TWM532593U - Voice-translation system - Google Patents

Description

Speech translation system

This creation is about a translation system, especially about a speech translation system.

At present, the machine translation process of converting a country's voice to another country's voice is to first recognize the voice into a text, then translate the text into the target voice, and then use speech synthesis technology to synthesize the target voice's text into the target voice. The voice contains a lot of information that is far richer than text messages, such as laughter, sigh and other tone, as well as prosody information such as duration, pitch and energy of each word or syllable. These messages are very helpful in understanding the true meaning of the speaker. However, the speech synthesized by speech synthesis technology only relies on the translated text message, so that many messages behind the speech are lost, which will result in insufficient translation accuracy.

As mentioned above, the creators of this creation have designed a speech translation system after years of research, to improve the lack of existing technology, and to enhance the implementation and utilization of the industry.

In view of the above-mentioned problems, the purpose of the present invention is to propose a speech translation system for solving the lack of the prior art.

Based on the above purposes, the present invention provides a speech translation system including a text import module, a voice transceiver module, a database creation module, a target language selection module, and a translation processing module. The text import module provides a plurality of text materials that are fed into a plurality of national languages having a plurality of vocabularies. The voice transceiver module receives a plurality of reference voices; the plurality of reference voices correspond to the plurality of text materials and have a plurality of reference voices; and the voice transceiver module receives the to-be-translated voices having a plurality of to-be-translated voices. The database building module connects the text import module and the voice transceiver module; the database building module compares the plurality of words and the plurality of reference audio signals, and establishes a voice text corresponding database according to the comparison result. The target language selection module provides an interface; the interface has a plurality of language options corresponding to the language of the plurality of text materials, to provide a selection of the target language from the plurality of language options through the interface. The translation processing module is connected to the voice transceiver module, the database creation module and the target language selection module; the translation processing module compares the plurality of to-be-translated audio with the plurality of reference audio in the corresponding database of the voice text, and compares the ratio For a plurality of reference audios that are matched, a plurality of reference audios, a plurality of vocabularies or a combination thereof corresponding to the target language are extracted from the phonetic correspondence database, and the translation results are output.

Preferably, the target language selection module can provide a plurality of target languages for selecting a plurality of countries from a plurality of language options through the interface; the translation processing module can output a plurality of translation results corresponding to the plurality of target languages.

Preferably, the translation processing module can compare the pitch, loudness, timbre, length, or any combination thereof of the plurality of reference audio in the plurality of to-be-translated audio and speech-to-text databases in the speech to be translated.

Preferably, the translation processing module can store a plurality of location data corresponding to a plurality of countries in a plurality of national languages; the translation processing module can retrieve the location data of the corresponding target language from the plurality of location data, and then can be based on the location data. The translation result is output to an external electronic device located in the country corresponding to the target language.

Preferably, the server may include a live chat system connected to the voice transceiver module, and the instant chat system may receive the voice emitted by the user during the chat and may establish a corresponding voice to be translated, and then output the voice to be translated to the voice transceiver. Module.

Preferably, the voice to be translated received by the instant messaging system may be from an external voice transceiver held by the user.

Preferably, the external sound transceiving device can be a wireless earphone device.

Preferably, the server may further include a translation result display system connected to the translation processing module; the translation result display system may display the translation result from the translation processing module to the user.

As described above, the speech translation system of the present invention uses the text import module and the voice transceiving module to separately collect the multi-language text files and the corresponding voice file materials required for establishing the database, in particular, these The voice file data can be a voice file recorded by the user himself. Then, the database building module can be based on the data to establish a database that can accurately identify the user's voice color for subsequent querying and translating the corresponding text. Finally, the target language selection module and the translation processing module can perform translation functions using the database. In this way, the creation can accurately convert the voice with the characteristics of the individual pronunciation into the function of the voice, the text or both corresponding to the desired language.

The embodiments of the speech translation system according to the present invention will be described below with reference to the related drawings. For the sake of understanding, the same components in the following embodiments are denoted by the same reference numerals.

Please refer to FIG. 1, which is a block diagram of a speech translation system according to a first embodiment of the present creation. As shown, the speech translation system 10 includes a text import module 100, a voice transceiver module 200, a database creation module 300, a target language selection module 400, and a translation processing module 500.

The administrator or user of the speech translation system 10 can pre-collect a large number of articles with vocabulary that may be used in subsequent translations, and if necessary, collect a plurality of national language texts. Then, using the text file import function provided by the text import module 100, for example, the text import module 100 can receive text files in file formats such as doc, txt, html, rtf, etc., to collect the collected plurality of text materials 110. The text is imported into the module 100. The plurality of text materials 110 may have a plurality of vocabulary words 111, and the plurality of vocabularies 111 may be a plurality of terms in different countries, such as the United Kingdom, France, Japan, Korea, etc., which are merely illustrative and not limited thereto.

In addition, the user or manager may collect a plurality of reference audios 211 corresponding to the plurality of text materials 110. For example, the administrator seeks a professional recording person to record a reference audio 211 having a standard pronunciation for a plurality of words 111; or, customarily, the user separately records a reference audio 211 having a personal characteristic pronunciation for the plurality of words 111. Next, the article voice transceiver module 200 can receive a plurality of reference voices 210 provided by the administrator or the user, wherein the reference voices 210 have a plurality of reference voices 211 representing the pronunciations of the different words 111.

For example, the voice transceiver module 200 can include a mono or stereo speaker and a microphone, etc., for example, a Bluetooth headset or other hands-free device that can transmit and receive audio signals through communication technologies such as Bluetooth and WiFi. The file format of the reference voice 210 that can be received by the voice transceiver module 200 can be MP3, MIDI, AAC, WAV, etc., and is not limited thereto.

Further, the database creation module 300 is connected to the text import module 100 and the voice transceiver module 200 to receive a plurality of text materials 110 from the text import module 100 and corresponding to the plurality of voice transceiver modules 200. A plurality of reference audios 211 of the text material 110 (e.g., streamed reference audio 211). In this way, the database establishing module 300 can compare the plurality of vocabularies 111 in the plurality of text materials 110 with the plurality of reference audio 211 in the plurality of reference voices 210, and then establish the voice text corresponding database 310 according to the comparison result. That is, the pronunciation corresponding to each different vocabulary is collected.

After the voice text correspondence module 310 is created through the text import module 100, the voice transceiver module 200, and the database creation module 300, the user can perform voice using other modules included in the voice translation system 10. The function of translation is described in detail below.

The target language selection module 400 can provide a touch or non-touch interface 410. The interface 410 can have a plurality of language options 411 corresponding to the language of the plurality of text materials 110. Before the user performs the voice translation function, the target language 412 that is desired to be translated may be selected from the plurality of language options 411 through the interface 410 in advance. Preferably, the target language selection module 400 provides for selectively selecting a plurality of target languages 412 from the plurality of language options 411 to obtain translation results 510 of one or more different languages using the translation processing module 500 in subsequent programs.

Then, the voice transceiver module 200 can receive the to-be-translated voice 220 having a plurality of to-be-translated audios 221, that is, the user can emit a voice within the receiving range of the voice transceiver module 200, such as speaking or using a voice playback device (such as a speaker or The microphone, etc.) plays the audio file, wherein the audio to be translated 221 is a pronunciation of each of the words 111 before the translation. It should be understood that the language of the pre-translated audio 221 to be translated should be different from the translated target language 412, so that the translation is necessary.

Further, the translation processing module 500 is connected to the voice transceiver module 200, the database creation module 300, and the target language selection module 400. The translation processing module 500 should receive the voice text correspondence database 310 established by the database creation module 300 in advance. Therefore, when the translation processing module 500 receives the to-be-translated voice 220 from the voice transceiver module 200, the translation processing module 500 can input a plurality of reference audio signals in the plurality of to-be-translated audios 221 and the voice-word correspondence database 310. 211 for comparison.

Finally, when the translation processing module 500 compares the plurality of reference audios 211 that are consistent with the plurality of to-be-translated audios 221, the translation processing module 500 can correspond to the plurality of reference audios 211 that are matched by the comparison. The database 310 retrieves a plurality of reference audios 211, a plurality of words 111 or both corresponding to the target language 412, and integrates the audio 211, the vocabulary 111 or both to output the translation result 510.

Preferably, the translation processing module 500 can compare the pitch, loudness, timbre, length, or any combination of the plurality of reference audio 211 in the plurality of to-be-translated audios 221 and the speech-to-speech data library 310 in the speech-to-speech 220. In order to improve the accuracy of translation.

For example, when the speech to be translated 220 is English and the target language 412 is Japanese, the translation processing module 500 performs translation of the speech file of the English language (ie, the speech to be translated 220) into the Japanese speech file (ie, the translation result 510). The translation program, that is, only the translation result 510 of the audio file format is obtained.

However, those of ordinary skill in the art should understand that by simply changing the creation, the translation processing module 500 can be used to translate the English voice file (ie, the voice to be translated 220) into Japanese text according to the needs of actual use. File (ie, translation result 510); conversely, the English text file (ie, the voice to be translated 220) can be translated into the Japanese text file (ie, the translation result 510); or, the above translation method can be combined to obtain the sound file at the same time. The translation result 510 of the format and text file format is convenient for the user to understand and learn.

Please refer to FIG. 2 and FIG. 3 , which are block diagrams and schematic diagrams of the speech translation system of the second embodiment of the present invention. As shown in FIG. 2, the speech translation system 10 includes a text import module 100, a voice transceiver module 200, a database creation module 300, a target language selection module 400, and a translation processing module 500. In the present embodiment, the components of the same component symbols are similar to the embodiments described above, and will not be further described herein.

The server 30 can include a live chat system 31 that connects to the voice transceiver module 200. When the users in different countries chat on the instant messaging system 31 provided by the server 30 as the chat platform, the instant messaging system 31 can receive the voice emitted by the user and establish a corresponding voice to be translated 220.

Next, the live chat system 31 of the server 30 can output the speech to be translated 220 to the voice transceiving module 200 of the speech translation system 10 to perform the translation process as described above. For example, the voice to be translated 220 received by the live chat system 31 may be from an external voice transceiver 40 held by the user, such as a wireless headset device, but is not limited thereto.

The server 30 can further include a translation result display system 32 coupled to the translation processing module 500. The translation result display system 32 of the server 30 can display the translation result 510 from the translation processing module 500 of the speech translation system 10 to the user.

In detail, the translation processing module 500 can store a plurality of location data 530 corresponding to a plurality of countries in a plurality of national languages. After the translation processing module 500 of the speech translation system 10 generates the translation result 510, the translation processing module 500 can retrieve the location data 530 corresponding to the target language 412 from the plurality of location data 530.

Next, the translation processing module 500 can output the translation result 510 to the external electronic device 20 located in the country corresponding to the target language 412 according to the location profile 530. More precisely, the translation processing module 500 can provide the location information 530 for the user to pre-store or chat at the moment to instantly input the destination to be delivered.

As shown in FIG. 3, users in a specific location in countries such as North America, South America, Africa, etc., chat through the instant messaging system 31 provided by the server 30, and the users in the African countries speak to be translated. The voice 220 is, for example, Madagascar (for the South Island language), and the user can select the target language 412 in English and American to generate an English or American translation result 510 for two other users located in North America and South America. In this way, users from all over the world can communicate and learn effectively without additional translators to promote cultural exchanges.

The above is intended to be illustrative only and not limiting. Any equivalent modifications or alterations to the spirit and scope of this creation shall be included in the scope of the appended patent application.

10‧‧‧Voice translation system
20‧‧‧External electronic devices
30‧‧‧Server
31‧‧‧ Live Chat System
32‧‧‧ Translation result display system
40‧‧‧External sound transceiver
100‧‧‧Text Import Module
110‧‧‧Text materials
111‧‧‧ vocabulary
200‧‧‧Voice transceiver module
210‧‧‧Reference voice
211‧‧‧Reference audio
220‧‧‧voices to be translated
221‧‧‧To be translated
300‧‧‧Database building module
310‧‧‧Voice text correspondence database
400‧‧‧Target language selection module
410‧‧‧ interface
411‧‧‧ language options
412‧‧‧ Target language
500‧‧‧Translation Processing Module
510‧‧‧ translation results
530‧‧‧Location Information

Fig. 1 is a block diagram of a speech translation system according to a first embodiment of the present creation.

Fig. 2 is a block diagram of a speech translation system according to a second embodiment of the present creation.

Fig. 3 is a schematic diagram of a speech translation system according to a second embodiment of the present creation.

10‧‧‧Voice translation system

100‧‧‧Text Import Module

110‧‧‧Text materials

111‧‧‧ vocabulary

200‧‧‧Voice transceiver module

210‧‧‧Reference voice

211‧‧‧Reference audio

220‧‧‧voices to be translated

221‧‧‧To be translated

300‧‧‧Database building module

310‧‧‧Voice text correspondence database

400‧‧‧Target language selection module

410‧‧‧ interface

411‧‧‧ language options

412‧‧‧ Target language

500‧‧‧Translation Processing Module

510‧‧‧ translation results

Claims (8)

A speech translation system, comprising: a text import module, the text import module provides a plurality of text materials into a plurality of national languages having a plurality of vocabulary; a voice transceiver module, the voice transceiver module receives a plurality of a reference voice, the plurality of reference voices corresponding to the plurality of text data and having a plurality of reference audio, the voice transceiver module receiving a to-be-translated voice having a plurality of to-be-translated audio; a database establishing module, connecting the a text import module and the voice transceiver module, the database building module compares the plurality of words with the plurality of reference audio signals, and establishes a phonetic text corresponding database according to the comparison result; a target language selection module The target language selection module provides an interface having a plurality of language options corresponding to the language of the plurality of text materials to provide a target language from the plurality of language options through the interface; and a translation processing module a group, connecting the voice transceiver module, the database creation module, and the target language selection module, the translation office The module compares the plurality of to-be-translated audios with the plurality of reference audio signals in the corresponding phonetic text database, and compares the plurality of reference audio messages that are matched by the comparison to obtain a pair from the corresponding phonetic text database The plurality of reference audios, the plurality of words, or a combination thereof in the target language should be output and a translation result is output. The speech translation system of claim 1, wherein the target language selection module provides a plurality of the target languages selected from the plurality of language options through the interface, and the translation processing module outputs corresponding A plurality of translation results of the target language. The speech translation system of claim 1, wherein the translation processing module is different from the plurality of to-be-translated audios in the speech to be translated and the tones of the plurality of reference audio in the corresponding database of the speech text. , loudness, tone, length, or any combination thereof. The speech translation system of claim 1, wherein the translation processing module stores a plurality of location data corresponding to a plurality of countries in a plurality of national languages, the translation processing module extracting from the plurality of location data The location data corresponding to the target language is then output based on the location data to an external electronic device located in the country corresponding to the target language. The voice translation system of claim 1, wherein the server comprises a live chat system connected to the voice transceiver module, and the instant chat system receives the voice emitted by the user during the chat and establishes the corresponding waiting The voice is translated, and then the voice to be translated is output to the voice transceiver module. The speech translation system of claim 5, wherein the to-be-translated speech received by the live chat system is from an external sound transceiving device held by the user. The speech translation system of claim 6, wherein the external sound transceiving device is a wireless earphone device. The speech translation system of claim 5, wherein the server further comprises a translation result display system connected to the translation processing module, the translation result display system displaying the translation result from the translation processing module to user.