Badwin !Ie ort No. Atc - 23

1;FGAII 0.-- m LA. C CMo? T UiCrxROPHONE FO M~Hiw.€. AMIEW i" OIrSEMS

Roger E. Kirk

The Baldwin Piano Company

Cincinnati 2, Ohio

October 1957

NOT REPRODUOthLE

Project No. 7(15-4300)

-~ Contract No. M33 (616) - T - rved

Task 4 dishibu>u its

Communication and Navigation Laboratory

Vright Air Development Center

Air Research and Development Command

Reproduced by theUnited States Air Force CLEARINGHOUSE

for Federal Scientific & Technical

Wright-Patterson Air Force Base, Ohio

NWIN

BestAvai~lable

Copy

F ~~NOT Rk~P.flthTO T1bTD

IL ~ ~Notice

ifien Government draig, specifications, or other data

are used for aypurpose other then in connection with a

definitely related Government procurement operation, the United

obligation wahatsoever; and the fact that the Government meay

hav fomulted funisedor in any vray Supplied the said

drawngs spcifcatonsor other data, is not to be regarded

ky nplcatonorotherwise as inany manner licensing the

holer r ay oherperonor corporation, or conveying any

righs o penisionto anuactreuse, or sell any

patented invention that nay in any %way be related thereto.

4

'6

Foreword

This report was prepured by the Engineering and Research

Deparrtnnt of the ±!aldwin Piano Company, under United States

Air Force Contract :o. A.F33 (616) - 3323, Task No. 4. The

task was administered under the direction of Messrs. E. Lazur

end T. P. Mountz, Jr., ;!NE-3, Comnication and Navigation

Laboratory, Iright Air Development Center, PiPB, Ohio.

The contract was terminated for the convenience of the

government on September 15, 1957 by the contracting officer

at WADC. This report summarizes all research completed up to

the time of contract termination.

Q*

4l

-Abstract

Tjhe vibration spectrum and level of speech froan different

anatomaical locations was measured using contact microphones

adiapted for tnis purpose. This information was uscd to select

the optimum -anatomical locations for picking up speech in ex-

tremely noisy (140 db re. 0.000% dyne/cm2', sound fields near

jet engines. On the basis of this preliminary research several

microphone-noise shield systems for use at selected anatomical

locations were develoved. Articulation t~ests of contact micro-

Dhono-noise shield systems were conducted in a 120 db jet inoise 'spectrum. Unsaitisfactory articulation scores were obtained with

these systems even with optimum electrical equalization of speech

:4 s~nal, an~tomical location and seal around the microphone.

A noise cancelling system empoloying a pressure gradient t

Amicroohone (M-76) located in close proximity to the skin pro-

vided satisfactory speech intellibibility in a 120 db jet noise

spectrum. The contract was cancelled by the contracting officer

at aADC before an evaluation of this noise cancelling system was

completed. On the basis of available data it appears that this

noise cancelling microphone system located on the forehead or

neck area would provide satisfactory communicat ion in a 140 db jet

noise.

THlE BW'iN PIANO COMPANYCINCDCIIATI 2, OHIO AF 33 (616)-3323

iv

, . Table of Contents

P~rt section -Paragraph Title .e

A I Purpose I

S Identciftoaton of Poronnel

C Survey or1 Literature Direcl.]yApplIcbl'i to Prob'e.n

" H~oi:se Lev-el!s .Kncount-ired -In Vie] n [ty

or Jot AircraFt 4

4 Previous Research On fHody ConductedSpeech 5

5 Previous Research On The Attenuationof Sound At The Ear 7

Ii A Experiaental Procedure 8

! instrumentation 8

la Voice absorber 8

lb Contact sicr6hone 10

ic Contact -icrophone calibration de-uice 10

I 6cc coupler unit 12

:ie Speech equalizer device 14

If Low noise transistorized amplifier 14

2 Speech Spectrt.m Measurement 16

Subjects 20

III A Results 20

1 Speech Spectrma At Six Anatomical

Locations 20

2 Speech Level At Four AnatoviicalLocations 2.

3 Articulation Tests At Three Anato:nicalLocations 26

4 Articulation Tests In A 120 db Xet2Noise'" 28

V

4k i.I-'

'Nb'e of Contentis(Contiiiued)

Pa; Section, ?oragraiph A -Ltle Ti,

-~ IV A AltermitTO IJethoris yor 1?ikpnsdo Speeh From 7%.e Arnity 3

A 4 Systoii A 32

4. Sys.e 8

v A Conclusions 40%

'Bibitography 4

o"1

List Or.Figures '-

iF'igur6 No. Page1Ite

Ieilbl_ il~ In Voice- Abuorbar 1

3 6cc. .Goupli 3 irnt Acce:norv -1

4- Speech E quaiz~erlllbtwork 1

5Circuit, blagran, at' Lowj la Transiustor

6 Bloc20rc Ibiar.--,q -Out iteit 1

7 Anatomaicai Libcations Used-In Investigation 19

a Sjiech. Spectrm At- Th re Ahato~il cal Locations 21

9 pechSpedtruwi At.-Thred Ahitoliic41 locations 22

10. SystemA, Pxoppsod Nloise Canciieling S,-,ystem 3

System C', -Propose'd Noise6 Ciafceling. system-

13 S.*4tom D, Acoustic, 'TicrophoneSystem 3

14 System E, NToise-Cah~ce11i,, Sys-vea U.4ing,

Pre.sure Giradient iMicrophoiie II

vili

List -OfT'ables9

'!able, 110. title

I Attenu.6tion Appl-ioi] To Spqebh- Signals. ForEqutl Lou'dnes2

II Attenuation App]led To Four -Speeich -Equali-zation.Slopes for Eual Loudnes.'l 265

III Ilesults of -t.clainTests A+. TlireeAnatomnical Lo'cations. "Three EqualizationSlopes Ablied to- Speec(.h. Specdtrum.. Listenersand Talker In Qjuiet. 27

IV Resuilts of j-Artculat-loh Tests, 'Talker -In120 db Tet Noise, Listeners In quiet. 2g-

v Results of Airticulation Tests. With 1-r55Microphone IhIA-34 Microphone Noise Shield.Talker In 120 db, 3et Noise, * Listenrs InIQuiet 3

AA

~ .;T I: Lv3SP

1 i?~or r

IPART,ict'Lonr A 'r~;

0' 4m r Ilk I c' .) 11 'U Tj I bot' To tr~ ~n';

r~~~,.Ji)~h ii'c-~Ir'it. (r~) no' r4 tL (, e .0

- . trj- -,Po 11,, reei~ o" Wi3 id 1v~ .0ie~U

c- 2JIc for ttic -. 3uciu e-et th, vibi-atinn s- ,c.ctru %m I. -I-vv. '.

tions of' thr' JIml -ir cAr.atomoy. CL, tl, basi; ot . t iCuir.%ion ttodt.3, te

of~timllU! loec. ',ol:.r -j contacs i.vonihonv in It be et-'i..I

nowl.-I o" ;"-ch level L, t,,, oot.itv-i ncaic- io

1,., '10~t sr-n es s t bo itin a;u%: 1X: notit; ~ st i.t-o 7

yi brat-lor'±~os trr1I; lucersi t the urn~o :f of' ._cki-- un n P

'isi 1±101-t atibn "vil 13o a d in t'- d S 1n o~ n r. IS E

3 hie1-'As to b3 use in -c-injunction zith the traxsducetr. r O*'r. the ~ ni

of. the :-esults obtujined,- with a -alcro:Ahone pand noise shie_-. Iocz te'i,

.t the ooti:mm tanato-opicU Ilodation, ten TU cronohone -iodelo 1erc tc bea

~iuvlo~]foi- lirfitc trial use by Ucai t se rvicp pexsonnel.

;q,bld t~f',;ts are to bce ccnducterl -ita the 7~tcrrjphonc %odels 'Iuxiig

jfAt cyr"ne t-rZL (lD(3..tAions to 'lctorviine theL'- a1.f'ctvenr:is In rio-

*v !:ipi comauinctknl'n between thf - en-tIne trim :-cherni' o.-wl the

* :~cchohc witi~iithre &~ca~

CIIEs Bf.L _PL 0 C'J:T.AYCr qu" -..a- 0:, 1

K1

COHiTACT '.IC(!O!O. tiTRFTIGATION.Final Igineering Report 6n

Task 4, asil.fned underContract No. !d,':5( 618)-52:5

Sectlon B - Ide,.tification of- Perionnei ('Total manhours expeni'ed on theproject jioWn in -arentheses)

2. a) A. 13. Bereskin, professor of eledtrical engineering,, University

of Cincinnati and electronic consultant for Baldwin. (14)

b) Harold A. Conklin, Tr., senior electronic engineer, B.S.EE.,

University of Toledo; seven years experience in. companications,

ordnance electronics research and development, and acoustical

research and developmeAt at Philco Corp., National Bureau1-of

Standards ( Diamond Ordnance ruze Labs.)., and Baldwin. (88)

c) Terry Dempsey, acoustical laboratory techrifian. (52)

d)Y Robert K. Duncan, supervising engineer, acoustical research

section, B.S.EE. Purdue University, Lt. Cmdr. U.S. §.R. with

experience. in Sonar and electronic equipment. Registered

Professional Engineer;- LLB.,- Attdrney-at La*, Ohio Bar.

Twelve years industrial experience as acoustical and electrical

engineer- at RCA, Cpehart- Farnsworth, and Baldwin. (2?)

e) Roger E. Kirk, senior psychoacoustic engineer, Ph. D. in expi-

mental psychology, Ohio State University; two years experience

in psychoacoustics and human engineering. (584),

4') D. W. Martin, assistant chief engineer and director of research,

Ph. D. in physics, sixteen years experience in-Industrial acous-

tical research and development at RCA and El dwin, and as con-

sultant, author and editor in fields of acoustics and audio. (112)

g) R. L. M4urphy, intermediate electrical engineer, B.S., American

Television Institute of Technology, six years industrial experience

in measurement of acoustical transducer and audio system per-

formance. (467)

CflCIrATI 2, OHoO

i.ha I ngineerinR Recpokt onTask -4, assigne'd undier

Contract Ie. M335(6l6)-33P23

Do Lnald Patrick, acoustical laboratory technician; 2. ye-ars pre-

enginceridg 6t-University lof Kentuckf and-University ot' Gin-

cIhnnti .fk ye!ars IndIustrial laboratory experience. (36)

S1), Arthur N~ttit, acocusticil Thborjttory technic-an. (,?I)

- k) ,Th~ folown~-enginearing -aoiodll shop technic Iansz worked, on this

ads k.- P., knge (,12)- I. Reidinger (13.5)-

-Foitner (4) - Thuenezian (9.5)

- - 1) the f.rt-'tes0t.crew cons4isted of the folw b tu den-t fra;;P

theC~legeCoseratoy'of LMusic:

- - 11 Ho~ds ().L. Radabautih (3)

- -J.I~kson ()I., Surface (3),

A. Loon 05), T.- -Vian, -t31

a)Th scod-tet re cnsst 6 f the folloin atdnsfom the

Cdollege Goserviitory of .Music and the Unlvelpsity of Cincinnati

.0-0-. Eich j (49y A.. McClellan (44.5)

iHartFi, (-) .- ~l~ 49)H. A,116

*S. Izenrson -(49)' 3. Lankston (49)

-- R. Lipl(& (49) T. Surface (46),I

- - - THE BLDI PIAI40 COMP"TYCINCIN14ATI 2, 0OHIO

I----

-~ ~REPRODuc

;-t re . of TrA _L -a-

c.r-i' sic-cr- of- ro -,_rztO

Ct arm.-.-__ fi-e

~ . -*i--0: ~ -u'tic -oa r-i %to~c thda7-r

opea-',os.'.h, a~ of -roatie devce: -su.e: ac, 1 ort

i, -' "2 *~ .'ra- t n-a postions L±- '.cc--or_4;n ts ac - ~ by

Adre: , ";anno- r . vo 17-r~ew ) thre- u c o f -5. _i.' er ear I

- ~ ~ ~ ~ ~ ~ ~ s '71r -is;~ i.it :U:. cn ~-- ee o~i -r-

wereros.l .:h~ueo ~oetv eie u! s'' 'tO--' erx-

i prer-lisile er;x-pozure ti-ie in 4. 130) db jet. noise Lr7iv;y

)D minutrjs a3 cowvared to -33 aewmlls for unproteclte- earz.

A -AZ.

NOREPRODUCL~

: *i a C-u =irm' vej nnli le~

_rn

bcine os2I:;J tr a~e' t~ an- - - - ~

-i --- '- t

s;3.C n z~~f war- iora Cut ir i ~~

OHM

- - U tNOT REPRODUCLBL~

S. *

- i.-- ~A - -

I

1

- - . ;t~ -- - 1.~.--... -.- - *---~-.,----- -~ ~-~--.&A

-.. .

- . . -t - -- - ---

A

- ~-i -n- -

-, .- - ,- -.

-, - ~ -...-..--- - - 4-

k - - i~- --

12* -- -- -A------ -- -- A-----

s.~ - - - F& - -

N-.- - -- - -a. a-..

- 4-- - -.g- Al - ~.J ~-..: - - -. --

- - - -,-4-...-.- ~ - --.. * - . --

- -; --. ~*

I- ~ - 5-* -- - - -- ~ - 14..-~ .4-- 0. ~e

-- -A t.. ~

-- --- .-

* - -. - ~.. N ~ 4- -~. -'... - -- .---. -- v--~~

-- 4- -

-A- -~-----, -~

I.

OT -REPRODUCIBLE

"" S .. '-.. " .. .+ ' -- .. ?~ n - ." - :.; + "" -

t -bthIone

in"

. .. ......- = 4 . a. -, : ._ . .:.

- -L

S - - ";- .. . ... L -

-. 5 I

.-startion~1~'~

CC!ZTAGT

Zinn!Enielir Retort o'Task 4, a~si, ned under

P ;ar II

Sfetion ;1 - Ernerkeantal Pr weure

:) Voice b:.orber

T. order to flle_3tlate the Vibration spetrum end level ofesc~e 9V-p s picked up froS loicaliy chosen loct oas om

the an-atony, it was necedsury to, provide mans fIr tsIoiag

the talker's bone conducted speech from his Lir conuctedseoch. To ths end a voice absorber w adatad Oa rll-

able box. The design of -whe voice absorber is shao in 'Figure

1. T overall-dimns ions of the vo"le absorter were 60 x 36 x 22

incbes. Ite interior of the voice absorber l lined with 3. layems

of Aerocor fibre glass.. The JAdtb piece for the voice absorber 1s-

desilpned from a modified A13A Oygs mak. _4A. adju3table -head -r-est

hold the subject's head- firily In place. A sound level -iter raicro-

phone vas mounted inside the voice absorber approximately 13

inches from the talker's lips. A VU meter was uounted on the

front of the voice absorber so that the talker could *onitor

his speech level while talking into the Voice absorber. The

voice absorber provided attenuation of sound of the order of 15

to 20 db. It was believed that the acoustic, 12pedance offered

to the voice mechanism by the voice absorber ws not signift-I

,cantly different than that offered by a free field. In order to

ascertain if talking into the voice absorber produced a change

in the speech spectrum, the spectrum ror-a typical talker was

e-asured in an anecsoic clisxmber u.ing is 6401U, microphone. Fol-

TIM IMI)NdflN PIAN4O %'AIYCINCDINATI 29 01110O,

-: ~ IA

Io,*s

let

lid

IleV

did

10' .60, gz

CCC i~e 04.. igp

I.e is6go$ too

3 .i If e ee u i *o f fe~g -so#

ii' to'i se ;L.

EgG lot

ICC a Of

of

*i~E ALDD;PUNO' COWIPNYCINCINUATI O

-NOT RODUna.

a a at c ha-,--:rs S1

.4 - ~ - 11 0

cr--el _ r-t -' _-: Xbot-v . comans -p -3 - t.o ' f

use %-it un-,- -- sds.4i oicr h

-. -.mvOf C --

*~*j b ~U NOT REAPRODUCML

(71-

-17

0 U.

*7-

o - I 1 I - 1.*

- - OT RPWRODUCIL

-S..' L~ t

;.J~

~*%.~r vm- -

* on~n :bao~vtebrsls.n

I-.C tA 'h n ,: -irnos -, n -b

SCpU' er IY A , ;iffl i It %. netxe *hFt the;G~ -1.to o !up

V) flW at; "-st by -. " l g soun

Lr~ t. - - . i!~ -~ 3' ii.. uC .~ n: .s ~n; c~ijA:i')4l o

FF

-11.0

UEDWIN.'t,, CqlAN

UI-we--- /!

-14-

COMTAC1! j1ICRO;7d01!oi! IMWESTIGAT [OiFinal Enginerin, "Report on

Task 4, aSsigned underContract No. A35 (616 )-:3:53

levels in the 6 c c. coupler. The .'/. E. amplifier was removed'

from the equimnent setup and an Altec power supply used to provide jthe polarizing voltage for the 640C/M microphone. The microphone out-

put was led directly to the RCA booster amplifier., Measurement of

coupler sound pressures at different locations was not completed,

but the high levels observed suggested the feasibility of microphone -

noise shield Srstems- D and E later tried.

e) Speech equalizer device

It has been observed by numerous investigators that body conducted

speech Sounds differ markedly in spectrum from airborne speech

sound. emitted from the mouth (2, 3, 8). In general, body con-

ducted speecn -sounds contain less high frequency energy relative,

to low frequency energy than do air conducted speech sounds. In

order to render body conducted speech sounds-maximaly intelligible,

electrical equalization of the signal after it has: been picked up

frot the anatomy is necessary. To this end a passive equalizer

network was developed which provided means for obtaining the fol-

lowing speech equalization slopes: 6 db/octave, 6 db/octave to

1000 cps ,,nd 12 db/octave thereafter and 12 db/octave. A circuit

diagram of the equalizer network is shown in Figure 4., i,

f) Low noise transistorized amplifier -

The vibration amplitude of speech signals from certain anatomical

location, for example the forehead area, is relatively low-in

level. When equalization of speech signals from the forehead

ThE Bui3ADf' PIANO, COviof

u=CININTI 2, OHIO ,-

W401

............ ........

41,'

620 62- L

P051i~r) 1.12 Cb/octave risirig froma 0-10) yC

Po-ition 2.-6 db/octave risingrm 0-1 and'1 dh ,.vebeginning about I KC0

Position 31. 6 db/octave rising from 0-4-10~ KC

Positdon 4.; No equalizatj icn

Ty. wzywl~ PIVIO mTIClkl'NATL 2i,,OI,04'6

-16-

CON-TACT" SUIROP-HONE r,.-VBSIG.,T1CFina Englneerimg Report on

Task 4, assigned underContract No. AF3 (616) - 33. 3

are -was attempted, considerable difficulty with electrical

noise, T.V. and radio signal interference, was encountered.

The T. V. sifnals, radio signals and hum were found to be entering the

system primarily through the contact microphone and microphone cable.

In order to minimize these interference erfects, a low noise transis-

torized amplifier was constructed. The circuit developed separately

by Bereskin (3) is shown in Figure 5. A block diagram showing all

the combinations of equipment used in picking up speech vibrations

from different anatomical locations is shown in Figure 6. Tha

system shown in Figure 6 allows for considerable flexibility in

the types of microphones used.

2. Speech Spectrum Measuremont

On the basis of preliminary listening comparisons of the speech

spectrum from all locations on the head and neck area, six anatomical

locations were chosen for more extensive investigation. The six 10-

cations were (1) right and left side of neck on plane zith larynx,

(2) one inch above nasalis, (3) across coronal suture on top and

along midline of skull, (4) one inch below occiptial protuberance,

(5) great wing of sphenoid and (6) one inch back and below mental

protuberance of mandible. Figure 7 shows the 6 locations used in the

investigation.

The contact microphone was attacba to The anatomical location

under investigation by means of an adjustable strap.

The subject'n head was pooltioned in front of the voice absorber

,so that a good seal around the chin, cheeks and nozie was obtainAd.

THi E BALDWIN PIANO CO!rHiYCflJClJ' 4ATI 29 OHIO

1r

12I1

iti

:e].m-mtl ndi q-3r-ofl

drI ers

12)R- is :Poxia- ar -~q-3.mpicatdn o !

.1. __________gain_1_____il.

(5 oe e rfrne3

Flt-OCICo~li- OHIO.

MWU

II

I!

mI

3

iII .. ~ I]

-, IC IC--. I

I,o I,~i.II..-4

-~

ju . c.~3-3*~ -~

S

*'~ L -

.3. -.

ii*~1-E

- - -4

~- .~ -

4. 4- I

- - -

- --

F-~.. :-. LP

- r

.. -LLi. :4

4

400 00

[JZ1J 0

-ZI~, B~d.D~fl~ PIAIIO CC!~PANYOfliCINNATI 2, OHIO .~

I

4 d:~ ~L~ ,'atY @ !YS~:~O

I - 5 __

1 -6 61

Y,

I Oight ad left side of ne-k on plane with lar'nx

2 One inch~ above nasalis

3 Acrnss croonal suture on top and along =idJ ine of skull

. One inch below occiptial. protubcrance

5 'reat ving of sphenoid

o One inch back ara below mental protuberance of mandible

'ih.JYIN P1.40:0 CU YANYCINCfl-ATI 2, OHIO

~I

- - NOT REPRODUCIBLE

U, *- -test

. - o- - .

--J . .. - - ,-- - . .- fi ". - -- - .t -

S- "t t es

S - - -- .-. nt - - r.

o1.

9. Of . , . - F 9 -

"...... . c

-tI lI , • firs

NOT REPRODUCIBLE

P-4

E4 04~ 3

0 ..3

0 0

4 -P;4

.4.

%* . "-0~

ri i .4- ' .. ...

CC

E-'O.4f

C, CCTTA-GA40J

"W ....

NOT REPRODUCIBLE:

00

44 .o

0 0

-£0 di00

03~

4J40 =114

fi. >

11

TQU 0A'

(x)TkT ? z~o~io~T:NSST rGATI0O;Fir,%- : inecrinr~ Reort on

Task4, ineJunder

energy below 1000 cps. This is of interent since freque ncles

above 1000 cps contribute lisproportionately to speech intelli-

C-ibility. it should be noted that the data -.-.as obtais'-d far one

subject only.

j ~The lf2;t Kontak M'icroohone wa used to pick. u-) the body

conductedI speech vibrations. The freqiuency response of* this

microphone has not been deter:-iined due to termninat ion of' the

contract and therefore only relative coioaarisons aon.7 the curves

can be made. On the basis of linterning comparisons of speech

from the six anatomical locations, it appeared tha.t best intelli-

gibility could be achieved from-. the forehead region.

2. Speech L7evel At Four Anatomical Locations

It is evident from Figures 8 and 9 that there is a consider-

able difference in speech level at different anatomical locations.

In order to ma~ze a valid con'narison of speech intelligibility

among the locations it is necessary to equate the speech levels

so that they s;ound equally louI. This .-.as accomplished by havingI.the subject3 monitor their spreech on the sound Level meoter while

maintain a constant vocal effort. Loudness balance comoarisons

were made while the contact microphone was switched to each

of te aatomcallocations under investigation. The a'-ount of

Speech signal attenuation necessary for equal loudness was

determinod while the subJects repeated the"Joe . . .. Lawn"test

sentence. The results of the loudness balance tests are shown

CINCININATI L, 01110

n.] "n ineerinp .eport onTras" -, aa:;ir nd un.',er

Contract No. ;(G1h-.2.

in Tabl,- f. ,\-o 3ets or loudness balance tests were made for

each subject. In this way thp variability attributable to

factors such as olacelent of the microphone and errorr. in loud-

ne:;!3 balancin;, the signals could be deterzained. The average

di:'f-re: c bewe,' tno fir:;' ar I second trials was 1.,, db. This

ir. :i:,t thut t... reliability of t!:. exnertmrental procedure is

B:.-t ifi fac tory.

TABLE I

Attenuation Apnied To Speech Signals For Equal Loudnes3

ight Side ol' One InchSubject Neck On Plane Below Occiptial One Inch Great Wing

Wfith Lar-nx Protuberance Above Nasalis Of Sphenoid

E -21 -7 -l -2-24 -9 -2 -3

;Th 1 -30 -7 -3 -2-27 -10 -5 -6

.;1 -27 -9 -2 -3-13 -5 -4

Jl 1 -27 -8 -3 -3-30 -6 -3 -2

RL 1 -2b -8 -3 -24 -z. -6 -, 0

i.' 1 -23 -7 -3 -32 -26 -7 -3 -2

JS 1 -27 -9 -4 -5-29 -11 -8.,..5

Ml -26.1 -7.9 -2.7 -2.9

M 2 -26.7 -8.9 -3.8 -3.1

1 2.53 .83 88

C 2.37 2.47 2.17 1.88

,,Ih. Plt.OWfIM 1'IAJ'JO CO'PANY

C [NCIL ArI 2, OHIO

co; r:'ACT :!CEOPHOi NV I~J 3T P';T!O!'Fina! Engincering Report on

Task ,, as-:ign,:d underContract No. AFZ:"(616)-n323

It is evIdent froin Table I that speech signals frou the

throat arearequired the most attenuation for equal loudness

followed by the occiptial protuberance, nasalis and great wing

of sphenoid in that order. The results of these loudness

balance tests are in agreement with the findings of nrevious

investig,=tors (2, 4, 9 & 10) to the effect thiat the highest

level speech sounds are those in the region of the larynx. In

all of the articulation tests reported in the results section,

the speech signals were adjusted for equal loudness.

When electrical equalization was applied to the soeech

signals to render them more intelligible, it was necessary

to change the system gaia in order to ac ieve equal speech

loudness. The amount of attenuation used with the four speech

equalization slopes is shown in Table II. The amount of attenu-

ation applied to each slope for equal loudness was determined

I ' by the loudness baleaice procedure described above.

TABLE II

Attenuation Applied To Four Speech EqualizationSlopes for Equal Loudness

Microphone0 6 db 6-12 db 12 db

Electro Voice 29 12 3 0

'V Amperite Kontack 30 14 2 0

For a description of the different slopes see Part II Section A le

qqi BALDWIN PIANO C0o::tpANY

CINCMH\NATI 2, OHIO3'k

f

-26-

CONTiACT IA fOROPli0I'~T~ STIGATIONFinal Engineering Report onTask 4, assigned under

Contract No. IiF33 (616)-323

3. Articulation Tests At Three Anatomical Locations

Articulation tests were conducted to determine which ana-

tomical locations provided the best speech intelligibility.

Three anatomical locations were used; one inch above the nasalis,

one Inch below the occiptial protuberance and the right side of (Ithe neck on plane with the larynx. An Amperite Kontack micro-

phone and an Electro-Voice microphone were used to pick up the

speech vibrations. The listeners and talker were located in

auiet. The contact microphones were connected to the AIC-1O

Intercommunication System by means of a 60011 to 51L matching

transformer at the output of the speech equalizer device as

shown in Figure 6. The microphone input to the AIC-10 System

was used so that this part of the system could remain constant

regardless of the type of microphone used. The subjects listened

to PB word I ists over ?79/AIC earphones. The results of these

articulation tests are shown in Table III.

It is evident from Table III that the best speech intelligibility

is obtained Vrom the forehead area. It is interesting to note the

equalization slopes which provided the best intelligibility at the

three anatomical locations. This is an indirect indication of the

relative distribution of speech energy along the frequency continuum.

We would expect the throat area to require the most equalization

since as shown in Figures 8 and 9, it had less high frequency energy

relative to low frequency energy than either the back of the head or the

forehead. This expectation is confirmed by the data in Table iII.

THJE PIANO COMPANYC-INCfINATI 2, OHIO~, ~ ~ ___

~- 2°7 -

' ,, ..,-.-T.,%;!: ACT ............ :: r-~T*D ES I ATr,"rinel Engineering Renort on

Task 4, assigned underContract 1.o. AF33(Gl6)-3:'23

TABLE; III

lesults of Articulation Tests At Three natomi cal Locations.

Three F auivstion Slopes Applied To 3pecch Spectrum*

Listeners & Talker In ,uIet

Rligt Side Of One InchTalker M.icrophone Neck On Plane Below Occiptial One Inch Above

With Larynx i Protuberance Nasalis

0 6 6-12 12* 0 6 6-12 12 0 6 6-12 12

J.S. Azmperite 88 90 91 84 90 94 96 94 97 96 97 93

J.S. Armperite 84 91 91 96 92 97 97 91 97 96 93 97

J.S. Electro-Voice 82 94 78 90 88 91 96 86 93 91 96 94

J.S. Electro-Voice 81 83 84 87 91 92 96 87 97 98 97 941.S. Electro-Voice 81 84 85 85 87 95 95 89 97 93 93 91

AVE. 83 86 86 88 90 94 96 89 96 95 95 92

[ : "i.E Anmperite 7? 83 85 89 88 93 9? 92 99 93 91 89

Amperite 85 91 90 93 88 91 96 91 99 98 95 96

*';.E. Electro-Voice 84 92 88 80 90 87 87 85 97 92 92 81

- . Electro-Voice 8," 79 85 89 91 84 82 81 90 95 93 80

cW.E. Electro-Voice 80 93 79 89 79 85 90 90 95 95 87 80

ILVE.~ 82 88 85 88 72 88 90 88 96 95 92 85~

* AVE.(To j 83 87 86 88 81 91 93 89 96 95 94 88TALKiMS)

* For description of equalization slopes see Section le titled Speech equalization device.

THI E B Ai"'D P [A: CO,.PANYCrN~h1.AT 2,OHIO

-28-

CC,'fAcT :.1ICROPi.Oi:r, I"IEsTICATIOciFinal Engineering Report on

Taskl 4, assigned underContract -o. PY33 (61C)-3323

TP differences shown in Table III between the equalization

slopes are snall. e would expect speech intelligibility

differences among the slopes to be more evident under con-

ditions of nigh ambient noise at the talker end of the system.

4. Articu3ation Tests In A 120 db Jo Noise

Articulation tests were conducted with the talker located Iin the contractor's noie room with a 120 db jet noise spectrum. I

Listeners were located in the control room in a relatively low

ambient noise field (88 db re. 0.002 dyne/an 1. Noise shields

for use with the contact microphones were obtained by modifying I

existing ear-muffs. Two types of ear-muffs were used for this

purpose; 3traight Away jound Protector AO 372-8 and Safe-T-Zar

.iuff No. 255. It was possible to achieve ambient noise attenuation

of the order of 20 db by using these shielding devices around the

microphones. The results of articulation tests conducted under

the ab,)ve conditions are shovm in Table IV.

C~t~, T GT ', - ....

Task 4, a ,-- un.erContract No. " ""! ,-3

:-.sut:., C. ArtP-ulation Tests

Talker in 1:,- -!b Jet N:oise

L!stoners In .ulet"

Itight .3ide Of One inch

Talker Microphone Nleck On Plane .;low Occinti.I One: Inch AboveSLarynx Protuberance ..fasal is

3 6 6-12 12 0 6 6-12 1% 0 6 6-12 12

W. E, Amerite Kontak 15 19 12 8 11 14 8 6

e. Aaperite Kontak 20 18 14 6 22 -2 1i 8

J.S. Amperite Kontak .i 20 14 29 43 32 30 28

J.S. Amperite Kontak 11 19 17 22 30 ,5 22 20 F3..;. Aiperite Kontak 30 33 37 28 39

J.3. Ampcrite Kontak 22 31 33 30 4X0

W.E. '.1-76 mike (mois- 12 2l 19 28 22 22 1.P 17

ture barrier 1ii pressed against skin),

i 14-76 (moisture 14 16 21 20 P5 39 25 22barrier pressedagainst skin)

T. S. M-76 mike (mois- 24 35 37 23

*ture barrier

pressed against si.in)

T.S. I .- 76 (moisture 14 39 40 25barrier pressedapainst skin)

T.S. M-76 (moisture 22 25 19 31 29 41 31 22barrier removed,mike against skin)

J.S. M-76 (moisture 29 40 39 23barrier removed,mike against skin)

i Data not complete due to ter- ination of contract.

In

!no! _'Engineerinv' Reort on

IWsk 4, assirnt.-d undIer

7:0.

It is evident fra-z thze results shown in Table ri that speech

intelli 'ibility usin,. a C('ntLaCt rMicrophone in a 120 db jet noise

v:as not setisr.actory. As -ih be expecte-, the problen was

poor SI:; r~st:o. It will lie reca:1ed that arti-ilation scores in

vitt~t (Table I11) were quite stfaor.Attempts to achieve

£:o~tnn'n :1b of' ambient noise attenuation by improving, the

rnicropine nt'Ase si.~elds were unsuccessful.

The mirronhone currently in use by -echanics while servicing -

jet aircralt in a 140) db of noise i3 the LI-55 in a modified 'M-34

rnicroubone noise shield. A microphone ani noise shield typical

of units currently in use zias obtained and articulation tests

conductei. The results of tnese tests are showi.n in Table 11.

TABU, V

Results OP Articulation Tests With M-55 'Microphone

In .1-34 Microphone Nqoise Shield

Talker In 120 db Jet Noise

Listeners In ',uiet

Talker IArticulationScr

J.3. 86

J.3. 86

3.S. 89

J.S. 81

T.S. 79

L~fAN84.2

TKIEi LALD.;D:1; IANO COipaIu~YCflm~yh~ ~,OHIO

II

ZX-WCT :HCROPE~ v~:na1 . ineerinv Report on

Tas, 4, assi iend underContract io -'' 1 32

.Then th.e results of artictilation tests usil- a contact

microohone an! a .- b5 microphone are compFrrel It is e-vilen t

atht the. -r..rancc o:! the ctrntct m:icroihone under o.imum

Y cona:itionzs (e-iualization, anat-.nica1 Location, 4'q,od s-,round

shli!d, utc.J is far Itifrior t the r."ce articulation

test scores usins- a co:tact maicrophone i In !;;l b jet noiz e ere

not satisfactory, it *','as aVDarent that little chance fiJr success

could be exnected in 140 db noise as outlin-id i:. the task as-

sig-.ent. In view of these fndings, it .-.-as deci'Ied t . try a

different approach to the problem of ricking up speech fr: -he

anatomy. These approaches as ;:ell as a qualitative evaluation

of the two approaches which were reduced to practice are ies-

cribed in Part l7 of this report.

PART IV

A. Alternative :iethods For Picking Up Speech From The Aiuuiomy

The results presented in Part III of this report indicate that

the chief problem encountered with contact microphones in a 120 db

jet noise is poor signal to noise ratio. Attempts to achieve greater

than 20 db ambient noise attenuation by shielding means around the

microphone were unsuccessful. Since adequate shielding around the

microphones could not be obtained, the following five alternative

approaches to improving S/N ratio were considered. Four of the ap-

proaches considered woul1 involve some method of' cancelling the noise

which entered the systea in addition to the use of shieldin! around

the :i crophone.

Tili'. 1A\1A1h J PIAN0 C0O1iPANYC fNC/INATI 2, OHIO

~ ~ 4~ 4.~. ~!

-32-

CONTAC1 .lC:rROOP 1fETIGTIONFinal Lngineering Report on

Task 4, assigned underContract No. AF33(616j-33Z3

1. System A

The first system is illustrated in Figure 10. This system con-

sists of two contact micropt.one3 and noise shields located at adjacent

places on tne i nt my. This system requires that the microphones be

located on ureas which have the same noise level and spectrum at the

surface of the skin but different speech levels. Two possible locat-

ions would be the forehead area and the cheek. The ambient noise

levels would probably be very similar at both locations. In addition,

the noise wiich is radiated from these areas is probably similar in 1spectrum. The signal outputs from the two microphones would be led

to a nixing circuit. The output of the mixing circuit would be proport-

ional to the difference in signal levels between the microphones.

Although the noise levels at the microphones would be similar, the

phases would be different at high frequencies, so that the noise

signals would not be completely cancelled. This approach was neither

reduced to practice nor considered promising.

2. System B.1

The second system is shown in Figure 11. This system con-

sists of a contact microphone located on the skin to pick up

speech, and an acoustical microphone located in the same enclo-

sure to pick up noise not attenuated by the noise shield. The acou-

stic micronhono would feed the noise 1800 out of phase to a

small loudspeaker located in the enclosure. Such a system would

effectively cancel noise i-i the enclosure around the contact

micropbonc. The feasibility of cancelling noise in a small en-

TME !4kLD;'7N IANO.1 CO"_'-MY

CJ '1GJ:".%,11.J 2,OFI

- - !i.44

7 1 I-rYN~

-,.~ 7~;NN

.3. 13'. -~i V Ip -. I I./ f I . - A

- L - 2I' ....2~

AN

'-N'

2-1

~! ~ cNi~f. . .-

-U-

0 / IIL J I I

i-/

'I* Ii

JJN

V ~'

I, ... J

'S I 0*--y I

.\N'-

~ NN

r '

fl

ThE EALD~3IIN PIANO CO~M~Y- CinCINNATI 2, OHIO

_________________ ~ *'*~ '-!' ______________

"y.r

- ICI T 2 HO-tt

GI: , (;~~~OiT.AC(,i LC.i{C;i:u[,.. I..iG,,,.' .

iFinal ;ngineerini Renort onh Task 4, assi",ned under

Contract :;o. .,33 (616)-33.3

c]osure wis been previously demonstrated by Olson (12). Pre-

lirminary work on this system was begun but not completed nrior

to termination of the contract.

,molification o tr.in afproach would involve a bridge inV the electronic circuit :o thait the acoustic microohone (21 wo:i

alvay.i drivfn in the null dircetion. This aoproach was not

tried due to terrmination of the contract.

3 . Sy:tem C

The third system is shown in Figure 12. Thi3 system uses

i a contact microphone located on the skin. This microphone is

activated by speech signals as well as noise in the enclosure.

[ The output of the microphone is fed 180 ° out of phase to a loud-

speaker in the enclosure which cancels the noise. The system was

not reduced to practice.

4. System D

The fourth system in shown in Figure 13. This system con-

sists of an acoustic microphone in close proximity to the skin.

(2: The microphone would be located so as to achieve a good seal

with the skin. Im experimental model of this system was con-

structed. Preliminary tests with this system have shown promise

of providing good speech intelligibility in a noise level of

120 db. It is believed that this system would not be satisfact-

ory in a noise level of 140 db.

THE BDL;JIrN PI./ CO,,AIM4fxcrni;A.'L4TI 2, O2IO

i' A

- -t

* ''*''.'~'1 j'~e ( r.r,. * r ~'

I.,-'

7 {

7

-if'4

A A

- 7- 'if

.1- -4

-I' A

-'/

7 '.4

[L~~ _

v~e ~>A-NJ

/ /I.' / // ' i 'die 4

/1 / /1/ / ~/ / A

/ 'if.

'I'

4: ' i~'is r -. lt '.er

3 .I~3O De'r~:i-'~~3 . 1'ter

THE I3ALDWfl~ PIANO COMPANY if

-' Cfl{Ofl~NATI 2, OHIQ,~ *.,, - .-.- ~ ~'

44A ~ -______________________________________

1~

-

I:

{I - 7' NIl'

( -'

/7 \. \ I-

L"ia :~I~I " - -.

-- -~

~ icrcphon~"lose

TIi~ Bd~U 1?fl~ ~iO COMPANY

I:

_____ _____ ___ I;___________________________________________ __________________ ___________ ~.,iJ94

& ________________________________________ ________________ * i t.,~

_____________________________________ _______________ __________ ,Z~VA._____________ ~-1~

Finul ;ngvineerinfP Rcport onTask 4, assignod under

Contract No. A?35(6l6)-3323

n.System E

The fifth systei is shown in Figure 14. This system con-

qistsj of' an '1-76 pressure gradient microphone in a noise shield.

One side of the noise cancellin,', microphone was sealed off from

the otntr :-flle as snown in Fig~ure 14. TPhe diaphragmis of' the nois~e

cancellinfg niicrophoric were ortenl1d at right angles to the skin. 1Speech was permitted to enter one side of the enclosure but was

prevented from~ entering the other side by a rigid metal plate.

Thus the speech signals activated only one diaphragm of the pres-

sure gradient nmicrophona. PNoise on the other hand could activate

both diaphragmqs of the microphone. The effect of this arrange-

ment was to cancel the noise entering the enclosure While trns-

mitting the speech signal. T'his system was reduced to practice

using an early -iodel of the '.1-76 microphone in a Safe-T-Muff ear

defender. The enclosure was separated into two sections by a

half inch sheet of neoprene rubber. The inside of the enclosure

was shaped with clay to reduce the internal volume and to produce

a horn effect from the skin to the microphone diaphrapa. The

inside surface of the metal plate,which prevented the speech sig-

nal froma entering one side ofl the enclosure,was covered with

1/16 inch neoprene to simulate skin.

On the basis of listening, comparisons of this system with

the M-55 in a M-34 shield, it is believed that intelligible comn-

munication in a 140 db jet noise could be obtained. At the timi

the experimental model of this systeri was completed, the con-

tractor's articulation test crew was on vacation. It was planned

Mp, BkYD%7 PfA!I0 COIPANYCfl'Cfl'ATI 2, OHIO

7Z;________________________________________________ . ~

4

II - .

7,. -

-~ .

4

I jAI~

I

ill:

ii!

I ~1

I 1" II I f

I'

- / /i /I .-- ' II:- A

/

I~A

i/ -. ~

i --- s-I I /:1' '1" Ti- /

- / / ,,/' 1~ // /I- / C , '/-

-j

~ Press ir- ;~.:V - ,fc. . -'V

I-,

A

1-.7

TiTh. BALTY.Vfl-I PIANO COMPANY~4. ~ ~74,. - ~.,C --

F~inal Znginocring' Roport onTask 4, aL'si trnud under

Contract No. AF (6 J ~ ?

to conduct articulation tests on this noise cancelling microphone

system when the articulation test crew returned,h!at the cofitroct

was termninated before tis could be accomplished.

The signal to noise ratio of tine mi-crophonfe systen- was +18' db

in a 1:v. dt; jet noi.-e.. This figure was obtained with the micro-.

phone syste-i Iodated on the side of -the throat. This i's not the

optimum Iodation for the synte.;a but was dictated by the .shApeof

tte. shield enclosure. The speech quality of the- sys tem wan, judged-

good even in this throat position.

On the -basis of the data available it appears that -this miicro-_

phone systdm Would meet the req4irenionts set forth in the task as-I

sign.'n6nt.. Specifically these requirements were to: (1l) provide

intelli-ible dornunication in a 140 db jet noise and (2) jprovide

a minimum of kestriction of head and body movements of the wearer.-

A. Conclusions

The following' conclus ions, have been drawn from the data Oresented.

in this report.

1.- The higthe'st intensity level dpeech sounds are those radiated

fron-the larynx area.

2, Speech intelligibility is best at anatomiical locations on the

forehead (I inch abbv6 nasalis) and temple area (great wing

of .sphenoi'd).

3. There appears to be an iaverse relationship for areas on the q

head between speech, intelligilbility and speech, intelisity level.

THE~ BALII'4 PIANO 'OXIPAtM1Y A-ATI 2 OHIO

- 41 -

WncurAr ;.ICROPION % INIVSTIGATI01-lnal Engineering Report on,

Task 4, azsigned-underContract No. AF33(616)-3323

4. A microphone located on the throat requires more electrical

equalization of the speech spectrum than the same microphone

located on the forehead or back -of the head (one inch below

the occiptial protuberance) to achieve maximum intelligibility.

5. The maximum overall noise attenuation which can be achieved

by, shielding :eani around amIcrophbne located on the skin

is of the order of 20 db.

6. Conventional contact microphones under optimum conditions of

shieldihni, anatomical location,, seal between shield and skin

and electricei equalization will not provide satisfactory

articulation scores in a 140 db jet not e field.-

7. Poor signal-to-noise ratio is the chief factor responsible

for unsatisfactory peftormiance~ of contact microphones in-a

120- db jet notse system.

8., A noise cancelling system consisting of a pressure gradient.

microphone (M-?6) 6d6id in, close proximity to the skin in

a special Shielding device -can provide satisfactory speech

itelligibility in a 120 db jet noise. It is believed 6n the

basis of S/N measurement in a 120 db noise field that this

type of system could provide satisfactory intelligibility in

a 140 db jt -noise.

9. Some system of-noise cancelling is necessar, either by the

microphone or in the cavity enclosing the microphone if

satisfactory speech communication is ,to be carried on in

jet noise fields -of 140 db.

ThE IM iLf'IRI PLIO CO.PTANYOflCrATI 2, :OHIO

I ~ -------- ~ ----------------.

-42-

CONTACT 1AICROPiiOiE flWESTIGATIONFiial Engineering -Report on

Task 4, assiened underContract No. AFZ3(616)-3323

B. Bibliography

(1) Atkinson, Chester 3., An Estimate of the Spectrum of Body-

conducted Sidetone for the Vowel (i) State University of Iowa,

Supplementary Report, No .3, 1956.

(2) bekeey, Georg V., The Structure of the Middle Ear and the Hearing

of One's Own Voice by Bone Conduction, Journal of the Acoustical

Society of America, 21, 21752, 1949.

(3) Bereskin, A. B., A Transistorized Decade Aiplifer for Low Level

Audio Frequency Applications, IRE Transistors on ,Adio, September -,

October, 1957.

(4) Black, Raymond, Ear-insert Microphone, Journal of the Acoustical

Society of Amuerica, 29, 260-264, 1957,

(5) Dolch, John P., Schubert, Earl, D., Study of Body-conducted

Sidetone State University of Iowa, Supplementary Report 1o. 6, 1954.

(6) Hershkowitz, Joseph, Levine, Leo U:., Attenuation of Ear Protectors

by Loudness Balance ahtUThresh0ld Methods, Journal of the A~dus-

tidal Society of America, 49-889-894, 1957.

(7) Eldred, Kenneth 14., Noise Characteristics of Air Force Turbojet

Aircraft, ;right Air Development Center Technical Note, 56-280.

(8) Eldred, Kenneth M., -Cannon, William T., Von Clarke, Henning,

Criteria for Short Time Exposure of Personnel to High Intensity- Tet

Aircraft Noise, Wright Air Development Center Technical liote, 55-355.

(9) Martin, Daniel W., Mapmetic Throat Microphones of High Sensitivity, iii I

Journal of the Acouutical Society of America, 19, 43-50, 1947.

THE PALDWfL PIANO COPANY

CINCINATI 2OHIO

CONTACT IldiCrOPONE fl11ESTIGivrimi-Final Engineerinkg Report on

Task 4, assigned underContract two. AF33(616)-5323

(10) Moser, 11. mI., Dreher, 3. J., Oyer, H. . The Relative Intensity

-Of Sound At Various Ainatomical Locationq On the Hend and Neck

During Phonation of Vowel Sounds,, Air ForceCambridge Renearch

Center, Technical Report 31.

(11) .1ullendore, 3. U.,. An, Eperimental Study of the'Vibration of

the Bones of the Head and Chest. During Sustained Vowel. Sounds,

'ppeech 1Monogrgpbs, 16, 163-177, 1949.

A1l2) Olson, Harry F., and flTay, Everett, Electronic Sound Abqorbe r,

Journal of the Acoustical Society of America, 25, 1130-1136, p1953.J

-77r