United States Patent [19]

Bull et al.
[54] APPARATUS FOR NON-INVASIVE

Jun. 15, 1982

Waveforms and Nasalization, Stevens et al, Journal of

MEASUREMENT AND DISPLAY

NASALIZATION IN HUMAN SPEECH

Speech and Hearing Disorder, vol. XXXVII, 3.

both of Charlottesville; Milton T.

Contingencies for Bioelectronic Modi?cation of Nasal

ity, Fletcher, Quan-Tech, Reprint from Journal of
Speech and Hearing Disorder, Aug. 1972, vol. 37, No.

Edgerton, Timbercreek, all of Va.

3.

The University of Virginia,

Chu, et al, An Electro-Acoustical Technique etc.,

Medical Research Eng, vol. 12, No. 1, pp. 18-20.

[75] Inventors: Glen L. Bull; Wesley E. McDonald,

[73] Assignee:

4,335,276

[11]
145]

Charlottesville, Va.

[21] Appl. No.: 140,951

Primary ExaminerMark E. Nushaum

Assistant ExaminerE. S. Kemeny

[22] Filed:

Attorney, Agent, or Firm-Oblon, Fisher, Spivak,

Apr. 16, 1980

[51]
[52]

Int. Cl.3 .............................................. .. G10L U110

.... .. 179/1 SP; 179/1 SC

[58]

Field of Search .............. .. 179/1 SC, 1 SP, 1 SE;

[56]

128/10, 731, 732, 635, 773; 434/319, 321

References Cited

McClelland & Maier

[57]

An apparatus for the acquisiton of a raw speech signal

and the essentially simultaneous acquisition of a trans
form of the speech signal, wherein said transform co

U.S. PATENT DOCUMENTS

2,416,353

2/1947

3,281,534

10/1966

3,383,466

varies as a function of changes in one or more parame

179/] SC

ters in the speech signal and is indicative of a predeter

mined selected speech characteristic, such as nasaliza
tion, pitch or intensity, The apparatus includes a micro

Shipman et a1. ............ .. 179/1 SP X

Dersch

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

ABSTRACT

. . . . . . . ..

5/1968 Hillix et a1. .

128/630

Brady ....... ..

179/1 SC

phone for producing ?rst signals representative of raw

3,646,576 2/1972 Griggs et al.

179/1 SC

3,752,929

179/1 SC

speech, and a second transducer, such as, for example,

an accelerometer for generating second signals essen

3,483,941 12/1969
8/1973

Fletcher .... ..

3,846,586 11/1974 Griggs

179/1 SC

3,855,416 12/1974

179/1 SP

Fuller .... ..

3,881,059

4/1975

Stewart

3,906,936

9/1975

Habal

4,015,088

3/1977 Dubnowski et a1.

4,061,041 12/1977

tially simultaneous to the production of the ?rst signals,

with the second signals being indicative of a selected

.. 179/1 SP

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

. . . ..

179/1 SC

Fletcher et a1.

4,074,069

2/1978 Tokura et al. ..

4,187,396

2/1980

parameteric characteristic of the human speech, such as,

for example, nasalization. The ?rst and second signals
are applied to data processing circuits which analyzes

128/ 724

..... .. 73/646

179/1 SC

the ?rst and second signals to produce transform signals

Luhowy ......................... .. 179/1 SC

based on arithmetic combinations thereof. The appara

OTHER PUBLICATIONS

tus further includes display means for providing video

The Effects Of Feedback Filtering On Nasalization,

Sharon R. Garber, Ph.D., Presented at Convention of
the American Speech and Hearing Association, Hous
ton, Texas, Nov. 21, 1976.
A Miniature Accelerometer For Detecting Glottal

graphic and alphanumeric display of the transform sig

nals accompanied by synchronous audio display of the
raw speech.

20 Claims, 20 Drawing Figures

F ______ _ZNZIIIEPEE~PROCES___S_IE_____II
Am 1 HP

,4

EMS is

ms

@_l_FILTER ATEIJCONV _ ?i
MIC

'
'

HP

I FILYER

u/v

22 001w

'

01v

as
_

L06

[- ___1H4;OE'?05I15L1|
r

40
T

4
a

|
1

MU) _ 3/ _ V9

52

1
I

LP 2
28 FILTER

AUX 1

~05

(34) :

AM 56

l_

A66 A10
i

___

l
{

L- ----------- ___-J

PS

,56

DISPLAY

II,

gisliqic i2
cavmaLER
1

D/A

______________ __

{_
l
e

5
'

1/0

44
CPU _

7
m i
Emmi :

58
CONTROL

I
|
L

INTERHUPT
TIMERS
\48
MICROCOMPUTER

i
|
_J'

4 US. Patent Jun. 15, 1982

Sheet 1 of 17

NORMAL

4,335,276

H YPEWASAL

QENASAL

OFNIREASODNEXLC

SUBJECTS

H611

$5

$6

$7

U.S. Pateznt

Jun. 15, 1982

Sheet 3 of 17

4,335,276

Ey

?
_

E_

__

1324

2634 MS

1; +1234

> TAKEDATA

A: _ 234

Trim
FIG. 3

US. Patent

Jun. 15, 1982

Sheet 4 of 17

4,335,276

WM

m?lI

'lllllllllllkvmh

FIG. 4

US. Patent

Jun. 15, 1982

Sheet 5 of 17

COMMAND
paocessoa
' FIG. 5A

"''-"

4,335,276

COLLATERAL
PROCESSOR

MAIN DISPLAY

DATA
A U

co ISITION

& SPEECH

PLAYBACK

GET

KEYBOARD
comma

FI6. 58
VALID

COMgfAND
PRINT ERROR
MESSAGE

TRANSFORM
TAKER

LOOK UP
COMMAND IN
JUMP TABLE
8 GO THERE

SPEECH
TAKER

U.S., Patent

Jun. 15, 1982

NA/N LOOP

4,335,276

Sheet 6 of 17
TRANSFORM

CURSOR

ROUTINE

_QL_, ggaggggq

4I
I

CURSOR

,5

LOOPS

5
5

MOVEMENT

FIG . 5D

I
I

SYNCHRONOUS

SPEECH

PLAYBACK

LOOP

INTERRUPTS

(~ MAIN LOOP )

FIG. 6A

CLEAR SPEECH
\
TURN ON TIMERS
a ENABLE
lNTERRUPTS

RECORD
FULL FLAG
,

SET
INTERRUPT
VECTORS

WAIT FOR
INTERRUPT

TRANsFORN
REAOY FLAG

sPEEcR
STORAGE RECORD

FULL?

REAL-TIME
D'SPLAY
PLOT TRANSFORM
POINT ON SCREEN

YES TuRN OFF TIMERS

a DISABLE

INTERRUPTS

EXIT TO MAIN
DISPLAY a

$P5gU?gBACK

US. Patent

Jun. 15, 1982

Sheet 7 of 17

4,335,276

INTERRIIPT
FROM SPEECH
k

TINIR

SET UP MULW
PLEXER 8 ACOUIRE
A RAW SPEECH
vALuE FROM AN)

YES

S'LENT

SILENT

M55??? FLAG

INTERVAL ~

C AR SILENT

INLTEERVAL FLAG

SET SILENT
INTERvAL
FLAG

INCRENENT
SPEECH STOR
ACE POINTER

INCRENENT SPEECH
STORACE POINTER a

STORE
RAw SPEECH

SET MEMORY

VALVE

LOCATION = PF
7

INCRENENT SPEECH
STORACE POINTER

ENABLE

MEMORY LOCATION
V

c'?i?smoi

SPEECH STORAGE
LOCATION

'

RETURN

FIG- 6B

'NTERRUPTS

wHERE INTERRuPTEO

BY SPEECH TAKER

U.S. Patent

Jun. 15, 1982

Sheet 8 of 17

4,335,276

TRANSFORM
INTERRUPT

NOTE. SYSTEM TIMING IS CRITICAL

ENABLE
INTERRUPTS

I. TRANSFORM ACQUISITION OCCURS

BETWEEN SPEECH INTERRUPTS

2. TRANSFORM CALCULATIONS,

STORAGE 8 DISPLAY MUST OCCUR

BEFORE NEXT TRANSFORM
INTERRUPT

I
WAIT FOR SPEECH INTERRUPT TO MAINTAIN SYN
CHRONY WITH SPEECH

INCREMENT
TRANSFORM

k STORE TRANSFORM

STORAGE POINTER

VALUE 0F F

SET UP MULTIPLEXER
8 ACOUIRE

TRANSFORM VALUE(S)
I

PERFORM TRANSFORM
CALCULATIONS

FIG. 6C

(IF ANY)

INCREMENT
TRANSFORM
STORAGE POINTER

STORE

TRANSFORM

VALUE

SET

* TRANSFORM-READY

FLAG

>

RETURN

Mm!

Patent

Jun. 15, 1982

Sheet 9 of 17

4,3352?

INTERRUPT FROM
RANSFORM TIMER '

WAIT FOR SPEECH

INTERRUPT TO MAIN
TAIN SYNCHRONY
WITH SPEECH
7

SET UP MULTIPLEXER
8 ACOUIRE MIC
LOGARITHMIC RMS
VALUE

FIG. 60

SET UP MULTIPLEXER
a ACOUIRE ACCL
LOGARITHMIC RMS
VALUE

MIC
RMS VALUE

RMS VALUE

< THREIPSHOLD

< THREJSHOL

CCL

AT

LgITFFELRLENOE

5ET TRANSFORM

OF MIC HACCL

VALUE = FF FF

LOGARITHMS

SET TRANSFORM
READY FLAG IN
MAIN LOOP

INCREMENT
TRANSFORM
STORAGE POINTER

STORE
TRANSFORM

RETURN

VALUE

U.S. Patent

Jun. 15, 1982

Sheet 10 of 17

4,335,276

REAL-TIME
DISPLAY
V

GET
TRANSFORM
POINTER

GET

TRANSFORM
VALUE

MORE
POINTS ON
SCREEN

CLEAR SCREEN

?
T

MAP TRANSFORM
VALUE INTO Y-AXIS

THROUGH LOOKUP TABLE

INITIAL/2E

X -AXIS COUNTER

INCREMENT X-AXIS
COUNTER a LOOK

UP COORDINATE

F I 6. 6 E

PLOT TRANSFORM
POINT ON SCREEN

CLEAR
TRANSFORM-READY
FLAG

>

U.S. Patent

Jun_:15',1982

Sheet 11 of 17

4,335,27

OUTPUT AVERAGE OF

( MAIN' Loop

W VALUES
ALL TRANSFORM
r0 SCREEN

'

-' SET

'

SET cunson

INVTEEQZLQT

HALF FLAG

1!

CALL
TRANSFORM
CURSOR LOOP

CLEAR SCREEN
V

SET UP POINTERS.

TRANSFORM
STORAGE, E TC.

"

FORMATa

DRAW PAGE
0~ SCREEN
!

PLOTSCREEN-

CL EAR SCREEN

NOTE. 7711's esfab? MANIPULM _

Iishes the pain! in

transform storage CUgglN TERS RE

of which new plot
gfRLgTVgfvE/VT
of transform
values begins and
adjusts the other
. f

ffjfegs ppm

FORMATS

FULL OF TRANS -

DRAW PAGE

FORM POINTS

0N SCREEN

1!
AVERAGE ALL VA LUES IN TRANSFORM
STORAGE RECORD

PLOT NEW SCREEN

FULL OF TRANS
FORM POINTS

60 TO
A

GO TO
A

US. Patent

Jun. 15, 1982

Sheet 12 of 17

4,335,276

TRANSFORM
CURSOR LOOP

PLOT CURSOR
AT CURRENT

POSITION

CURSOR

CURSOR

TOGGLI; HALT

roam; LEFT

A 1' ENTRY POINT TO TRANSFORM CURSOR LOOP

B = CURSOR HALT ROUTINE

C:

1'

RIGHT

"

FIG. 7B

N0

U050 Patent

Sun. 15, 1982

4,335,276

Sheet 13 of 17

TURN OFF TIIIIERS

S: DISABLE
INTERRUPTS

CURSOR
LEFT FLAG
s51 ?

RES YNCH SPEECH 8

TRANSFORM POINTERS AFTER LEI-T
MOVEMENT
y

CLEAR CURSOR -

LEFT FLAG

RESYNCH SPEECH 8
TRANSFORM POINT
ERS AFTER RIGHT/NO
MOVEMENT

'

CLEAR CURSOR

RIGHT FLAG

INITIALIZE
SPEECH-SYNCH
COUNTER; SET CUR
SOR- HALT FLAG

FIG. 7C
I

UPDATE CURSOR
POSITION DISPLAY

I IN msec)
I

UPDATE CURRENT
TRANSFORM
VALUE DISPLAY

EXIT TO
COMMAND
PROCESSOR

A = ENTRY POINT TO
TRANSFORM - CURSOR LOOP

U.S . Patent

Jun. 15, 1982

4,335,276

Sheet 14 of 17

WAIT FOR

TRANSFORM
INTERRUPT

ENABLE
INTERRIIPTG

GLEAR
CURSOR-HALT
FLAG

TURN OFF TIMERS

a DISABLE
INTERRuPTs

WAIT FOR
SPEECH
1N TERRUPT

RESYNCH SPEECH 8
TRANSFORM POINTERS AFTER RIGHT
MOVEMENT

INITIAL/Z5
SPEECH-SYNCH
COUNTER

GLEAR CURSOR-

DECREMENT
TRANSFORM

RIGHT FLAG

sToRAGE
POINTER

INITIAL/2E

SET CURS0R_

COUNTER

TURN ON TIMERS
a ENABLE
INTERRuPTs

I 60 TO

A = ENTRY

POINT T0
TURN OFF TINERG
a MA BLE

FI 6 . 7D

INTERRUPTS

TRANSFORM -

CURSOR LOOP

U.S. Patent

Jun. 15, 1982

4,335 ,276

Sheet 15 of 17
WA IT FOR

TRANSFORM
IN TERRUPT
v

ENABLE
INTERRUPTS

CLEAR

TURN OFF TIMERS

CURSOR-HALT

& DISABLE

FLAG

WAIT FOR
. SPEECH

[NTERRUPTS

INTERRUPT

RESYNCH SPEECH

I
a
TRANSFORM POINT-

,
5PC;,_$mcH

ERS AFTER LEFT

MOVEMENT

COUNTER

I T AL/ZE

INCREMENT
CLEAR CURSOR

TRANSFORM

LEFT FLAG

STORAGE
POINTER

INITIALIZE
SPEECH-SYNCH

SE7- CURS0R_
LEFT FLAG

COUNTER
I

TURN ON TIMERS
8 ENABLE

INTERRUPTS

E
A

A = ENTRY

POINT TO

FIG. 7E

TURN OF TIMERS
8 DISABLE
IN TERRUP TS

TRA NSFORiL/I >

CURSOR LOOP

US. Patent

Jun. 15, 1982

Sheet 16 of 17

SPEECH
IN TERRUPT

INCREMENT
SPEECH -SYNCH
COUNTER

FIG. 7F(iI
TOFIG. TFIII)

NEW

SILENT
INTERVAL .9

YES

COPY VALUE INTO

SILENT INTERVAL
PLA YBACK
COUNTER

DECREMENT
SILENT INTERVAL

PLAYBACK
COUNTER

OUTPUT RAW
SPEECH VALUE
TO DA C
I

INCREMENT
SPEECH STORAGE
POINTER

ENABLE
INTERRUPTS

4,335,276

U.S. Patent

Jun. 115, 1982

Sheet 17 of 17

4,335,276

FROM
COPY VALUE INTO
SILENT INTERVAL
PLAYBACK

COUNTER
T

DECREMENT
SILENT INTERVAL
PLAYBACK

COUNTER

DECREMEN T
SPEECH STORAGE
POINTER
V

ENABLE

INTERRUPTF

RETURN

FIG. 7F (ii)

F16. 7F ( i 1

4,335,276

APPARATUS FOR NON-INVASIVE

MEASUREMENT AND DESPLAY NASALIZATION
IN HUMAN SPEECH
BACKGROUND OF THE INVENTION
l. Field of the Invention
This invention relates to an apparatus for the non

ing nasal vibration during speech to obtain a quantita

tive measure related to nasality. Stephens et al ?lters,
recti?es, and time averages the output of the accelerom
eter. Then, with the aid of a computer, the smooth
signal is'sampled, log converted and displayed on an
oscilloscope to provide a visual display of nasalization.
In a related development, Garber et al, The Effects
of Feedback Filtering on Nasalization in Normal and

invasive detection and treatment of speech disorders,

Hypernasal Speakers, J. Speech Hearing Res. 22

especially disorders effecting speech nasalization, and

(1979), 321333, in order to investigate the effect of

auditory feedback on vocal production and nasalization
in particular, tested the effects on the nasalization of
various subjects who listened to their speech filtered at
various frequencies. Thus, Garber et al have investi

more particularly to such an apparatus for generation of

quantitative predictive information related to underly

ing physiological and perceptual correlates of nasal
resonance.

15 gated whether production of nasal quality would

2. Description of the Prior Art
Early efforts at diagnosis and treatment'of disorders
change when subjects hear their voices ?ltered. In im
of nasal resonance have been based on perceptual as
plementing their study, Garber et al used the output of
sessments of the patients speech by the clinician. This
an accelerometer of the type employed by Stephens et
approach has suffered for several reasons. Consistency
al placed on the nose to obtain a measure of nasalization.
of judgments among clinicians, dependent upon exten
The output of the accelerometer was first routed to a
sive clinical training, is often lacking. The subjective
tape recorder. The recorded signal was later transferred
judgment is an assessment of the overall quality of the
to a graphic level recorder and analyzed through mea
patients speech, and therefore de?nition of speci?c
surement of peaks in the signal with respect to a pre
attributes which give rise to the problem is poor. Feed
recorded calibration tone. The arithmetic average of
back to the patient is delayed rather than immediate. 25 measured peaks constituted the nasalization score.

Therefore, recent efforts have focused on development

of methods which provide consistent, repeatable results

with greater immediacy, and greater speci?city with

To validate the measurement, a preliminary study

was conducted in which subjects were requested to

speak at various intensity levels. The Pearson product

momentcorrelation between accelerometer output and
perceptual judgments of nasality was 0.77. A correction
process and apparatus in which electrical signals repre
factor was then introduced to compensate for intensity
sentative of the sounds- emitted from the nose and
differences between the various conditions by subtract
mouth are utilized to determine the degree of nasalance
ing each subjects vocal level from an arbitrary refer
of speech. In this apparatus, a pair of sound-isolated
microphones are carried in the housing adapted to be 35 ence level, dividing this value by two, and adding it to
the subjects nasalization score. After adjustment of
brought into place about the face of the patient in order
scores in this manner, the correlation reported between
to respectively measure sounds emanating from the
accelerometer output and perceived nasality was 0.82.
nasal and oral cavities. The outputs of the microphones
In this manner it was determined that the nasalization
are ?ltered for respective frequency bands thought to
have high nasal and oral content, and a ratio of the 40 score accounted for 67% of the variance in nasality,
provided that intensity level was held constant. An
?ltered microphone outputs computed to obtain a quo
attempt was made to hold the intensity level constant in
tient signal which is then threshold detected against a
respect to de?nition of the problem.

In US. Pat. No. 3,752,929 to Fletcher is described a

reference representing a known degree of nasality.

the main study described in the preceding paragraph by

requesting subjects to speak at a constant vocal effort. A

a visual display such as a lamp by which the patient can 45 visual display of vocal intensity was provided to facili
tate maintenance of constant vocal effort.
determine whether or not a given sentence contains

Then the output of the threshold detector is applied to

The measurement technique developed by Garber et

more or less nasalance relative to the reference estab
a1 lacks instantaneous quanti?cation and therefore lacks
lished by the threshold detector.
the immediate feedback necessary for efficient immedi
The approach outlined in the Fletcher patent, which
represented a major advance in providing a practical 50 ate modi?cation of speech production. In the form im
quantitative measure of disorders of nasal resonance,

plemented, the technique also requires that subjects

nevertheless requires that the patient place his face in

maintain constant vocal effort to maximize accuracy of

the mask which provides acoustic isolation between the

the measure.

microphones and thereby permits separation of the oral

SUMMARY OF THE INVENTION

and nasal acoustic signals. Unfortunately, the use of the 55
facial mask requires that the patient place his head in a
Accordingly, it is an object of the present invention
stable position, and further limits interaction between
to provide a novel apparatus for non-invasive measure
the patient and the clinician. This may present severe
ment and display of nasalization in human speech which
difficulties with young children or paralyzed patients
provides immediate feedback by which a patient can
who comprise a large percentage of the population seen 60 monitor, evaluate and modify his speech for nasaliza
in the clinic for defective velopharynseal valving. Fur
tion.
thermore, the degree of separation and acoustic isola
Another object is to provide a novel apparatus which
tion between the microphones has been questioned.
can provide feedback facilitating second language
An alternate approach devised by Stephens et al, A
learning in instances in which the set of nasal phonemes
Miniature Accelerometer for Detecting Glottal Wave
in the second language differs from those of the speak

forms and NasalizationfJ. Speech Hearing Res. 18

(1975), 594-599, utilized a light-weight accelerometer
attached to the external surface of the nose for measur

ers native language.

Another object is to provide a novel apparatus of the
type noted above capable of deriving a measure which

4,335,276

samples from the two RMS channels. Thus a new ratio

is formed every two milliseconds. The measure ac

provides predictive information with respect to related

physiological events and perceptual correlates of nasal

quired, therefore, consists of a ratio of vibration at the

resonance.

novel apparatus which provides diagnostic information

nasal wall transduced by an accelerometer over the

combined oral and nasal acoustic outputs transduced by

about the relative severity of disorders of nasal reso

nance and sorts patients into diagnostic categories based

rithm of each ratio acquired is formed to facilitate rec

Yet another object of this invention is to provide a

the microphone. In this mode of operation, the loga

on the range of the measures obtained for productions

ognition of patterns present in a graphic display of the

of nasal and non-nasal phonemes.

Yet another object of this invention isto provide a
novel apparatus which permits identi?cation of the

ratios.
A ratio of the two linear RMS signals is formed by
means of a divider circuit in hardware. The digital pro

phonemic content of speech associated with speci?c

cessor then acquires the signal formed by the output of

sections of a static graphic display of the measure of

the divider circuit at a 500 Hz rate as the raw speech

signal is sampled at an eight KHZ rate. Selection of a

Yet another object of this invention is to provide a 15 linear or logarithmic ratio is controlled through com

nasalization over time.

novel apparatus which permits identi?cation of the rate

mands input through the command keyboard on the

and slope of the transition from a nasal to a non-nasal

control panel.

phoneme.

The ratios over time are plotted as a line on a display.

ingful basis for comparisons among patients as well as a

An upward or downward shift represents a proportion

ately greater or lesser degree of nasalization. The arith
metic average of all ratios formed for the utterance
recorded is displayed in the lower right-hand corner of

basis for recording progress within a given patient with

the screen.

Yet another object of this invention is to provide a

portable, easily implemented apparatus which provides

consistent, repeatable measures which provide a mean
a disorder of nasal resonance.

20

The digitized signal from the raw speech channel is

Another object of this invention is to provide a novel 25 stored concurrently with the ratios formed from the
sampled RMS channels in such a manner that the rela
apparatus which permits identi?cation of the phonemic
tive relationship in time between the ratios and the
content of speech associated with speci?c sections of
digitized audio signal is preserved. A moving cursor can
static graphic displays of measures of other transforms
be advanced across the graphic plot synchronously
of speech such as intensity and pitch over time.
These and other objects are achieved according to 30 with the replayed audio signal, permitting identi?cation
of the phonemic content associated with a given seg
the invention by providing a new and improved appara
ment of the plot. This is accomplished by means of a
tus for non-invasive measurement and display of nasal
toggle with three positions: cursor right, cursor left, and
ization in human speech including the following sec
halt. A binary code corresponding to each position of
tions: two transducers (an accelerometer and a direc
tional microphone), an analog proprocessing section, an 35 the toggle is sent to the digital controller, which directs
the movement of the cursor accordingly.
analog-to-digital converter, a digital data processor, a
When the cursor is halted, the instantaneous value of
display section, and a control panel.
the ratio and the time in milliseconds at that point in the
The accelerometer is mounted on the external nasal
utterance are displayed in the lower and upper right
wall for measurement of nasal wall vibration, while
hand corners of the screen respectively. Thus the abso
airborne sound consisting of combined nasal and oral
lute value of a ratio formed at a speci?c vtime in the
output is transduced by the directional microphone.
The microphone is mounted on a headset to maintain a

utterance can be determined, as well as the arithmetic

constant position with respect to the subjects lips. In

the analog preprocessing section the accelerometer and
microphone outputs are ampli?ed, RMS averaged and

average of all ratios formed for the entire utterance.

transferred to a multiplexer in the analog-to-digital con

version section. A 30 Hz highpass ?lter with a 12 dB per
octave slope on the output of the accelerometer can be
enabled to compensate for artifacts associated with

turning and other movements of the head which would

otherwise be recorded by the accelerometer. The am

pli?ed output of the raw speech signal is also trans

ferred to the multiplexer to provide a record of the

speech associated with time-varying ratios formed from

Digitization of the audio signal from the raw speech

45 channel at an eight KHz rate requires one byte of mem

ory for each digitized sampled stored. Thus direct stor

age of a signal sampled at an eight kHz rate for one

second would require 8000 bytes of memory. The eight

bit microprocessor utilizes a 16-bit address bus, permit
ting a maximum of 64 kilobytes of memory to be ad
dressed, placing the upper limit on the duration of the
speech signal which can be stored. To conserve mem
ory and extend the maximum length of utterances
which can be recorded, the duration of silent intervals

the two RMS signals. An AGC circuit on the output of 55 in perceptually continuous speech is coded, rather than
storing each sample with a value of zero as a separate
the raw speech channel can be enabled to improve the
byte. During playback of the digitized speech signal, a
?delity of transient consonants such as voiceless /th/
series of zeros is then sent to the digital-to-analog con
which have an inherently low relative intensity level.
verter for the duration coded at that point in the stored
The two RMS signals are provided in two forms:

linear and logarithmic. The two logarithmic RMS sig

signal. This results in an appreciable savings in memory

nals are sampled at a 500 Hz rate by vthe analog-to-digi

tal converter as the raw speech signal is sampled at an
either kHz rate. The digital processor, which utilizes an

required for storage of the digitized speech signal.

eight-bit microcomputer of the 8080/8085/280 family

consonants such as /P/ creates a slight but rapid move

Contraction of the musculature associated with lip

movement which accompanies production of labial

of microprocessors, controls the multiplexing and ana 65 ment of the nasal wall in some individuals. When the
speed of this movement exceeds the frequency of the 30
log-to-digital conversion of the respective signals. The
Hz ?lter on the output of the accelerometer, an artifact
digital processor forms a ratio of accelerometer output
consisting of a sharp spurious peak in the graphic dis
over microphone output for each successive pair of