CN108415904B - A dual-channel real-time translation method - Google Patents

Abstract

The invention discloses a dual-channel real-time translation method, which includes the following steps: 1) a communication connection between a smart wearable device, a smart terminal and a server; 2) setting the translation language; 3) the smart wearable device preferentially judges whether its microphone has pickup; 4) The intelligent terminal sends the voice file to the server; 5) The MCU controls the arithmetic module to control the voice recognition module to convert the voice file into a text file; 6) The MCU controls the arithmetic module to control the translation module to translate the text file to obtain the translated text file; 7) The MCU controls The operation module controls the speech synthesis module to synthesize the translated voice file according to the content of the translated text file; 8) The translated voice file is played through the speaker of the smart terminal or played through the speaker of the smart wearable device. The invention can utilize the voice collection and playback channels of the smart wearable device and the smart terminal to realize two-way real-time translation, and the translation cost is low, which meets the needs of users.

Description

A dual-channel real-time translation method

technical field

The invention relates to the technical field of translation, in particular to a dual-channel real-time translation method.

Background technique

As more and more people travel abroad, consumers are increasingly demanding real-time translation. At present, there are fatal problems in the design of translation smart wearable devices on the market. Either the TWS headset communication method is required, but the translation cost is extremely high; Either need to use professional translation tools to communicate with each other, but this way of translation language is very limited, it is difficult to achieve the purpose of accurate and rapid communication.

SUMMARY OF THE INVENTION

The technical problem to be solved by the present invention is to provide a dual-channel real-time translation method in view of the above-mentioned deficiencies in the prior art, which can fully utilize voice collection and processing methods such as speakers and microphones of smart wearable devices and speakers and microphones of smart terminals. The playback channel realizes free translation in multiple languages, which is simple and efficient, and can best meet the needs of users.

For solving the above-mentioned technical problems, the technical scheme of the present invention is:

A dual-channel real-time translation method is based on an intelligent wearable device, an intelligent terminal, a server and an application, wherein the intelligent wearable device and the intelligent terminal are provided with speakers and microphones, the application includes a translation language setting module, and the server includes an MCU Control arithmetic module, speech recognition module, translation module and speech synthesis module, described speech recognition module, translation module and speech synthesis module are all connected with MCU control arithmetic module, described translation method comprises the following steps:

1) The smart wearable device and the smart terminal are communicated and connected through the first wireless communication network, and the intelligent terminal is communicated and connected with the server through the second wireless communication network;

2) Open the app and set the translation language through the translation language setting module;

3) The smart wearable device preferentially judges whether its microphone picks up sound. If so, the smart wearable device stores the voice input content as the first voice file, and then sends the first voice file to the smart terminal through the first wireless communication network. No, the smart terminal judges whether its microphone picks up sound, and if so, the smart terminal stores the voice input content as the second voice file, and if not, returns to the smart wearable device to preferentially judge whether its microphone picks up sound;

4) The intelligent terminal sends the first voice file or the second voice file obtained in step 3) to the MCU control computing module of the server through the second wireless communication network;

5) The MCU control operation module controls the voice recognition module to convert the first voice file or the second voice file obtained in step 4) into the first text file or the second text file;

6) The MCU controls the arithmetic module to control the translation module to translate the first text file or the second text file obtained in step 5) to obtain the third text file or the fourth text file;

7) The MCU control operation module controls the speech synthesis module to synthesize the third speech file or the fourth speech file according to the content of the third text file or the fourth text file;

8) The MCU controls the computing module to send the third voice file or the fourth voice file to the smart terminal through the second wireless communication network, and then the third voice file is played through the speaker of the smart terminal or the smart terminal sends the fourth voice file through the first voice file. The wireless communication network is sent to the smart wearable device and played out through the speaker of the smart wearable device.

As a preferred solution, step 3) the microphone of the smart wearable device picks up the sound to form a first voice file, and then the smart wearable device sends the first voice file to the smart terminal through the first wireless communication network, and the microphone of the smart terminal sends the first voice file to the smart terminal. Pick up the sound to form a second voice file.

As a preferred solution, step 4) the intelligent terminal sends the first voice file and the second voice file to the MCU control computing module of the server through the second wireless communication network, and the MCU control computing module performs sound source filtering on the second voice file, A second voice file filtered by the audio source is obtained.

As a preferred solution, step 5) the MCU controls the arithmetic module to determine whether the first voice file obtained in step 4) is silent, if so, the MCU controls the arithmetic module to ignore the first voice file, if not, the MCU controls the arithmetic module to control the voice The recognition module converts the first voice file into a first text file; the MCU controls the calculation module to determine whether the second voice file filtered by the sound source obtained in step 4) is silent, if so, the MCU controls the calculation module to ignore the second voice file, if If not, the MCU controls the computing module to control the speech recognition module to convert the second speech file into the second text file.

As a preferred solution, the audio source filtering method in step 4) includes the following steps:

1) The MCU control operation module calculates the estimated value of the voice signal of the voice of the smart terminal microphone picking up the sound source of the smart wearable device microphone through mathematical modeling;

2) The MCU control operation module calculates the first voice signal value of the second voice file, and then the MCU control operation module subtracts the voice signal estimated value obtained in step 1) from the first voice signal value to obtain the second voice signal value;

3) The MCU controls the arithmetic module to re-convert the second voice signal value into a voice file and save it, so as to obtain the second voice file filtered by the audio source.

As a preferred solution, the first wireless communication network is a Bluetooth wireless communication network or a WIFI wireless communication network.

As a preferred solution, the second wireless communication network is a WIFI wireless communication network or a 4G communication network.

The beneficial effects of the present invention are as follows: the present invention can fully utilize voice collection and playback channels such as speakers and microphones of smart wearable devices and speakers and microphones of smart terminals, which can not only achieve the purpose of two-way real-time translation, but also reduce translation costs. And it can realize free translation in multiple languages, which is simple and efficient, and can best meet the needs of users.

Description of drawings

1 is a schematic structural diagram of an intelligent wearable device, an intelligent terminal, a server and an application of the present invention;

Fig. 2 is the translation flow block diagram of Embodiment 1 of the present invention;

Fig. 3 is the translation flow block diagram of Embodiment 2 of the present invention;

FIG. 4 is a block diagram of the audio source filtering process for the second audio file in Embodiment 2 of the present invention.

In the picture: 1-smart wearable device, 2-smart terminal, 3-server, 31-MCU control operation module, 32-speech recognition module, 33-translation module, 34-speech synthesis module, 4-speaker, 5-microphone.

Detailed ways

The structural principle and working principle of the present invention will be further described in detail below with reference to the accompanying drawings.

Figure 1-2 shows Embodiment 1 of the present invention, a dual-channel real-time translation method, based on a smart wearable device 1, a smart terminal 2, a server 3 and an application, the smart wearable device 1 and the smart terminal 2 are both set There are speakers 4 and microphones 5, the application includes a translation language setting module, the server 3 includes an MCU control arithmetic module 31, a speech recognition module 32, a translation module 33 and a speech synthesis module 34, the speech recognition module 32, translation The module 33 and the speech synthesis module 34 are both connected with the MCU control arithmetic module 31, and the translation method includes the following steps:

1) The smart wearable device 1 and the smart terminal 2 are communicatively connected through the first wireless communication network, and the smart terminal 2 is communicatively connected with the server 3 through the second wireless communication network;

2) Open the app and set the translation language through the translation language setting module;

3) The smart wearable device 1 preferentially judges whether its microphone 5 picks up sound, and if so, the smart wearable device 1 stores the voice input content as the first voice file, and then sends the first voice file to the smart terminal through the first wireless communication network 2, if no, the smart terminal 2 judges whether its microphone 5 picks up sound, if so, the smart terminal 2 stores the voice input content as a second voice file, if not, returns to the smart wearable device 1 to preferentially judge whether its microphone 5 is There is a sound pickup step; wherein, when the microphone 5 of the smart wearable device 1 is in a sound pickup state, the microphone 5 of the smart terminal 2 is in a dormant state; when the microphone 5 of the smart terminal 2 is in a sound pickup state, the smart wearable device 1 The microphone 5 is dormant.

4) The intelligent terminal 2 sends the first voice file or the second voice file obtained in step 3) to the MCU control computing module 31 of the server 3 through the second wireless communication network;

5) The MCU controls the arithmetic module 31 to control the voice recognition module 32 to convert the first voice file or the second voice file obtained in step 4) into the first text file or the second text file;

6) The MCU controls the arithmetic module 31 to control the translation module 33 to translate the first text file or the second text file obtained in step 5) to obtain the third text file or the fourth text file; wherein, the first text file and the second text file are are the languages before translation, and the third character file and the fourth character file are languages after translation.

7) The MCU controls the arithmetic module 31 to control the speech synthesis module 34 to synthesize the third speech file or the fourth speech file according to the content of the third text file or the fourth text file;

8) The MCU controls the arithmetic module 31 to send the third voice file or the fourth voice file to the smart terminal 2 through the second wireless communication network, and then the third voice file is played out through the speaker 4 of the smart terminal 2 or the smart terminal 2 sends the fourth voice file to the smart terminal 2. The voice file is sent to the smart wearable device 1 through the first wireless communication network and played out through the speaker 4 of the smart wearable device 1 .

The first wireless communication network is a Bluetooth wireless communication network, and the second wireless communication network is a WIFI wireless communication network. It should be noted that the first wireless communication network may be a WIFI wireless communication network, and the second wireless communication network may be a 4G communication network.

As shown in FIG. 3, Embodiment 2 of the present invention, a dual-channel real-time translation method, is based on a smart wearable device 1, a smart terminal 2, a server 3 and an application, and the smart wearable device 1 and the smart terminal 2 are both provided with speakers 4 and a microphone 5, the application includes a translation language setting module, the server 3 includes an MCU control arithmetic module 31, a speech recognition module 32, a translation module 33 and a speech synthesis module 34, the speech recognition module 32, the translation module 33 and the speech synthesis module 34 are all connected with the MCU control arithmetic module 31, and the translation method comprises the following steps:

1) The smart wearable device 1 and the smart terminal 2 are communicatively connected through the first wireless communication network, and the smart terminal 2 is communicatively connected with the server 3 through the second wireless communication network;

2) Open the app and set the translation language through the translation language setting module;

3) The microphone 5 of the smart wearable device 1 picks up the sound to form a first voice file, and then the smart wearable device 1 sends the first voice file to the smart terminal 2 through the first wireless communication network, and the microphone 5 of the smart terminal 2 picks up the sound. sound to form a second voice file;

4) The intelligent terminal 2 sends the first voice file and the second voice file to the MCU control computing module 31 of the server 3 through the second wireless communication network, and the MCU control computing module 31 performs sound source filtering on the second voice file to obtain the sound source filtering The second voice file after; wherein, the microphone 5 of the smart wearable device 1 adopts the dual-mic Beamfoming technology, which can ensure that only the voice of the person who speaks to the microphone 5 of the smart wearable device 1 is in the first voice file. The microphone 5 of 2 picks up sound in various ways, so it is necessary to filter the sound source of the second voice file.

As shown in FIG. 4 , the sound source filtering method includes the following steps: 1) The MCU control operation module 31 calculates, through mathematical modeling, a speech signal estimation value of the voice picked up by the microphone 5 of the smart wearable device 1 and the microphone 5 of the smart terminal 2 picked up by the sound source; 2 ) The MCU control operation module 31 calculates the first voice signal value of the second voice file, and then the MCU control operation module 31 uses the first voice signal value to subtract the estimated value of the voice signal obtained in step 1) to obtain the second voice signal value 3) The MCU controls the arithmetic module 31 to re-convert the value of the second voice signal into a voice file and save it, so as to obtain the second voice file filtered by the audio source.

5) The MCU control operation module 31 judges whether the first voice file obtained in step 4) is silent. If so, the MCU control operation module 31 ignores the first voice file. If not, the MCU control operation module 31 controls the speech recognition module 32 to The first voice file is converted into a first text file; the MCU control operation module 31 judges whether the second voice file filtered by the sound source obtained in step 4) is silent, if so, the MCU control operation module 31 ignores the second voice file, if not , then the MCU controls the arithmetic module 31 to control the speech recognition module 32 to convert the second voice file into a second text file; since this method is a real-time translation method when two people who speak different languages are in dialogue, generally speaking, at the same time There is only one voice file in the first voice file and the second voice file filtered by the audio source.

6) The MCU controls the arithmetic module 31 to control the translation module 33 to translate the first text file or the second text file obtained in step 5) to obtain the third text file or the fourth text file; wherein, the first text file and the second text file are are the languages before translation, and the third character file and the fourth character file are languages after translation.

7) The MCU controls the arithmetic module 31 to control the speech synthesis module 34 to synthesize the third speech file or the fourth speech file according to the content of the third text file or the fourth text file;

8) The MCU controls the arithmetic module 31 to send the third voice file or the fourth voice file to the smart terminal 2 through the second wireless communication network, and then the third voice file is played out through the speaker 4 of the smart terminal 2 or the smart terminal 2 sends the fourth voice file to the smart terminal 2. The voice file is sent to the smart wearable device 1 through the first wireless communication network and played out through the speaker 4 of the smart wearable device 1 .

The first wireless communication network is a Bluetooth wireless communication network, and the second wireless communication network is a 4G communication network. It should be noted that the first wireless communication network may be a WIFI wireless communication network, and the second wireless communication network may be a WIFI wireless communication network.

The above are only preferred embodiments of the present invention, and any minor modifications, equivalent changes and modifications made to the above embodiments according to the technical solutions of the present invention fall within the scope of the technical solutions of the present invention.

Claims (3)

1. A double-channel real-time translation method is characterized in that: based on intelligent wearing equipment, intelligent terminal, server and application, intelligence wearing equipment and intelligent terminal all are equipped with loudspeaker and microphone, use including translation language setting module, the server includes MCU control operation module, speech recognition module, translation module and speech synthesis module all are connected with MCU control operation module, the translation method includes following step:

1) the intelligent wearable device is in communication connection with the intelligent terminal through a first wireless communication network, and the intelligent terminal is in communication connection with the server through a second wireless communication network;

2) opening an application to set translation languages through a translation language setting module;

3) the intelligent wearable equipment preferentially judges whether a microphone of the intelligent wearable equipment picks up sound, if so, the intelligent wearable equipment stores the voice input content as a first voice file, and then the first voice file is sent to the intelligent terminal through a first wireless communication network, if not, the intelligent terminal judges whether the microphone picks up sound, if so, the intelligent terminal stores the voice input content as a second voice file, and if not, the intelligent wearable equipment returns to the step that the intelligent wearable equipment preferentially judges whether the microphone picks up sound; in the step, a microphone of the intelligent wearable device picks up sound to form a first voice file, then the intelligent wearable device sends the first voice file to the intelligent terminal through a first wireless communication network, and the microphone of the intelligent terminal picks up sound to form a second voice file;

4) the intelligent terminal sends the first voice file or the second voice file obtained in the step 3) to an MCU control operation module of the server through a second wireless communication network; in the step, the intelligent terminal sends the first voice file and the second voice file to an MCU control operation module of the server through a second wireless communication network, and the MCU control operation module filters a sound source of the second voice file to obtain a second voice file after the sound source is filtered; the sound source filtering method comprises the following steps: 1) the MCU control operation module calculates a voice signal estimation value of voice of a microphone pickup sound source of the intelligent wearable equipment picked up by the intelligent terminal microphone through mathematical modeling; 2) the MCU control operation module calculates a first voice signal value of a second voice file, and then the MCU control operation module subtracts the voice signal estimated value obtained in the step 1) from the first voice signal value to obtain a second voice signal value; 3) the MCU controls the operation module to convert the second voice signal value into a voice file again for storage, and then a second voice file with a filtered sound source can be obtained;

5) the MCU control operation module controls the voice recognition module to convert the first voice file or the second voice file obtained in the step 4) into a first character file or a second character file; step 5) the MCU control operation module judges whether the first voice file obtained in the step 4) is silent, if so, the MCU control operation module ignores the first voice file, and if not, the MCU control operation module controls the voice recognition module to convert the first voice file into a first character file; the MCU control operation module judges whether the second voice file filtered by the sound source obtained in the step 4) is silent, if so, the MCU control operation module ignores the second voice file, and if not, the MCU control operation module controls the voice recognition module to convert the second voice file into a second text file;

6) the MCU control operation module controls the translation module to translate the first character file or the second character file obtained in the step 5) to obtain a third character file or a fourth character file;

7) the MCU control operation module controls the voice synthesis module to synthesize a third voice file or a fourth voice file according to the content of the third text file or the fourth text file;

8) the MCU control operation module sends the third voice file or the fourth voice file to the intelligent terminal through the second wireless communication network, and then the third voice file is played through the loudspeaker of the intelligent terminal or the intelligent terminal sends the fourth voice file to the intelligent wearable device through the first wireless communication network and then the fourth voice file is played through the loudspeaker of the intelligent wearable device.

2. The dual-channel real-time translation method according to claim 1, wherein: the first wireless communication network is a Bluetooth wireless communication network or a WIFI wireless communication network.

3. The dual-channel real-time translation method according to claim 1, wherein: the second wireless communication network is a WIFI wireless communication network or a 4G communication network.

Images