James R. Young Stanford Research Institute SUMMARY Psychological tests were conducted to determine the ability of various physical units to predict the judged perceived noisiness of the sounds from a variety of aircraft operating under landing and take - off power. The best agreement between objective measures and subjective judgments of the noisiness or unacceptability of aircraft noise of all types is generally found by calculating the tone - corrected effective perceived noise level in- EPNdBt from 1/3-octave band spectra taken every 1/2 second during the noise cycle.
When low - frequency aircraft noise is compared with high - frequency aircraft noise, a systematic overestimation of the perceived noisiness of the lower frequency noise is obtained by the various PNdB units.
INTRODUCTION Over the past 20 to 30 years, research has led to the development of a number of ways of measuring noise that purport to be correlated with the measurement or response of man's auditory system to noise. These methods run the gamut from simple peak sound - level meter readings to weightings of 1/3-octave band spectra taken every 1/2 second during the duration of a noise occurrence. Although the measurement procedures vary, the purpose and goal of making these measurements is to estimate or predict how unwanted, unacceptable, or noisy the sounds being measured are perceived to be by people.
Most of the psychological tests concerned with establishing the relationships between perceived noisiness and the physical aspects of aircraft noise have been conducted in the laboratory wi€h recordings of the noise; relatively small groups of subjects and often a restricted variety of aircraft noises were used. The present tests were designed to per - mit as valid an examination as possible of the ability to predict from physical measure - ments the judged acceptability to people of a wide variety of actual aircraft noise when heard in or outside typical homes. In addition, a number of ancillary questions related to the generation and propagation of the noise and the reaction of house structures to the noise were to be investigated. These tests were recently conducted at the NASA Wallops Station.
The present report is concerned primarily with the relation between the psychologi - cal judgment of people outdoors and physical measurements made from recordings of the noise from microphones located outdoors. Measurements of house vibrations at Wallops Station are reported in reference 1. Additional technical reports by the National Aero - nautics and Space Administration and Stanford Research Institute on various aspects of the studies will be prepared later as further analyses of the data a r e completed.
PROCEDURE A frame house, a brick - veneer house, and a large yard near one of the houses, all located in a residential area of Wallops Station, were chosen as the test sites. The sub - jects were adults, primarily housewives, selected from communities in the local area.
Figure 1 is a photograph of some of the subjects as seated for the tests. The subjects w e r e tested with an audiometer and all were found to have normal hearing (*15 dB from audiometric zero). The subjects were paid and given careful instructions prior to and during the tests as to the importance of the tests and the nature of the task they were to perform .
The fundamental task of the subjects was to mark on an answer sheet which of two aircraft sounds presented to them in a brief period of time they considered to be the least acceptable, assuming these noises were heard in o r near their home 20 to 30 times per day. The subjects also rated each noise on a scale ranging from completely acceptable to completely unacceptable.
Because of the very large number of aircraft tested, it was not possible to pair, for the judgments, each aircraft noise with every other aircraft noise. Instead, two of the aircraft were chosen to provide a reference, or standard, noise, and this noise was paired with the noise from each of the other aircraft when operating under landing and take - off power. The reference, or standard, aircraft chosen were the turbojet 880 and the 1049G (Super Constellation), a propeller - driven aircraft with reciprocating engines (recip - prop).
The aircraft were flown so that about 5 minutes elapsed between pairs of noises, and usually 1 minute elapsed between each member of a pair. The altitude for the oper - ational conditions for each aircraft was carefully monitored and controlled for all flights.
An attempt was made to operate the reference aircraft at an altitude that would be reason - able for that aircraft at about 2 to 3 miles from an airport. Figure 2 is a schematic illustration of the flight paths followed by aircraft for the tests. Table I summarizes the aircraft tested and number of overflights made per aircraft.
Physical Measurements Tape recordings made of the aircraft noises reaching the ground were played through 1/3-octave band filters. Each filter output was passed through an envelope detector, smoothed, and sampled every 1/2 second. These samples were digitized and stored on magnetic tape and were used for subsequent calculation of the following units: (1) Max dB(A), dB(B), dB(C), dB(N) (2) Max phons (Stevens), PNdB, PNdBtl, and PNdBt2 (3) Peak phons (Stevens), PNdB, PNdBtl, and PNdBt2 (4) E (effective) dB(A), dB(N), PNdB, PNdQ1, and PNdBtZ (5) EE (estimated effective) PNdB, PNdBtl, and PNdBt References 2 and 3 give a detailed description of these units. These units of noise mea - surement appear at the present time to be the best available units for the evaluation of the perceived noisiness of the sound from aircraft. An additional method is that of cal - culating loudness levels, in phons, by the 1/3-octave band method of Zwicker (ref. 4 ) . It is planned to obtain max phons (Zwicker) for these aircraft noises for their inclusion in the final report of the tests conducted at Wallops Station.
One - third - octave band spectra present outdoors when the overall sound pressure level (SPL) was at max dB(C) a r e shown in figure 3 for representative flights for each of the aircraft tested.
Psycho logic a1 Measur e ment s The paired - comparison tests which are believed to provide the most essential psychological data from this study are scored and interpreted as follows: Step 1. - The percent of listeners in a group (for this report a group of 33 people) who preferred the reference aircraft noise when it appeared first in a given pair and when it appeared second in the same pair are averaged.
Step 2.- The percent obtained in step 1 is plotted against the level, measured by a given physical unit, of the comparison aircraft noise. Inasmuch as the level of the com - parison noise was systematically varied, the percent of people, in general, who preferred the reference noise increased as the level of the comparison noise increased. An attempt was made to have the comparison noise vary over a range that caused the percent of peo - ple preferring the reference aircraft noise to change from near zero percent to near 100 percent. Sample plots of the data are shown in figure 4.
Step 3. - On each function, such as those shown in figure 4, a perpendicular is dropped to the abscissa from the point where the 50 - percent line crosses the curve drawn through the data points.
Step 4.- The value obtained in step 3 is taken as the level, for the given unit of mea - surement, required for the comparison noise if and when it is to be perceived as equal to the reference aircraft noise in unacceptability o r noisiness.
Step 5. - The difference, if any, between the reference and comparison noises when judged to be equal is taken as the index of the ability of each of the physical units to prop - erly measure or indicate the perceived noisiness of each pair of sounds.
RESULTS m D DISCUSSION Table 1 1 presents a summary of averaged data obtained when the noise from jet air - craft was judged against that from other jet aircraft or from turboprop aircraft, recip - prop aircraft, or helicopters. Two equally important indicators that show the accuracy with which an objective measure predicts the subjective judgment data are shown in table 1 1 and are described as follows: (1) The average of the differences between an objective measure for the reference and comparison aircraft noise when they a r e judged to be equally noisy. l f the physical units were perfectly correlated with the psychological data, the average of the differences would be zero for each pair; that is, when the aircraft noises w e r e judged to be equal, they would be measured physically as being equal.
(2) The range of the differences between an objective measure for the reference and comparison aircraft noise when they are judged to be equally noisy. It is important that an objective measure be reasonably accurate with respect to the most common or most prevalent aircraft noises evaluated. This accuracy is reflected in the average of the dif - ferences between the objective measures of any two noises judged to be subjectively equal.
Two particular conclusions can be drawn from table I I . First, the average differ - ences for the better units are significantly smaller when the comparison and reference noises are both high - frequency jet noises than when one of the noises is a low - frequency propeller - aircraft noise. These differences are about 1.5 dB when the comparison and reference noises are both high - frequency jet noises and about 3.0 dB when one is a low - frequency propeller - aircraft noise. Appropriate modifications to procedures for mea - suring noise in the lower frequency bands should reduce the differences between the phys - ical measures and judged noisiness and thereby provide an effective perceived noise level that would predict, with an average accuracy of about 1.0 dB, judgments of the noisiness or unacceptability of aircraft noise regardless of its source, spectral complexity, and, Second, when all types of aircraft noises and both the average within limits, duration.
differences and the total range of differences between the values for the references and comparison noises are considered, the most accurate units of measurement are usually EPNdBt.
The relative accuracy of the various objective measures is illustrated in table 111, which shows how each unit ranks with regard to the average of the average differences and the range of differences. A unit with the rank of one would have the best agreement with the subjective judgments.
Two other recent studies (see refs. 5 and 6) have been made of subjective judgments of aircraft noise in which all, o r nearly all, of the objective measures used in the present study were evaluated. Table IV shows that EPNdB has the highest rank in each of the
studies. It should be pointed out, however , that the practical significance between some
of the differences in these measures, an amount equivalent to less than 1.0 dB in level in some cases, is perhaps questionable.
CONCLUSIONS The following conclusions a r e based on the results from judgment tests of aircraft noise : 1. The best agreement between objective measures and subjective judgments of the noisiness or unacceptability of aircraft noise of all types is generally found by calculating the tone - corrected effective perceived noise level in EPNdl3t from 1/3-octave band spec - tra taken every 1/2 second during the noise cycle.
2. When aircraft noise containing its energy predominately in the lower frequencies is compared with predominately high - f requency aircraft noise, a systematic overestima - tion by about 3.0 dB of the perceived noisiness of the lower frequency noise is obtained by the various PNdB units.
REFERENCES 1. Mayes, William H.; Findley, Donald S.; and Carden, Huey D.: House Vibrations Signifi - cant for Indoor Subjective Response. Conference on Progress of NASA Research Relating to Noise Alleviation of Large Subsonic Jet Aircraft, NASA SP - 189, 1968.
(Paper No. 39 herein.)
Concepts of Perceived Noisiness, Their Implementation and Application.
2. Kryter, K. D.: J. Acoust. SOC. Amer., vol. 43, no. 2, Feb. 1968, pp. 344 - 361.
3. Stevens, S. S.: Procedure for Calculating Loudness: Mark VI. J. Acoust. SOC. Amer., vol. 33, no. 11, Nov. 1961, pp. 1577 - 1585.
4. Zwicker, E.: A Graphic Method of Determining Loudness and Loudness Level From the t*Third-Octave't Level Diagram. Frequenz, vol. 13, no. 8, Aug. 1959, pp. 234 - 238.
5. Hecker, Michael H. L.; and Kryter, K a r l D.: Comparisons Between Subjective Ratings of Aircraft Noise and Various Objective Measures. Tech. Rep. NO - 68 - 33, FAA, Apr. 1968.
6. Kryter, K. D.; Johnson, P. J.; and Young, J. R.: Psychological Experiments on Sonic Booms. Annex B of Sonic Boom Experiments at Edwards Air Force Base, NSBEO - 1 - 67 (Contract A F 49(638)-1758), CFST1, U.S. Dep. Com., July 28, 1967.
TABLE 1.- NUMBER OF OVERFLIGHTS FOR JUDGMENT TESTS p o t a l of 189 pair4 Reference aircraft: Turbojet. 880 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 132
Recip.prop. 1049G . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 80
Comparison aircraft: Turbofan (front). 727 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 22 Turbofan (front). C - 141A . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 22 Turbofan (aft). 990 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 28 Turbojet. 720 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 20 Turbojet. 1329 Jet Star . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 18 Turbojet (afterburner). F - 106 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 12
Turboprop (STOL). CV-7A . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 16
Turboshaft (helicopter). 204B . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 20
Turboshaft (helicopter). CH - 47 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8 TABLE I I . - COMPARISON OF OBJECTIVE MEASURES AND SUBJECTIVE JUDGMENTS Best units Indicator Aircraft of noise Max dB(B) Max dB(C) accuracy 2.0 2.1 Average difference 2.9 Jet v s jet 5.5 7.5 Range of differences 11.0 Jet vs turboprop, 7 . 0 - 7.9 Average difference 5.6 reeip - prop, o r 20.0 25.0 Range of differences 17.0 helicopter Best units Worst units Indicator Aircraft I I I of noise
EPNdB 1 EPNdBtl /EEPNIB
accuracy Peak MaxdB(B) MaxdB(C) EdB(N) EPNdBt2
PNdBtll 1 1
Average difference Grand average f o r all aircraft Range of differences 10.3 11.0 20.0 25.0 TABLE ID.- RELATIVE ACCURACY WITH WHICH OBJECTIVE MEASURES PREDICT JUDGMENT DATA OBTAINED AT WALLOPS STATION
BO units u s e d
Unit name Rank Unit name Rank EPNdB 1 Max phons 8 EdB(N) 1 Max dB(A) 8 EPNdBtl 2 Max PNdBt2 9 EPNdQ2 2 Max PNdB 9 EEPNdB 3 Max dB(N) 9 EdB(A) 4 Peak PNdBt2 10 EEPNdBtl 5 Max PNdBtl 11 EEPNdBt2 6 Peak PNdBtl 12 Peak phons 7 Max dB(B) Peak PNdB 8 Max dB(C) 14 I TABLE 1V.- BEST AND WORST PREDICTORS OF JUDGMENTS OF AIRCRAFT NOISE Edwards AFB FAA NASA Predictors field tests laboratory tests field tests (ref. 5) (ref. 6) EPNdBtl EPNdBtZ EPNdB, EdB(N) Best EEPNdBtl EPNdB EPNdBtl, EPNdBtZ EEPNdB EEPNdB M a x dB(N) Max dB(A) Peak PNdBtl Peak PNdBtl Worst Max dB(C) Max PNdBt1 M a x dB(B) Max dB(C) Max dB(B) M a x dB(C) Figure 1 L - 68 - 8585 AIRCRAFT F L I G H T PATH
-
-
/ - - - - - .
/-- ) COMPARISON / I-
/ ---- AIRCRAFT
/ /
- -
-/----- /
- - / ---
.I
-IL_ /----<- REFERENCE I
1 ’ AIRCRAFT
// O z r s G ) / ’
< - - - - - ’ . . OUT DO OR SUB J ECTS Figure 2 I/SOCTAVE BAND SPECTRA PRESENT OUTDOOR§ WHEN OVERALL SPL IC WEIGHTING ) REACHED ITS MAXIMUM I / 3 - OCTAVE BAND CENTER FREQUENCY, Hz l l l l l l l l l l l l l l l l l l l l l l l f - 10 - 2 0 - 30 m, -40
- h
a - 50
z
(3 s o U -I a - 10 v) -I zi - 20 u W > 0 - 30
e
TURBO FAN (FRONT!, W - 40 - TURBOFAN (AFT), L C-141A tJT3D-8) 990 (CJ805-23) G I I I I I I I I I I I I I I I I I I I I I I I I -1 - 50 W a -I w
e
- I o I I I I I I I I I I I I I I I I I I I I I I I I I I - 2 0 - 30 - 7 2 7 (JT8D-I) - 40 - 50 Figure 59 8 I/3-OCTAVE BAND SPECTRA PRESENT OUTDOORS WHEN OVERALL SPL (c WEIGHTING ) REACHED TO MAXIMUM II3-OCTAVE BAND CENTER FREQUENCY) H Z 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 I I I I I HELICOPTER,204B - 10 ( T 5 3 - L - 13) CRUISE, 1100' - 20 - 30 TURBO JET (AFTERBURNER), m F-106 (J75-P-17) - -o - 40 n I I I I I I I I I I I I I I I I I I I I a - 50 z
g o
V w -I a - 10 v) -I
d - 20
IL W > 0 - 30 I - W - 40
f
a 1 - 50 W IL -I W HELICOPTER,CH- 47 - Figure 3(bl EXAMPLES OF PAIRED - COMPARISON JUDGMENTS OF SUBSONIC NOISE BY OUTDOOR LISTENERS AT WALLOPS STATION
-
100 r n REFERENCE REFERENCE - 1049G AT 97 EPNdB AT 92 MAX dB(C) 60 - 50 50 40 40 - 1.0 EPNdB, - 1.0 EPNdB, - 12.0 MAX dB(C)
t
1 1 20 2o
t
0 I I I I I 82 87 92 97 102 107 112 77 82 87 92 97 102 107 EPNdB, OF COMPARISON (C-141A) MAX dB(C) OF COMPARISON (C-141A) Figure 4