JPS6324360A - Electronic translator - Google Patents

Abstract

PURPOSE:To make effective translation by making electronic translation analyzing input voice, detecting silent state using a part of output when uttering translated language from a speaker, and making translation command and input control when the time proportion level arrived at a fixed level. CONSTITUTION:Output of an input amplifier 13 is converted to character signals of original language by a voice analyzing circuit 14, and sent to an electronic translating circuit 15, and converted to character signals of translated language through a translation controlling circuit 22. The character signals are voice synthesized 17, amplified 18, and voice of translated language is uttered from a speaker 19, and at the same time character displayed 16. Time of silent state in which original language is lower than a fixed level is measured by a part of output of the amplifier 13 in a blank detecting circuit 21. Output of this circuit 21 increases in proportion to the silent time, and when exceeded a specified value, the circuit 22 is operated, and a translation command is given to the circuit 15. At the same time, whitches S2, S3 moves automatically to a contact (b), and checks input, and displays 23 that translation is going on.

Description

[Detailed description of the invention] The present invention relates to an electronic translation machine.

In this specification, translation refers not only to converting the language of one country into the language of another country, but also to converting letters and symbols into sounds, and sounds into letters and symbols in the language of the same country. Also say. Therefore, (hereinafter referred to as margin) translation is also the act of converting a dialect into a standard language, and conversely, converting a standard language into a dialect, even if the language is from the same country.

Furthermore, from this, calculation is also defined as a type of translation.

The source language refers to the language before translation, and the translated language refers to the language after translation. For example, in an electronic translator set to translate Japanese into English, "Good morning" is the source language and "GOODMORNING" is the translated language. As shown in FIG. 1, the electronic translator proposed by the present inventor in Japanese Patent Application Laid-Open No. 56-103765 applies an audio signal from a microphone 1 to an input amplifier 3 via an external input terminal 2. , is amplified to a level that allows speech recognition, is analyzed and recognized by the speech analysis circuit 4, is converted into a character signal, and is stored in the memory of the electronic translation circuit 5.

Next, by turning on the translation control switch S1 and emitting a translation instruction pulse, the character signal of the original language stored in the memory is translated into a character signal of the translated language by the electronic translation circuit 5 and output. The output of the translated language character signal is converted into a translated language audio signal by the speech synthesis circuit 7, and then applied to the output amplifier 8, and further applied to the speaker 9 via the external output terminal 10. speaker 9
converts the applied audio signal of the translated language into speech and pronounces it.

Furthermore, the original language and the translated language are configured to be displayed in characters by the character display circuit 6.

Next, an electronic translation configured like this The operation of the machine will be explained using FIG.

FIG. 3(A) shows the original language input signal, and FIG. 3(B) shows the original language input signal.
This is a translation instruction pulse for causing the electronic translation circuit 5 to perform a translation operation, and is generated by turning on the translation control switch S1.

Further, (C) is an output signal of the translated language.

That is, by applying the translation instruction pulse of the same (B) to the input signal of the original language of the same (A) "Anatawada Redesuka", the translated language of the same (C) "WHO ARE Y" is applied.
OU? '' output signal is sent out.

Now, after applying the audio input signal "Ana Tawada Redesuka" through the microphone 1, until the next audio input signal "Watashiwa Tomdesu" is applied, that is, between times t1 and t4, the translation control switch is switched. S1 is conducted, a translation instruction pulse is generated, and "anatawa daredesuka" is translated into one sentence (translation unit).
It is necessary to instruct the translation as follows. Otherwise, the next signal, ``watashiwa tom desu'', would also be a single sentence, ``ana tawa daredesu kawatashi wa tom desu''.
Translation is incomplete because it tries to translate as one sentence,
Or it becomes untranslatable. With this configuration, the operator had to carefully operate the translation control switch S1 to generate a translation instruction pulse for each category of the source language to be applied, which was very inconvenient because it was time-consuming. . Furthermore, if the operation is incomplete or forgotten, the translation will be incomplete or impossible.

Furthermore, if the operator is not familiar with the translation language, he or she will not be able to know that the translation is incomplete or impossible, so the operator will continue to operate assuming that the translation is complete. As a result, there was a drawback that misunderstandings occurred between the translator and the translator, and the original purpose of the translator could not be achieved. In addition, since the operator is preoccupied with speaking, operating the translation control switch S1 at each sentence break is a heavy burden and has the disadvantage of making the conversation unnatural. there were.

The present invention provides an electronic translation machine that eliminates these drawbacks and enables a more natural translation that is easy to use and easy to operate.

An embodiment of the present invention will be described below with reference to FIGS. 2 and 4. FIG. 2 is a block diagram of an electronic translation machine which is an embodiment of the present invention, and FIG. 4 is a timing chart of outputs from each part of the same machine.

In the figure, 11 is a microphone, and the microphone 11 is connected to an input amplifier 13 via an external input terminal 12 and a switch S2. The input amplifier 13 amplifies the audio signal applied via the microphone 11 to a level that allows the audio analysis circuit 14 to analyze and recognize the audio. Here, the speech analysis circuit refers to a circuit that performs spectrum analysis on an audio input signal, performs speech recognition, and converts it into a character signal. The output of the input amplifier 13 is converted into a character signal in the original language by the speech analysis circuit 14, and then sent to the electronic translation circuit 15 which performs translation according to instructions from the translation control circuit 22, and the character signal is converted into a corresponding translation. translated into written signals of the language. Next, this translated character signal output is converted into a speech signal of the translated language by the speech synthesis circuit 17, amplified by the output amplifier 18, and then applied to the speaker 19 via the external output terminal 20. . The speaker 19 pronounces the voice of this translated language. Further, the character display circuit 16 displays characters in the original language and the translated language.

On the other hand, a part of the output of the input amplifier 13 enters the blank detection circuit 21. The blank detection circuit 21 detects a state in which there is no audio input signal of the original language applied through the microphone 11, that is, a silent state, and measures the duration of the silent (no signal) state of the audio input signal. Note that the silent state refers to a state in which the level of the input signal is below a certain value. That is, in this embodiment, as shown in FIG. 4(B), since the output level (DC voltage) of the blank detection circuit 21 is configured to increase in proportion to the silent time, the longer the silent time is, the The output level of the blank detection circuit 21 becomes high.

However, the translation control circuit 22 has an input level (DC voltage)
It is activated when the value exceeds the specified value, and as shown in Figure 4 (C),
The electronic translation circuit 15 is configured to apply a translation instruction pulse. That is, this corresponds to turning on the translation control switch S1 of the conventional electronic translator, as shown in FIG. 3(B).

Therefore, as shown in FIG.
．． 2 seconds), the output level of the blank detection circuit 21 does not reach the specified value, so the translation control circuit 22 does not operate. Therefore, at this stage, the electronic translation circuit 15 does not perform a translation operation.

However, during a long silent period such as between sentences, for example from time t3 to t7, the output level of the blank detection circuit 21 reaches a specified value. Therefore, at time t4, the translation control circuit 22
is activated, and the electronic translation circuit 15 receives a translation instruction pulse (time t
4 to t5), the switches S2 and S3 are turned on to setting b, and the translation display lamp 23 is turned on as shown in FIG. 4(E). This allows the operator to know that the electronic translation circuit 15 is in translation operation.

Furthermore, since the switch S2 is electrically connected to the contact b, the translated language spoken from the speaker 19 is no longer applied to the electronic translation circuit 15 via the microphone 11 and translated. In addition, switches S2 and S3 are interlocked,
In addition, both switches are always electrically connected to contact a. Next, after the electronic translation circuit 15 completes the translation, the speaker 1
When the translated language is finished being pronounced from 9, the translation control circuit 22 returns both switches S2 and S3 to contact a. As a result, the translation display lamp 23 is turned off, and the input signal applied via the microphone 11 can be applied to the input amplifier 13.

Next, the operation will be explained in more detail.

As shown in FIG. 4(A), when the audio input signal "Anatawa Daredesuka" is applied from the microphone 11, the input signal enters the input amplifier 13 via the external input terminal 12 and the switch S2. . In the input amplifier 13,
The voice analysis circuit 14 spectrally analyzes this voice signal and amplifies it to a level sufficient to recognize the voice.

The output of the input amplifier 13 is converted into a character signal by a speech analysis circuit 14, and then sent to an electronic translation circuit 15 where it is temporarily stored. During this time, a part of the output of the input amplifier 13 is sent to the blank detection circuit 21. As shown in the same figure (A)
The audio signal of “Anatawa Daredeska” is “Anatawa” and “
The time t1 to t2 between "Ana Tawa" and "Dare Desuka" is composed of two voice signal parts: "Dare Desuka". Since there is usually a blank time of about 0.2 seconds in a conversation, the blank detection circuit 21 , same figure (B)
However, this level of output does not reach the specified value sufficient to operate the translation control circuit 22. Therefore, the translation control circuit 22 does not send out a translation instruction pulse to only translate the "anatawa" part.

Next, when the input signal "DAREDESKA" is input, the output of the blank detection circuit 21 returns to zero. However, since a silent state continues after this "daredesuka", the output of the blank detection circuit 21 increases almost in proportion to time, and after a certain period of time (about 1 second in this embodiment), the output of the blank detection circuit 21 increases to a specified value. , and then returns to its original state.

However, at this time, the translation control circuit 22 operates and sends a translation instruction pulse to the electronic translation circuit 15 as shown in FIG. Translation is performed as one translation unit, switches S2 and S3 are connected to contact b, the translation display lamp 23 is turned on, and the input terminal of the input amplifier 13 is connected to ground, and the connection between the microphone 11 and the input amplifier 13 is established. Cut off.

As a result, the electronic translation circuit 15 converts the character signal of the original language into "WHO ARE YOU?" as shown in FIG.
The character display circuit 16 translates into a character signal of the translation language
to be displayed. Furthermore, this output is transmitted through a speech synthesis circuit 17, an output amplifier 18, and an external output terminal 20.
and is applied to the speaker 19 through. Then, the speaker 19 pronounces the sound "Who are you".

When the output of the translated language is finished, the translation control circuit 22 returns to its original state, turns the switches S2 and S3 on to the contact a side, and displays the translation display lamp as shown in FIG. 23 is turned off.

Furthermore, regarding the audio input signal "Watashiwa Tom Death", there is a silent period of about 0.2 seconds between "Watashiwa" and "Tom Death", but the output of the blank detection circuit 21 is The specified value is not reached. However, since there is a long silent period after "Tom Death", the blank detection circuit 21
The output of reaches the specified value, activates the translation control circuit 22,
Sends a translation instruction pulse. As a result, "watashiwa tomdesu" can be translated as one translation unit.

Thereafter, through similar operations, the translated language "I am Tom" is pronounced from the speaker 19.

In this way, according to this embodiment, translation can be performed automatically by simply providing a silent period for each sentence, so that the effort of (1) operation can be saved.

Moreover, since such silent periods are created unconsciously during normal conversation, translation control can be performed without requiring any special force. In addition, (1) operations are not required, so even people with hand disabilities or those who are unable to operate manually can not only perform translation operations without any burden, but also eliminate translation errors due to operational errors. Furthermore, during translation, the connection between the input terminal of the input amplifier and the microphone is cut off, so the audio from the speaker is applied to the electronic translation circuit again via the microphone and is not retranslated. The operator can know the translation via the translation display lamp, so if he or she speaks while looking at it, he or she can enjoy the conversation with peace of mind, even if he or she does not have sufficient knowledge of the translation language.

Embodiments of the present invention include the following.

(i) An electronic translation machine that performs automatic translation every time there is a near-blank (silence) state for a certain period of time or more in the input signal of the source language.

(ii) An electronic translation device that is configured to translate the input signal of the source language applied between a blank period of a certain time or more that exists in the input signal of the source language and the next blank period as one unit of translation. (iii) A device comprising a means for temporarily storing an input signal in the source language and a means for detecting a blank (silent) part in the input signal in the source language, and which exists in the input signal in the source language for a certain period of time or more. An electronic translation machine that translates temporarily stored input signals for each blank space as one unit of translation.

(iv) comprising means for temporarily storing an input signal in the source language, means for detecting a blank portion in the input signal in the source language, and a switch circuit for connecting/disconnecting the input circuit of the electronic translation circuit; The temporarily stored input signal is translated as one unit of translation for each blank part of a certain time or more in the input signal, and the switch circuit is cut off from the blank time until the translation is completed. An electronic translator.

Furthermore, among the electronic translation machines currently on the market, there are some in which the type of translation language can be easily changed by replacing the translation LSI module, and the present invention can also be implemented with these.

Note that in this specification, the warning that input is not allowed in the warning means includes the following.

A. If the warning means is a visible display means; a. Turning off lamps, etc. when the lamp, etc. lights up when input is possible and turns off when input is not possible.

b. Lighting of lamps, etc., which turn off when input is possible, and turn on when input is not possible.

(Hereafter, margin) c. If the blue lamp etc. lights up when input is possible, and the red lamp etc. lights up when input is not possible, the red lamp etc. lights up or changes color to red.

d. Marks, symbols, characters, etc. are displayed or blinking when nothing is displayed when input is possible and marks, symbols, characters, etc. are displayed when input is not possible.

e. Non-display, where marks, symbols, characters, etc. are displayed when input is possible, and nothing is displayed when input is disabled.

B. If the warning means is an audible sounding means; a. A short sound (beep, beep, beep...
··Such).

b. There is no sound when input is possible, and a sound is emitted when input is disabled (continuous or discontinuous sound).

In addition, similar to the above, warnings that can be entered include turning off, lighting, blinking, discoloration of lamps, etc. in each location, displaying marks, symbols, letters, etc., displaying no display, pronunciation, voice display, etc.

Furthermore, synchronization in this specification includes not only synchronization in a strict sense, but also almost synchronization and following with a slight delay.

As described above, according to the present invention, while the translation means is performing a translation operation, the input blocking means is automatically operated to block input, and the warning means is synchronized with the input blocking means to automatically prevent input. Since the input control means is provided to issue a warning that input is not possible, unnecessary information (including noise) is input to the translation means via the input means during translation, and is mixed with necessary information to be translated, resulting in mistranslation. This can be automatically and reliably prevented without any operation of a key switch or foot switch, and a warning indicating whether input is possible or not (for example, a lamp) can be issued in synchronization with this. Anyone can obtain accurate translations in a timely manner, without any special force or attentiveness, just by looking at the lights (e.g., lights on) and typing.

Further, according to the present invention, since a warning that input is not possible is given from the start of translation to the end of translation, the user can know that the translation operation has started, and can eliminate anxiety. Furthermore, during the period, you will be warned that input is not possible, so you can easily time your input and operate with peace of mind.

If the present invention is applied to a voice input type electronic translator, it will be even more effective.

[Brief explanation of the drawing]

Figure 1 is a block diagram of a conventional electronic translation machine, and Figure 3 is a timing chart of the output of each part of the machine. The instruction pulse (C) in the same figure is the output signal of the translated language. FIG. 2 is a block diagram of an electronic translator which is an embodiment of the invention, FIG. 4 is a timing chart of the output of each part of the same machine, and FIG. (B) shows the output level of the blank detection circuit, (C) shows the translation instruction pulse, (D) shows the output signal of the translated language, and (E) shows the operating state of the translation indicator lamp. 1.11...Microphone, 2.12...External input terminal, 3.13...Input amplifier, 4.14...
...Speech analysis circuit, 5.15...Electronic translation circuit, 6
．． 16...Character display circuit 7.17...Speech synthesis circuit, 8.18...Output amplifier 9.19...Speaker, 10.20...External output terminal, 21...・
・Blank detection circuit, 22... Translation control circuit, 23.
...Translation display lamp, S1...Translation control switch S2, S3...Switch.

Claims (2)

[Claims] (1) an input means for inputting information about the source language to be translated; a translation means for translating into information about the corresponding translation language based on the information about the source language to be translated input into the input means; output means for outputting information regarding the translated language translated by the translation means; input blocking means for preventing information input to the input means from being input to the translation means; a warning means for warning whether input is allowed or not; and while the translation means is executing a translation operation, the input blocking means is automatically operated to block input, and the warning means is set to the input blocking means. An electronic translation device comprising input control means that synchronizes and automatically issues a warning that input is not possible. (2) The warning means is a visible display means, and
2. The electronic translation machine according to claim 1, wherein the input control means continues to maintain a warning display indicating that input is not possible while the translation operation is executed by the translation means.