RU2571588C2 - Electronic device for automatic translation of oral speech from one language to another - Google Patents

Abstract

FIELD: physics.

SUBSTANCE: device comprises a housing in which there is a computer with an independent power supply and software which enables, through a computer input device, to input into the computer acoustic signals in different languages, translation thereof using the corresponding software and emitting in the form of acoustic translated signals through an output device. The acoustic signal input and output devices are designed and arranged such that it is possible to perform spatial selection of acoustic signals generated by speakers and to suppress external acoustic interference and noise by active noise reduction techniques using multi-band signal processing.

EFFECT: high accuracy and speed of translation in external noise and interference conditions.

15 cl, 2 dwg

Description

Technical field

The invention relates to the field of electronics, in particular to electronic devices for organizing communication between people speaking different languages.

State of the art

Currently, various systems and devices for recognizing voice messages are known, for example, devices for implementing voice control of modern cell phones, smartphones, tablet computers, laptops, as well as systems and devices for translating acoustic signals from one language into another language, implemented in the form of various handheld electronic dictionaries and translators (ITRAVL, Partner, SpeechGuard jetBook e-Book Reader Multilux 2, BRM Electronics, etc.).

Currently, there are a large number of different ways of organizing the process of recognition and translation from one language to another language using databases of speech sounds, words or phrases.

For example, RF patent application No. 2008114596/09 (priority date 10/17/2006, published 06/27/2010, G10L 15/14) describes a speech recognition method and an electronic device in which, by frame-by-frame analysis of sound messages and their multi-parameter analysis for coincidence of characteristic signs with stored in the database signs of words or phrases perform speech recognition.

Various speech recognition systems are well known:

- Speech Recognition (SR) speech recognition system, available in 22 languages, which can be integrated into various electronic devices;

- Text-To-Speech (TTS) - speech synthesis, literally - “from text to speech”. Presented in almost all devices and programs produced by Ectaco. Allows you to hear the pronunciation of any words, phrases, numbers or letters written in a language that is understandable for this system - there are currently 17 such languages.

- LingvoSoft - software for translating texts. Combines various linguistic programs for translating texts and teaching foreign languages. Among LingvoSoft programs there are speaking and non-speaking dictionaries, phrasebooks, training cards, text translators. LingvoSoft applications exist in versions for different platforms.

The disadvantage of well-known pocket translators is the inconvenience of using them when communicating. To do this, it is desirable to have a similar translator for each person, since for a correct translation one has to speak very close to the microphone mounted in the translator's case. When using a single translator, the communication process drastically slows down, because it is necessary to sequentially enter phrases in one or another language with a high quality of the input acoustic signal and listen to the translated phrases or words sequentially from one language to another, for example, as is done in the Multilux 2 translator and BRM Electronics, manufactured by US companies Ectaco and BRM Corp.

http://www.youtube.com/watch?v=nsnIpZepbgw

http://www.youtube.com/watch?v=0SUICVEKS18

The prior art electronic transceiver system and corresponding devices with the function of simultaneous translation of oral speech from one language to another (RF patent application No. 2008137394/22, 09/19/2008, G06F 17/28).

The system includes at least one transceiver communication terminal of an incoming analog audio signal (AA) in speech form. This terminal contains: non-volatile storage device, such as flash memory; speech input means AA in the input language to the communication terminal, which is functionally transmitting; means for analog-to-digital conversion and processing of the speech form AA into a code text form in the input language; means for translating the code form of the text (CFT) in the input language into the CFT in the output language; means of digital-to-analog conversion and processing of QFT in the output language in AA in the same language in speech form, as well as means for outputting the speech form AA in the output language through the communication terminal, which is functionally receiving.

Distinctive features are the following. The communication terminal, which is functionally transmitting, is equipped with a means of verifying the spoken language of the owner of the terminal regarding its written form, structured by means of this communication terminal. The structure of CFT translation tools in the input language in the CFT in the output language is organized two-level. Moreover, the first level is part of the communication terminal, which is functionally transmitting, and is equipped with a means for converting CFT in the input language into CFT in an intermediate language generally accepted in world practice, for example, English. The second level is part of the communication terminal, which is functionally host, and is equipped with a means of converting CFT in an intermediate language into CFT in the output language.

The disadvantage of such a system is the inconvenience of using this method of communication, which requires the convenience of each user having an appropriate electronic device (transceiver terminal) with the function of a simultaneous interpreter and the presence in this place of an appropriate wireless communication line to realize the possibility of synchronizing the operation of devices and exchanging messages via a special a radio channel and a specific protocol, for example, using cellular telephone equipment.

A known group of inventions (RF patents No. 2320012, 2235370, 2344479, 2288550, 2326495, 2211491, 2145446, 2038704, 2106075, 2106073, 2106074), which describe systems for transmitting messages with feedback and multi-band processing of electrical signals of the transmitted message and feedback signals communication in a special processing unit.

The signal processing unit contains two of the most integrated functional units: a) a coordinated optimal filter; b) signal generating units for active noise reduction.

The processing unit provides a comparative analysis of all possible parameters of the signal components from the output of the message source, the signals of the noise sources and feedback signals generated at the point of receipt of the message. At its output, component-wise optimally corrected (predistorted) signals of the transmitted message and component-wise optimized signals are designed to actively reduce the noise components so that the distortion of the transmitted signal at the point of receipt of the message is minimized as much as the specific conditions for the organization of the communication or control system allow. The nodes of such a processing unit can be made in the form of a specialized analog or digital device or implemented virtually - in the form of a computer program.

A large group of inventions is known (US2002126854 (A1), 2002-09-12 H04R 19/00; H04R 19/02; BR 9913954 (A), H04R 19/00, 1998-01-07; BR 9710335 (A), 2000- 01-18, G10K 15/02; AU 2004302819 (A1), 2005-03-10, H04R 25/00; AU 2002306653 (A1), 2003-09-29, H04B 11/00; AU 7333000 (A), 2001 -03-19, H04R 3/00; AU 6776100 (A), 2001-03-13, H04R 19/02; AU 6497299 (A) 2000-04-10, H04R 3/00; AU 5902099 (A), 2000 -04-10, H04R 17/00; AU 3889897 (A), 1998-02-09, G10K 15/02; AU 3790901 (A), 2001-05-14, 2001-05-14; AU 2107099 (A) , 1999-07-26, H04R 19/00), which describe various technical solutions related to the method of transmitting sound vibrations from one point in space and time to another point using special ultrasonic vibrations.

The method consists in the formation of air vibrations at the point of receipt of a message in the sound frequency range of the waves due to exposure to air in this area of space with high-power ultrasonic signals, but of different frequencies. At certain ratios of vibration frequencies of ultrasonic frequencies and at high power, the air medium from a linear medium turns into a nonlinear region of space in which the combination (total - difference) harmonic components of these ultrasonic signals appear. If one of the ultrasonic signals has a constant level and is emitted at the same frequency, and the other ultrasonic signal is an ultrasonic signal modulated by the amplitude of the sound signal (for example, an SSB with an upper sideband), then the beats of such signals may appear at the frequencies of the sound spectrum . These components can be heard by the human hearing without special receivers or devices. A similar system for generating sounds in the above patents is called a parametric sound reproducing system.

In the indicated group of inventions, technical solutions are described that relate mainly to schemes of ultrasonic generators, modulators, distortion correctors, and ultrasonic emitters for implementing the method described above for generating sounds near human hearing organs in mono and stereo.

The advantage of such a system can be considered the possibility of highly directional transmission of an audio message to a specific person in a localized area of space, so that outsiders outside the diagram of the emitter will not be able to hear the transmitted message. By adjusting the power of ultrasonic signals, the level of modulation of signals, changing the direction of radiation or the directional pattern of ultrasonic signals, it is possible to widely change the position of the space region for receiving sound signals for certain users, its size and the volume level of the sound signal at a particular point in space.

Disclosure of invention

The problem solved by the invention is the creation of a mobile, easy-to-use electronic device (translator) designed to realize the possibility of communication between a person speaking one language, equipped with an electronic device, with a person speaking another language that does not have such a device, or when this device is turned off, in the mode of simulating communication between people in a language they understand in real time, so that everyone speaks and hears answers mainly in a language that he understands.

EFFECT: increased communication distance between people when using an electronic device for automatic translation of oral speech from one language to another language, increased accuracy and speed of translation in conditions of external noise and interference.

An additional technical result is an increase in the secrecy of communication from the possible eavesdropping of a conversation by outsiders. The possibility of operational spatial reorientation of the communication channel to another person - an interlocutor or a group of people, the ability to switch to communication mode via radio, ultrasonic or infrared communication channels.

In a known electronic device for automatically translating oral speech from one language to another, comprising a housing in which at least one computer is located with an autonomous power source and software that allows the computer to carry out the input of an acoustic signal in an input language, processing this signal in A computer and converting it into an output acoustic signal in an output language; an input device for an acoustic signal in an input language in a computer connected to the first input of the computer; the output device from the computer of the acoustic signal in the output language connected to the first output of the computer. According to the invention, the computer software is configured to input and process acoustic signals in the input language and the output language into a computer and to translate at least one input language into at least one output language and vice versa to realize the possibility of human communication, speaks an input language equipped with the above electronic device, with a person who does not have such a device or has a device in the off state, and who speaks the output language, wherein the electronic device is provided with at least one additional input device for the computer of the acoustic signal, which is all kinds of sounds, the sources of which are acoustic signals created by a person speaking an input language, sounds created by a person speaking an output language, and possible acoustic noise and noise from third-party sound sources, at least one additional input device of the acoustic signal in the computer is in the form of a microphone, place it in the housing or other nodes of the electronic device on the body or clothing of man, the speaker in the source language, and it is connected to the corresponding auxiliary input computer; at least one additional device for outputting from the computer an acoustic signal in the input language, which is performed in the form of at least one earphone, designed to be placed on the head of a person who speaks the input language and is transferred to the hearing organ of this person, transferred to the computer from the output language to the input language of the acoustic signal generated by a person speaking the output language, an additional output device from the computer of the acoustic signal in the input language is connected to the corresponding additional output A VM, a computer input device of an acoustic signal in an input language, and at least one additional input device of an acoustic signal in a computer are spaced apart from each other in space, and an input device of an acoustic signal in a computer in the input language is configured to receive an acoustic signal generated mainly a person speaking an input language, for example, perform it in the form of a headset microphone, laryngophone or microphone in contact with a person’s neck or head or placed in a short In close proximity to the mouth of this person, the output device from the computer of the acoustic signal in the output language is placed in the body of the electronic device for automatically translating oral speech from one language to another or make it portable with the possibility of attaching it to the clothes or body of a person who speaks the input language. Possible implementations of the method such that:

- at least one additional input device for the computer of the acoustic signal is performed with a narrow radiation pattern to realize the possibility of receiving acoustic signals coming mainly from the location of the person speaking the output language, for example, by placing a narrowly focused microphone on the earphone or clothing of the person speaking in the input language, and the orientation of its radiation pattern in the direction of the orientation of the head or body of a person speaking the output language;

- at least one additional input device for the computer of the acoustic signal is performed with a wide radiation pattern to realize the possibility of receiving acoustic signals mainly from all sides, to realize the possibility of receiving audio messages from a person speaking the output language, if his location is unknown, for example, his it is not visible at night, or when a person speaking the input language, blind or several people speaking the output language participate in communication;

- an additional input device for the computer of the acoustic signal is performed with a changing radiation pattern, for example, in the form of two microphones with a narrow and wide radiation pattern, and computer software is configured to switch the radiation pattern of an additional device, for example, by selecting a signal from additional computer inputs of the signal with a microphone with a narrow or wide radiation pattern, for example using voice control commands;

- the device for outputting from the computer an acoustic signal in the output language as an emitter comprises a loudspeaker operating at sound frequencies;

- the loudspeaker is performed as an active device comprising a signal power amplifier, a speaker and an autonomous power source;

- the output device from the computer of the acoustic signal in the output language contains the nodes of the ultrasonic parametric system with emitters of ultrasonic signals, which are designed to generate sound signals in space in the region of the head of a person speaking the output language as a result of beating of signals at high-power ultrasonic frequencies, for which air the medium is a medium with nonlinear properties;

- the computer software is configured to automatically select the input and / or output language of the translation depending on the results of the analysis of the acoustic signal in the input language supplied to the input of the computer from the output of the device for input into the computer;

- the computer software is configured to process the signals received by the computer from the input device of the acoustic signal in the input language and / or at least one additional input device of the acoustic signal to the computer to control at least the translation start and end processes at least one input language, at least one output language, and vice versa;

- computer software is capable of processing signals received by a computer from an input device of an acoustic signal in the input language and / or at least one additional input device of the acoustic signal to the computer for controlling in time the processes of generating signals on the computer outputs, incoming to the output device from the computer of the acoustic signal in the output language and signals arriving at least one additional output device from the computer of the acoustic signal in the input language;

- to improve the accuracy of translation from the output language into the input language, the computer software is capable of processing electrical signals received by the computer from the input device of the computer an acoustic signal in the input language, and signals received by the computer from at least one additional device input into the computer an acoustic signal to compensate for the spectrum of the signal received by the computer from an additional input device, components of the acoustic signal generated by a person speaking an input language, and use Hovhan thus obtained is purified by interference electric signal for its subsequent transfer via a signal output from the computer language to language input;

- to increase the accuracy of translation from the output language into the input language, the computer software is capable of processing signals received by the computer from at least one additional input device of the acoustic signal to the computer, and the signal from the computer output intended for its supply to the device the output from the computer of the acoustic signal in the output language to compensate for the spectrum of the signal received by the computer from an additional input device into the computer of the acoustic signal of the components of the acoustic signal generated by the device ystvom acoustic signal output in the target language;

- to increase the signal-to-noise ratio in the region of the hearing organ of a person speaking an input language, a signal supplied to an additional output device from the computer of an acoustic signal in the input language is generated in the computer as a result of summing up the electrical signal received during the translation of the message from the output language to the input language and the components of the processed electrical signals contained in the spectrum of the signal supplied to the computer from an additional input device into the computer of the acoustic signal;

- as an electronic device for the automatic translation of oral speech from one language to another, use a cell phone, smartphone, tablet computer, laptop, netbook or other portable, mobile electronic device;

- as an electronic device for the automatic translation of oral speech from one language to another, a mobile electronic device, mainly a cell phone, smartphone, tablet computer, laptop, netbook, can be used, and the software of this device can be configured to synchronize the operation of such devices between two or more users and the transmission of acoustic signals generated by people not through an open acoustic communication channel, but through a radio channel, ultrasound howling or infrared communication channel with the subsequent reception of these signals and simultaneous translation of received messages from one language to another in each of the electronic devices.

These advantages, as well as features of the present invention will become apparent during consideration of the following possible embodiments of the invention with reference to the accompanying drawings.

A brief description of the drawings.

In FIG. 1 shows the main components of an electronic device for the automatic translation of oral speech from one language into another language.

In FIG. Figure 2 shows one of the possible algorithms for the processing of signals received by an electronic device for automatically translating oral speech from one language to another language.

The best embodiment of the invention.

A variety of machine translation algorithms from one language to another language and portable devices for carrying out such a translation are currently developed. They have a common drawback, namely, that when using only one electronic device, the process of communication using such a device is not entirely convenient, since it is necessary to pronounce phrases loudly enough while communicating. And in essence, it is possible to take turns entering information and translating it from one language to another, and vice versa, which slows down the process of communication between people speaking different languages, and in many situations it is almost impossible.

The present invention allows to eliminate this drawback and provide a mode of communication between people in the most natural way, when people speak a language they understand and hear the answers of the interlocutor also in a language they understand in real time, as in the usual communication of people speaking the same volume same language they understand. In this case, the electronic device provides:

- reception of acoustic signals of speaking people in different languages;

- special processing of these signals for the organization in time of the processes of translation and output of these signals into space;

- the formation of acoustic signals near people’s hearing organs, mainly in a language they understand.

The mode of imitation of communication between people speaking different languages, using one electronic device in comparison with analogues is achieved by:

1) introducing into the electronic device a number of new functional (additional) nodes for input and output of acoustic signals generated by speaking people in different languages (in input and output languages);

2) certain conditions for the spatial arrangement of these devices for input and output signals on the clothes or body of at least one person who is equipped with an electronic device;

3) forming in a certain way radiation patterns for signal input and output devices and providing spatially selective modes of input and output of acoustic information;

4) new algorithms for organizing the process of outputting generated signals in time, taking into account the analysis of the current parameters of the input information signals from various input devices;

5) new algorithms for organizing the processing of received and output signals that allow electronic suppression of various sources of acoustic noise and interference in relation to useful, informative signals generated for communication and / or in relation to electrical signals intended for their subsequent recognition and translation into another language;

6) due to a new type (physical nature) of signals output to users, which allow to realize a number of new functional capabilities of the device, as well as new communication channels when using the device, and thereby increase the usability of the device with the ability to use various messaging channels .

To illustrate the invention, we use FIG. one.

In FIG. 1 shows the main components of an electronic device for automatically translating oral speech from one language to another language, as well as the sound signals received and generated by this device.

An electronic device for automatically translating oral speech from one language to another comprises a housing 1, in which at least one computer 2 with an independent power supply 3 and software is located, which allows the computer to carry out the input of acoustic signal 4 into the input language, processing this signal 4 in the computer 2 and converting it into an output acoustic signal 8 in the output language.

As an electronic device for the automatic translation of oral speech from one language to another, a specialized, narrowly functional device can be used, for example, like an electronic translator like Multilux 2 or almost any modern cell phone, smartphone, tablet computer, laptop, netbook or other portable, mobile an electronic device containing a computer 2, configured to download the appropriate software, power supply 3 and other nodes, which will be discussed below.

Like the prototype (RF patent for application No. 2008137394/22, 09/19/2008, G06F 17/28), an electronic device for automatically translating oral speech from one language to another also contains an acoustic signal input device 6 for inputting 4 computers into the input language, connected to the first input of the computer 2, the output device 7 from the computer 2 of the acoustic signal 8 in the output language, connected to the first output of the computer 2.

In contrast to the prototype, the computer software 2 of the device is configured to input and process in computer 2 in real time various acoustic signals (4, 12, 13) and translate the received acoustic signals and output them from at least one input language to at least one output language, and vice versa, as they are input, to realize the possibility of communication between a person 9 who speaks an input language, equipped with the above electronic device, with a person 10 who does not have such a device and speaks in the output language. Or when person 10 has a similar device, but it is in the off state at the time when person 9 is accessing it.

According to the invention, the electronic device is equipped with at least one additional device 11 for inputting into the computer 2 an acoustic signal, which is all kinds of sounds, the sources of which can be acoustic signals 4 created by a person 9 speaking an input language, sounds 12 created by a person 10, speaking the output language. As well as possible acoustic interference and noise 13 from third-party sources of 14 sounds.

In the simplest embodiment of the device, at least one additional device 11 for inputting an acoustic signal into the computer 2, made in the form of a microphone, can be used.

This microphone can be placed in the housing 1 of the electronic device (for example, use the standard microphone of a smartphone, iPhone or tablet computer).

The microphone 11 can be structurally located on other nodes of the electronic device (as shown in Fig. 1) or mounted on the body or clothing of a person 9 (for example, using clothespins) speaking an input language.

In more complex embodiments, there may be several such additional devices 11. For example, some of them can be used as sounding devices for receiving predominantly noise signals 14 from sources 13 of possible acoustic noise and interference, and some can serve as sounding microphones necessary for implementing algorithms for automatically actively reducing noise and interference against useful acoustic signals and structurally located as close as possible to the user's hearing organ (15) (to the eardrum).

In an embodiment, at least one additional input device 11 for inputting an acoustic signal into the computer 2 can be made with a narrow radiation pattern to realize the possibility of receiving acoustic signals 12 coming mainly from the location of a person 10 who speaks the output language. This is one of the essential features that allows you to increase the communication distance between people and reduce the proportion of noise and interference in the received acoustic signal.

For example, the device 11 can be placed on the earpiece or clothing of a person 9 who speaks an input language. And his radiation pattern can be oriented towards the head or body of a person 10 who speaks the output language, when person 9 turns his head or body towards the interlocutor 10.

An embodiment of an electronic device is also possible, in which at least one additional input device 11 of the acoustic signal computer 2 is performed with a wide radiation pattern for realizing the possibility of receiving acoustic signals mainly from all sides, for realizing the possibility of receiving audio messages from a person 10, who speaks the output language, if his location is unknown, for example, he is not visible at night, or when person 9, speaking the input language, is blind. Or when several people who speak the output language participate in the communication, and it is desirable to organize the process of receiving messages from all participants in the communication 10 with person 9.

The additional input device 11 for input into the computer 2 of the acoustic signal can also be made with a changing radiation pattern, for example, it can be made in the form of two microphones with a narrow and wide radiation pattern, and the computer software 2 is configured to switch the radiation pattern of the additional device 11, for example, by selecting the acoustic signal introduced into the computer 2 either from a microphone with a narrow or wide radiation pattern. Such a switching of the radiation pattern of the device 11 can be carried out using, for example, voice control commands or by pressing the corresponding button in the preset menu of device operation modes.

The additional device 11 is connected to the corresponding at least one additional input of the computer 2.

The electronic device is also equipped with at least one additional device 14 for outputting from the computer 2 an acoustic signal 5 in the input language, which is performed in the form of at least one earphone. It is placed on the head of a person 9 who speaks an input language, and with his help they transmit to his organ of hearing 15 a signal 12 translated into a language that he understands, created by person 10 who speaks the output language.

An additional device 14 for outputting from the computer 2 an acoustic signal in the input language is connected to the corresponding additional output of the computer 2.

The input device 6 in the computer 2 of the acoustic signal 4 in the input language and at least one additional input device 11 in the computer 2 of the acoustic signal is spaced relative to each other in space. Moreover, the input device 6 in the computer 2 of the acoustic signal 4 in the input language is configured to receive the acoustic signal 4, created mainly by a person 9 who speaks the input language. These are also essential features of the invention, which allow to implement a number of new modes of operation of the device compared to analogues and solve the problem of the invention.

The device 6 can be made in the form of a headset microphone (as shown in Fig. 1), a laryngophone or microphone in contact with the neck or head of a person 9, or placed in the immediate vicinity of the mouth of that person 9. The directional pattern of the microphone 6 can be oriented to the side human mouth 9.

There are various options for the placement and execution of the output device 7.

The device 7 of the output from the computer 2 of the acoustic signal 8 in the output language as the emitter 17 may include a loudspeaker operating at audio frequencies, for example, in the form of a regular smartphone loudspeaker.

If an electronic device, such as an iPhone or a smartphone, does not contain a powerful loudspeaker that provides a sufficiently loud and high-quality sound signal, then the loudspeaker can be made as a separate active device containing a signal power amplifier 16, a speaker with increased power 17 and, for example, an autonomous power source 18, resembling a sound column for a smartphone, but in a flat case and with elements of its attachment to clothing or the human body 10.

As an embodiment of such a device 7, a more powerful device 7 can be used - a powerful megaphone, with which you can communicate with a large enough remote group of people, for example, foreign tourists, football fans, refugees during sports, peacekeeping, military or police operations.

In an embodiment, the device 7 for outputting from the computer 2 the acoustic signal 8 in the output language may contain special nodes of the ultrasonic parametric system with emitters of high-power ultrasonic signals instead of sound speakers. Such ultrasonic emitters are designed to generate sound signals in the region of the head of a person 10, speaking an output language as a result of the beating of two signals of high-power ultrasonic frequencies, for which, at a certain signal power (power flux density of the ultrasonic signal), the air in the region of the head of a person 10 with nonlinear properties.

Such an ultrasonic emitter can be made flat and placed in the side pocket of a shirt, jacket, or attached to clothes or headgear with a clothespin, elastic or strap.

When using such a parametric ultrasonic emitter, it is possible to form a fairly narrow sector of the voiced space in which one or another person can hear the signal generated for him in the output language. At the same time, people who are outside the zone (diagram) of the signal formation will not be able to hear the transmitted message in the output language. The transmission range of such a signal can be adjusted and range from half a meter to 3-5 meters.

A powerful ultrasonic megaphone working on the same principle can be used to ensure a long range of signal transmission. Using devices of this type 7, it is possible to provide highly directional signal translation in the output language by tens or even hundreds of meters and organize a series of fundamentally new communication options. For example, a foreigner with such a device may covertly seek help from a police officer or other person if, for example, he is in danger of being surrounded by hostile persons. Police officers will be able to covertly warn hostages in their native language of their further actions, etc.

To receive a response message from long distances as an additional device 11, you can use a microphone gun.

If a person 10 who speaks the output language has the same electronic device, then when it is turned on, it is possible to organize communication in various additional modes that further expand the functionality of the claimed electronic devices:

- messaging with receiving messages via the ultrasonic communication channel. In this case, the device should contain an ultrasonic signal presence detector (microphone) and, for example, an SSB receiver of ultrasonic signals, which is switched on, for example, if the level of the ultrasonic signal with SSB modulation exceeds a predetermined threshold;

- messaging can be organized as in an analogue (RF patent for application No. 2008137394/22) over a radio channel using, for example, a cellular system of any well-known Wi-Fi or Bluetooth communication standard. In this case, the regular (transceiver) nodes of the smartphone with software that synchronizes the operation of electronic devices can be used.

The operation of an electronic device for the automatic translation of oral speech from one language into another language can be organized by various algorithms.

To implement a mode of simulating communication between people, as if they are communicating in their native language, it is advisable to use auxiliary algorithms for preliminary analysis and processing of input signals received in computer 2 from devices 6, 11. And to organize your own simultaneous process of simultaneous translation from one language into another language, you can use well-known algorithms for recognizing sound speech signals and translating them into another language and synthesizing the translated signal.

It is possible to use a number of service functions and additional signal processing algorithms in the software that allow the organization of communication in conditions of strong external noise and noise in the modes of active noise suppression and cleaning of received signals from extraneous signals, which are interference and increase the likelihood of failure in recognizing signals and their subsequent translation.

It is also possible to use algorithms in the signal processing process to generate automatic prompts from the computer (computer 2), about the need to repeat your message for a person 9 or 10, if the level of external noise and interference is very large.

The software may include routines for analyzing voice commands for controlling various functional capabilities of an electronic device, programs for organizing preliminary voice calibration of various signal input and output devices, setting matched filters necessary for implementing algorithms for active reduction of noise or interference, and verbal verifications speech of the owner. As well as programs for using certain databases of electronic dictionaries and how to access them from an electronic device. For example, in the form of access to internal or external memory cards or in the form of wireless access programs for devices to virtual - cloud databases or to a remote powerful computer - translator (server).

To explain the principle of operation of one of the possible options for an electronic device, we use FIG. 2.

The operation of an electronic device begins with its inclusion, for example, by launching the corresponding program. For example, programs in the form of an application to a smartphone or cell phone with initial, factory settings for the parameters of the analyzed signals, possible service modes and languages for translation.

In particular, for the convenience of using an electronic device, the computer software 2 can be configured to automatically select the input and / or output language of the translation depending on the results of the analysis of the acoustic signal in the input language 4 received at the input of the computer 2 from the output of the device 6 to enter it into the computer 2.

Such voice control commands can be typical phrases, for example, “Translation from Russian into English” or “Translation from Russian into German”. These commands received by computer 2 will be perceived as programming the catch of an automatic translator, in which the input language is Russian, and the output language is English or, for example, German.

The computer software 2 can be arranged to process the signals received by the computer 2 from the input device 6 of the computer 2 of the acoustic signal 4 in the input language and / or at least one additional device 11 of the input of the acoustic signal to the computer 2 to control time processes of the beginning and end of the translation of at least one input language to at least one output language and vice versa.

For example, in the process of processing signals received in computer 2 from device 6 and device 11, an automatic analysis of the levels of these signals averaged over some analysis time can be carried out.

If at time moment T1 the signal level changes during the AT analysis from the corresponding microphone (for example, in 0.1-0.2 seconds) over a certain threshold (for example, above the level of 40 dB), then a decision is made to launch the appropriate translation program from input to output language or vice versa. And this signal recorded in the pipelined memory of the device begins to be immediately translated or examined for compliance with its control command. In an embodiment, the translation can be carried out after filtering and cleaning the signal from noise (in Fig. 2 these operations are not shown).

It is clear that there can be various options for constructing signal generation algorithms so as to minimize mutual interference between the emitted sound signals and the received sound signals. The presence of at least two input devices for acoustic signals reduces the time required for their translation, so communication between people will occur with shorter pauses, more realistic.

For example, if at some point in time T1 the signal level from the input device 6 in the computer 2 (during the analysis time of the order of 0.1-0.2 sec) sharply increased, and then after a certain analysis time ΔT = T2-T1 (for example, through 0.5-1.5 sec) sharply decreased, the analysis program for control commands is turned on. If the AT signal received during this time can be identified, for example, as one of the control commands, for example, as a translation command from Russian into German, then the possible next phrase of a person 9 who speaks the input Russian language and the translation of this phrase into German language without another language type recognition.

If the analysis of the message cannot be identified as a control command, then the message recorded in the memory is transferred to the computer 2 in accordance with the previously set translation direction command, which was previously identified or set in manual settings.

This translated phrase is formed in the form of an audio signal 8 in the output language and, using, for example, a loudspeaker 17, reaches the hearing organs of a person 10 who speaks the output (German) language.

Having thus received a message in the output language, a person 10 can respond in the output language and his signal 12 will be received using the device 11. This signal will be processed in the computer 2 and translated from the output language into the input language according to a similar scheme. And with the help of headphones 14 transferred to the organ 15 of the ear of a person 9 with the minimum possible delay. In the variant, practically in simultaneous translation mode. And so on.

If the implementation of the sound signal from the output of the device 6 or 11 could not be translated, then it may be decided that this was a hindrance. For example, person 9 accidentally sneezed or coughed. Or a person 10, speaking the output language, answered too quietly or inarticulate against the background of noise and interference. In this case, a request signal can be generated to repeat the corresponding phrase, which can be automatically generated in the computer 2 and transmitted via headphones or a speaker 17 to a person 9 or 10 (these operations are not shown in Fig. 2). And so on.

If a person 9, while broadcasting his translated phrase in the output language, without waiting for its end, starts talking again, or vice versa, starts to do something additional, not speak one person at a time, then the corresponding signal from the output of device 6 or 11 sharply increased in computer 2, a ban is formed on the emission of the corresponding signal in the output language, so that it does not create additional interference with the useful - received acoustic signals 4, 12. A blocked piece of the phrase will be played when the level of the corresponding signal ala decreases below the threshold.

The above energy features of acoustic signals generated by people during a conversation, allow us to formulate the following group of essential features of the invention: computer software 2 can be configured to process signals received in computer 2 from input device 6 in computer 2 of acoustic signal 4 in the input language and / or at least one additional device 11 for inputting into the computer 2 an acoustic signal for controlling in time the processes of generating at the computer outputs 2 signals received 7 and the device output from the computer 2 8 acoustic signal in the target language, and signals to at least one additional output device 14 of the computer 2 to 5 of the acoustic signal input language.

To improve the accuracy of translation from the output language to the input language, the computer software 2 can be configured to process electrical signals received by the computer 2 from the input device 6 of the computer 2 acoustic signal 4 in the input language and the signals received by the computer 2 s, at least one additional device 11 for input into the computer 2 of the acoustic signal to compensate in the spectrum of the resulting processed signal received in the computer 2 from the additional device 11, the components of the acoustic signal generated by a person 9 who speaks an input language, and using the electrical signal thus obtained, cleared of interference, to implement, with its help, a subsequent translation into the computer 2 of the signal from the output language into the input language.

To improve the accuracy of translation from the output language to the input language, the computer software 2 is capable of processing signals received by the computer 2 from at least one additional device 11 for input into the computer 2 an acoustic signal and a signal from the output of the computer 2, intended for its supply to the device 7 output from the computer 2 acoustic signal 8 in the output language to compensate for the spectrum of the signal received by the computer 2 from the additional device 11, the components of the acoustic signal generated by the device 7 output acoustic signal in the output language.

To increase the signal-to-noise ratio in the region of the hearing organ of a person 9 who speaks the input language, the signal supplied to the additional device 14 for outputting from the computer 2 an acoustic signal in the input language is formed in the computer 2 by summing the electrical signal received during the translation of the message from the output language to the input language, and the components of the processed electrical signals contained in the spectrum of the signal supplied to the computer 2 from the additional input device 11 of the computer 2 acoustic signal, which essentially y are noise and interference.

Devices for automatically reducing the noise components in the signal spectrum are described in detail in the above group of patents (for example, in patent No. 2326495 C2, H04B 1/00, 01/06/2004). The essence of these algorithms is that, using multi-band processing of the extracted noise signal and the useful signal, noisy by the components of the noise signal, they automatically select the levels and phases of the noise components so that they, when combined with the components of the noisy signal, compensate for the noise components in it.

As an electronic device for the automatic translation of oral speech from one language to another, a mobile electronic device, mainly a cell phone, smartphone, tablet computer, laptop, netbook, can be used. And the software of this device can be configured to synchronize the operation of such devices between two or more users and transmit acoustic signals generated by people not through an open acoustic communication channel, but through a radio channel, an ultrasonic or infrared communication channel, with the subsequent reception of these signals and simultaneous translation of received messages from one language to another in each of the electronic devices.

Industrial applicability

The invention is applicable in the electronic industry as automatic, mobile translators from one language to another.

Claims (15)

1. An electronic device for automatically translating oral speech from one language to another, comprising a housing in which at least one computer is located with an autonomous power source and software that allows you to perform the functions of inputting an acoustic signal into the computer in the input language, processing this a signal in a computer and converting it into an output acoustic signal in the output language; an input device for an acoustic signal in an input language in a computer connected to the first input of the computer; a computer output device of an acoustic signal in an output language connected to the first computer output, characterized in that the computer software is configured to input and process acoustic signals in the input language and output language into the computer and to translate from at least one input language into at least one output language and vice versa to realize the possibility of communication between a person speaking an input language equipped with the above electronic device and a person who does not have such a device your TV or having a device in the off state and speaking the output language, while the electronic device is equipped with at least one additional input device for the computer of an acoustic signal, which is all kinds of sounds, the sources of which are acoustic signals created by a person speaking in the input language, sounds created by a person speaking the output language, as well as possible acoustic noise and noise from third-party sound sources, at least one additional A new input device for the computer of the acoustic signal is performed in the form of a microphone, placed in a case or other components of an electronic device, on the body or clothes of a person speaking an input language, and connected to the corresponding additional input of the computer; at least one additional device for outputting from the computer an acoustic signal in the input language, which is performed in the form of at least one earphone, designed to be placed on the head of a person who speaks the input language, and transfer to the hearing organ of this person transferred to the computer from the output language to the input language of the acoustic signal created by a person speaking the output language, an additional output device from the computer of the acoustic signal in the input language is connected to the corresponding additional output A computer, an input device for a computer of an acoustic signal in an input language, and at least one additional input device for a computer of an acoustic signal are spaced apart from each other in space, and an input device for an acoustic signal in a computer in an input language being configured to receive an acoustic signal generated mainly a person speaking an input language, for example, perform it in the form of a headset microphone, laryngophone or microphone in contact with the neck or head of a person or placed in edstvennoy vicinity of the mouth of the person, an output device from the acoustic signal of computers in the target language are placed in the electronic device housing to automatically translate speech from one language to another, or make it a remote chance of its fastening to the clothes or the human body, the speaker in the input language. 2. The electronic device according to claim 1, characterized in that at least one additional input device of the acoustic signal to the computer is performed with a narrow radiation pattern to realize the possibility of receiving acoustic signals coming mainly from the location of the person speaking the output language, for example, by placing a narrowly focused microphone on the earphone or clothes of a person speaking the input language, and orienting his radiation pattern in the direction of the orientation of the head or body of the person shouting in the output language. 3. The electronic device according to claim 1, characterized in that at least one additional input device of the acoustic signal to the computer is performed with a wide radiation pattern to realize the possibility of receiving acoustic signals mainly from all sides, to realize the possibility of receiving audio messages from a person who speaks the output language, if his location is unknown, for example, he is not visible at night, or when a person who speaks the input language, is blind, or several people participate in communication in the output language. 4. The electronic device according to any one of paragraphs. 1-2, characterized in that the additional input device in the computer of the acoustic signal is performed with a changing radiation pattern, for example, in the form of two microphones with a narrow and wide radiation pattern, and the computer software is configured to switch the radiation pattern of an additional device, for example, by selecting signal from additional computer inputs of a signal from a microphone with a narrow or wide radiation pattern, for example, using voice control commands. 5. The electronic device according to claim 1, characterized in that the device for outputting from the computer an acoustic signal in the output language as a radiator comprises a loudspeaker operating at sound frequencies. 6. The electronic device according to claim 1, characterized in that the loudspeaker is performed as an active device comprising a signal power amplifier, a speaker and an autonomous power source. 7. The electronic device according to claim 1, characterized in that the device for outputting from the computer an acoustic signal in the output language contains nodes of an ultrasonic parametric system with emitters of ultrasonic signals, which are designed to generate sound signals in space in the region of the head of a person speaking the output language, as a result of beating of signals at high-power ultrasonic frequencies, for which the air medium is a medium with nonlinear properties. 8. The electronic device according to claim 1, characterized in that the computer software is configured to automatically select the input and / or output language of the translation, depending on the results of the analysis of the acoustic signal in the input language supplied to the input of the computer from the output of the device for entering it into a computer. 9. The electronic device according to claim 1, characterized in that the computer software is capable of processing signals received by the computer from the input device of the acoustic signal in the input language and / or at least one additional input device into the acoustic computer a signal for controlling in time the processes of the beginning and end of translation from at least one input language to at least one output language and vice versa. 10. The electronic device according to claim 1, characterized in that the computer software is capable of processing signals received by the computer from the input device into the computer of the acoustic signal in the input language and / or at least one additional input device into the acoustic computer a signal for controlling in time the processes of generating at the computer outputs the signals arriving at the output device from the computer an acoustic signal in the output language, and the signals arriving at at least one additional output device and h computer acoustic signal in the input language. 11. The electronic device according to p. 1, characterized in that to improve the accuracy of translation from the output language into the input language, the computer software is capable of processing electrical signals received by the computer from the input device of the acoustic signal in the input language into the computer, and signals, coming into the computer from at least one additional input device to the computer of the acoustic signal to compensate in the spectrum of the signal entering the computer from the additional input device, the components of the acoustic signal, creating consumed by a person speaking the input language, and using the electrical signal thus obtained, cleared of interference, for the subsequent translation of the signal from the output language into the input language using a computer. 12. The electronic device according to claim 1, characterized in that to increase the accuracy of the translation from the output language into the input language, the computer software is capable of processing signals received by the computer from at least one additional input device of the acoustic signal to the computer, and a signal from the output of the computer, intended for its supply to the output device from the computer of the acoustic signal in the output language to compensate for the spectrum of the signal received by the computer from the additional input device to the computer of the acoustic signal, components of the acoustic signal generated by the output device of the acoustic signal in the output language. 13. The electronic device according to claim 1, characterized in that in order to increase the signal-to-noise ratio in the region of the hearing organ of a person speaking an input language, a signal supplied to an additional output device from the computer of an acoustic signal in the input language is formed in the computer as a result summing the electrical signal received in the process of translating a message from the output language into the input language and the components of the processed electrical signals contained in the spectrum of the signal received by the computer from an additional input device in Computer acoustic signal. 14. The electronic device according to claim 1, characterized in that as an electronic device for automatically translating oral speech from one language to another, use a cell phone, smartphone, tablet computer, laptop, netbook or other portable mobile electronic device. 15. The electronic device according to claim 1, characterized in that as an electronic device for automatically translating oral speech from one language to another, a mobile electronic device can be used, mainly a cell phone, smartphone, tablet computer, laptop, netbook, and software of this device can be configured to synchronize the operation of such devices between two or more users and transmit acoustic signals generated by people not through an open acus an optical communication channel, and through a radio channel, an ultrasonic or infrared communication channel, with the subsequent reception of these signals and simultaneous translation of received messages from one language to another in each of the electronic devices.

Images