US3074025A - Automatic amplitude control of speech sound waves - Google Patents

Description

Jan. 15, 1963 M. v. KALFAIAN AUTOMATIC AMPLITUDE CONTROL 0F SPEECH SOUND WAVES Filed Nov. 1o, 195s 2 Sheets-Sheet 1 www INVENToR.

4A fw Jan. 15, 1963 M. v. KALFAIAN r3,074,025

AUTOMATIC AMPLITUDE CONTROL oF SPEECH souNn wAvEs Filed Nov. 1o, 1958 2 sheets-sheet 2 ORM/NAL M o AHPL/Tuos EQuA L/zED con/LEX vvE PULSES AT MAJOR PE'AKS o I 60A/TRW` wAvf. "m k' AT cAPAc/Ton cl2 V AIN CONTRlLBLE T u AMPLIFIER OUTPUT PEAK 5 9 Bus /NcoM/Ma 8 COMPLEX 2 wA v5.5 cA/N Rsnuc//va u/v/DlREeT/o/AL PfEwBAcK .STORAGE HAOR PEAK r PULSE T r D/SC//ARGER .ssLfcToR Pannen INV EN TOR.

United States Patent O 3,074,025 AUTOMATIC AMPLiTUDE CONTRGL F SPEECH SUND WAVES Meguer V. Kalfaian, 962. Hyperion Ave., Los Angeies 29, Calif. Filed Nov. 10, 1958, Ser. No. 773,065 1 Claim. (Cl. S30-131) This invention relates to automatic amplitude control systems, and particularly to a system for normalizing the amplitude of propagated speech sound waves to a constantpeak amplitude level. Its main object is to transform an incoming speech sound wave having wide variation in amplitude into an outgoing speech sound wave having a uniform amplitude level, without loss of intelligibility and with minimum sacrifice in quality.

In various forms of speech sound wave production and reproduction, it is often desired that the wide variations of amplitude in speech sound waves, as originated in the vocal system, are synthetically reduced to a standard peak amplitude. Such standardization of peak amplitude will eect loss of voice quality, but the intelligibility will be preserved, and accordingly, the said modified speech sound waves may lbe utilized for transmission through various forms of transmission links where the signal-tonoise ratio is very low. In ordinary transmission through such low signal-to-noise ratio link, the transmission of the phonetic sound may be received intelligibly, because of its inherent high amplitude, but the phonetic sounds, such as. v or s, may be completely lost in noise, because of their inherent weaknesses in amplitude. In another form of speech sound wave reproduction, peakamplitude standardization is desirable for the analysis of amplitude ratios between various resonances in different phonetic sounds of the speech, for example, in the system as disclosed in my patent application, Serial Number 723,510, filed March 24, 1958, now Patent No. 2,921,133. Ordinary methods and systems of amplitude control, as practiced in the art of electronics, however, are not adequate for theparticular purpose contemplated herein, as they are not capable of following the sudden changes in amplitude Ifrom one phonetic sound to another. As each phonetic sound comprises repetitious wave trains, and the peak amplitude of these wave trains varies widely, both by reason of the inherent characteristics of the vocal system, and by environmental changes, the object of the present invention is, accordingly, to provide an amplitude control system which shifts the peak-amplitude of the propagated speech sound waves to a standard level in stepwise shifts train after train. In its broader aspects, the fundamental frequencies of the propagated speech sound waves are iirst selected; narrow pulses are produced at these fundamental frequencies; and these narrow pulses are used to drive the amplitude-control system. The nature of such a system will be apparent from the following specification, and in connection with the accompanying schematic arrangements, wherein, FIG. 1 is a schematic diagram of the system embodying the invention; FIG. 2 is a series of waveforms illustrating the operation of the invention; and FIG. 3 is a block diagram illustrating the principles of the invention.

In ordinary speech, each phonetic sound consists of a train of substantially replica waves, formed by fairly regular puffs of air from the glottis, which are set into vibrations in the momentarily formed resonant cavities of the vocal system. As each puff of air enters these cavities, an initial surge of pressure is formed, and accordingly, these vibrations are commenced by a high peaked wave; the series of occurrences ofwhich determine the fundamental (pitch) frequencies of the speech sound wave at any given period of time. It is then possible 3,074,025 Patented Jan. 15, 1963 to select these high peaked Waves, or, major peaks, as named in the present disclosure, and produce at these peak periods marker pulses. A detailed description on systems for selecting these major peaks, and producing marker pulses thereat, may be referred to a disclosure in my U.S. Patent No. 2,673,893, March 30, 1954. A section of several Wave trains (just mentioned) is illustrated at A in FiG. 2. The major peaks of these wave trains are indicated by the numerals 1 through 3 in the drawing. It will be noted that the amplitudes of wave trains a and b may be different than the amplitude of wave c. 'Ihe purpose herein contemplated is, accordingly, to change the amplitudes of these wave trains, in a manner that, the major peaks have constant reference amplitudes, without distorting the original wave form; thus keeping distortion in quality of the original speech sound wave at a minimum. This is accomplished by passing the original sound wave through a gain controllable amplifier, the gain of which is shifted stepwise at said major peaks, such as shown by the square waves at D of FIG. 2. These gain shifting points are abrupt, but in practice, a lag of substantially short pulse period is suflicient to obviate any distortion effects upon the gain-controlled sound wave. Such a short pulse is shown at C of the drawing in FIG. 2, which causes pulse waves 4 to appear upon the gain-controlled trains of waves at B. Due to the high frequency characteristics of these pulses, however, they can be easily eliminated in an outgoing circuit having a band pass width of only covering the spectrum band of the original speech sound waves, in a conventional mode.

The system actually used is described by Way of a block diagram, as shown in FIG. 3, wherein, the original speech sound Wave is produced in block 5. The output of block 5 is applied to the input of a gain controllable amplifier l6, and a major peak selector block 7. The output of the gain controllable amplifier block 6 is fed back to its input in degenerative mode by way of a unidirectional storage device, for example, a rectifier and a capacitor connected in series, in block 8. This storage device is reference biased by way of block 9, in such a manner that, the storage device 8 is energized only when the magnitude of amplified wave of block 6 exceeds in magnitude of the reference bias in block 9. Thus, the peak amplitude of amplified wave of block 6 is substantially equal to the reference magnitude of bias in block 9. The storage device in block 8 is normally non-dischargeable, and ac-` cordingly, after it has been charged to a controlling value, it holds the gain of amplifier `block 6-insteady state. A normally inoperative discharger element in block 10 is provided. The major peak selector in block 7 produces a pulse signal in block 11, at the arrival of a major peak, and this pulse signal is applied to the discharger in block 10, for momentarily discharging the storage device in block 8, so that it may recharge to a new level for the control of amplifier block 6. Thus the input of amplifier -in block 6 sees the original wave at Ain FIG. 2, and thel output of said amplifier is transformed into a wave having uniform amplitude of major peaks, as shown at B in FIG. 2.

Having described the broader aspects of the present invention, detailed specification is now made by way of the schematic arrangement in FIG. 1, which consists of ve dual control tubes V1 to V5, and four phaseA inverter tube V6 to V9. The dual control tubes V1 to V5 are of the type having two separate control grids, either one of which is capable of controlling the grain of the tube. Thus, the speech sound waves are applied to one of the control grids, for example, the rst grid G1, of dual control tube V1. The control grid G1 of V1 is normally biased through load resistor R1 (as indicated in the drawing) to a conductive point of the tube,

so that the voltage variations of the speech sound waves appearing at grid G1 are transmitted faithfully to the anode circuit resistor R2. These corresponding voltage variations across R2 are produced in reversed polarity with respect to the voltage variations appearing at control grid G1. Accordingly, the voltage variations across R2 are coupled to the control grid of phase inverter tube V6, through coupling capacitor C1, so that these latter voltage variations are now transmittedY to the anode circuit resistor R3 of V2in phase .polarity corresponding to'the phase polarity of the voltage variations appearing at the control grid G1 of the dual control tube V1. This process is repeated through several stages comprising dual Acontrol tubes V2 to VS, and intermediate phase inverter stages comprising tubes V6 to V9, so that the first control grids G1 of dual control tubes V1 to VS receive Athe voltage variations of the speech sound wave in the same polarity. The rst control grids G1 of dual controly tubes V2 to V5 are biased through load resistors R4 to R7, respectively; the control grids of phase inverter tubes V6 to V9 are biased through load resistors R8 to R11, respectively; the anode circuit resistors of dual control tubes V2 to V5 are designated as R12 to R15, respectively; the anode circuit resistors of phase inverter tubes V7 to V9 are designated as R16 to R18, respectively; and the coupling capacitors from stage to stage of alternate dual control and phase inverter tubes V1 to V9 are designated as C1 to C8.

Continuing with the function of dual control tube V1 and the .phase inverter -tube V6, it was stated that each time the voltage variations of the speech sound wave are applied upon the first'control grid of a dual control tube, forexample, V1, vthese voltage variations are transmitted to the anode circuit of that tube in phase inverted polarity. Also, by applying these phase inverted voltage variationsto the control grid of an intermediate stage, for cxample, tube V6, the original polarity of said voltage variations is reinstated in the anode circuit resistor, for example, resistor R3. Thus by the use of intermediate phase inverter stages, the first control grids G1 yof the dual control tubes V1 to V5 see the voltage variations of the speech sound waves inthe same polarity.

The `voltage variations of the' specchsound waves in the anode -circuit resistor `R of the last dual control tube V5 are coupled to the vcontrol grid of cathode kfollower tube V10, through coupling capacitor C9. The voltage variations developed Vacross load resistor R21 are charged in` storage capacitor C12 through series connected diode D1. 'Ihe closed circuit loop of storage capacitor C12; the diode D1; and load resistor R21, however, is made in series with a reference bias voltage ofrsource B1, vthe magnitude ofthe later of which determines the peak-.amplitude ofthe speech sound vwaves to be standardized. The diode D1 is so polarized that only signals of negative potential developed across load resistor R21,is transmitted to becharged in storage capacitor C12. Also, the reference bias potential of B1 is so. polarized that, the negative signal potential developed in the load resistor R21 does not pass through the diode D1 until the amplitude of said negative potential is higher than theA potential of B1. Thus, the storage in capacitor C12 remains constant until a negative potential of higher amplitude than the, potential of B1 is developed in the load resistor R21, at which point the capacitor C12 becomes negatively charged and reducesjsimultaneously the gain of dual control tubes V1 to V5; by virtue of,V

capacitor C12 remains constant4 thereafter, determinipg the standard amplitude level of the speech sound waves at a suitable output terminal, for example, taken from across the anode circuit resistor R15 of dual control tube V5.

In accordance with the characteristic nature of the fundamental frequencies of speech sound waves, as de scribed in the foregoing by way of the graphic illustration in FIG. 2, and the block diagram in FIG. 3, it is then only necessary to lproduce narrow pulses at the major peaks (fundamental frequencies) of the propagated speech sound waves, and discharge the storage capacitor C12 during these narrow pulse periods, so that the gains of dual control tubes V1 to V5 may be shifted stepwise from major peak to major peak. The discharge of storage capacitor C12 is achieved by the discharger tube V11, which is connected electrically in parallel with the capacitor C12. This latter tube, however, is normally rendered non-conductive by a high negative bias upon its control grid, as indicated in the drawing. The marker pulses at fundamental frequencies are applied upon the control grid of discharger tube V11 in positive polarity, so that the tube V11 becomes conductive at these short pulse time periods, discharging the capacitor from its previous storage. This discharge, however, is not complete, as during the discharge period the voltage across the load resistor R21 is at a new amplitude level, and the voltage across capacitor C12 assumes this amplitude level immediately after the discharging pulse has ceased to zero value. Thus, the gains of all dual control tubes V1 to VS are varied in steady state steps at the repetition rate of'said fundamental frequencies, and the amplitude of the original speech sound waves is shifted to a standard level, where all said major peaks have the same height.

Referring again to the modulator tubes V1 to V5, it is essential that during gain shiftingperiod the first control grids G1 of these tubes are impressed upon by signal voltages of llike polarity, so that the gain of one tube will' notoset the gain of another tube. For this reason,

the phase inverter stages comprising tubes V6 to V9` are included intermediate to the dual control tubes. In reference to the phase inverter and cathode follower tube V10, this latter tube may be eliminated in conjunction with the component yparts comprising resistors R19, R23 and C10, byA connecting the coupling capacitor C9 from the anode circuit of dualcontrol tube V5 to the resistor R21. In this case; the function of the larrangement wouldfbe the same as described previously,e.g., the gains of dual control tubesy V1 to V5 being shifted during said narrow pulses produced at said major peaks.

`It will be4 noted in the arrangementgof FIG. l, that, a single dual control 'tube will su'ice for the purpose cou-A templatedherein. The reason that tive dual control tubes (V1-V5) are used in the arrangement, to perform the function of a single dual control tube, is `to reduce the burden imposed on each tube, and thereby reduce non-- linear distortion ofthe original sound wave. Since non-- linear parts will inherently distort the admitted vwave-- form, regardless how small it may be, the phase inverter tubes V6 to V94 are chosen of the pentode type, so that` of the distortion present.

of dual control stages may be used, according to the: particular use and requirements involved.

When V`the controlled amplitude of the speech sound wave is to be made at a highly amplified level, then the] signal voltage Ideveloped across plate circuit resistor R15 of dualcontrol tube VSfmay `be first amplified across an impedance means, and a fractional tap taken therefrom to becoupled to the control grid of tube V10. The first control grids G1 of dual control tubes V1 to V5, and the control grids vof phase inverter tubes V6 to V9 are shown, connected in parallel to a-common bias source. Such connection is shownfor convenience, however, and

the biases on vdifferent tubes may vary, if so desired,

Without aiecting the function of the arrangement. The circuit arrangement in FIG. 1 is not limited to vacuum tubes, as transistors may also be utilized to obtain similar performance. The tetrode transistors, available commercially, are examples in substitution of said dual control tubes.

What I claim is:

In complex waves having contiguous Wave trains of random amplitudes, the beginning and ending of each wave train being determined by major peaks of the complex wave, the system of equalizing the amplitudes of these wave trains stepwise in steady states comprising means for producing said complex waves; a major peak selector; means for applying the produced complex waves to said selector for producing pulse waves at said major peaks; a gain controllable amplifier having irst and second inputs and an output; means for applying the produced complex waves to the first input of said amplifier, for amplifying same at said output; a series connected circuit comprising a reference bias source, a rectifier means, and a capacitor means; means for applying said amplified output Waves to last said circuit for rectifying same and charging said capacitor by the rectified waves proportionally when larger in magnitude than said bias source; means for coupling the charge of said capacitor to said second input in gain reducing direction, thereby reducing the gain of said amplifier to a point where the peak of the amplified wave train is substantially equal to that of the reference bias; a normally inoperative discharger means in parallel with said capacitor; and means for applying said pulse waves upon said discharger means for momentarily discharging the capacitor, and thereby resetting for the said steady state gains of the amplifier.

References Cited in the file of this patent UNlTED STATES PATENTS 2,316,354 Moritz Apr. 13, 1943 2,713,620 Tilley July 19, 1955 2,799,735 Breckman et al Aug. 16, 1957 2,902,548 Moeller Sept. 1, 1959 2,925,476 Atlas Feb. 16, 1960

Images