Published in: Proceedings of the 3rd International Meeting on Low Frequency Noise and Vibration, London, September 12-13, 1985

Transcription

1 Aalborg Universitet Annoyane from audible infrasound Møller, Henrik Published in: Proeedings of the 3rd International Meeting on Low Frequeny Noise and Vibration, London, September 12-13, 1985 Publiation date: 1985 Link to publiation from Aalborg University Citation for published version (APA): Møller, H. (1985). Annoyane from audible infrasound. In Proeedings of the 3rd International Meeting on Low Frequeny Noise and Vibration, London, September 12-13, 1985 (pp ) General rights Copyright and moral rights for the publiations made aessible in the publi portal are retained by the authors and/or other opyright owners and it is a ondition of aessing publiations that users reognise and abide by the legal requirements assoiated with these rights.? Users may download and print one opy of any publiation from the publi portal for the purpose of private study or researh.? You may not further distribute the material or use it for any profit-making ativity or ommerial gain? You may freely distribute the URL identifying the publiation in the publi portal? Take down poliy If you believe that this doument breahes opyright please ontat us at vbn@aub.aau.dk providing details, and we will remove aess to the work immediately and investigate your laim. Downloaded from vbn.aau.dk on: April 29, 217

2 Proeedings of the 3rd International Meeting on Low Frequeny Noise and Vibration, London, September 12-13, 1985, pp ANNOYANC FROM AUDIBL INFRASOUND Henrik M 11er Institute -of letroni Systems Aalborg University Strandvejen 19, DK-9 Aalborg, Denmark INTRODUCTION For nearly 2 years researhers and environmentalauthoritiesha"\e been worried about possible extra-auditory effets of infrasound, suh as disturbane of equilibrium and influene on the irulatory system. xperimental findings are not very onordant, but in general the effets seem to have been exaggerated (1). However, lak of diret physiologial effets from infrasound does not mean that infrasound is insignifiant from an environmental point of view. Infrasound an be deteted by the human ear, and when it beomes suffiiently loud, it an be annoying. Some investigations indiate that a possible "threshold of annoyane" would be only slightly above the hearing threshold. The hearing threshold at infrasoni frequenies has been determined in several experiments and also the growth of loudness above threshold has been investigated (2, 3). Whether knowledge about the loudness of infrasound an be used in assessment of annoyane from infrasound is however unertain and the present study was arried out. Four experiments were inluded in the study. xperiment I overed annoyane from pure tones. Curves of equal annoyane were determined in the frequeny range Hz. Referene was made to an otave noise band at 1 Hz. In xperiment II the signifiane of the exposure time was investigated. xperiment III was a study of the annoyane from non-sinusoidal infrasoni noise, while xperiment IV overed annoyane from ombinations of audio and infrasoni noise. xperiment I is desribed in detail in an artile in the Journal of Low Frequeny Noise and Vibration (4). xperiment II-IV have previously been presented at Internoise 84 (5). This presentation is an extended version of the Internoise paper. It will over all four experiments, though the most detailed desription will be given for xperiment I, while for the other experiments only hanges in method will be mentioned. XPRIMNT I Subjets. 18 university students aged between 2 and 25 partiipated as subjets. An audiometri test ensured normal hearing. Sound onditions. Pure tones at the following frequenies and levels were used: 4 Hz: 12 and 124 db; 8 Hz: 19, 114, 119 and 124 db; 16 Hz: 95, 12, 19 and 116 db; 31.5 Hz: 75, 84, 93 and 12 db. A 1 Hz otave filtered pink noise presented at four levels (2, 4, 6 and 8 db) served as referene. This made a

3 2 ANNOYANC FROM AUDIBL INFRASOUND total of 18 different sound onditions. Apparatus. The experiments were performed in a 16 ubi metre pressure hamber (6). The infrasound was emitted via 16. eletrodynami loudspeakers driven by a B & K 2712 power amplifier. The 1 Hz noise was emitted via an equalized Hi-Fi sound reprodution system with the loudspeaker plaed 14 m from the subjet. A omputer ontrolled the experimental session. xperimental design. ah subjet was exposed to the whole range of stimuli. The o rder in whih a subjet reeived the 18 stimuli was determined from a latin square design that balaned out both order and arry-over effets. ah subjet was exposed to only one stimulus a day 18 days and at the same hour every day. Proedure. A session lasted 2 minutes during whih the subjets were reading newspapers. After an initial 5 minutes period o f silene the sound was presented for 15 minutes. Following this the subjet was asked to indiate on a graphi sale the degree of annoyane that he would probably feel at home if his neighbour p r odued the same s ound for two hours. The sale was a 15 mm horizontal line of whih the left e nd was marked "not at all annoying" and the right end "ver y annoying", see Figure 1. not at all annoying very annoying F i gur e 1. The gra~:~ ~ sale used by the s ubjets to ind iate degre e of annoyane. Results. Degree o f annoyane was measured i n mm from the "no t at all annoying " end, and the means for eah s o und ondition are shown in Figure 2. The relationship between sound p ressure level and annoyane rating is linear for t he infrasoni frequenies, and regressio n lines are inluded in the figure. In Figure 2 points of equal annoyane are represented by horizontal lines. From eah of the f our 1 Hz points horizontal lines have been drawn, and t he points where they interset the regression lines have been determined. These points an be shown graphially a s the equal annoyane ontours in Figure 3. Dis ussion. The equal annoyane urves demonstrate that t he l ower the frequeny the greater the sound pressure must be to ause a given amount of annoyane. Compared wi th 1 Hz the ur ves lie muh loser in the infrasoni range. This hange is already seen at 31.5 Hz, but it beomes even more pronouned with d ereasing frequeny. The same general pattern is seen f o r the equal loudness urves (2), and the pres ent results support the theror y that the annoya~e of infrasound is losely related to the loudness sensation. The loseness of the urves in the infrasoni region i mpl i es

4 J ANNOYANC FROM AUDIBL INFRASOUND that relatively small hanges in sound pressure may ause large hanges in annoy ane. From an environmental point of view this is important sine a modest redution in sound pressure will in some ases be enough to alleviate annoyane aused by infrasoni noise oi 1 -, I.. (1) u, >. 5 <{ Sound pressure level (db) Figure 2. Annoyane ratings fo r pure infrasoni tones obtained i n xperiment I. F i lled irles r epresent means of 18 subjets, full l i nes a re r egr ession l ines. 14 XPRIMNT II This experi ment was arried out to show the effet o f exposure time on t he annoyane rat ings. Subjets, sound onditions, apparatus and experimental design were the same as in experiment I. The proedure was hanged only with respet to exposure times. The entire e xperiment was repeated for ea h of the following exposure times: a) 1 5 minutes preeded by 5 minut es of silene b) 3 minutes pr eeded by 1 minute of s ilene ) 3 seonds preeded by 1 seonds o f silene. Results in a ) were o btained from xperiment I. In b) the sound ondition were given on one day and irrunediately following eah other. The same proedure was used in), exept that no newspapers were availabl e beause of the short exposure time.

5 ANNOYANC FROM AUDIBL INFRASOUND 14 -al 'U - QJ > QJ 1 8 QJ L.. :::, Vl Vl QJ L.. a. 'U : :::, (/') O l.-~...l-~-'-~--"~~-'--~~~ ----~~'--~~~~~ ~ Frequeny (Hz) Figure 3. qual annoyane urves for pure infrasoni tones based on results from xperiment I. Results. A signifiant effet of exposure time was seen (.1% level in an analysis of variane). Mean values for all sound onditions are shown in Figure 4. It is obvious that the exposure time should be speified for results obtained with the present proedure and rating sale.. There was no signifiant interation between sound ondition and duration. Disussion. The absene of interation between sound ondition ardduration means that although a variation with exoosure time is present, this variation is the same for all sounds.onsequently, the proedure and rating sale are useful for omparative measurements, and the results will be independent of exposure time. Therefore, the use of shorter and resoure saving experiments an be justified. For the two remaining experir.~nts an exposure time of 3

6 ANNOYANC FROM AUDIBL INFRASOUND minutes preeded by 1 minute of silene was hosen. When results from a) and ) are reported in the following, a orretion is made to refer these results to an exposure time of 3 minutes. _ 9 \ : 8 a Time (min l Figure 4. Dependene of annoyane rating on exposure time. The filled irles represent means of all sound onditions XPRIMNT III In this experiment the annoyane of one-third-otave noise at i nfrasoni frequenies was rated. 16 sound onditions were used and onsequently only 16 subjets partiipated. The infrasound exposures were: 8 Hz: 1, 15, 11, 115 db; 16 Hz: 88, 97, 16 and 115 db; 31.5 Hz: 7, 8, 9 and 1 db. The referenes were as in previous experiments. The 16 subjets were randomly hosen from the original 18 subjets. In the reported results a minor orretion is made in order to refer the means to the original group. The proedure NdS as in xperiment II b). Results. The ratings for one-third-otave r.oise bands are shown as unfilled irles in Figure 5. The ratings are in very lose agreement with ratings for pure tones obtained in xperiment I and II (filled irles). Disussion. The lose agreement means that the annoyane from a pure infrasoni tone is the same as from a one-third-otave band at the same frequeny and at the same sound pressure level. This is in ontrast to what is vali~ a t higher frequenies where normally several db must be added to the A-weighted sound level of a pure tone in order to give a reasonable measure o f annoyane. XPRIMNT IV This experiment was designed to show what effet the presene of audio frequeny noise has on the annoyane from infrasound. The exposures were ombinations of audio frequeny and infrasoni noise. The audio frequeny noise was a 1 Hz otave-filtered pink noise that ould either be absent or appear at the three levels: 3, 55 and 8 db. The infrasoni noise was a pure tone at 16 Hz

7 ANNOYANC FROM AUDIBL INFRASOUND that ould either be absent or appear at the three levels: 95, 15 and 115 db. All ombinations were used ma king a total of 16 different sound onditions. The proedure was as in xperiment II b) and the subjets as in xperiment III. a., u g_ 5 <( 31.5 Hz 16 Hz 8 Hz O -l-~~~-,- :_...:..;..:..::,...--=----=-~--=-,.-=-~ ~--,~~~--; ' Sound pressure level (db) Figure 5. Annoyane ratings for one-third-otave noise bands (unfilled irles) and pure t o nes (filled ir les). Ratings fo r pure tones are pooled means from xperiment II a)-). ommon regression lines are shown, r 2 =.98 (8Hz), 99 (16 Hz),.995 ( Hz). Results. The annoyane ratings obtained are shown in Figure 6. ah of the four diagrams shows the results for a fixed value of the 16 Hz noise. It is seen that the addition of the 1 Hz noise hanges the annoyane rating. All the signifiant hanges appear as inreases in annoyane. An inrease is seen for all levels of the 16 Hz tone, a lthough the needed level of 1 Hz noise is different at different levels. Disussion. A loser l?ok at the figure will unveil that the annoyane rating of a omposite noise is equal to or slightly above the annoyane rating of the most annoying of the individual noises. It is above only when the two noises are omparable in annoyane. This observation agrees well with existing experiene for audio frequeny noise. The theory has been proposed that an unbalaned spetrum (a spetrum with unusually high low frequeny ontent) should be espeially annoying (7). The spetrum of pure infrasound is extremely unbalaned, and if the theory were true, the addition o f audio frequeny noise would redue the annoyane. This aoes not happen, a nd the theory is not supported by our results.

8 7 ANNOYANC FROM AUDIBL INFRASOUND 15 No 16 Hz exposure 15 16Hz = 15 db e ~ 1 ~,oo O'I.!:... Cl... -O'I : '- QI Cl >,. >,. : < < t t No 1kHz 1 khz level No 1 khz 1kHz level 15 16Hz = 95dB QI so 15 16Hz = 115d8 ~ 1..S 1 en en... Cl Cl '- QI Cl >,. >,. < <,f, i BO I eo No 1kHz 1kHz level No 1 khz 1kHz level '- QI 5 Figure 6. Annoyane ratings for ombinations of audio and infrasoni noise obtained in xperiment IV. The four diagrams show results for different levels of the 16 Hz tone. The points indiate means! one standard error of mean. US OF WIGF.TING CURVS G-Weighting. Two weighting urves have reently been suggested for measurements of infrasoni noise (8). Both of them over the frequeny range 1-2 Hz and they have a gain of O db at 1 z. The only differene is that they have different slopes, namely 12 db per otave (G1-urve) and 6 db per otave (G2-urve). The urves in Figure 3 have a mean slope of 11.7 db per otavein the frequeny range Hz. The equal loudness urves that were previously determined (3) had a mean slope of 12.3 db per otave in the range Hz. These findings suggest that

9 ANNOYANC FROM AUDIBL INFRASOUND measurements with the G1-urve that has a slope of 12 db per otave would give a fair indiation of the annoyane and loudness assoiated with infrasound. In Figure 7 mean annoyane rating is shown versus G1-weighted infrasound level for all infrasoni exposures in xperiment I through III (filled irles). The figure shows -a very lose linear relationship (oeffiient of orrelation r!=.93). 15-,-~~~~~~~~~~~~~~~~--, --1 O'\ C1, u g_ 5 <{ 4 14 Figure 7. Annoyane rating versus G1-weighted infrasound level. Filled irles r epresent mean values for frequenies at 16 Hz and below. Unfilled irles represent mean values at 31.5 Hz. Values were obtained from xperiment I-III. r 2 =.93 for the filled irles. Figure 8 shows the same results versus the G2-weighted infrasound level. Here r 2 =.77 and it is learly illustrated that the G2-urve prov ides a measure of the annoyane nat is muh inferior to that of the G1-urve. A-weighting. For the 1 Hz otave-filtered pink noise that served as r e ferene, Figure 9 shows the relation between A-weighted sound level and annoyane rating (filled irles). The oeffiient of orrelation r 2 =.97. Low audio frequenies. In the Figures 7-9 values for 31.5 Hz a re shown as unfilled irles. From Figure 7 and 8 it is obvious that the G- urves do not provide values that an be used for assessment of annoyane from 31.5 Hz. This agrees well with the

10 9 ANNOYANC FROM AUDIBL INFRASOUNO intention behind the 6-urves. These urves were given a sharp utoff above 2 Hz so that possible restritions on G-weighted levels should not interfere with restritions in the audio range Hz is within the audio range and is thus supposed to be overed by the A-urve. However, Figure 9 learly shows that the annoyane from 31.5 Hz (unfilled irles) does not follow the same line as the annoyane from 1 Hz. The annoyane from 31.5 Hz rises muh steeper than that from 1 Hz. The two regression lines interset at approximately 45 db. This result might have been predited from Figure 3 where the narrowing of the urves for dereasinq frequenies is oresent already at 31.5 Hz. The origin of the A-urve also explains this. The A-urve is approximately the reiproal of the 4 phon urve. Assuming a lose relationship between loudness and annoyane, then A-weighted levels will reflet the annoyane of sounds with levels around 4 phon. For low frequenies at levels well below 4 phon the annoyane is expeted to be lower than that predited by the A-weighted level. At levels muh above 4 phon the annoyane is expeted to be higher than that predited ~y the A-weighted level. This is exatly what an be seen in Figure n 1 - '- a, v >- 5 : <( Infrasound level {db (G 2)) Figure 8. Annoyane rating versus G2-weighted infrasound level. Filled irles represent mean values for frequenies at 16 Hz and below. Unfilled irles represent mean values at 31.5 Hz. Values were obtained from xperiment I-III. r 2 =. 77 for the filled irles.

11 I () ANNOYANC FROM AUDIBL INFRASOUND A-, B- and C-weighting. Originally the intention was that the A-urve should be used only at levels around 4 phon, while the B and -urves should be used at higher levels. This proedure is almost never used in real life and this is most probably the reason why it has been so diffiult to obtain a good orrelation between objetive measures and subjetive ratings for noises ontaining onsiderable low frequeny energy. Relation between G1- and A-numerials. Figure 7 showed a good orrelation between G1-weighted infrasound levels and annoyane rating. Thus if a "one-figure" measurement is wanted for infrasound, the G1 urve might be a good hoie. However, this urve provides only a frequeny weighting and G1-weighted infrasound levels do not reflet the fat that the annoyane inreases steeply above the t hreshold. Thus the onversion shown in Figure 1 may be useful. For a given G1-weighted infrasound level it an be read whih A-weighted level auses t he sane rating of annoyane en 1 : _. ~ i ~ I g_ 5 ~ g : < I Sound level (db (A)) Figure 9. Annoyane rating versus A-weighted sound level. Filled irles represent mean values for 1 Hz noise bands. Unfilled irles represent mean values a t 31.5 Hz. Values were obtained from xperiment I-III. r 2 =.97 for the filled irles,.99 for the unfilled.

12 // ANNOYANC FROM AUDIABL INFRAS OUND,..., -<{..., en :g. Cl, > Cl, 5 - ::, If) Q -+-~~~~~~~~~,--~~~~~~~~ Infrasound level (db (G1)) Figure 1. Conve rsio n of G1 - weighted infrasound level to the A weighted level of an audio frequeny noise that auses the same rating of annoyane. The figure is obtained from the regression line in Figure 7 and 1 Hz ratings in Figure 2. r 2 =. 97. SUMMARY Contours of equal annoyane were determined for pure tones in the frequeny range Hz. The urves show a narrowing of the dynami range of the ear at low frequenies. The same pattern is seen for the equal loudness urves, and the resul ts support the theory that the annoyane of infrasound is related to the l oudness sensation. An noyane ratings of 1/3 otave noise did not deviate from ratings of pure tones with the same sound pressure l evel. Combinations of audio and infrasoni noise were in general given a rating lose to o r slightly above the rating of the most annoying of the individual noise onditions. The proposed ISO G1-weighting urve provides an objetive me asure that orrelates very well with subjetive annoyane ratings 15

13 /~ ANNOYANC FROM AUDIABL INFRASOUND for infrasoni frequenies. Values obtained with the proposed G2- weighting urve do not orrelate nearly as well. Beause of the low dynami range of the ear at infrasoni frequenies, are should be taken when evaluating G1-weighted levels. The numerial values should not be diretly ompared to A weighted levels. Low audio frequenies - in this investigation represented by pure tones and noise bands at 31.5 Hz - are not overed by the proposed G-weighting urves and they are insuffiiently overed by the A-urve. A possible solution might be the originally intended level dependent use of the A, Band C urves. Further researh is needed in this area. RFRNCS 1. Panel disussion of Conferene on Low Frequeny Noise and Hearing, 7-9 May 198 in Aalborg, Denmark. Aalborg Universi ty Press, L. D. Whittle, s. J. Collins and D. W. Robinson: The audibility of low frequeny sounds. Journal of Sound and Vibration, vol. 21, no. 4, , Henrik M ller and Jente Andresen: Loudness of pure tones at low and infrasoni frequenies. Journal of Low Frequeny Noise and Vibration, vol. 3, no. 2, 78-87, Jente Andresen and Henrik M ller: qual annoyane ontours for infrasoni frequenies. Journal of Low Frequeny Noise and Vibration, vol. 3, no. 3, 1-9, Henrik M ller, Peter Henningsen and Jente Andresen: Annoyane from infrasound - alone and in ombination with audio frequeny noise. Proeedings of Internoise 84, , Henrik M ller: Constrution of a test hamber for human infrasound exposure. Journal of Low Frequeny Noise and Vibration, vol. 1, no. 3, , M.. Bryan: Low frequeny noise annoyane. In w. Tempest: Infrasound and Low Frequeny Vibration. Aademi Press, International Standardization Organization, ISO/ DIS 7196: Aoustis - Methods for. desribing infrasound.