Multiple speakers in reverberation room measurements

Publisher’s version / Version de l'éditeur:

Vous avez des questions? Nous pouvons vous aider. Pour communiquer directement avec un auteur, consultez la première page de la revue dans laquelle son article a été publié afin de trouver ses coordonnées. Si vous n’arrivez pas à les repérer, communiquez avec nous à PublicationsArchive-ArchivesPublications@nrc-cnrc.gc.ca.

Questions? Contact the NRC Publications Archive team at

PublicationsArchive-ArchivesPublications@nrc-cnrc.gc.ca. If you wish to email the authors directly, please see the first page of the publication for their contact information.

https://publications-cnrc.canada.ca/fra/droits

L’accès à ce site Web et l’utilisation de son contenu sont assujettis aux conditions présentées dans le site LISEZ CES CONDITIONS ATTENTIVEMENT AVANT D’UTILISER CE SITE WEB.

Building Research Note, 1982-05

READ THESE TERMS AND CONDITIONS CAREFULLY BEFORE USING THIS WEBSITE.

https://nrc-publications.canada.ca/eng/copyright

NRC Publications Archive Record / Notice des Archives des publications du CNRC : https://nrc-publications.canada.ca/eng/view/object/?id=f47e2724-5689-4c92-add5-52ee0c119295 https://publications-cnrc.canada.ca/fra/voir/objet/?id=f47e2724-5689-4c92-add5-52ee0c119295

NRC Publications Archive

Archives des publications du CNRC

This publication could be one of several versions: author’s original, accepted manuscript or the publisher’s version. / La version de cette publication peut être l’une des suivantes : la version prépublication de l’auteur, la version acceptée du manuscrit ou la version de l’éditeur.

For the publisher’s version, please access the DOI link below./ Pour consulter la version de l’éditeur, utilisez le lien DOI ci-dessous.

https://doi.org/10.4224/40000527

Access and use of this website and the material on it are subject to the Terms and Conditions set forth at

Multiple speakers in reverberation room measurements

1

I

I h Division of Building Research, National Research Council of Canada I

- - -. - -> -. --- -- - -- " -

W T I P L E SPEAKERS IN RF;VERBERATZON ROOM MEASUREMENTS

1

I/

MULTIPLE SPEAKERS IN REVERBERATION ROOM MEASUREMENTS

by

A.C.C. Warnock

In reverberatlon room measurements af transmission loss and absorption coefficients, the sound s o u r c e is generally one or more

loudspeaker systems. Both ASTH methods dealing wlth these measurements (E90 and C423) (1,2) allow b u t do not actually recommend the use of

m u l t i p l e sound sources, nor d o they say anything about the methods of e x c i t a t i o n of multiple sources. In the current d r a f t revisian of IS0 R354(3), the I S 0 equivalent of C423 which a l s o deals with the

measurement of absorption, m l t i p l e speaker p o s i t i o n s are required. Also, measurements are to be repeated by operating each loudspeaker system coasecutively or by f e e d i n g each speaker independently ~ 5 t h

uncorrelated random noise. The current version of IS0 140/111(4)

actually a d v i s e s against the u s e of m u l t i p l e speakers. This Note

presents some measurements made f n the reverberation roams at the

D i v i s f o n of Building Research (DBR), National Research Council of Canada

(NRCC) t o examine the effects of multiple loudspeakers fn transmission l o s s and absorption measurements.

GENE= PROCEDURES

For all measurements, four loudspeakers were p l a c e d in the corners of the reverberatlon room being excited, Each speaker had its

a m power amplifier and independent pseudo-random n o i s e generator. Thus it was p o s s i b l e t o use each speaker on I t s own, all four together with the same and therefore correlated input, or a l l four together with

independent, uncorrelated inputs. The speaker cones, nominally 300 mm

in diameter, were mounted in horn enclosures wlth no separate high- frequency d r i v e r units. All measarements were made in acc~rdance with

the requLsements of ASTM E90, C423 and I S 0 140/~11, using a computer- c o n t r o l l e d testing apparatus which incorporated a real time analyser. Nine microphones were used in each raom to measure the space-averaged sound pressure level. Decays were averaged at each microphone in the r e c e i v i n g room to o b t a i n the mean room decay rate which in turn

d e t e d n e d room absorption,

TRPLNSMT SSION LOSS MEASUREHENTS

Transmission l o s s measuremeats w e r e made through a wall and a Eloar specimen to o b t a i n additional Information. k t a i l s of t h e

specimens used are n o t important for t h i s work and are not given here.

Transrnissfon lass measurements in this laboratory have routinely been made in one d i r e c t i o n only; t h a t is, one room fs designated as the

source room and t h e other a s t h e receiving room, For wall trausmfss;ion loss measuremeats the source room is normally the smallest room wfth a

volume of 6 5 m3 and the receiving room, also used for saund absorption measurements, is the largest room with a volume sf about 250 m3

.

transmission l o s s measurements the smallest room becomes the receiving room and the room above, with a volume of 158 m3, becomes the source

room. This room, which will be referred to as the upper room, d i d not

have any diffusing elements. Far these experiments, the measurements

were made in b o t h directions to observe the differences, if any.

However, it was not possible t o make r e l i a b l e measurements Ln the

reverse d i r e c t f o n through t h e floor because of excessive n o i s e i n t m s i o n i n t o the upper room.

Figure 1 and T a b l e 1 show variations i n transmfssion loss

through the wall specimen wfth each speaker excited in turn and measured

from the _{small t o the large room. F i g u r e 2 and Table 2 show the}

corresponding d a t a for measurements in the reverse d i r e c t i o n . Table 3 shows the variations observed f o r a floor transmission loss measurement

from the upper room to the lower.

In,

considerable variation in the transmfssion loss values obtained wfth single speakers, fndtcating that a s i n g l e speaker of t h i s type does not

uniformly excite the room, The tables show, however, that there was almost no effect on sound transmfssion class (STC)(5) f o r these

specfrnens, although it is clear that for a specimen with an STC sating

determined by its low-frequency transmfssfon l o s s values and t h e 8dB rule(5), the STC could change by a f e w points.

T a b l e 4 compares the transmission loss values for both

correlated and uncorrelated e x c i t a t i o n of the loudspeaker s y s t e m , The differences between these two methods of excitatfon are not very great

and occur mainly at t h e l o w e r frequencies. Figure 3 , however, shms t h e

data for transmission from the small to the large room. The increase i n eransmFssion l o s s values at l m frequencies when correlated sources are

used indicates t h a t interference affects are causing spuriausly high results. Table 4 a l s o p e m d t s comparison of the transmisston l o s s

valves for the uncorrelated sources wfth those obtained by averaging the values i n Tables 1 t o 3 f o r the f o u r Individual speakers. It can be seen that in all three cases the agreement is e x c e l l e n t .

In p r i n c i p l e , if one accepts the need for m u l t i p l e speaker

systems, uncorrelated n o i s e sources are to be preferred since

interference effects are thus aveided. One also expects t o prodoce a

more zandon field in the source room and thus approach more c l o s e l y the ideal of a d i f f u s e sound f i e l d . Table 5 supports t h i s i d e a . It shows

t h e spatial standard deviation of the sound pressure l e v e l in the three source rooms when the four sources a r e uncorrelated and when they are correlated. Using uncosrelated excitation of multiple speakers clearly

leads to a more unfform sound field, thereby increasing the p r e c i s i o n of

the measurements.

Flgure 4 compares the measured transmission lasses f a r the two directions of measurement through the wall. One possible reason for the large discrepancies at low frequencies shown in t h i s graph is that t h e dimensions of the small room become comparable to the wavelength of

sound at

I w

decay rates measured in the small room were being increased by sound

t ransmlssion through t h e specimen from the large room, in which sound decays more slowly, thus t e n d i n g to increase t h e adjustment f o r room absorption and the calculated values of transmission loss. To test t h i s

second hypothesis, a l l significant transmission from t h e large room

through the specimen was eliminated by c l o s i n g the a w i l i a r y lead-cored

door l o c a t e d between the room. T a b l e 6 shows the change in measured reverberation times and t h e resulting correction t h a t should be a p p l i e d to the calculated transmission loss.

It

differences in reverberation t i m e are s t i l l not sufficient to account

for the d i f f e r e n c e s between the transmission losses measured in the t w o directtons.

CONCLUSIONS I

These results demonstrate that there are good reasons for using

multkple speaker system in reverberation room transmission l o s s measurements and t h a t independent, uncorrelated n o i s e generators to power the speakers are a l s o to be preferred. The transmission l o s s

differences observed for single loudspeakers at the lover frequencies

can be explained q u a l i t a t i v e l y as inadequate excdtation of the lmer room modes. There are, however, some differences at the higher

f r e q u e n c i e s t h a t can perhaps be attributed to beaming from the

laadspeakers which, as has been stated, had no high-f requency driver u n i t s and were therefore poor approximations to p o i n t sources. Thus the recornendation i n IS0 140/111 that only a s l n g l e speaker be used does

n o t represent s t a t e o f - t h w a r t aeasurement procedures, e s p e c i a l l y in view of the comparatively low cost of pseudo-random noise generators, pmer amplfffers and speakers.

These results have led to the d e c i s i o n t o use multlple

independent speaker s y s terns in all f utuse transmission loss measurements in the reverberation rooms at DBR. They have also shown the need to exadne more c l o s e l y the transmission in both directions through t e s t

specimens. This vdll require constructing a new w a l l . in the upper room t o prevent the i n t r u s i o n of external n o i s e by way of t h e floor hatch

from the sample preparation area. Diffusing devices ~3.11 a l s o be i n s t a l l e d i n t h e upper room to increase the diffusivity of the sound f l e l d s

.

ABSORPTION MEASUREMENTS

UsPng the same four loudspeakers as in t h e transmission loss

measurements, a series of absorption measurements was carried o u t on a reference absorbing specimen. This specimen consisted of glass fibre batts 50 mm thick encased in an alurnfnum frame and faced on both sides with diamond-patterned, expanded aluminum mesh

-

committee 620 (now E33) and used in laboratory t e s t s performed accordfng

to ASTM C423. The specimen was supported on a wooden frame so t h a t its

(formerly identified as the Number 7 mounting) which is described in

ASTM E795 ( 6 ) .

Fifty sound decays w e r e ensemble averaged at each of the nine microphone positions. The program calculated t h e decay rate and

curvature f a r each ensemble average as well as the absorption

coefficients from the mean reverberation tlmes. Curvature is deffned a s (RbJRt-1) x 100% where Rb is the reverberation t h e calculated for t h e

bottom half af t h e decay and Rt is that for the t o p h a l f . The

repeatability 05 t h i s system permits one t o discern quite small differences between different experimental conf iguratiens w i t h a hPgh degree of confidence.

The differences in the mean room decay rates were generally small when the room was excited with I n d i d d u a l speakers or with

multiple speakers. The largest differences were observed in t h e lower

frequency bands. The mean reverberation times f o r all t e s t s in the

empty room are given in Table 7.

Since the room reverberation times, both w i t h and without the

specimen, are uaed to calculate the absorption coefficients, Z r is

necessary to look at these coefficients t o d e t e d n e the final effect of

the changes in procedure described above. Figure 5 and T a b l e 8 shaw the mean values of the absorption c o e f f i c i e n t obtained when the four

speakers are e x c i t e d i n d i v i d u a l l y . The differences are t y p i c a l l y greater than the repeatability of the measurements, e s p e c i a l l y in the m i d d l e and lower frequencies.

It must be borne in mind t h a t the use of nine microphones

i n

reverberation room i n t r o d u c e s a consfderable amount of averaging.

Figure 6 displays the range af absorption coefflclents measured for two of the loudspeakers. In t h i s case the absorption coefficient for t h e

specimen was calculated for each microphone in turn and stored by the computer. F i g u r e 7 shows the a b s o l p t i s n coef ffcf e n t s for a single d+crophone when each of the four loudspeakers was used in turn.

Although these p l o t s do not r e a l l y have a d i r e c t bearing on the matter at hand, they do illustrate t h a t averaging over all the variables in reverberation room i s important and that microphone position is one of

the most sign5ficant of rhese variables.

T a b l e 9 compares the mean coefficient calculated f o r the

specimen using the data for each of the four speakers excsted

i n d i v i d u a l l y and compares thgs with the correlated and the uncosrelated cases. T h e r e are no significant differences between these values.

Table 10 shows the e f f e c t of using uncorrelated noise on the measured values of s p a t i a l standard devjlatf on of reverberation t i m e , as well as on the percent curvature. The data in the table were obtained

by averaging the results of 20 sets of data, 10 w i t h correlated

excitation of the speakers and 10 with uncasrelated excitation.

It

be seen that although there are some advantages in the use af the

uncorrelated multiple sources in conjunction w i t h a r o t a t i n g diffuser, they are quite small,

These experiments verify t h a t in a b s o r p t i o n measurements one

sees changes i n measured coefficients when the p o s i t i o n of any of the loudspeakers used in the measurements is changed. It is to be expected t h a t if the loudspeakers were a c t u a l l y moved, the changes observed would be at least as great, if not greater, since the room would be altered by

the movement. In these experiments, however, there was no physf cal movement of the loudspeakers. It is already known t h a t measured

absorption coefficients depend on the specimen and microphone positions.

The dependence on loudspeaker position is thus s i m p l y one more effect

t h a t m s t be averaged to improve t h e precision of the measurements.

The use of m l t i p l e speakers or speaker p o s i t i o n s as s p e c i f i e d

in the current draft of I S 0 354 t s thus j u s t f f i a b l e , On the b a s i s of the measurements in this experiment, one could perhaps argue that although m u l t i p l e speaker p o s f t i o n s are d e s i r a b l e , it is not r e a l l y necessary to use the _{speakers consecutively, and that the use of}

uncorrelated noise excitation f o r m u l t i p l e speakers is not r e a l l y

essential. However, s i n c e t h e e f f e c t s of t h e m u l t i p l e speakers will be most noticeable at t h e lower frequencies and in reverberation soam t h a t are substantially smaller than the 250 rn3 of t h e DBR room, it is likely t h a t t h e requirements in I S 0 3 5 4 will a s s m e mch more importance.

OVERALL CONCLUSIONS

These measurements provide a justiftcatfan f o r the i n s t a l l a t i o n

in the NRC reverberation room of new speaker s y s t e m , each p r o v i d e d with low-frequency and high-frequency driver units to excite the room modes e f f e c t i v e l y at all frequencies. Four speakers are currensly planned for each room; however, is is n o t known from t h i s experiment

whether u s i n g more than f o u r speakers would result: in a significant irnprovcrnenr in p r e c i s i o n - Since It is r e l a t i v e l y sfmple to construct

pseudo-random noise generator;, each o f t h e speakers will be provided

w i t h i t s m n o i s e generator and power amplifier,

REFERENCES

(1) ASTM E90-75, "Standard method far laboratory measurement of

airborne sound transmisaton l o s s of b u i l d i n g p a r t i t i o n s " , American

S o c i e t y f o r Tesifng and Haterials, Philadelphia, 1975. (2) ASTM C423-81, "Standard test method f o r sound absorption

c o e f f i c i e n t s by the reverberarion room method", kmerican Society

f o r Testing and Materials, P h i l a d e l p h i a , 1981.

(3) I S 0 R354-1963, "Meesurernent of absorption c o e f f i c i e n t s in a reverberation room", International Organization f o r

Standardizatiod, Geneva.

( 4 ) IS0 140/3/111-1978, "Laboratory measurements of airborne sound ~nsulatlon of building elements", Geneva.

(5) EP43-73, "Standard classification f o r d e t e r m i n a t i o n of sound

transmission class", Amerfcan Society f o r Testing and Materials,

Table 1 : Transmission loss values f o r measurements through the w a l l from the small t o the large room using single speakers

Frequency Speaker number

(Hertz) 1 2 3 4

m

STC 24

T a b l e 2: Transmission loss values f o r measurements through the w a l l from t h e large to t h e small room u s i n g s i n g l e speakers

Frequency Speaker number

(Hertz) 1 2 3 4 MEAN

Table 3: Transmission l o s s values f o r normal measurements through a flour using i n d i v i d u a l speakers

Frequency Speaker number

(Hertz) 1 2 3 4 MEAN 8 0

100

l G O O

2000

Table 4 : Transmissian loss values for measurements made with four loudspeakers using uncorselated

(u)

( c )

-

-

Table 7 : Reverberation t i m e s (seconds) in the empty room for s i n g l e speakers and for correlated I C ) and nncorrelated

(u)

(Hertz) 1

Speaker number

2 3

Table 8: Mean absorption coefficient for individual excitation of the loudspeakers (average of nine microphone pos it ions)

Frequency (Hertz) 1

100

1600

ZOO0

Table 9: Measured absorption coefficients using correlated (C) end uncorrelated (U) sound compared with the mean far the

four i n d i v f dual speakers f rorn T a b l e 8 Frequency

(Hertz) C

Table 10: Effect of using multiple speakers and correlated ( C ) and

uncorrelated (U) sound on standard deviation 0 5 sound

pses sure l e v e l and percent curvature

Frequency s.d. SPL percent

(Hertz) (dB ) curvature

l [ l l l , t ~ l r ! ! ~ , ~ : # , !

100 200 500 1K Z K 5 K 113 OCTAVE CENTRE FREQUENCY, Hz F I G U R E I WALL TL FOR I N D l Y r D U A L S P E A K E R S - 65 m 3 TO 250 rn3 F R E Q U E N C Y , H z F I G U R E 3 W A L L T l F O R b 5 m 3 T O 250 m 3 F O U R C O R R E L A T E D S P E A K E R S

-

---

-

0.5

5

113 OCTAVE CENTRE FREQUENCY, Hz

F T G U R E 5 A B S O R P T I O N C O E F F I C I EMJS FOR I H D I V 1 D U A L S P E G K E R S (CIVERRGE O F N I N E M I C R O P H O N E P O S I T I O N S ) F R E Q U E N C Y . Hz A 8 5 0 R P T t O N C O E F F E C t ENTS FROM A S I N G L E M l C R O P H O N E F O R E A C H O F THE FOUR L O U D S P E A K E R S FREQUENCY. Hz M A X I M U M AND M l N l M U M V A L U E S OF A B S O R P T I O N C O E F F I C I E N T MEASURED I N THE ROOM FOR T W O L O U D S P E A K E R S - 20 400 200 500 1 K 2 K 5 K F R E Q U E N C Y , H z F I G U R E 8 D E C A Y C U R V A T U R E W l T H C O R R E L A T E D -I 3 A N D U N C O R R E L A T E D (---? E X C f T A T I O k ! O F F O U R S P E A K E R S