MUSIC 318 MINI-COURSE ON SPEECH AND SINGING

NONLINEAR SOURCE-
FILTER COUPLING IN
SPEECH AND SINGING

“Modeling source-filter interaction in belting and high-pitched operatic male

singing” (I. Titze and A. Worley, JASA 126, 1530 (2009))
“Source-Filter Interaction in Speaking and Singing is Nonlinear” (I.Titze, ECHOES
Summer 2007)_
“The Human Instrument” (I.Titze, Scientific American Jan. 2008)
 NONLINEAR vs LINEAR MODELS OF PHONATION

THE LINEAR MODEL OF SPEECH AND SINGING ASSUMES THAT THE SOURCE (GLOTTIS),
FILTER (VOCAL TRACT) AND RADIATOR (MOUTH, NOSE) ACT INDEPENDENTLY.

THE LINEAR MODEL SUCCESSFULLY EXPLAINS MANY ASPECTS OF SPEECH AND

SINGING. AS LONG AS THE DOMINANT SOURCE FREQUENCIES ARE WELL BELOW THE
FORMANT FREQUENCIES OF THE VOCAL TRACT (GENERALLY TRUE IN MALE SPEECH),
THE SOURCE IS INFLUENCED ONLY SLIGHTLY BY THE FILTER. IN FEMALE AND CHILD
SPEECH, HOWEVER, THE INTERACTION IS GREATER.

RECENT RESEARCH SUGGESTS THAT HUMANS CAN OPERATE THEIR SOURCE-FILTER

SYSTEMS WITH EITHER LINEAR OR NONLINEAR COUPLING. FOR LINEAR COUPLING THE
SOURCE IMPEDANCE (TRANSGLOTTAL PRESSURE DIVIDED BY GLOTTAL FLOW) IS KEPT
MUCH HIGHER THAN THE INPUT IMPEDANCE TO THE VOCAL TRACT. THIS IS
ACCOMPLISHED BY ADDUCTING THE VOCAL FOLDS FIRMLY AND WIDENING THE
EPILARYNX TUBE SO THAT GLOTTAL FLOW IS DETERMINED BY AERODYNAMICS, AND
ACOUSTIC PRESSURES ABOVE AND BELOW THE GLOTIS HAVE LITTLE INFLUENCE.
NONLINEAR SOURCE-FILTER COUPLING IN PHONATION

IN NONLINEAR COUPLING THE ACOUSTIC AIRWAY PRESSURES CONTRIBUTE TO THE

PRODUCTION OF FREQUENCIES AT THE SOURCE. TRANSGLOTTAL PRESSURE INCLUDES A
STRONG ACOUSTIC COMPONENT, MUCH AS IN WIND INSTRUMENTS.

FOR NONLINEAR COUPLING, THE GLOTTAL IMPEDANCE IS ADJUSTED TO BE

COMPARABLE TO THE VOCAL TRACT INPUT IMPEDNACE, MAKING THE GLOTTAL FLOW
HIGHLY DEPENDENT ON ACOUSTIC PRESSURES IN THE VOCAL TRACT. THIS IS
ACCOMPLISHED BY SETTING ADDUCTION LEVELS OF THE VOCAL FOLDS THAT MATCH A
NARROWER EPILARYNX TUBE.

EVIDENCE OF NONLINEAR COUPLING IS THE PRODUCTION OF NEW FREQUENCIES IN

THE FORM OF DISTORTION PRODUCTS, LOWERING OF THE OSCILLATION THRESHOLD
PRESSURE, PRODUCTION OF SUBHARMONICS, AND SUDDEN JUMPS AS EITHER VOWEL
OR F0 ARE CHANGED.
 POSITIVE FEEDBACK TO VOCAL FOLDS

POSITIVE FEEDBACK FROM THE VOCAL TRACT TO THE VOCAL FOLDS CAN INCREASE SOUND
PRODUCTION. (THIS IS SIMILAR TO THE POSITIVE FEEDBACK FROM A BRASS INSTRUMENT
TO THE PLAYER’S LIPS). THE IDEAL TIMING OF THE “KICK” COMES WHEN THE MOVEMENT
OF THE AIR IS DELAYED WITH RESPECT TO THE MOVEMENT OF THE VOCAL FOLDS. THE AIR
COLUMN THEN HAS NEGATIVE INERTANCE. INERTIVE REACTANCE HELPS TO SUSTAIN THE
FLOW-INDUCED OSCILLATION OF THE VOCAL FOLDS. (See Titze, 2008).

THE SINGER’S TASK IS TO ADJUST THE SHAPE OF THE VOCAL TRACT (BY CAREFULLY
SELECTING FAVORABLE “SINGING” VOWELS) SO THAT INERTIVE REACTANCE IS
EXPERIENCED OVER MOST OF THE PITCH RANGE—NO EASY TASK.
 SOURCE-FILTER INTERACTION IN BELTING AND
OPERATIC SINGING

BELTERS USE VOCAL TRACT RESONANCES (FORMANTS) DIFFERENTLY FROM

CLASSICALLY-TRAINED (OPERA AND ART SONG) SINGERS. THE SECOND HARMONIC,
FOR EXAMPLE, RECEIVES STRONG REINFORCEMENT FROM THE FIRST FORMANT IN
BELTING. (COMPARE THIS TO PAVAROTTI WHO USES THE SECOND FORMANT TO
REINFORCE THE THIRD HARMONIC IN OPERA SINGING).

TITZE AND WORLEY (2009) OBSERVE THAT THIS FORMANT TUNING CAN BE AIDED BY
UTILIZING SUPRAGLOTTAL INERTIVE REACTANCE TO REINFORCE VOCAL FOLD
VIBRATION BY CHOOSING PITCH-VOWEL COMBINATIONS THAT KEEP SEVERAL
HARMONICS IN FAVORABLE REACTANCE SIMULTANEOUSLY.
D5 SUNG BY BARBRA STREISAND ON THE WORD “STRONG” SHOWING A STRONG 2 ND
HARMONIC (Fig. 11.1 in Miller)
 SOURCE-FILTER INTERACTION

DIAGRAM OF VOCAL FOLDS

AND LOWER VOCAL TRACT
 SOURCE-FILTER INTERACTION IN BELTING AND
OPERATIC SINGING
BELTERS USE VOWELS THAT ARE MODIFIED TOWARD A “MEGAPHONE” MOUTH SHAPE.
BOTH THE FUNDAMENTAL AND THE SECOND HARMONIC ARE THEN KEPT BELOW THE
FIRST FORMANT.

OPERA SINGERS, ON THE

OTHER HAND, USE VOWELS
MODIFIED TOWARD AN
INVERTED MEGAPHONE
MOUTH SHAPE FOR
TRANSITIONING INTO THE
HIGH-PITCH RANGE. THIS
ALLOWS ALL THE HARMONICS
EXCEPT THE FUNDAMENTAL
TO BE “LIFTED” OVER THE
FIRST FORMANT.
 MOUTH AREA OF SINGERS

MOUTH AREA AND SOUND

SPECTRUM FOR PAVOROTTI
SINGING /α/ VOWEL AT A4

MOUTH AREA AND SOUND

SPECTRUM FOR CAB
CALLOWAY SINGING /α/
VOWEL AT A4

(Titze and Worley 2009)

 ACOUSTIC IMPEDANCE

ACOUSTIC IMPEDANCE IS SOUND PRESSURE DIVIDED BY FLOW VELOCITY

ACOUSTIC ADMITTANCE (1/IMPEDANCE) IS FLOW VELOCITY DIVIDED BY SOUND

PRESSURE. CONDUCTANCE IS THE IN-PHASE (REACTIVE) PART OF ADMITTANCE;
INERTANCE IS THE OUT-OF-PHASE PART
 VOCAL TRACT INPUT IMPEDANCE

VOCAL TRACT CARICATURES (left) AND CORRESPONDING INPUT IMPEDANCES (right) AS A

FUNCTION OF FREQUENCY (THICK LINES ARE REACTANCES AND THIN LINES ARE
RESISTANCES) (Titze and Worley, 2009)
 COMPUTER SIMULATION OF GLOTTAL AIRFLOW

COMPUTER SIMULATION OF GLOTTAL AIRFLOW WITH A SELF-SUSTAINED OSCILLATION

VOCAL-FOLD MODEL THAT INTERACTS WITH: A UNIFORM TUBE OF DIFFERENT AREAS (left)
AND A NEUTRAL TUBE WITH DIFFERENT EPILARYNX AREAS (right) (Titze and Worley, 2009)
 INERTANCE

(INERTANCE IS THE REACTIVE

PART OF ADMITTANCE; IT
INDICATES ENERGY STORAGE
DURING OSCILLATION)

INERTOGRAMS (INERTANCE vs FREQUENCY FOR SIX TUBE SHAPES (Titze and Worley, 2009)
 INERTOGRAMS OF SINGERS

VOCAL TRACT SHAPES DERIVED FROM MRI DATA OF A BARITONE SINGER (left) AND
CORRESPONDING INERTOGRAMS (Titze and Worley, 2009)
 MOUTH-TO-HEAD AREA RATIOS
Table I. Mouth-to-head area ratios.

Note Ratio Vowel

Male operatic (Luciano Pavarotti)

D #4
         0.0137 /e/

E4 0.0205 //

F#4 0.0288 //

G4 0.0290 //

A4 0.0291 //

Male belt (Cab Calloway)

D 0.0170 /u/

E4 0.0364 /o/

F 0.0614 /a/-/o/ (diphthong)

G4 0.0662 /a/

A4 0.0840 /a/