US3536837A

US3536837A - System for uniform printing of intelligence spoken with different enunciations

Info

Publication number: US3536837A
Application number: US713386A
Authority: US
Inventors: Ian Fenton
Original assignee: Individual
Current assignee: Individual
Priority date: 1968-03-15
Filing date: 1968-03-15
Publication date: 1970-10-27
Anticipated expiration: 1987-10-27

Description

United States Patent Office 3,536,837 Patented Oct. 27, 1970 3,536,837 SYSTEM FOR UNIFORM PRINTING OF INTELLIGENCE SPOKEN WITH DIF- FERENT ENUNCIATIONS Ian Fenton, 334 W. 87th St., New York, N.Y. 10024 Filed Mar. 15, 1968, Ser. No. 713,386 Int. Cl. G10l N US. Cl. 179-1 9 Claims ABSTRACT OF THE DISCLOSURE In a speech printing system for converting into singular discrete signals representative of phonetic sounds combinations of frequency components having determinable amplitude relations with each other representing each phonetic sound of a speech wave, a speech converter connected to a speech input converts speech signals at the speech input into such discrete signals. A reference system having an input connected to the speech input compares each of the discrete signals with each of a plurality of discrete reference signals representative of phonetic reference sounds to determine and select that one of the discrete reference signals which most closely compares with each of the discrete signals. A printer connected to the output of the reference system prints the selected discrete reference signal in place of each of the discrete signals.

DESCRIPTION OF THE INVENTION The present invention relates to a system for printing spoken intelligence. More particularly, the invention relates to a system for uniform printing of intelligence spoken with different enunciations.

In ordinary speech, each phonetic sound is produced is substantially replica wave trains that repeat successively at a fundamental pitch or frequency during propagation of the articulated sound. The succession of the wave trains is effected by substantially regular puifs of air from the glottis, which are set into vibration in the momentarily formed resonant cavities of the vocal system. Each Wave train includes all the necessary phonetic information and its specific waveshape is formed by the number of frequency components that are produced in such cavities,

frequency components composing a pure phonetic sound are independent of characterization components, which are produced principally by the larynx. The composite structure of the characterization components is inconsistent in form and varies in a complex manner with the varying pitch of the speakers voice. However, the frequency ratios of the basic components and their relative amplitudes remain substantially constant with respect to the fundamental frequency, although the frequency locations of all the basic components may change in the entire spectrum band of the voice. Thus, the human intelligence interprets phonetic sounds by measuring the ratios of basic frequency components, and their relative amplitudes, with respect to the fundamental frequency, without regard to the characterization components. The characterization components are interpreted as a form of voice quality. I

Systems for printing spoken intelligence are well known in the art. Such systems are disclosed in United States Pats. Nos. 2,705,260, issued Mar. 29, 1955, to M. V. Kalfaian, 2,708,688, issued May 17, 1955, to M. V. Kalfaian, 2,921,133, issued Jan. 12, 1960, to M. V. Kalfaian, and 3,225,883, issued Dec. 28, 1965, to W. A. Ayres, for example. A system for printing spoken intelligence, or speech, analyzes and selects various frequency components which occur simultaneously during the propagation of phonetic sounds, for the purpose of translating phonetic sounds into visible or printed intelligible indicia. The simultaneously selected frequency components are translated into discrete signals for the actuation of characterprinting keys such as, for example, the keys of a modified electric typewriter, or slotted code bars as in teletypewriters, so that spoken words may be translated into visual words.

There are many different dialects and accents in the language or speech of a single nation. Some accents are so pronounced that people of one section of a country may have considerable difficulty in understanding the speech of people of another section of the same country. Under these circumstances, the speech or word printing system may produce a completely, or at least partially, unintelligible printed record of spoken intelligence when the speaker has a strong accent or dialect. This constitutes a serious disadvantage of known systems for printing spoken intelligence.

The principal object of the present invention is to provide a new and improved system for printing spoken intelligence.

An object of the present invention is to provide a system for printing intelligence spoken with different enunciations.

An object of the present invention is to provide a system for printing understandable intelligence, although it is spoken with different enunciations, accents, dialects, and the like, with efficiency, effectiveness and reliability.

Another object of the present invention is to provide a system for printing intelligence spoken with different enunications which is of simple structure and inexpensive in cost and operation and which is readily incorporated with a known system for printing spoken intelligence.

In accordance with the present invention, a speech printing system for converting into singular discrete signals representative of phonetic sounds combinations of frequency components having determinable amplitude relations with each other representing each phonetic sound of a speech wave includes a speech input. A speech converter connected to the speech input converts speech sig nals at.the speech input into discrete signals representative of phonetic sounds. A reference system having an input connected to the speech input and an output compares each of the discrete signals with each of a plurality of discrete reference signals representative of phonetic reference sounds to determine and select that one of the discrete reference signals which most closely compares with each of the discrete signals. A printer connected to the output of the reference system prints the selected discrete reference signal in place of each of the discrete signals.

The reference system comprises a reference speech input for providing the phonetic reference sounds and reference speech signals corresponding to the phonetic reference sounds. A reference frequency converter connected to the reference speech input converts the reference speech sig nals corresponding to the phonetic reference sounds to the discrete reference signals. A comparator connected between the reference frequency converter, the speech converter and the printer compares each of the discrete signals provided by the speech converter with each of the discrete reference signals.

The reference speech input comprises a reference signal storage recording the phonetic reference sounds. A

readout head is in operative proximity with the reference signal storage. A readout control is connected to the speech input and is connected to and controls the operation of the readout head in accordance with speech at the speech input.

A coincidence gate and printing suppressor connected in series between the comparator and the printer gates the selected discrete reference signals and suppresses printing during transient periods between one phonetic sound and the next. A syllable and word separator connected between the speech input and the printer controls the spacing operation of the printer. The speech converter comprises a frequency converter for converting speech at the speech input to the discrete signals.

In accordance with the present invention, a method of printing intelligence spoken with different enunciations comprises the steps of converting speech into discrete signals representative of phonetic sounds. Each of the discrete signals is compared with a plurality of discrete reference signals representative of phonetic reference sounds. That one o'f the discrete reference signals which most closely compares with each of the discrete signals is printed in place of each of the discrete signals. The discrete reference signals are derived by converting the frequencies of recorded phonetic reference sounds. Selected ones of the discrete reference signals are gated after comparison with the discrete signals.

In order that the present invention may be readily carried into effect, it will now be described with reference to the accompanying drawing, wherein:

FIG. 1 is ablock diagram of an embodiment of a speech printing system including an embodiment of the reference system of the present invention; and

FIG. 2 is a block diagram of an arrangement which may be utilized as the reference speech input of the embodiment of FIG. 1.

The system and method of the present invention may be utilized with any suitable speech printing system for converting into singular discrete signals representative of phonetic sounds combinations of frequency components having determinable amplitude relations with each other representing each phonetic sound of a speech wave. Although the system of the present invent-ion is readily and facilely incorporated into any suitable speech printing system, to provided a uniform or completely comprehensible or intelligible printing of intelligence spoken with different enun'ciations, the system of the present invention is described herein, for illustrative purposes, as being included in the system disclosed in US. Pat. No. 2,705,260, and is not limited to the specifically disclosed speech printing system.

In FIG. '1, a speech input 11 may comprise any suitable speech input for providing speech or voice signals such as, for example, the original speech-wave 1 of FIG. 2 of U.S Pat. No. 2,705,260, hereinafter referred to as the patnt.'The output of the speech input 11 is connected to the input of a frequency converter 12 via a lead 13.

The frequency converter 12 may comprise any suitable means for converting speech signals from the speech input 11' to discrete signals representative of phonetic sounds. A suitable frequency converter may comprise, for example, the frequency-transposed speech-wave 2, the bandpass filter'f to 12;, the rectifiers and the cathode ray tube 3, of FIG. 2 of the patent. The discrete signals representative of the phonetic sounds of the speech input are provided at the output of the frequency converter -12 via a lead 14.

In a speech printing system of known type such as, for example, that of the patent, the discrete signals representative of phonetic sounds in the lead 14 are supplied to a phonetic printer 15. The phonetic printer 15 may comprise any suitable printer for printing indications of the discrete signals such as, for example, the phonetic printer 6 of FIG. 2 of the patent. The printer 15 may, of course, print ordinary words; rather than phonetic words, as in the arrangement described in US. Pat. No. 3,225,- 883.

A speech printing system of known typepreferably includes a syllable and word separator 16 having an input connected to the speech input 11 via a lead 17 and an output connected to the printer 15 via a lead 18. The syllable and word separator 16 may comprise any suitable means for providing separations between syllables and between words in the printed material such as, for example, the impedance Z, the rectifiers V5, V4, the RC network and the amplitude limiter 7 of FIG. 2 of the patent.

A printing suppressor 19 is preferably connected between the lead 14 which provides the discrete signals representative of phonic sounds and the printer 15. The printing suppressor 19 is connected to the printer 15 via a lead 20 and may comprise any suitable means for suppressing printing during transient periods between one phonetic sound and the next. A suitable printing suppressor may comprise, for example, the wave differentiating coupling condensers c1, c2, c3, 04, the mixer 8 andthe beam-control grid 9 of the cathode ray-tube 3 of the patent.

In accordance with the present invention, a reference speech input 21, having an input connected to the lead 13 via a lead 22 and an output connected to a lead 23. The re'ference speech input 21 comprises any suitable means for providing a plurality of phonetic reference sounds. A suitable reference speech input 21 may comprise, for example, a tape or other suitable record medium having recorded therein all the phonetic sounds which may be encountered in the language spoken into and printed by the speech printing system.

Considerable flexibility is accorded in the selection and recording of phonetic sounds for use as the reference speech input 21. Thus, for example, if the speech printing system is to be installed permanently in an area in which, specific sounds are improperly enunciated, such sounds may be solely recorded to the exclusion of sounds Which are usually properly enunciated or pronounced. The reference speech input 21 may be removed and replaced as desired, upon removal of the speech printing system to another area of the same language region. If the speech printing system is moved to another language area, or nation where another language is spoken, both the reference speech input 21 and the printer 15 are replaced.

A reference frequency converter 24 has an input connected to the output of the reference speech input 21 via the lead 23. The reference frequency converter 24 may comprise any suitable means for converting the reerence speech signals indicating the phonetic reference sounds to discrete reference signals representative of said phonetic reference sounds. The reference frequency converter 24 may thus be the same as the frequency converter 12.

A comparator 25 has one input connected to the output of the reference frequency converter 24 via a lead 26, a" second input connected to the output of the frequency converter 12 via the lead 14 and an output connected to'a lead 27. The comparator 25 may comprise any' suitable comparator means for comparing each of the discrete signalsprovided by the frequency converter 12 in the lead 14 with each of the discrete reference signals provided by the reference frequency converter 24 in the lead'2'6i'A;

suitable comparator 25 may comprise, for example, the comparator shown in FIG. 108 and described in lines 25 to 75 of column 188 and lines 1 to 17 of column 189 of U.S. Pat. No. 3,343,133, issued Sept. 19, 1967, to G. Dirks.

The comparator 25 functions, as described, and produces an output signal only when the discrete reference signal which is most similar to the reference signal in the lead 14 appears in the lead 26. When the other discrete reference signals, representing the other phonetic reference sounds, appear in the lead 26, no output signal is produced by the comparator 25.

A coincidence gate 28 has a first input connected to the output of the comparator 25 via the lead 27 and a second input connected to the output of the reference frequency converter 24 via a lead 29. The comparator 25 also has an output connected to the input of the printing suppressor 19 via a lead 30. The coincidence gate 28 may comprise any suitable coincidence gate such as, for example, and AND gate of any suitable type which functions to conduct the discrete reference signal in the lead 29 to the lead 30 only when a signal is provided simultaneously in the lead 27. The coincidence gate 28 thus supplies to the printer that one of the discrete reference signals which most closely compares to the discrete signal in the lead 14 in place of each discrete signal in said lead. A suitable AND gate may comprise, for example, that shown and described in a textbook entitled, Directory of Electronic Circuits, by Matthew Mandl, 1966, Prentice-Hall, Inc., Englewood Cliffs, NJ., pages 93 to 95.

FIG. 2 discloses an embodiment of a reference speech input 21. The reference input of FIG. 2 comprises a reference signal storage 31 which may comprise any suitable record medium, of the type hereinbefore described, having recorded therein all the phonetic sounds which may be encountered or mispronounced in the language spoken into and printed by the speech printing system, as hereinbefore described.

A readout head 32 is positioned in operative proximity with the reference signal storage 31. A readout control unit 33 has an input connected to the speech input 11 (FIG. 1) via a lead and an output connected to the readout head 32 via a lead 35. The electrical signals produced by the readout head 32 are supplied to the reference frequency converter 24 (FIG. 1) via a lead 35. The electrical signals produced by the readout head 32 are supplied to the reference frequency converter 24 (FIG. 1) via a lead 36.

When a speech signal is produced by the speech input 11 (FIG. 1), said signal actuates the readout control unit 33, and said readout control unit actuates the readout head 32 and causes said readout head to scan all the phonetic reference sounds recorded in the reference signal storage 31. This may be accomplished in any suitable manner such as, for example, by any suitable driving arrangement which moves the record medium of the signal storage 31 and the readout head 32 relative to each other in such a way that said head passes in operative proximity with each phonetic reference sound recorded in said record medium, in sequence.

The readout control unit 33 may therefore comprise a mounting arrangement for the readout head 32 and a motor drive coupled to said readout head or mounting arrangement for moving said readout head along the entire expanse of the tape or record medium. On the other hand, the readout control unit 33 may comprise a tape or record medium moving, stepping or driving arrangement and a motor drive coupled to said tape driving arrangement so that its entire expanse moves in the direction of its length past and in operative proximity with the readout head 32.

While the invention has been described by means of a specific example and in a specific embodiment, I do not wish to be limited thereto, for obvious modifications will occur to those skilled in the art without departing from the spirit and scope of the invention.

What is claimed is:

1. In a speech printing system for converting into singular discrete signals representative of phonetic sounds combinations of frequency components having determinable amplitude, relations with each other representing each phonetic sound of a speech wave,

a speech input;

speech converting means connected to said speech input for converting speech signals at said speech input into discrete signals representative of phonetic sounds;

a reference system having an input connected to said speech input and an output, said reference system comparing each of said discrete signals with each of a plurality of discrete reference signals representative of phonetic reference sounds to determine and select that one of said discrete reference signals which most closely compares with each of said discrete signals;

printing means connected to the output of said reference system for printing the selected discrete reference signal in place of each of said discrete signals;

syllable and word separating means connected between said speech input and said printing means for controlling the spacing operation of said printing means; and

printing suppressor means connected beltween said comparator means and said printing means for gating the selected discrete reference signals and supressing printing during transient periods between one phone tic sound and the next.

2. In a speech printing system as claimed in claim 1, wherein said reference system comprises reference speech input means for providing said phonetic reference sounds and reference speech signals corresponding to said phonetic reference sounds, reference frequency converter means connected to said reference speech input means for converting said reference speech signals corresponding to said phonetic reference sounds to said discrete reference signals and comparator means connected between said reference frequency converter means, said speech converting means and said printing means for comparing each of the discrete signals provided by said speech converting means with each of said discrete reference signals.

3. In a speech printing system as claimed in claim 2, wherein said reference speech input comprises reference signal storage means recording said phonetic reference sounds, readout means in operative proximity with said reference ssgnal storage means and readout control means connected to said speech input and connected to and controlling the operation of said readout means in accordance with speech at said speech input.

4. In a speech printing system as claimed in claim 2, coincidence gate means, and printing suppression means connected in series with said coincidence gate means between said comparator means and said printing means for gating the selected discrete reference signals and suppressing printing during transient periods between one phonetic sound and the next.

5. In a speech printing system as claimed in claim 3, coincidence gate means, and printing suppressor means connected in series with said coincidence gate means between said comparator means and said printing means for gating the selected discrete reference signals and suppressing printing during transient periods between one phonetic sound and the next.

6. In a speech printing system as claimed in claim 5, wherein said speech converting means comprises frequency converter means for converting speech at said speech input to said discrete signals.

7. A method of printing intelligence spoken with different enunciation, comprising the steps of converting speech into discrete signals representative of phonetic sounds;

the step of gating selected ones of the discrete reference signals after comparison with the discrete signals.

comparing each of the discrete signals with a plurality of discrete signals representative of phonetic reference sounds; printing that one of the discrete reference signals which most closely compares with each of said discrete signals in place of each of said discrete signals;

controlling the spacing operation of said printing means;

and

gating the selected discrete reference signals and suppressing printing during transient periods between one phonetic sound and the next.

8. A method of printing intelligence spoken with different enunciations as claimed in claim 7, wherein the discrete reference signals are derived by conversion from the frequencies of recorded reference sounds.

9. A method of printing intelligence spoken with different enunciations as claimed in claim 8, further comprising References Cited UNITED STATES PATENTS Davis et a1. 179-16 Biddulph et al.

Kalfaian.

Williams et al.

Kalfaian.

Bibbero.

Orthuber et al.

Johnson.

KATHLEEN H. CLAFFY, Primary Examiner C. W. JI RAUCH. Assistant Examiner