US2575333A

US2575333A - Noise suppression circuit

Info

Publication number: US2575333A
Application number: US24487A
Authority: US
Inventors: Toro Michael J Di
Original assignee: Federal Telecommunication Laboratories Inc
Current assignee: Federal Telecommunication Laboratories Inc
Priority date: 1948-05-01
Filing date: 1948-05-01
Publication date: 1951-11-20
Anticipated expiration: 1968-11-20

Description

NOV. 20, M. J. D! TORO NOISE SUPPRESSION CIRCUIT Filed May 1, 1948 2 SHEETS-SHEET 2 AUOIOFRfQl/E/YC'I' DISCRININATOR AMPLIFIER 10 14 15 12 @T 11 TLT 415 T4 "l/NVO/6'ED ,35 11 13 SOUNDBYPASS' X FIG 3 X is e9 FI6.2 BY

A TTORNE Y Patented Nov. 20, 1951 UNITED STATES PATENT 'OF'FIC E 2,575,353 fio 'lSEUr R SiQ out t. V MichaelJ. Di, Tore, Brooklyn, N. .assignor to Federal felecommunication Laboratoriea, 1310., New York, N. a corporation of Delaware libertarians? 1, 194s ,;saa1s ;-24,4n

17 Claims. (01. can) Thepresent invention relates-todevices for the suppression of noises accompanying a signaling wave, for example a speech wave.

In known devicesofthis" general character, the sig naleto-noise ratio is usually improved'byusing a normally blocked transmission path ior the speech wave which path becomes unblocked only whenthe inputamplitude exceeds a. predetermined level, so that only the spoken syllables but not-the inter-syllable noises willbe transmitted.

Such .a system, howeven-d'oes not rhscrimintrte against inter-syllable disturbances of high level, such as sharp impulseanor against low-amplitude disturbances coinciding with and superimposed upon the signals desired to be passed.

" Theprincipalobjectof the present invention is,

therefore, to provide a noise suppression circuit which-will be elfectiveat alltimes todiscriminate against all frequencies other than those characteristic of the particular-speech or other signalmg wave to be transmitted, whereby a large por: tion of the noisesaccompanying such wave-will be "substantially attenuated whether occurring during inter-syllableintervals or superimposed uponthe speech signals. l l

Iii-accordance with the-present invention, there is thus provided in a system of thecharacter described a wave filter having a plurality of resonance peaks at-a fundamental frequencyand harinonics thereof, said-filter being preferably controlled by a current-varying .with the pitch of the wave (if a speech wave) so as to .varythe fundamental frequency in accordance with the said pit h. 7 V. l,

The wave filter referred tomaybeanelectr-ical or mechanical resonator. preferably terminated at-bothends inra'very high impedan'ce. A suitable electrical resonator comprisesanopen-circuited artificial line whose. fundamental resonance frequency may be .varied,-.for example, by a system of pitch responsive relays arrangedto cut in-or out one Y or more sections of V the line as required. A suitable mechanical resonator co'mprises an elongated vibratory member under tension, said member being electromagnetically ex citable to oscillate at a fundamental frequency which is variable by adjusting the tension under which itis held. I

The above-mentioned and other featuresand object's' of this invention and'the mannerot-atr taining them will become moreappa-rentandthe m'venuon' itself will be better -understood by ref Rig s isYacircuit-arrangement showing an electric resonator represented by an-artifieial line-;

Figaiis a partial modification of the circuitarrange'ment of Fig. 3;- I ll Fig. 5 is a circuit arrangement showing! a me: chanical resonator represented by an electrically conductive ribbon under tension; r

Fig. 6 is a graph used to explain the operation of the arrangement of Fig. Stand Fig.7 is a further partial modification of the circuit-arrangement of Fig. 3.; a T 1 Referring to the drawing, there isshown in Fig. l the frequency spectrum of a speech wave l composed of a plurality of narrow frequency bands 2 of arying amplitudes Afthe peaks of these bands forming an envelope The-frequency bands occur at a fundamental frequencyio determining the pitch of the voice, anclat a plurality of low harmonics of said fundamental, indicated .for ex'- ample at -2fu, 5f and 10)o; The amplitudes A of the respective frequency bands 2 and,' therefore, the form of the envelope 3; varies cdnstantlyiin accordance with the sounds formulated by the speaker-;-at the same time, the pitch of the voice and, with it, the location of f0, 2ft] etc. in theirsquency band generally also varies, but at a relatively slow rate. 1 1

Let us now consider the transmission characteristics'of va line of eleetricallength l which is open-circuited at the far end (or terminated-in a very high impedance) fl heinput impedance of such a transmission line becomes- -a maximum for electrical lengths equal -to multiples of --half the wave length of th e-=incom-ing wave, thus Z= \/2, A, 3M2 2x etc. Since the wave length varies inversely with frequency, \=c/-,f (cbeing a constant), the input impedancewill .be a maxistituting thearnplifiera constant -current source. Resistor l ais -shunted by an artificial line -H comprising aplurality of sections Ha, llb flc.

lvld lle, .l lfnandiLlg. Theline H is open circuited ,atitsfar-end terminals [2, l 3-xand -i s normally brokenat -s'ections.l lb, -l l d-rand Iii, the breaks in themline: being bridgeable-shy-the armatures M, 15 and I'S:of respectiverelays ll,

cessively de-energized, .shortening the line H and raising its resonance Returning to Fig. 2, it will be seen that the shunting of load resistor H) by the input impedance Z1 of transmission line II will give rise to an output voltage'versus frequency characteristic 20 (dotted lines), E being the voltage developed across resistor [0 the magnitude of which is assumed to be large relative to the ohmic component of the characteristic impedance of line H. For an ideal line having zero attenuation (such as assumed in the graph 4 of Fig. 2) the value of IE0 at frequencies f0, 2ft), 3]0 etc. will rise to a peak 2! determined solely by the magnitude of load resistor Hi, this value being zero at odd multiples of frequency fo/Z when the input-impedance Z1 becomes negligible. In actual practice the shunting effect of line II will be more marked at the higher frequencies and the amplitudes of peaks 2| will decline as illustrated.

It will thus be seen from graph 20 in Fig. 2 that the output voltages developed across the load 10, II will be of appreciable magnitudes only in the neighborhood of resonance frequenpics f0, 21%, 3ft, etc. but will be more or less attenuated at all other frequencies. This is precisely the transmission characteristic required for the reproduction of the frequency spectrum ,of speech wave 1 (Fig. 1), provided the value of i0 is the same in both instances.

Since the fundamental of the speech wave varies, the electrical length of the line H and, thereby, the

resonance frequency in thereof should also vary. This can be accomplished by means of the relays l1, l8 and i9 which should be biased so as to .be responsive to currents of successively greater magnitude. It will be seen that relays l1, l8 and 19 in Fig. 3 are connected in series with one another and with the output of a device 22 vwave is lowered, and vice versa.

If, now, a wave of low pitch is applied to the input of amplifier and discriminator 22, a current will flow through the three relay windings connected in series which will be large enough to overcome the bias of all three relays i1, i8 and I9, attracting armatures l4, l5 and whereby-the line H will be extended to its full length; this will cause the fundamental resonancejrequency of the line to drop to the desired low value. As the pitch of the speech wave rises, relays l9, l8 and I! will become sucthereby progressively controlled by each being of course determined by the circumstances.

Fig. 4 shows a modification of that part of the circuit of Fig. 3 which is to the right of line X-X. Here the discriminator 22 is terminated a voltage divider 23 having

taps

24, 25 and 26 connected to the grids of

respective thyratrons

21, 28 and 29 energized by a source of alternating current 30. The cathodes of the '

thyratrons

21, 28 and 29 are biased by batteries 3|, 32 and 33, respectively. Relays H, is and I9, shunted by

respective condensers

34, 35 and 3B, are connected in the load circuits of

thyratrons

21, 28 and 29, respectively.

The device 22 may be an As the current through voltage divider 23 im creases, the tap 26 will first be suificiently positive to overcome the bias of battery 3! V and fire tube 21, thereby energizing the associated relay I 7. For lower pitch, again, the magnitude of the control current will rise, thus successively energizing relays l1, l8 and I9 in the order named, and again resulting in a progressive lowering of the resonance frequency To. It will be appreciated that the provision of thyratrons 2?, 28 and 29 permits a substantially instantaneous energization and de-energization of relays ll, l8 and I9, respectively, whereby the fidelity of the reproduction will be enhanced.

While in the foregoing there have been disclosed certain arrangements for varying the transmission characteristic of a wave filter by discrete stages, there will be now described a further embodiment of the invention wherein the location of the pass bands is continuously variable. Referring to Fig. 5, which also serves to illustrate the use of a mechanical in lieu of an.

electrical resonator, there is shown a permanent magnet 3'! and a metal ribbon 39 suspended for free vibratory motion transverse to the gap 39 of the magnet. Ribbon 38 is electrically connected across the input terminals 6, 1 by way of a coupling condenser 40 inserted between terminal 6 and the left-hand end 41 of the ribbon, the right-hand end 42 of. the ribbon being grounded as is the other terminal I.

The flow of audio frequency currents through the condenser 49 and ribbon 38 will set the latter in oscillatory motion transverse to thegap 39 of magnet 31, as indicated by the arrow 43. To provide a high-impedanc termination with respect to such oscillations, the end 41 is rigidly engaged by an anchorage 44 and the end 42 is held against displacement in the direction of arrow 43 by

flexible anchorages

45, 46. A contracting spring 4? applies tension to the end 42 through the intermediary of a rod 48, both the spring 47 and the rod 48 being secured to the upper extremity of the vertical arm 49 of a bell crank lever 50. Lever 59 is pivoted at 5| and carries an armature 52 at the extremity of its horizontal arm 53.

Audio frequency discriminator 22, connected as in Fig. 3 across the input terminals 6 and I, delivers a control current which energizes an electromagnet 54 attracting the armature 52 of lever 50. Thus the pull of magnet 54 opposes the force exerted by the spring 41 and tends to relax the tension in ribbon 38. "A conductor plate 55 is positioned close to the ribbon 38 adjacent the end 42 thereof, this plate being connected in series with a battery 56 and an output resistor 51 grounded at 58.

Output terminals

9, 9 are connected across the resistor 57 by way of a coupling condenser 59. It will be seen that the ribbon 38 and the plate 55 form a condenser connecting the positive terminal of battery 56 to ground at B0 and that, as the ribbon 38 vibrates, the capacitance of this condenser will vary, thereby setting up electrical oscillations across output re sistor 5i and terminals 8, 9.

The diagram of Fig. 6 shows, schematically, the ribbon 38 swinging in several modes of oscillation at a fundamental frequency having the Wave form BI and second, third and fourth harmonics indicated at 62, 63 and 64, respectively. The appearance of wave forms GI, 62, 63 and 64 are the result of excitation of the ribbon by an input wave containing the frequencies ft, 2fo, 3ft and H0, in being the frequency correspondingto the .to the right of line X-X.-

-fir tlm sle o escalatio a i dica ed a ;imped nce term tions provid the ends andAZ thereof etthe am time t :w zb e- .sirable ;to locate the exciting means 31 and the energy abstracting meansfiS rather closely to the ends of the ribbon, in order to avoid the -;nodal points of important frequencies. Hence it will be seen that the disposition of magnet 31 and condenser plate :55 within the first and the last .one-eighth of length (as indicated by lines Y.Y

and spe t ely) Will resu t n the t a mission of appreciable amplitudes of the fourth :harmonic and higher, whereas the arrangement of either ;-of said elements near the center of the ribbon would seriously reduce the amplitudes of all veven harmonics.

since an increase in the output current of discriminator 22 will act to reduce thetension in ribbon 3.3,suchan increase will lower thefundamental frequencyjo at which the ribbon goes into it sfirst mode of oscillation, indicated at 6 I. Since any change in the magnitude of the controlcurrent will be essentially continuous, the drop or rise in the resonance frequency of ribbon 38 .will also ,be gradual and may be made to correspond accurately to the variations in the pitch of the speech .wave.

pitch, in which ,case ;the position of the bell prank lever Ell should be reversed so that the pull of a magnet or solenoid acting thereon will re- ;inforce the tension ofspringe'l.

Although the provision ,of,-means for continu- ;ously varying the pass bands of a wave-filter has been describedwith reference to a mechanical resonator, it .is within the scope of the invention to provide means for continuouslyvvarying :the transmission characteristics of an electrical resonator of the general type illustrated in 'Figs. 3 and .4.or, if desired, to arrange for the stepwise variation of the fundamental frequency I of a vibrating body such as shown in Fig. 5. Continuouslyyariable electrical resonators may compri'se, for example, an artificial line similarto line H but incorporatingoneor more reactance {tubes ,of the type .well known conventional frequency modulators. 7 whichshows a further modification of that This I is illustrated in Fig.

triode, having its grid connected to an intermediate terminal of a voltage divider which con- 'sistsof a resistor 6'! and that portionoi resistor --2=3-which lies between tap Mandground, resistor 23 beir g again connected across the output of audiofrequency discriminator 22 (Fig. 3). The voltage divider referred to forms part of a differentiation circuit which further includes 0011-? dens er .68 connecting resistorfil to theplate of 'the' tube.

Hence the high Jfrequency voltage variationsappearing across the transmission line will be applied with a lead of substantially 90 degrees ,to the grid of

tubes

65 and 66, these tubes ithuslactinglas .capacitancesof a magnitude which -..will be variable.inaccordancewith the low fretquency bias derivedlfromresistor 23.

' 'COnly twoltubescareshownaby-way ofexamp'le. -.-Phe iiumber oftubesactuallyiused will -of. -;.co11rse a estates It will of course also be'possible :to let the control current decrease with falling ei ens ates-as wil th l si i the line. An obviousmodification sthe essuszf rsasiee s t es sf this s tive type in the series arms of the line, either r is list wh fi substa tial y d i $1 @indsn nqsn is ishsf in i '3 t es s sh rt a 's y a f wisounds yvhichmay include a wave filter of W isclosed hereinabove, e. g. comprising a sacsm ssisn l -1 which, ow willb l iixed electrical length. While the provision of ,s1. sbynsss wi l no generally r ul ndrs dec eas i e si na -te s ratio wil of course be possible to use ai freqjuency sensi- 2 iIlBWYQI Q zlsls k s t i by ss du in 1 21. when irs usnsi s cha cte ic :Q rtai y s s) a tran mit ed ib l vedit a fihsin e n w l arly usefulin connection withsystems for nsmission and reception of speech waves, itwil zbei u .toa ea pl sa to an kind of signalingsystem whereina periodic signal is t be-separatedfrom accompanying noises. --Such t riod svie a o u f ex ple n radar s sri tern rd ect on fin r navi a n uide n 5t 1 ;.Accordingly, lwhile the principles of the inven- ..tion hayebeen described in connectionwithcertain specifiaembodimentsitis to be clearly unclerstood that this description is madeonly by way ofexample andnot as a limitation upon the .scon mhths p esen i en on- 'f -:Wha cla med i '1. .A device forsuppressing noises accompanymg a si naling wave, c mprising. a s u c yof s nalin v wa e fil e mea h in a pl ra ty .of igesonancepealgsat a :f-undamental frequency n harmonic thereof to w ch sai i n in is applied, asource of control current... means app yin zth i l newav t i sq i t t -contro -it out u respons e t t .ir l nc spe t um 9 s id were and means ncludin said o rc fo a t n th aid fun am nta is s1u.en y torfo lpwthe hanes in? re sn n said-frequencyr ect um whi h recp n o har ic Ais ee i s i wed ,awQr netQ-ds 1 wh re n s fil e com ri res n o term n te n ighl mned nqeatbo h e d 5 -.,n' dev accordin t cl i 1, whe in sa d resonator is an electrical 1 transmission 7 line. .4. A device accordingto claim 3vwherein said -signaling-.wave;isla speech wave, further complisingcontrohmeans responsive to changes in athe pitc g fasa a pce h w e v r n th electrical length-of said transmission-linegso as to maintain the fundamental frequencies, of, said -line iand-iiofusaid.speech.wavesubstantiallyequal 5. A device according to claim 1 wherein said filter means comprises a resonator terminated in a high impedance at both ends, said resonator being an elongated resilient element under tension.

6. A device according to claim 5 wherein said signaling wave is a speech wave, further comprising control means responsive to changes in the pitch of said speech wave for varying the tension of said element so as to maintain the fundamental frequencies of said element and of said speech wave substantially equal to each other.

7. A system for transmitting a speech wave comprising a load across which said speech wave is developed and a transmission line in shunt with said load, said transmission line being substantially open-circuited at the far end thereof and being of such an electrical length that its fundamental resonance frequency substantially coincides with the resonance frequency of said speech wave, said transmission line being an artificial line composed of a plurality of sections, further comprising control means responsive to changes in the pitch of said speech wave for varying the number of sections in said line so as to maintain the fundamental frequencies of said line and of said speech wave substantially equal to each other.

8. A system according to claim 7 wherein said control means comprise a source of current and a plurality of relays respectively arranged to respond to different magnitudes of said current to connect additional ones of said sections to said line. A

9. A system according to claim 8 wherein said control means further comprise a plurality of thyratrons each having a load circuit including a respective one of said relays, voltage divider means connected across the output of said source, and means including said voltage divider means for selectively energizing said thyratrons in accordance with the magnitude of the current flowing through said voltage divider means.

10. A system for transmitting a speech wave comprising a resistor across which said speech wave is developed, an artificial line shunted across said resistor, the magnitude of 'said resistor being high relative to the ohmic component of the .characteristic impedance of said line, said line being composed of a plurality of sections and being terminated in a relatively high impedance at the far end thereof, and means for open-circuiting said line at certain of said sections for varying the electrical length of the line in such a manner that the fundamental frequency of the line corresponds substantially to the fundamental frequency of the said speech wave, said means comprising a source of control current responsive to the pitch of said speech wave and a plurality of relays selectively responsive to the magnitude of said current.

11. A system according to claim 1 wherein said filter is an artificial line composed of reactive elements, said elements including at least one reactance tube, said means for varying the fundamental frequency comprising control means responsive to changes in the pitch of said speech wave for varying the bias of said reactance tube so as to maintain the fundamental frequencies of said line and of said speech wave substantially equal to each other.

12, A system for transmitting a speech wave comprising an elongated vibratory electrical eonductor, magnetic means producing 'a flux-across said conductor, anchor means suspending said conductor for vibration transverse to said flux while opposing such vibration at both ends of said conductor, means for maintaining said conductor under tension, circuit means for setting up in said conduct-or electric currents corresponding to said speech wave, said conductor having a fundamental frequency of vibration substantially corresponding to the fundamental frequency of said speech wave, an output circuit including a conductive element positioned adjacent'said conductor to constitute said conductor and said element a capacitor, whereby electrical oscillations corresponding to the vibrations ofsaid conductor will be set up in said output circuit, and control means responsive to changes in the pitch of said speech wave for varying the tension of said conductor so as to maintain said fundamental frequency of vibration substantially equal to the fundamental frequency of said speech wave.

13. A system according to claim 12 wherein said control means comprise a source of current and electromagnetic means operable by said current to modify the action of said means for main taining said conductor under tension.

14. A system according to claim 13 wherein said last-mentioned means comprises a contracting spring, said electromagnetic means being operable to oppose the action of said spring.

15. A system according to claim 12 wherein said magnetic means and said conductive element are positioned adjacent opposite ends of said elongated conductor, respectively.

16. In a system for transmitting a signaling wave, in combination, a plurality of parallel transmission channels each comprising a wave filter having a plurality of resonance peaks at a fundamental frequency and harmonics thereof, a source of control current responsive to the frequency spectrum of said wave, means including said source for varying the fundamental frequency of at least one of said filters to follow the changes in one of the frequencies of said frequency spectrum, said one of the frequencies being'the same as, or harmonically related to, said fundamental frequency, and means for combining the outputs of said channels.

' '17. In a system for transmitting a signaling wave, in combination, a plurality of parallel transmission channels, a first wave filter in one of said channels adapted to pass frequencies corresponding to voiced sounds of said speech wave, a second wave filter in another of said channels adapted to pass frequencies corresponding to unvoiced sounds of said speech wave, a control network responsive to changes in the pitch of said speech wave, means including said control network for varying the location of the pass bands of said first wave filter to follow the changes in said pitch, and means for combining the output of said channels.

MICHAEL J. DI TORO.

REFERENCES CITED The following references are of record in the file of this patent:

UNITED STATES PATENTS Number Name Date 1,642,389 Shea Sept. 13, 1927 2,024,234 Kunze Dec. 1'1, 1935 2,033,231 Crosby Mar. 10, 1936 2,299,571 Dome Oct. 20, 1942 2,394,455 Koch Feb. 5, 1945