Sign up
Title:
SOUND-OPERATED, YES-NO RESPONSIVE SWITCH
United States Patent 3688126
Abstract:
A switch which may be placed in either of two positions in response to a voiced decision in the form of the two words "yes" or "no." The switch is activated by a spoken decision of predetermined minimum magnitude and senses the frequency distribution of the spoken decision. If the frequency distribution of the spoken decision indicates the presence of sibilant noise, as strongly present in the word "yes," the switch is placed in one position. If the frequency distribution of the spoken decision indicates the absence of sibilant noise, as is the case with the word "no," the switch is placed in the other position.


Inventors:
KLEIN PAUL R
Application Number:
05/110966
Publication Date:
08/29/1972
Filing Date:
01/29/1971
Assignee:
PAUL R. KLEIN
Primary Class:
Other Classes:
307/117, 367/198, 704/E11.003, 704/E15.045
International Classes:
G10L11/02; G10L15/26; (IPC1-7): H01H35/24
Field of Search:
340/148 179
View Patent Images:
Primary Examiner:
Jones H. O.
Assistant Examiner:
Smith, William J.
Claims:
1. A sound-operated, decision responsive switch comprising, in combination:

2. A switch according to claim 1 wherein said means for converting said spoken decision into a signal comprises:

3. A switch according to claim 1 wherein said selecting means comprises:

4. A switch according to claim 3 wherein said inhibiting means comprises:

5. A switch according to claim 3 wherein said converting means comprises:

6. A switch according to claim 3 wherein said first and second output means comprises first and second bistable flip-flops.

Description:
BACKGROUND OF THE INVENTION

1. Field of the Invention

The present invention relates to a sound-operated, decision responsive switch and, more particularly, to a switch which may be placed in either of two positions in response to a voiced decision in the form of the words "yes" and "no."

2. Description of the Prior Art

Voice or sound operated devices have been developed for various purposes. All known prior devices, however, may be broadly divided into two categories. The first category includes devices which are sensitive to and are operated by any audible sound, preferably the sound of the human voice. For example, U.S. Pat. Nos. 2,957,957 and 3,247,339 to Johnson and Miller, respectively, disclose sound-operated switches which are sensitive to particular frequencies within the audio spectrum. Other devices, within the same category, attempt to distinguish between the human voice and other sounds, such as background noise, data signals, etc. Typical of such devices are those disclosed in U.S. Pat. No. 3,270,216 to Dersch, U.S. Pat. No. 3,286,031 to Geddes, U.S. Pat. No. 3,321,581 to Zryd et al., and U.S. Pat. No. 3,448,215 to Engel. All of these devices simply detect the presence of a voiced sound, as contrasted to sound in general, and perform some operation in the presence of such sound. In this category, no attempt is made to differentiate between different voiced words.

The second category includes systems which differentiate between different voiced words. For example, U.S. Pat. No. 2,575,910 to Mathes discloses a voice-operated signalling system which is responsive to vocal command words and which may be employed in an automatic telephone exchange as a substitute for the present standard equipment which responds to the pulses generated by a finger dial. In other words, such a system purportedly distinguishes between many words independently of the phonetic characteristics of the individual speaker.

With the advent of the computer, many sophisticated attempts have been made to develop systems capable of recognizing human speech. A review of some modern systems is contained in an article entitled "Voice Recognition and Response Systems" by Cay Weitzman found on Pages 165 through 170 of the December, 1969 issue of "Datamation." As can be seen from Mr. Weitzman's article, existing systems are not only extremely complex, but extremely expensive. In addition, Mr. Weitzman points out that computer systems that can identify words in continuous speech of an unknown speaker are beyond the current state of the art in speech recognition and that only limited speech recognition systems have been developed and then only on an experimental basis.

It is therefore apparent that there are no simple, inexpensive, available voice or sound operated devices which are capable, even on a limited scale, of differentiating between different voiced words. More specifically, there are no available voice-operated devices which are capable of responding to a voiced decision, in the form of the words "yes" or the word "no" and operating upon such decision, in spite of the vast number of uses for such a device. For example, by means of a two track tape recorder and a mini-computer, it would be possible to carry out oral conversations with a machine. The first track would be used for messages to the human, and the second track would have markers used for addressing the messages, where the next message to the human would depend upon his previous answer to a yes-no question. One use of such a system would be in a protection system for the aged or infirmed where questions such as "Do you need help?" or "Shall I call a doctor?" could be addressed to the human and the response would be dictated by whether the answer was "yes" or "no." Another use would be in a burgler protection system where the human, upon new and unexpected entrance into a dwelling, could identify himself through answers to coded questions. Again, the device would expedite the setting up of complicated laboratory or industrial equipment by relatively untrained personnel by the question and answer technique. Similarly the device serves as an aid in complex laboratory and industrial tasks where a human operator is heavily involved with manual tasks. Again, the device, with the aid of the two track tape mentioned earlier, dispenses with the often costly and confusing array of switches which must be set in many scientific instruments. Another obvious use would be in an educational system where automation is becoming a common practice. More specifically, a student who is being fed individual instructional materials in recorded form could periodically be asked whether he or she understands the lesson up to that point. If the answer is "yes," the lesson would proceed. If the answer is "no," the machine would automatically backtrack by a predetermined amount.

In spite of all of these obvious uses for a sound-operated, decision responsive device, as well as a large number of additional uses, no such device is presently available.

SUMMARY OF THE INVENTION

In accordance with the present invention, there is provided a sound-operated, decision responsive switch and, more particularly, a switch which may be placed in either of two positions in response to a voiced decision in the form of the words "yes" or "no." Such a switch overcomes the problems inherent in the first category of known devices which are simply sensitive to audible signals and are incapable of differentiating between different voiced words. The present switch, while not having the ability to differentiate between many words, as in the second category of prior art devices discussed previously, eliminates the complexity and expense of such prior art devices and substitutes therefor a simple, inexpensive, uncomplex device having many practical uses.

Briefly, the present switch is activated by a spoken decision of predetermined minimum magnitude and senses the frequency distribution of the spoken decision. If the frequency distribution of the spoken decision indicates the presence of sibilant noise, as strongly present in the voiced word "yes," the switch generates a first output. If the frequency distribution of the spoken decision indicates the absence of sibilant noise, as is the case with the word "no," the switch generates a second output.

It is therefore an object of the present invention to provide a sound-operated, decision responsive switch.

It is a further object of the present invention to provide a sound operated, yes-no responsive switch.

It is a still further object of the present invention to provide a switch which may be placed in either of two positions in response to a voiced decision in the form of the words "yes" or "no."

It is another object of the present invention to provide a sound-operated, yes-no responsive switch which is activated by a spoken decision of predetermined minimum magnitude and which senses the frequency distribution of the spoken decision to determine the presence or absence of sibilant noise, a "yes" being indicated where sibilant noise is present and a "no" being indicated where sibilant noise is absent.

Still other objects, features and attendant advantages of the present invention will become apparent to those skilled in the art of a reading of the following detailed description of the preferred embodiment constructed in accordance therewith, taken in conjunction with the accompanying drawings wherein like numerals designate like parts in the several figures and wherein:

BRIEF DESCRIPTION OF THE DRAWINGS

FIG. 1 is a block diagram of the present invention;

FIG. 2 is a block diagram of a preferred physical embodiment of sound-operated, yes-no responsive switch constructed in accordance with the teachings of the present invention; and

FIG. 3 is a block diagram showing the details of a portion of the embodiment of FIG. 2 and a possible modification thereof.

DESCRIPTION OF THE PREFERRED EMBODIMENT

Referring now to the drawings and, more particularly, to FIG. 1 thereof, there is shown, in block diagram form, the basic elements of a sound operated, yes-no responsive switch, generally designated 10. Switch 10 operates on the known fact that the voiced word "yes" contains sibilant noise, present in all words containing the letter s, which results in an abundance of frequency components at the high end of the audio spectrum. On the other hand, the voiced word "no" contains no sibilant noise and no frequency components at the high end of the frequency spectrum. Accordingly, switch 10 includes a microphone 11 which responds to the voiced word and applies a corresponding electrical signal to an amplifier 12. Amplifier 12 operates as a threshold device so that switch 10 is only responsive to voiced decisions of predetermined minimum magnitude. The output of amplifier 12 is applied to a frequency analyzer 13. Frequency analyzer 13 may be any one of many known devices for sensing the frequency distribution of the decision. If the decision contains only frequency components at the low end of the audio spectrum, a signal is applied over a line 14 to activate a "no" output device 15 which generates a "no" indicative signal on a line 16. On the other hand, if the spoken decision, as sensed by frequency analyzer 13, contains frequency components at the high end of the audio spectrum, indicative of the presence of sibilant noise, a signal is applied over a line 17 to activate a "yes" output device 18 which generates a "yes" indicative signal on a line 19.

It is therefore seen that switch 10 is capable of being placed in either of two conditions in response to a voiced decision in the form of the words "yes" or "no." Simply by speaking one of these two words, switch 10 generates an output either on line 16 or line 19. After the appropriate output is generated, reset can be provided by any of several methods. For example, the response itself can be used to cause reset after performing a desired function. Alternatively, an external device, such as a computer, upon receiving the decision, can cause the reset. Of course, the reset can be generated manually. In either event, the output of switch 10 indicates whether the voiced decision was "yes" or "no."

Referring now to FIG. 2, there is shown a preferred physical embodiment of sound-operated, yes-no responsive switch 10. In FIG. 2, the output of amplifier 12 is applied to a frequency to pulse rate converter 20. Frequency to pulse rate converter 20 may be any one of many well-known devices for generating a uniform pulse of charge each time the output of amplifier 12 goes through zero amplitude. Thus, the output of converter 20 is a train of pulses whose rate is proportional to the frequencies present in the original voiced decision. A possible embodiment for frequency to pulse rate converter 20 is shown in FIG. 3. More specifically, the output of amplifier 12 may be applied to the input of a Schmitt trigger 21, a known device which changes state every time the input signal crosses a predetermined amplitude. The output of Schmitt trigger 21 may be applied to a differentiator 22 which produces a sharp pulse for every state change of Schmitt trigger 21. Since such sharp pulses are not suitable for the remainder of the circuit, they may be applied to a monostable multivibrator 23 which will then generate a train of uniform pulses, one pulse for each zero crossing of the original signal.

Referring now again to FIG. 2, the pulse train at the output of converter 20 may then be applied to a "yes" integrator 30 and a "no" integrator 31. Integrators 30 and 31 are conventional integration circuits, essentially RC devices, whose time constants and impedances are predetermined, but different. The outputs of integrators 30 and 31 are applied to first inputs of a "yes" flip-flop 32 and a "no" flip-flop 33, respectively. Flip-flops 32 and 33 may be conventional two state, J-K flip-flops, having J and K inputs and complementary outputs, the outputs of integrators 30 and 31 being applied to the J inputs of flip-flops 32 and 33, respectively. With such connection, if the spoken decision is "no," enough pulses are integrated by "no" integrator 31 to cause "no" flip-flop 33 to be set and to apply a true output on output line 35. Such output, indicating the presence of a "no," is applied to a first input of an AND gate 40. In the case of a "no" spoken decision, "yes" integrator 30 would not build up sufficient charge to fire "yes" flip-flop 32. As a result, "yes" flip-flop 32 would apply a true output on its complementary output line 34, which line is applied to the other input of AND gate 40. The result is an output from AND gate 40 on line 16 indicating that the spoken decision was a "no."

On the other hand, if a "yes" decision is received by microphone 11, "yes" integrator 30, as well as "no" integrator 31, will receive sufficient charge to fire flip-flops 32 and 33, respectively. In this case, a false output appears on complementary output line 34 of "yes" flip-flop 32 inhibiting the output of AND gate 40 on line 16. Simultaneously, "yes" flip-flop 32 provides a true output on output line 19 indicating that the spoken decision was a "yes."

Such "yes" and "no" outputs on lines 19 and 16, respectively, may be used to drive any additional equipment as may be desired and as briefly discussed previously. After a decision is reached, a reset pulse may be applied to each of the K input terminals of flip-flops 32 and 33 from any of the sources discussed previously with respect to FIG. 1. As soon as flip-flops 32 and 33 are reset, switch 10 is ready for the next spoken decision.

Referring again to FIG. 3, there is shown one possible modification to the embodiment of FIG. 2. More specifically, the output of converter 20 may be applied to a single integrator 50 whose output would then be applied to both of flip-flops 32 and 33. In this case, flip-flops 32 and 33 would be adjusted to have different triggering levels, flip-flop 33 triggering at a lower charge level than flip-flop 32.

Accordingly, it is seen that in accordance with the present invention, there is provided a simple, inexpensive uncomplicated device, capable of many uses, which may be placed in either of two positions in response to a voiced decision in the form of the words "yes" or "no." Such a switch may be used in protection systems for the aged or infirmed, in burglar protection systems, in complex laboratory or industrial installations, as a substitute for binary hand switches on scientific instruments, in an educational system, and in many other areas. The present system may also be used in combination with other known devices of the type previously described which will distinguish between voiced signals in the audio spectrum and other audio signals such as noise, data, etc.

While the invention has been described with respect to the preferred physical embodiment constructed in accordance therewith, it will be apparent to those skilled in the art that various modifications and improvements may be made without departing from the scope and spirit of the invention. Accordingly, it is to be understood that the invention is not to be limited by the specific illustrative embodiments, but only by the scope of the appended claims.