CN109446532A

CN109446532A - Translate bearing calibration, device and translation calibration equipment

Info

Publication number: CN109446532A
Application number: CN201811058673.7A
Authority: CN
Inventors: 陈良辉; 王文祺
Original assignee: Shenzhen Water World Co Ltd
Current assignee: Shenzhen Water World Co Ltd
Priority date: 2018-09-11
Filing date: 2018-09-11
Publication date: 2019-03-08
Also published as: WO2020052191A1

Abstract

This application discloses translation bearing calibration, device and translation calibration equipments, wherein, translate bearing calibration, it include: the selection instruction of the interpretive scheme for the translator transmission that management server receives pre-association, the interpretive scheme includes at least two in automatic translation mode, semi-automatic interpretive scheme and human translation mode；Corresponding translation channel is opened according to the interpretive scheme that the selection instruction selects；Receive the primitive sound signal to be translated that the translator is sent；The primitive sound signal to be translated is translated into targeted voice signal by the translation channel.The application, to improve the accuracy of translation, meets user to the use demand of accurate translation by increasing manual synchronizing option on the basis of former machine translation.The interpretive scheme of the application includes the semi-automatic interpretive scheme and human translation mode of free automatic translation, charge, and charge method is versatile and flexible, meets the use demand of different user.

Description

Translate bearing calibration, device and translation calibration equipment

Technical field

This application involves smart machine field is arrived, especially relates to translation bearing calibration, device and translation correction and set It is standby.

Background technique

Past 1 year China's departure population reached 1.3 hundred million times, and language obstacle is still one of their pain spot, translator or The market of mobile phone translation APP is still expanding, but existing translator or mobile phone translation APP are machine translation, and accuracy is low, Especially in important business meeting in use, translation accuracy is low to seriously affect commercial negotiation quality, therefore existing translator or hand The translation technology of machine translation APP could be improved.

Summary of the invention

The main purpose of the application is to provide a kind of translation bearing calibration, it is intended to solve existing translator or mobile phone translation The low technical problem of the translation accuracy of APP.

This application provides a kind of translation bearing calibrations, comprising:

The selection instruction for the interpretive scheme that the translator that management server receives pre-association is sent, the interpretive scheme include At least two in automatic translation mode, semi-automatic interpretive scheme and human translation mode；

Corresponding translation channel is opened according to the interpretive scheme that selection instruction selects, wherein interpretive scheme includes manual synchronizing Option, the translation channel include manual synchronizing channel corresponding with the manual synchronizing option；

Receive the primitive sound signal to be translated that the translator is sent；

The primitive sound signal to be translated is translated into targeted voice signal by the translation channel.

Preferably, the interpretive scheme that the selection instruction selects is semi-automatic interpretive scheme, the semi-automatic interpretive scheme Including manual calibration option, the manual synchronizing option includes one or two in identification correction, translation correction and synthesis correction , the translation channel includes manual calibration channel corresponding with the manual calibration option, described by the original to be translated The step of voice signal translates into targeted voice signal by the translation channel, comprising:

The primitive sound signal to be translated is sent to voice server by the translation channel；

Receive the corresponding first object language text of primitive sound signal to be translated that voice server is sent；

Judge whether the manual synchronizing option includes identification correction；

If being corrected including identification, the first object language text and the primitive sound signal to be translated are sent to The first customer service terminal that the manual calibration channel provides carries out identification correction, obtains correction first object language text；

Receive the correction first object language text of the first customer service terminal feedback；

Correction first object language text hair is translated into targeted voice signal by the translation channel.

Preferably, described that the correction first object language text is translated into target language message by the translation channel Number the step of, comprising:

The correction first object language text is sent to translating server by the translation channel；

Receive corresponding second target language text of correction first object language text that translating server is sent；

Judge whether the manual synchronizing option includes translation correction；

If being corrected including translation, second target language text and the primitive sound signal to be translated are sent to The second customer service terminal that the manual calibration channel provides carries out translation correction；

Receive the second target language text of correction of the second customer service terminal feedback；

The second target language text of the correction is sent to voice synthesizing server by the translation channel and carries out language Sound synthesis；

Receive the voice synthesizing server transmission obtains the corresponding target language message of second target language text Number.

Preferably, it is described judge manual synchronizing option whether include the steps that translation correction after, comprising:

If not including translation correction, second target language text is sent to voice synthesizing server and carries out voice Synthesis；

Receive the corresponding specified targeted voice signal of the second target language text that voice synthesizing server is sent；

Judge whether the manual synchronizing option includes synthesis correction；

If being corrected including synthesis, the specified targeted voice signal and the primitive sound signal to be translated are sent to The third customer service terminal that the manual calibration channel provides carries out synthesis correction；

Receive the targeted voice signal of the third customer service terminal feedback.

Preferably, the interpretive scheme that the selection instruction selects is artificial interpretive scheme, and the human translation mode includes Manual calibration option, the manual synchronizing option include identification correction, translation correction and synthesis correction, and the translation channel includes Manual calibration channel corresponding with the manual calibration option, it is described that the primitive sound signal to be translated is passed through into the translation Translate into the step of targeted voice signal in channel, comprising:

Receive the corresponding third target language text of primitive sound signal to be translated that voice server is sent；

The third target language text and the primitive sound signal to be translated are sent to the manual calibration channel The specified customer service terminal provided carries out identification correction；

Receive the correction third target language text of the specified customer service terminal feedback；

The correction third target language text is sent to translating server by the translation channel；

Receive corresponding 4th target language text of correction third target language text that translating server is sent；

4th target language text and the primitive sound signal to be translated are sent to the manual calibration channel The specified customer service terminal provided carries out translation correction；

Receive the 4th target language text of correction of the specified customer service terminal feedback；

The 4th target language text of the correction is sent to voice synthesizing server and carries out speech synthesis；

Receive corresponding first voice signal of the 4th target language text of correction that voice synthesizing server is sent；

First voice signal and the primitive sound signal to be translated are sent to the finger that manual calibration channel provides Determine customer service terminal and carries out synthesis correction；

Receive the targeted voice signal of the specified customer service terminal feedback.

Preferably, described that the primitive sound signal to be translated is translated into targeted voice signal by the translation channel The step of after, comprising:

Corresponding translation charging is generated according to preset rules；

The translation charging is sent to designated terminal, to receive the translation charging of the payment of user.

Preferably, described the step of generating corresponding translation charging according to preset rules, comprising:

The interpretive scheme that the selection instruction selects is judged whether for charge mode, wherein the charge mode includes The semi-automatic interpretive scheme and the human translation mode and corresponding first charging mode of the semi-automatic interpretive scheme and Human translation mode corresponds to the second charging mode；

If the interpretive scheme is charge mode, judge whether are first charging mode or second charging mode For timing translation；

It is translated if timing, then counting user is from starting primitive sound signal to be translated described in typing, until completing all Charging duration between the primitive sound signal to be translated；

According to the charging duration and the corresponding Freight Basis of first charging mode, the semi-automatic translation mould is calculated The corresponding first translation charging of formula；Or according to the charging duration and the corresponding Freight Basis of second charging mode, meter Calculate the corresponding second translation charging of the human translation mode.

The interpretive scheme that the selection instruction selects is judged whether for charge mode, wherein the charge mode includes The semi-automatic interpretive scheme and the human translation mode and the corresponding third charging mode of the semi-automatic interpretive scheme The 4th charging mode corresponding with human translation mode；

If the interpretive scheme is charge mode, judge whether are the third charging mode or the 4th charging mode For translation number of words charge；

If translation number of words charge, then the corresponding total number of word of the targeted voice signal is counted；

According to the total number of word and the corresponding Freight Basis of the third charging mode, the semi-automatic interpretive scheme is calculated Corresponding third translates charging；Or according to the total number of word and the corresponding Freight Basis of the 4th charging mode, calculate institute State the corresponding 4th translation charging of human translation mode.

Present invention also provides a kind of translation means for correctings, are integrated in management server, comprising:

First receiving module, the selection instruction for the interpretive scheme that the translator for receiving pre-association is sent, interpretive scheme Including at least two in automatic translation mode, semi-automatic interpretive scheme and human translation mode；

Module is opened, the interpretive scheme for selecting according to the selection instruction opens corresponding translation channel, wherein institute Stating interpretive scheme includes manual synchronizing option, and the translation channel includes that manual synchronizing corresponding with the manual synchronizing option is logical Road；

Second receiving module, the primitive sound signal to be translated sent for receiving the translator；

Translation module, for the primitive sound signal to be translated to be translated into target language message by the translation channel Number.

It is described to apply journey present invention also provides a kind of translation calibration equipment, including memory, processor and application program Sequence is stored in the memory and is configured as being executed by the processor, and the application program is configurable for executing Above-mentioned translation bearing calibration.

The application on the basis of former machine translation by increasing manual synchronizing option, to improve the accuracy of translation, Meet user to the use demand of accurate translation.The interpretive scheme of the application include free automatic translation, charge it is semi-automatic Interpretive scheme and human translation mode, and charge method is versatile and flexible, meets the use demand of different user.

Detailed description of the invention

The flow diagram of the translation bearing calibration of one embodiment of Fig. 1 the application；

The structural schematic diagram of the translation means for correcting of one embodiment of Fig. 2 the application；

The structural schematic diagram of the translation module of one embodiment of Fig. 3 the application；

The structural schematic diagram of second transmission unit of one embodiment of Fig. 4 the application；

The structural schematic diagram of second transmission unit of another embodiment of Fig. 5 the application；

The structural schematic diagram of the translation module of another embodiment of Fig. 6 the application；

The structural schematic diagram of the translation means for correcting of the another embodiment of Fig. 7 the application；

The structural schematic diagram of the generation module of one embodiment of Fig. 8 the application；

The structural schematic diagram of the generation module of another embodiment of Fig. 9 the application.

Specific embodiment

It is with reference to the accompanying drawings and embodiments, right in order to which the objects, technical solutions and advantages of the application are more clearly understood The application is further elaborated.It should be appreciated that specific embodiment described herein is only used to explain the application, not For limiting the application.

Referring to Fig.1, the translation bearing calibration of one embodiment of the application, comprising:

S1: the selection instruction for the interpretive scheme that the translator that management server receives pre-association is sent, the interpretive scheme Including at least two in automatic translation mode, semi-automatic interpretive scheme and human translation mode.

The management server linking translator, each server in translation process and the customer service manually participated in of the present embodiment Terminal meets the translation demand of user to put up corresponding translation channel in time according to the user's choice.The present embodiment Management server is connected translator, each server in translation process and the customer service terminal manually participated in, can be each by identifying The correlation signal in corresponding device number, or identification Handshake Protocol is held to establish incidence relation.The interpretive scheme of the present embodiment is not only It further include the semi-automatic interpretive scheme and human translation mode manually participated in, to meet different user including automatic translation mode The use demand of group.

S2: corresponding translation channel is opened according to the interpretive scheme that selection instruction selects.

The present embodiment is turned over by increasing the link manually participated in, including manual synchronizing option in translation channel with improving The accuracy of translation part.The manual synchronizing option of the present embodiment includes identification correction, translation correction and synthesis correction, by artificial Correction improves the accuracy of translation, more can be close to current specifically used occasion, such as specific commercial meeting to make to translate content View, professional domain meeting etc..

S3: the primitive sound signal to be translated that the translator is sent is received.

The present embodiment receives the interpretive scheme and primitive sound signal to be translated that user selectes by translator, to mention High user uses the convenience of translator.

S4: the primitive sound signal to be translated is translated into targeted voice signal by the translation channel.

The translator of the present embodiment has the function of intertranslation, and above-mentioned primitive sound signal to be translated refers to the common-use words of user's typing Kind voice, so that common languages voiced translation is exchanged languages voice (i.e. targeted voice signal) at current zone, to realize reality When link up；Above-mentioned primitive sound signal to be translated or the languages voice that user is of little use is received, to translate into use in time The common languages voice (i.e. targeted voice signal) in family, is that user timely receives corresponding information.

Further, the interpretive scheme that selection instruction selects is semi-automatic interpretive scheme, the semi-automatic interpretive scheme packet Manual calibration option is included, the manual synchronizing option includes one or two in identification correction, translation correction and synthesis correction, The translation channel includes manual calibration channel corresponding with the manual calibration option, described by the primitive sound to be translated Signal translates into the step S4 of targeted voice signal by the translation channel, comprising:

S40: the primitive sound signal to be translated is sent to by voice server by the translation channel.

Primitive sound signal is sent to voice server by management server by the present embodiment, to complete to pass through after translating Translated targeted voice signal is sent to translator and played out by management server.The application other embodiments can also pass through Translator is directly associated with voice server, only when needing manual synchronizing to access, is associated with by management server artificial Correction.

S41: the corresponding first object language of the primitive sound signal to be translated that the voice server is sent is received Text.

The link that the present embodiment is intervened in prosthetic correction, it is identical with the realization process of former machine translation and principle, it does not go to live in the household of one's in-laws on getting married It states.

S42: judge whether the manual synchronizing option includes identification correction.

The selection information that the interpretive scheme of the user's selection sent by identification translator carries judges.For example it identifies Include " identification correction " keyword into above-mentioned selection information, then determines to include identification correction, otherwise do not include.

S43: if including identification correction, the first object language text and the primitive sound signal to be translated are sent out The the first customer service terminal provided to the manual calibration channel is sent to carry out identification correction.

It include identification correction in the selection information that the interpretive scheme of the present embodiment carries, then management server is then passing through Manual identified correction is accessed after machine recognition, i.e., in corresponding first mesh of primitive sound signal for receiving voice server feedback After marking language text, above-mentioned first object language text and the primitive sound signal to be translated are sent to manual identified correction Corresponding first customer service terminal, be bonded the conversion of the speech-to-text in speech recognition process more currently Specific usage scenario improves translation accuracy.

S44: the correction first object language text of the first customer service terminal feedback is received.

S45: correction first object language text is translated into targeted voice signal by above-mentioned translation channel.

The present embodiment is associated with machine translation and manual synchronizing by management server, by server each in machine translation process The file of completion is sent to directive property the corresponding customer service terminal of specified manual synchronizing according to the selection of user, and by customer service The file for completing manual synchronizing that terminal is sent is resend to corresponding in machine translation process according to machine translation process The corresponding server of next link, to keep entire translation flow more smooth.The machine translation process of the present embodiment includes three A link, speech recognition converts speech into the process of text, original language text is translated into target language text process and By the process of target language text synthesis target voice.In above three link, each link can access the mistake of manual synchronizing Journey into before next machine translation link, accesses manual synchronizing link that is, after upper machine translation link.For example voice is known The mistake that manual identified corrects, original language text is translated into target language text is added after not converting speech into the process of text Human translation correction is added after journey, artificial synthesized correction etc. will be added after the process of target language text synthesis target voice, this Primitive sound signal of being subject in the manual synchronizing of embodiment is corrected.

Further, described to turn over the correction first object language text by described in another embodiment of the application Translate the step S45 that targeted voice signal is translated into channel, comprising:

S450: the correction first object language text is sent to by translating server by translation channel.

S450a: corresponding second target of the correction first object language text that the translating server is sent is received Language text.

S451: judge whether the manual synchronizing option includes translation correction.

The selection information that the interpretive scheme of the user's selection sent by identification translator carries judges.For example it identifies Include " translation correction " keyword into above-mentioned selection information, then determines to include identification correction, otherwise do not include.

S452: if including translation correction, by second target language text and the primitive sound signal to be translated It is sent to the second customer service terminal that the manual calibration channel provides and carries out translation correction.

S452a: the second target language text of correction of the second customer service terminal feedback is received.

The present embodiment human translation correction and manual identified is corrected to the same customer service terminal, with save human resources, Simplify the related information in translation channel.In the application other embodiments, manual identified correction, human translation correction and artificial conjunction It can be respectively different customer service terminals at correction, to improve translation efficiency and translation quality.

S453: by the translation channel by the second target language text of the correction be sent to voice synthesizing server into Row speech synthesis.

S453a: the corresponding target voice of second target language text that the voice synthesizing server is sent is received Signal.

The present embodiment with joined simultaneously in machine translation process the half of human translation correction and manual identified correction from For dynamic interpretive scheme, be described in detail information according to the user's choice joined in former machine translation channel it is specified artificial The realization process of correction.The semi-automatic interpretive scheme of the present embodiment refers to that manual synchronizing option includes identification correction, translation correction With at least one of synthesis correction or two, user can unrestricted choice according to their needs, management server is according to the choosing of user It selects to build and meets the translation channel that user selects intention.

Further, it is described judge the manual synchronizing option whether include the steps that translation correction S45 after, comprising:

S456: if do not include translation correction, by second target language text be sent to voice synthesizing server into Row speech synthesis；

S456a: the corresponding specified target of second target language text that the voice synthesizing server is sent is received Voice signal.

S457: judge whether the manual synchronizing option includes synthesis correction.

The selection information that the interpretive scheme of the user's selection sent by identification translator carries judges.For example it identifies Include " synthesis correction " keyword into above-mentioned selection information, then determines to include identification correction, otherwise do not include.

S458: if including synthesis correction, by the specified targeted voice signal and the primitive sound signal to be translated The third customer service terminal for being sent to the offer of manual calibration channel carries out synthesis correction；

S459: the targeted voice signal of the third customer service terminal feedback is received.

The present embodiment with joined simultaneously in machine translation process the half of artificial synthesized correction and manual identified correction from For dynamic interpretive scheme, be described in detail information according to the user's choice joined in former machine translation channel it is specified artificial The realization process of correction.A certain single manual synchronizing option is added in other, or human translation correction and artificial conjunction are added simultaneously At correction, and the forming process in the corresponding translation channel of semi-automatic interpretive scheme being respectively formed, translation process are for example above-mentioned specific Embodiment is similar, does not repeat.

Further, the interpretive scheme that the selection instruction selects is artificial interpretive scheme, the human translation mode packet Manual calibration option is included, the manual synchronizing option includes identification correction, translation correction and synthesis correction, the translation channel packet Manual calibration channel corresponding with the manual calibration option is included, it is described to turn over the primitive sound signal to be translated by described Translate the step S4 that targeted voice signal is translated into channel, comprising:

S401: the primitive sound signal to be translated is sent to by voice server by the translation channel.

S402: the corresponding third target language text of primitive sound signal to be translated that voice server is sent is received.

S403: third target language text and the primitive sound signal to be translated are sent to the manual calibration channel The specified customer service terminal provided carries out identification correction.

S404: the correction third target language text of the specified customer service terminal feedback is received.

S405: correction third target language text is sent to by translating server by translation channel and is translated；

S405a: corresponding 4th target of the correction third target language text that the translating server is sent is received Language text.

S406: the 4th target language text and the primitive sound signal to be translated are sent to the manual calibration The specified customer service terminal that channel provides carries out translation correction.

S407: the 4th target language text of correction of the specified customer service terminal feedback is received.

S408: the 4th target language text of correction is sent to voice synthesizing server and carries out speech synthesis.

S408a: the 4th target language text corresponding first of the correction that the voice synthesizing server is sent is received Voice signal.

S409: first voice signal and the primitive sound signal to be translated are sent to manual calibration channel and provided Specified customer service terminal carry out synthesis correction.

S409a: the targeted voice signal of the specified customer service terminal feedback is received.

The present embodiment joined manual identified correction, human translation correction and artificial conjunction simultaneously in machine translation process For the human translation mode of correction, information according to the user's choice is described in detail joined in former machine translation channel The realization process of specified manual synchronizing.The human translation mode and semi-automatic interpretive scheme of the present embodiment, difference are to be wrapped The manual synchronizing option item number included is different, and charge is also different, but the realization of manual synchronizing is added in original machine device translation flow Process is similar, does not repeat.

Further, described that the primitive sound signal to be translated is translated into target language message by the translation channel Number step S4 after, comprising:

S5: corresponding translation charging is generated according to preset rules.

The automatic translation of the translator of the present embodiment is free, but needs to need to collect the charges when manual synchronizing, And according to user's selection be semi-automatic interpretive scheme or human translation mode, the corresponding translation charging method of institute is not yet Together, above-mentioned preset rules are not only related with the interpretive scheme of user's selection, and with the specific of the manual synchronizing option of user's selection Option is related, and the corresponding expense of different manual synchronizing options is different, select the semi-automatic interpretive scheme an of manual synchronizing with Select the charging of the semi-automatic interpretive scheme of two manual synchronizings different.And above-mentioned preset rules further include the charging of user's selection Mode be Fare determined by travel time still press translation number of words charge it is related.

S6: translation charging is sent to designated terminal, to receive the translation charging of the payment of user.

After the completion of the translation charging of the present embodiment is by management server, and it is sent to translation generator terminal, to facilitate user to carry out Payment.The translator of the present embodiment can detect whether before user starts charge mode there are arrearage automatically, and arrearage is then if it exists It issues arrearage to remind, and limitation continues to use the semi-automatic interpretive scheme or human translation mode of charge simultaneously, control semi-automatic Interpretive scheme or human translation mode are in disabled status.

Further, the step S5 that corresponding translation charging is generated according to preset rules, comprising:

S501: interpretive scheme that selection instruction selects is judged whether for charge mode, wherein the charge mode includes institute State semi-automatic interpretive scheme and the human translation mode and corresponding first charging mode of the semi-automatic interpretive scheme and people Work interpretive scheme corresponds to the second charging mode.

The difference for the manual synchronizing option that first charging mode of the present embodiment is selected according to user and it is different, citing ground, Select a manual synchronizing option charges for 1 yuan/min (charging unit price), two manual synchronizing option charges are 1.2 yuan/min (charging unit price), that is, after selecting semi-automatic interpretive scheme, expense is risen with 1 yuan/min (charging unit price) starts charging, increases another Item manual synchronizing option is with 1.2 yuan/min (charging unit price) charging.The expense of second charging mode of the present embodiment with 1.5 yuan/ Min (charging unit price), which rises, starts charging." first " of the present embodiment, " second " etc. are not used in restriction for being only used for distinguishing, The such term effect in elsewhere is identical, does not repeat.

S502: if the interpretive scheme is charge mode, judge first charging mode or second charged party Whether formula is timing translation.

In the charge mode of the translator of the application, including two kinds of statistics total cost sides of timing translation and translation number of words charge Formula, the present embodiment is by taking timing is translated as an example.

S503: translating if timing, then counting user is from starting primitive sound signal to be translated described in typing, until completing Charging duration between all primitive sound signals to be translated.

After user's selection starts charge mode, when management server receives first voice to be translated of user Start to translate timing, and record the time started, until user is when mobile phone is translated and clicks translation END instruction on APP or translator Translation timing, and record end time are terminated, counts above-mentioned charging duration, i.e., " when charging duration=end time-beginning Between ", and calculate the corresponding translation expense of above-mentioned charging duration and send mobile phone translation APP or translator, it is next after user's clearing It is secondary just to continue to use semi-automatic interpretive scheme or human translation mode.The present embodiment is forgotten to translate in mobile phone in order to avoid user Translation END instruction is clicked on APP or translator, management server is set preset time such as 5 seconds, is receiving user's typing Translation sentence after, automatically into timing wait state, with etc. lower sentence to be translated of user's typing to be received, sentence in real time Whether disconnected timing wait state is more than preset time, when the preset time is exceeded, do not receive user's typing lower it is to be translated Sentence, and do not receive user and translate click translation END instruction on APP or translator in mobile phone, then automatic pause translation is counted When.After automatic pause translates timing, whether starting pause translation duration statistic processes is within a specified time monitored in real time and is received User continues typing sentence to be translated and starts timer if receiving and continue to build up timing, if not receiving, statistics is simultaneously Record pause translation duration judges whether pause translation duration is more than specified time, such as 1 hour, if being more than specified time The sentence to be translated of user's typing whether is received after judging pause translation timing, if otherwise terminating to translate timing, and is recorded End time, and calculate the expense of the period and be sent to mobile phone translation APP or translator, " the charging duration of the period Duration is translated in time started=end time-- pause ".

S504: it according to the charging duration and the corresponding Freight Basis of first charging mode, calculates described semi-automatic The corresponding first translation charging of interpretive scheme；Or according to the charging duration and the corresponding charging mark of second charging mode Standard calculates the corresponding second translation charging of the human translation mode.

First charging mode of the present embodiment includes charging type, such as Fare determined by travel time or by translation number of words charge, is also wrapped Include Freight Basis i.e. charging unit price.Semi-automatic interpretive scheme is obtained according to " translation charging=charging duration * charging unit price " respectively Corresponding first translation charging and the corresponding second translation charging of human translation mode.

S505: interpretive scheme that selection instruction selects is judged whether for charge mode, wherein the charge mode includes institute It states semi-automatic interpretive scheme and the human translation mode and the semi-automatic interpretive scheme corresponds to third charging mode and people Work interpretive scheme corresponds to the 4th charging mode.

The difference for the manual synchronizing option that the third charging mode of the present embodiment is selected according to user and it is different, citing ground, Select a manual synchronizing option charges for 100 yuan/K word (charging unit price), two manual synchronizing option charges are 150 yuan/thousand Word (charging unit price), that is, after selecting semi-automatic interpretive scheme, expense is risen with 100 yuan/K word (charging unit price) starts charging, increases Another manual synchronizing option is with the charging of 150 yuan/K word (charging unit price).The expense of 4th charging mode of the present embodiment with 200 yuan/K word (charging unit price) are risen and start charging.

S506: if the interpretive scheme is charge mode, judge the third charging mode or the 4th charged party Whether formula is translation number of words charge.

For the present embodiment is by translation number of words charge, specific explanations realize process.Above-mentioned translation number of words feeling the pulse with the finger-tip poster message Number corresponding total number of word.

S507: if translation number of words charge, then the corresponding total number of word of the targeted voice signal is counted.

The present embodiment obtains target voice by executing word counting to the text before synthesizing the targeted voice signal The corresponding total number of word of signal.

S508: according to the total number of word and the corresponding Freight Basis of the third charging mode, described semi-automatic turn over is calculated Translate the corresponding third translation charging of mode；Or according to the total number of word and the corresponding Freight Basis of the 4th charging mode, Calculate the corresponding 4th translation charging of the human translation mode.

The third charging mode of the present embodiment includes charging type, i.e., by translation number of words charge, further include Freight Basis i.e. Charging unit price.Corresponding first translation of semi-automatic interpretive scheme is obtained according to " translation charging=total number of word * charging unit price " respectively Charging and the corresponding second translation charging of human translation mode.

The present embodiment on the basis of former machine translation by increasing manual synchronizing option, to improve the accurate of translation Degree, meets user to the use demand of accurate translation.The interpretive scheme of the present embodiment include free automatic translation, charge half Automatic translation mode and human translation mode, and charge method is versatile and flexible, meets the use demand of different user.

Referring to Fig. 2, present invention also provides a kind of translation means for correctings, are integrated in management server, comprising:

First receiving module 1, the selection instruction for the interpretive scheme that the translator for receiving pre-association is sent, interpretive scheme Including at least two in automatic translation mode, semi-automatic interpretive scheme and human translation mode.

Module 2 is opened, the interpretive scheme for selecting according to selection instruction opens corresponding translation channel.

Second receiving module 3, the primitive sound signal to be translated sent for receiving the translator.

Translation module 4, for the primitive sound signal to be translated to be translated into target voice by the translation channel Signal.

Referring to Fig. 3, in one embodiment of the application, the interpretive scheme that selection instruction selects is semi-automatic interpretive scheme, described Semi-automatic interpretive scheme includes manual calibration option, and the manual synchronizing option includes identification correction, translation correction and synthesis school One or two of center, the translation channel includes manual calibration channel corresponding with the manual calibration option, translates mould Block 4, comprising:

First transmission unit 40, for the primitive sound signal to be translated to be sent to voice by the translation channel Server.

First receiving unit 41, it is corresponding for receiving the primitive sound signal to be translated that the voice server is sent First object language text.

First judging unit 42, for judging whether the manual synchronizing option includes identification correction.

Identification correction unit 43, if for including identification correction, by the first object language text and described wait turn over The primitive sound signal translated is sent to the first customer service terminal that the manual calibration channel provides and carries out identification correction, obtains correction the One target language text.

Second receiving unit 44, for receiving the correction first object language text of the first customer service terminal feedback.

Second transmission unit 45, for the correction first object language text to be translated into target language by translating channel Sound signal.

The present embodiment is associated with machine translation and manual synchronizing by management server, by server each in machine translation process The file of completion according to user selection directive property be sent to the corresponding customer service terminal of specified manual synchronizing, and by customer service end Hold the file for completing manual synchronizing sent according to machine translation process, resend to it is corresponding in machine translation process under The corresponding server of one link, to keep entire translation flow more smooth.The machine translation process of the present embodiment includes three Link, speech recognition convert speech into the process of text, original language text are translated into the process of target language text and is incited somebody to action The process of target language text synthesis target voice.In above three link, each link can access the process of manual synchronizing, I.e. after the completion of upper machine translation link, into before next machine translation, manual synchronizing link is accessed.For example speech recognition will Manual identified correction is added after being converted to the process of text in voice, original language text is translated into the process of target language text after Human translation correction is added, artificial synthesized correction etc., this implementation will be added after the process of target language text synthesis target voice Primitive sound signal of being subject in the manual synchronizing of example is corrected.

Referring to Fig. 4, in one embodiment of the application, the second transmission unit 45, comprising:

First transmission sub-unit 450, for being sent the correction first object language text by the translation channel To translating server.

First receiving subelement 450a, the correction first object language text sent for receiving the translating server This corresponding second target language text.

First judgment sub-unit 451, for judging whether the manual synchronizing option includes translation correction.

Second transmission sub-unit 452, if for include translation correction, by second target language text and it is described to The second customer service terminal that the primitive sound signal of translation is sent to the offer of manual calibration channel carries out translation correction.

Second receiving subelement 452a, for receiving correction the second object language text of the second customer service terminal feedback This.

Third transmission sub-unit 453, for being sent the second target language text of the correction by the translation channel Speech synthesis is carried out to voice synthesizing server.

Third receiving unit 453a, second target language text sent for receiving the voice synthesizing server The corresponding targeted voice signal.

Referring to Fig. 5, in another embodiment of the application, the second transmission unit 45, comprising:

4th transmission sub-unit 456, if being sent to second target language text for not including translation correction Voice synthesizing server carries out speech synthesis.

4th receiving unit 456a, second target language text sent for receiving the voice synthesizing server Corresponding specified targeted voice signal.

Second judgment sub-unit 457, for judging whether the manual synchronizing option includes synthesis correction.

5th transmission sub-unit 458, if for include synthesis correction, by the specified targeted voice signal and it is described to The third customer service terminal that the primitive sound signal of translation is sent to the offer of manual calibration channel carries out synthesis correction.

5th receiving unit 459, for receiving the targeted voice signal of third customer service terminal feedback.

Referring to Fig. 6, in another embodiment of the application, the interpretive scheme that selection instruction selects is artificial interpretive scheme, described Human translation mode includes manual calibration option, and the manual synchronizing option includes identification correction, translation corrects and synthesis correction, The translation channel includes manual calibration channel corresponding with the manual calibration option, translation module 4, comprising:

Second transmission unit 401, for the primitive sound signal to be translated to be sent to language by the translation channel Sound server.

Second receiving unit 402, the primitive sound signal pair to be translated sent for receiving the voice server The third target language text answered.

Third transmission unit 403, for sending out the third target language text and the primitive sound signal to be translated The specified customer service terminal provided to the manual calibration channel is sent to carry out identification correction.

Third receiving unit 404, for receiving the correction third target language text of specified customer service terminal feedback.

4th transmission unit 405, for being sent to the correction third target language text by the translation channel Translating server is translated.

4th receiving unit 405a, the correction third target language text sent for receiving the translating server Corresponding 4th target language text.

5th transmission unit 406, for sending out the 4th target language text and the primitive sound signal to be translated It send the specified customer service terminal provided to manual calibration channel to carry out translation correction, obtains the 4th target language text of correction.

5th receiving unit 407, for receiving the 4th target language text of correction of specified customer service terminal feedback.

6th transmission unit 408, for by the 4th target language text of the correction be sent to voice synthesizing server into Row speech synthesis.

6th receiving unit 408a obtains the 4th target of the correction for receive that the voice synthesizing server sends Corresponding first voice signal of language text.

7th transmission unit 409, for first voice signal and the primitive sound signal to be translated to be sent to The specified customer service terminal that manual calibration channel provides carries out synthesis correction.

7th receiving unit 409a obtains targeted voice signal for receive specified customer service terminal feedback.

Referring to Fig. 7, the translation means for correcting of the another embodiment of the application, comprising:

Generation module 5, for generating corresponding translation charging according to preset rules.

Payment module 6: translation charging is sent to designated terminal, to receive the translation charging of the payment of user.

Referring to Fig. 8, the generation module 5 of one embodiment of the application, comprising:

Second judgment unit 501, for judging interpretive scheme that selection instruction selects whether for charge mode, wherein described Charge mode includes the semi-automatic interpretive scheme and the human translation mode and the semi-automatic interpretive scheme corresponding the One charging mode and corresponding second charging mode of human translation mode.

The difference for the manual synchronizing option that first charging mode of the present embodiment is selected according to user and it is different, citing ground, Select a manual synchronizing option charges for 1 yuan/min (charging unit price), two manual synchronizing option charges are 1.2 yuan/min (charging unit price), that is, after selecting semi-automatic interpretive scheme, expense is risen with 1 yuan/min (charging unit price) starts charging, increases another Item manual synchronizing option is with 1.2 yuan/min (charging unit price) charging.The expense of second charging mode of the present embodiment with 1.5 yuan/ Min (charging unit price), which rises, starts charging.

Third judging unit 502, if for the interpretive scheme be charge mode, judge first charging mode or Whether second charging mode is timing translation.

First statistic unit 503, for translating if timing, then counting user is from starting primitive to be translated described in typing Sound signal rises, until completing the charging duration between all primitive sound signals to be translated.

After user's selection starts charge mode, when management server receives first voice to be translated of user Start to translate timing, and record the time started, until user is when mobile phone is translated and clicks translation END instruction on APP or translator Translation timing, and record end time are terminated, counts above-mentioned charging duration, i.e., " when charging duration=end time-beginning Between ", and calculate the corresponding translation expense of above-mentioned charging duration and send mobile phone translation APP or translator, it is next after user's clearing It is secondary just to continue to use semi-automatic or human translation.The present embodiment is forgotten to translate APP or translator in mobile phone in order to avoid user END instruction is translated in upper click, and management server is set preset time such as 5 seconds, in the translation sentence for receiving user's typing Afterwards, automatically into timing wait state, with etc. lower sentence to be translated of user's typing to be received, real-time judge calendar scheduling wait for Whether state is more than preset time, when the preset time is exceeded, does not receive lower sentence to be translated of user's typing, and do not connect It receives user and translates click translation END instruction on APP or translator in mobile phone, then automatic pause translates timing.Automatic pause turns over After translating timing, start pause translation duration statistic processes, within a specified time monitors in real time and whether receive user's continuation typing Sentence to be translated starts timer and continues to build up timing if receiving, if not receiving, counts and records pause translation Duration judges whether pause translation duration is more than specified time, such as 1 hour, and pause translation is judged if being more than specified time The sentence to be translated of user's typing whether is received after timing, if otherwise terminating to translate timing, and record end time, and is counted Calculate the expense of the period and be sent to mobile phone translation APP or translator, the period " charging duration=end time-opens Duration is translated in m- pause when the beginning ".

First computing unit 504 is used for according to the charging duration and the corresponding Freight Basis of first charging mode, Calculate the corresponding first translation charging of the semi-automatic interpretive scheme；Or according to the charging duration and second charged party The corresponding Freight Basis of formula calculates the corresponding second translation charging of the human translation mode.

Referring to Fig. 9, the generation module 5 of another embodiment of the application, comprising:

4th judging unit 505, for judging interpretive scheme that selection instruction selects whether for charge mode, wherein described Charge mode includes the semi-automatic interpretive scheme and the human translation mode and the semi-automatic interpretive scheme corresponding the Three charging modes and corresponding 4th charging mode of human translation mode.

5th judging unit 506, if for the interpretive scheme be charge mode, judge the third charging mode or Whether the 4th charging mode is translation number of words charge.

Second statistic unit 507, for if translation number of words charge, then counting the corresponding total word of the targeted voice signal Number.

First computing unit 508, for calculating described half according to total number of word and the corresponding Freight Basis of third charging mode The corresponding third of automatic translation mode translates charging；Or according to the total number of word and the corresponding charging of the 4th charging mode Standard calculates the corresponding 4th translation charging of the human translation mode.

It is described to apply journey present invention also provides a kind of translation calibration equipment, including memory, processor and application program Sequence is stored in the memory and is configured as being executed by the processor, and the application program is configurable for executing Translation bearing calibration in any of the above-described embodiment.

It will be understood by those skilled in the art that step counting equipment of the present invention and above-mentioned involved for executing the application Described in one or more equipment in method.These equipment can specially design and manufacture for required purpose, or It also may include the known device in general purpose computer.These equipment have the computer program being stored in it or using journeys Sequence, these computer programs are selectively activated or are reconstructed.Such computer program can be stored in equipment (for example, meter Calculation machine) in readable medium or it is stored in and is suitable for storage e-command and is coupled in any kind of medium of bus respectively, The computer-readable medium includes but is not limited to any kind of disk (including floppy disk, hard disk, CD, CD-ROM and magneto-optic Disk), ROM (Read-Only Memory, read-only memory), RAM (Random Access Memory, random access memory), EPROM (Erasable Programmable Read-Only Memory, Erarable Programmable Read only Memory), EEPROM (Electrically Erasable Programmable Read-Only Memory, Electrically Erasable Programmable Read-Only Memory), Flash memory, magnetic card or light card.It is, readable medium includes the form by equipment (for example, computer) can read Any medium of storage or transmission information.

The foregoing is merely preferred embodiment of the present application, are not intended to limit the scope of the patents of the application, all utilizations Equivalent structure or equivalent flow shift made by present specification and accompanying drawing content is applied directly or indirectly in other correlations Technical field, similarly include in the scope of patent protection of the application.

Claims

1. a kind of translation bearing calibration characterized by comprising

The selection instruction for the interpretive scheme that the translator that management server receives pre-association is sent, the interpretive scheme include automatic At least two in interpretive scheme, semi-automatic interpretive scheme and human translation mode；

Corresponding translation channel is opened according to the interpretive scheme that the selection instruction selects；

2. translation bearing calibration according to claim 1, which is characterized in that the interpretive scheme that the selection instruction selects for Semi-automatic interpretive scheme, the semi-automatic interpretive scheme include manual calibration option, and the manual synchronizing option includes identification school Just, one or two in translation correction and synthesis correction, the translation channel includes corresponding with the manual calibration option Manual calibration channel, it is described that the primitive sound signal to be translated is translated into targeted voice signal by the translation channel Step, comprising:

Receive the corresponding first object language text of the primitive sound signal to be translated that the voice server is sent；

If being corrected including identification, the first object language text and the primitive sound signal to be translated are sent to described The first customer service terminal that manual calibration channel provides carries out identification correction；

The correction first object language text is translated into targeted voice signal by the translation channel.

3. translation bearing calibration according to claim 2, which is characterized in that described by the correction first object language text The step of this translates into targeted voice signal by the translation channel, comprising:

Receive corresponding second target language text of the correction first object language text that the translating server is sent；

If being corrected including translation, second target language text and the primitive sound signal to be translated are sent to described The second customer service terminal that manual calibration channel provides carries out translation correction；

The second target language text of the correction is sent to voice synthesizing server by the translation channel and carries out voice conjunction At；

Receive the corresponding targeted voice signal of second target language text that the voice synthesizing server is sent.

4. translation bearing calibration according to claim 3, which is characterized in that described whether to judge the manual synchronizing option After including the steps that translation corrects, comprising:

If not including translation correction, second target language text is sent to voice synthesizing server and carries out voice conjunction At；

Receive the corresponding specified targeted voice signal of second target language text that the voice synthesizing server is sent；

Judge whether the manual synchronizing option includes synthesis correction；

If being corrected including synthesis, the specified targeted voice signal and the primitive sound signal to be translated are sent to described The third customer service terminal that manual calibration channel provides carries out synthesis correction；

5. translation bearing calibration according to claim 1, which is characterized in that the interpretive scheme that the selection instruction selects for Human translation mode, the human translation mode include manual calibration option, and the manual synchronizing option includes identification correction, turns over Correction and synthesis correction are translated, the translation channel includes manual calibration channel corresponding with the manual calibration option, described to incite somebody to action The step of primitive sound signal to be translated translates into targeted voice signal by the translation channel, comprising:

Receive the corresponding third target language text of the primitive sound signal to be translated that the voice server is sent；

The third target language text and the primitive sound signal to be translated are sent to the manual calibration channel to provide Specified customer service terminal carry out identification correction；

Receive corresponding 4th target language text of the correction third target language text that the translating server is sent；

4th target language text and the primitive sound signal to be translated are sent to the manual calibration channel to provide Specified customer service terminal carry out translation correction；

Receive corresponding first voice signal of the 4th target language text of the correction that the voice synthesizing server is sent；

First voice signal and the primitive sound signal to be translated are sent to the finger that the manual calibration channel provides Determine customer service terminal and carries out synthesis correction；

6. translation bearing calibration according to claim 1, which is characterized in that described by the primitive sound signal to be translated After the step of translating into targeted voice signal by the translation channel, comprising:

Corresponding translation charging is generated according to preset rules；

The translation charging is sent to designated terminal, to receive the translation charging of user's payment.

7. translation bearing calibration according to claim 6, which is characterized in that described to generate corresponding turn over according to preset rules The step of translating charging, comprising:

Judge interpretive scheme that the selection instruction selects whether for charge mode, wherein the charge mode include described half from Dynamic interpretive scheme and the human translation mode and corresponding first charging mode of the semi-automatic interpretive scheme and human translation Mode corresponds to the second charging mode；

If the interpretive scheme is charge mode, judge whether first charging mode or second charging mode are meter When translate；

It is translated if timing, then counting user is from starting primitive sound signal to be translated described in typing, until completing all described Charging duration between primitive sound signal to be translated；

According to the charging duration and the corresponding Freight Basis of first charging mode, the semi-automatic interpretive scheme pair is calculated The the first translation charging answered；Or according to the charging duration and the corresponding Freight Basis of second charging mode, calculate institute State the corresponding second translation charging of human translation mode.

8. translation bearing calibration according to claim 6, which is characterized in that described to generate corresponding turn over according to preset rules The step of translating charging, comprising:

Judge interpretive scheme that the selection instruction selects whether for charge mode, wherein the charge mode include described half from It moves interpretive scheme and the human translation mode and the corresponding third charging mode of the semi-automatic interpretive scheme and manually turns over Translate corresponding 4th charging mode of mode；

If the interpretive scheme is charge mode, judge whether the third charging mode or the 4th charging mode are to turn over Translate number of words charge；

According to the total number of word and the corresponding Freight Basis of the third charging mode, it is corresponding to calculate the semi-automatic interpretive scheme Third translate charging；Or according to the total number of word and the corresponding Freight Basis of the 4th charging mode, calculate the people The corresponding 4th translation charging of work interpretive scheme.

9. a kind of translation means for correcting, which is characterized in that be integrated in management server, comprising:

First receiving module, the selection instruction for the interpretive scheme that the translator for receiving pre-association is sent, the interpretive scheme Including at least two in automatic translation mode, semi-automatic interpretive scheme and human translation mode；

Module is opened, the interpretive scheme for selecting according to the selection instruction opens corresponding translation channel, wherein described turn over The mode of translating includes manual synchronizing option, and the translation channel includes manual synchronizing channel corresponding with the manual synchronizing option；

Translation module, for the primitive sound signal to be translated to be translated into targeted voice signal by the translation channel.

10. a kind of translation calibration equipment, including memory, processor and application program, the application program is stored in described It in memory and is configured as being executed by the processor, which is characterized in that the application program is configurable for right of execution Benefit requires 1 to 8 described in any item translation bearing calibrations.