KR101336716B1 - Listen and write system on network - Google Patents

Abstract

The present invention relates to a listening dictation system online, and more particularly, when a value of a time length of a non-speech interval is derived, a probability distribution map is obtained according to the time length, and sentences, idioms, and words are respectively calculated according to the probability. To store data by separating the data, the data is provided at the time of listening and writing of the user to read and cut the module to enable the separation of words, idioms and sentences, and the user to search the listening data of the desired level, If the dictation data retrieval module that retrieves the voice data that you want to hear and dictate by keyword or request, and the dictation that enables to analyze the ability to receive, store and analyze the user input data after the dictation, and to listen and write Evaluation module and temporary after input of a certain time interval or amount A server equipped with a data recovery module that enables data recovery in case of an emergency, and a playback control unit for repeatedly playing a specific section during keyboard input or easily moving back and forth between sections, and listening and writing input. Collect user input data for and store it on the server from time to time, and if the input is finally sent to the server after the input is completed, the user raw data collection unit for evaluation, dictation data upload to upload voice data and text for listening and writing A listening and dictation system on-line including a client having a sync and annotation editor for sync editing and annotation between audio and text data and text data.
According to the present invention, first, there is an effect of quantifying the ability to hear and dictate in multiple ways.
Second, analyzing the raw data for multi-faceted evaluation can increase the probability of checking whether a human input or a non-human program has been entered, and thus, it is effective to verify whether the evaluator is a human in the online evaluation.
Third, by using the repeat play control function and the automatic cut-out function, the recording work time is reduced to less than four times the length of the recorded file even for a general person who has no problem using the keyboard.
Fourth, by using the sync editing function, multiple levels of explanation or translation can be added to a specific location of a voice file, which can be used as a teaching aid or a tool for translation education.
Fifth, the recording work can be done jointly, and a new market can be created by providing dictation materials and using processors, dictation materials users, and server operators.
Sixth, it can create a new market consisting of call requesters, record performers, and recording server operators, and create a new market consisting of voice lecture commentaries, voice lecture comment users, and voice lecture comment file server administrators. There is a new market creation advantage for book or audiovisual lectures.

Description

Listen and write system on network

The present invention relates to a listening dictation system online.

In general, a predetermined listening and dictation software is used for listening and dictating with a personal computer.

In addition, the sync editor software that synchronizes the movie screen and the subtitle has a disadvantage in that the dictation function such as result comparison is insufficient because the program is made for the purpose of producing the movie subtitle.

The conventional listening dictation service is composed of one sentence dictation or a specific phrase filling, so that a function for continuously listening and dictating several sentences as a whole is not implemented.

Therefore, there is a demand for the development of a system that provides a user with a variety of convenient functions for reading and dictating the actual sentence (reading in syntax units, comparing and checking answers, multi-face evaluation method, etc.).

The object of the present invention is to quantify the ability to listen and dictate in multiple ways, and analyzing raw data for multi-faceted evaluation can increase the probability of checking whether a human input or a non-human program is input. By using the repeat play control function and the automatic break reading function, the recording time can be reduced to less than four times the length of the recording file even for a general user who has no problem using the keyboard. Sync editing allows you to add multiple levels of explanation or translation to a specific location in a voice file, which can be used as an aid to voice lectures or as a tool for translation training, to collaborate on recordings, and to dictate material. A new city of providers and processors, dictator users, and server operators Create a new market with voice requesters, record performers, and record server operators, and create a new market with voice lecture commentators, voice lecture comment users, and voice lecture comment file server administrators. It is possible to provide an online listening and dictation system that can create new markets for audiobooks or audiovisual lectures.

The present invention devised to achieve the above object, when the value of the time length of the non-speech interval is derived, obtain a probability distribution according to the time length, and according to the probability to separate the sentences, idioms and words, respectively, the data Save, but the data is provided at the time of listening and writing playback of the user to cut words, idioms and sentences, and the reading module and the user to search the listening data of the desired level, the user's keywords or requests A dictation data search module in which voice data desired to be listened and dictated is searched, and a dictation evaluation module and a predetermined time, which enable to analyze the ability to receive, store, analyze, listen to and write user input data from a client after dictation. Recovery after emergency by saving as temporary data after interval or certain amount of input A server having a data recovery module for enabling a user and a playback control unit for replaying a specific section repeatedly or easily moving back and forth between sections during keyboard input, and collecting user input data for multi-faceted evaluation during listening and writing input. If the data is stored in the server from time to time, and the data is finally transmitted to the server after the input is completed, the user raw data collection unit for evaluation, the dictation data upload unit for uploading the voice data and the text for listening and writing, and the sync between the voice data and the text data Includes a client with sink and annotation editor for editing and annotating.

The server may further include a joint recording data splitting module for automatic coordination between workers for recording collaboration and a recording result integration module for integrating the collaboration recording results.

The client may further include a recording request posting unit for posting a recording request and payment cost and selecting a recording author.

The client may further include a recording result verification unit providing an environment for verifying the recording result.

When the client presses a predetermined key while listening, the client may further include a voice annotation playback unit provided with a description related to the point in time.

The effects of the present invention are as follows.

First, it has the effect of quantifying the ability to listen and dictate in multiple ways.

Second, analyzing raw data for multi-faceted assessments can increase the likelihood of determining whether a human or a non-human program has been entered, and whether the evaluator is a human, purely listening or dictating, or using other negative methods in online assessment. It is effective to confirm.

Third, by using the repeat play control function and the automatic cut-out function, the recording work time is reduced to less than four times the length of the recorded file even for a general person who has no problem using the keyboard.

Fourth, by using the sync editing function, multiple levels of explanation or translation can be added to a specific location of a voice file, which can be used as an auxiliary material for voice lecture materials or as a tool for translation education.

Fifth, the recording work can be done jointly, and a new market can be created by providing dictation materials and using processors, dictation materials users, and server operators.

Sixth, it can create a new market consisting of call requesters, record performers, and recording server operators, and create a new market consisting of voice lecture commentaries, voice lecture comment users, and voice lecture comment file server administrators. There is a new market creation advantage for book or audiovisual lectures.

BRIEF DESCRIPTION OF THE DRAWINGS The accompanying drawings, which are incorporated in and constitute a part of the specification, illustrate preferred embodiments of the invention and, together with the description, serve to further the understanding of the technical idea of the invention, And shall not be construed as interpretation.
1 is a block diagram showing an online listening dictation system according to an embodiment of the present invention;
2 is a conceptual diagram for explaining a listening dictation system online according to an embodiment of the present invention;
3 is a conceptual diagram showing online recording in a listening dictation system online according to an embodiment of the present invention;
4 is a conceptual diagram illustrating a voice lecture annotation material business model in an online listening dictation system according to an embodiment of the present invention;
5 to 9 are conceptual diagrams illustrating a server-client interaction order in a listening dictation system online according to an embodiment of the present invention.

Hereinafter, a preferred embodiment of the listening dictation system online according to the present invention will be described in detail.

1 is a block diagram showing a listening dictation system online according to an embodiment of the present invention. Referring to FIG. 1, the system for listening and dictating online according to a preferred embodiment of the present invention includes components including a server 100 and a client 200, which will be described in detail as follows.

The server 100 includes a broken read module 110, a dictation data storage module 120, a dictation multifaceted evaluation module 130, and a data recovery module 140.

The reading module 110 is a software module that implements an algorithm for classifying words, idioms, and sentence units. When a listening MP3 file is provided, the voice reading module 110 analyzes the voice activity of the MP3 file into a voice section and a non-voice section. Divide.

When the value of the time length of the non-speech interval is derived, the reading module 110 obtains a probability distribution map according to the time length, and stores information by dividing the sentence into a sentence, an idiom, and a word according to the probability. do. Then, the server 100 stores information divided into respective sentences, idioms, and words corresponding to the listening MP3 file.

Here, the divided information is provided to the client 200 when the user listens and writes and reproduces the words, idioms, and sentences.

However, since the information generated in the module is classified based on the non-voice interval, further reconfirmation is required, which can be refined more precisely through the sync editing function between the voice file and the text file.

In other words, the sync editing function between the voice file and the text file is a function that allows the user to easily dictate and break the voice file in the non-speaking parts, at appropriate time intervals, or in words, idioms, contexts, and sentences.

Conventionally, while the dictation is repeatedly performed in sentence units, the present invention avoids the simple repetition of such sentence units, and the information is measured in advance by the sentence division, the idiom division, and the word division. Save and cut off using this information. Therefore, the server 100 previously stores the division unit information of the sentences, idioms, and words, and after the client 200 requests the listening MP3 file from the server 100, the listening MP3 from the server 100. When the file and the division unit information are received, the user may provide a service to cut and play a sentence, an idiom or a word by using the listening MP3 file and the division unit information.

In addition, it provides a convenient dictation environment by adding a function of replaying only by user specific input or playing back and forth by user specific input, not continuous repeat playback.

In addition to the voice file and the corresponding basic text to use the dictation function, the side (data provider) that provides time information on the time point at which each word is played in the voice file and information that distinguishes them by the semantic unit of the sentence. It provides the ability to broker transactions between the parties who listen and dictate by using.

It also acts as a mediator between the provider of the processed information and the user of the data using these services, adding not only dictation material but also interpretation information in units of words, idioms, phrases and sentences to use the voice annotation function. Provides various usage environments of users and data users.

The dictation data storage module 120 is a software module that finds a user's desired level of listening material. In the case of audio data of listening and dictation, the level or field may be distinguished according to the speed of reproduction or the word or content used.

It acts as a search engine that finds voice data that you want to hear and dictate based on a keyword or request that a user presents.

For this function, the server needs to analyze the related information in advance and make a database. It performs the functions of such a proactive database and search engine.

Conventional dictation programs or services are written and listened to for short sentences or a few sentences, so there is no need to move the voice file back and forth.

However, in order to dictate a voice file of several minutes or more, it is necessary to repeat playback of the current playback portion and to move forward and backward as the user intends.

In addition, in order to facilitate such playback control while dictating what is heard, the present invention forcibly assigns a keyboard key, which is not well used for recording, to use a break reading function and a convenient repeat playback control function. It can greatly reduce the trouble of listening and dictation.

As an example of being used as a playback control key, when the Shift + Tab key is input, the previous part is played back in time (it may be a sentence, a phrase, or a word, depending on a user's designation). The part is repeatedly played back, and when the shift + enter key is input, the next part in time is played.

In addition, the backslash key performs a pause or replay function during playback, and the enter key performs a function of replaying the next part in time as the cursor moves to the next line.

Since Shift + Tab, Tab, Backslash, and Shift + Enter are not used for actual recording, the text editing functions are limited and used only for playback control. It is an object of the present playback control to provide a function such that an ordinary user can intuitively and naturally dictate by such an implementation.

The dictation evaluation module 130 is a software module that receives the dictation and stores the user input information from the client and stores the data on the server, and analyzes the ability to analyze and hear the dictation. Various information that can be measured at the client terminal (for example, the time interval from the point of time to the dictation after the playback of a specific section, the number of repetitions until the dictation, the text of the dictation, and the length of the sentence dictated once) The user can objectively evaluate the ability to hear and dictate, and all information that can be measured and recorded on the client terminal) is transmitted to and inputted into this module.

However, it is designed so that other information can be added in addition to the four input information presently presented, and a mapping function for quantifying the information in one dimension can be selected from various types.

At this time, the mapping function is an element other than the IT technology, and the application program interface (API) is disclosed to be accessible from the aspect of language education, so that the mapping function can be made and applied individually.

The input information is defined as the time interval from the time of playback to the time of dictation, the number of repetitions of playback until the time of dictation, the dictation result text, and the length of sentence dictation at one time. Includes all additional information that can be objectively assessed and measured and recorded on the client terminal.)

Conventional dictation programs or web services only judge noon by simply comparing the resulting text.

However, the present invention allows comprehensive determination using all the input information generated during listening and dictation.

For example, even if the input text and the answer are correct, the ability to listen and dictate may be evaluated differently according to the number of repeated listening, and even if the listening and dictation is performed at one time, the sentence length may be evaluated differently.

On the other hand, it provides a function that can develop and apply a mapping algorithm for dimensionalizing the information of the four different criteria that occurs after the collection of the raw data for the multi-faceted evaluation.

Mathematically, this function can be expressed as a mapping function that represents a 4-dimensional array as a real number (one-dimensional).

There can be several such functions, which will be numerous because they can be applied in many ways after being analyzed pedagogically.

In general, if a long sentence is written at a time within a short time even if the text results are the same by giving different weights for the time interval from the point of time when the specific section is played to the point of dictation, the number of repeated listening and the length of the dictated sentence. The higher scores are common sense, but these functions can vary depending on the type of weight.

In this case, it is possible to develop other evaluation methods that take into account language education aspects, and the user may choose to complement or compete with the evaluation method itself.

The data recovery module 140 is a software module that stores a temporary file after a certain time interval or a certain amount of input and restores it in an emergency. The current data after the key input of a specific time interval or a specific number of times is generated by the server. It is automatically stored in the storage area and used as data that can be recovered later.

The client 200 includes a playback controller 210, a user raw data collector 220 for multifaceted evaluation, a dictation data uploader 230, and a sink and annotation editing unit 240.

The playback control unit 210 is a JavaScript code for playback control and keyboard input. The playback controller 210 stores keyboard usage optimized for listening and dictation, and various user input information generated while listening and dictating sections through such usage.

The reproduction control unit 210 is configured to a multi-dimensional array for the continuous storage of the user input information and the user input information is stored.

The multi-faceted evaluation user raw data collection unit 220 is a JavaScript code for storing the user input information from time to time during listening and writing input, and finally transmitting the information to the server after the input is completed. It is stored in the client browser area in the form of a multi-dimensional array, which is sent to the server for temporary recovery and temporarily stored for emergency recovery.

At this time, when the user finally selects the work completion approval, the temporarily stored contents are deleted and only the final version remains.

The dictation data uploading unit 230 is a voice file and text upload JavaScript code for listening and writing. The dictation data uploading unit 230 may register and use a user arbitrary file.

In other words, the MP3 file containing the voice content and the text content of the voice file are uploaded to the server side at once. In this case, when there is no text and only a voice file, only the voice file is transmitted and stored to the server so that the text can be added later or dictated using this service.

The sink and annotation editing unit 240 is a JavaScript code for sync editing and annotation between the voice file and the text file, and is synchronized with only the voice file and the text file.

In other words, through sink editing, information is processed and stored in a special format. In addition to the sink editing feature, it is implemented in JavaScript code that includes text editing to add comments such as words, idioms, and sentence-by-phrase interpretation.

Sink editing and annotated information are sent to the server for storage. The server is stored in pairs with the corresponding MP3 file.

On the other hand, the server 100 may further include a recording result integration module 160 for integrating the joint recording data splitting module 150 and the collaboration work recording results for automatic coordination between the workers for the recording collaboration work.

The joint recording data dividing module 150 is an automatic coordinating software module for recording a collaborative work, and divides a file to be recorded by a predefined method so that one or more recording workers can work online or offline. do.

At this time, a function of dynamically rescheduling the allocation of each split worker is included. In addition to reducing the work time through the splitting operation, the function of lowering the error through the redundant operation is included.

When used for the purpose of recording in the present invention, as an online environment that can be shared by several people, that is, a function that can be used overlapping or independently of the duplicate recording function, recording is generally based on the time of the actual voice file Since it takes at least about two to three times as much time, it is possible to shorten the working time by dividing the work by several people. This recording collaboration is then built online.

When recording using the co-recording function, it is a function to broker a transaction between the person who wants to record and the person who is recording. (A side: A group of one or more recording workers who listen to and dictate voice files).

For example, if side B uploads a voice file, posts a recording work completion time, or is willing to pay, the side A is assigned a task through a selection process such as bidding or first-come-first-served basis, and is provided online or offline. A recording work program is provided and completed either alone or in collaboration.

After completion of the recording work, B side checks through the inspection process such as the comparison function of duplicate recording results and completes payment.

At this time, the web service brokers transactions by providing on-line and off-line environments necessary for work occurring at each stage.

In addition, the recording result integrating module 160 is a software module for integrating the collaborative recording result, and is a module in which a function of combining or selecting the recording result of the divided work or the recording result through the duplicate work is implemented.

When each worker uploads the data to the server as a result of working online or offline, the server arranges them so that they do not overlap by using the worker's recording timestamp information, so that the original MP3 files are aligned.

If there is a duplicate work part, the comparison function can be divided into the same case and the other case, and the user (recording result inspector) can select the case.

Meanwhile, the client 200 may further include a recording request posting unit 250, a recording result verification unit 260, and a voice annotation playback unit 270.

The recording request posting unit 250 is a bulletin board code that can be selected by the recording request posting and payment costs and the recording author, the recording request side should upload the prepared MP3 file, this function is implemented in the module.

In addition to uploading MP3 files, the list of requirements, such as information about the maximum cost you are willing to pay, will be made available to the recording worker for viewing.

Here, the expenses to be paid are implemented in a manner that is deposited to the server administrator, and related work is implemented through this module. Recording workers begin work after receiving approval through the process of confirming work through this module.

At this time, the recording work environment is provided by on-line or off-line software in this service, and the work result data is transmitted to the server and merged.

The recording result verification unit 260 is a JavaScript code that provides an environment for verifying the recording result. When the recording operation is completed, an environment in which the recording operation requester can check the recording result as a sampling or the whole through the module is described. to provide.

The recording requester is satisfied and the transaction between the recording requester and the worker is terminated through the work completion confirmation process.

After the transaction ends, the cost and result information presented by the requester is sent to the worker.

It also includes a function that arranges on the screen to compare or select the voice part and the recording result when multiple people record repeatedly.

Through the code, different recorders duplicate the same voice part to store the result, thereby improving the recording quality.

That is, in the case of a text file without text, at least one or more people hear and dictate the result text, or check the text and the original voice content again, thereby providing an environment for determining an error.

And, if there is no corresponding text in the voice file, it is a function to compare the result of listening and dictation by several people, and the time information of the voice file is recorded together in the text while listening and writing. This makes it easy to compare people's dictation results at the same location in the voice file.

The voice annotation playback unit 270 is a JavaScript code that appears when the help key is pressed during listening, and the description is related to the point in time. This is the part where the function to listen to the explanation of the corresponding part is implemented through the specific input (mouse click or keyboard input in the browser, specific touch in the smart phone, etc.).

For example, in the case of English listening, explanations can be given in the order of words, idioms, and sentences according to the level after a particular input.

During the description, English playback is stopped, and the description content is played back in the MP3 file for annotation, in which the contents of the words, idioms (contexts), and sentences entered by the provider are stored separately.

In other words, while listening to and writing a foreign language such as English, or only listening, the function may be interpreted by dividing it into words, phrases, or sentence units through a "help" input.

This function helps you to interpret the lecture in a foreign language and interpret it in the uninterpreted part. You can use the data that has been pre-populated with words, idioms, phrases and sentences. If you do, it is a function that shows or speaks the information.

In general, lecture materials are one-dimensional and passive.

One-dimensional means that you can only move over time, and it's passive, so the listener needs to keep listening.

This function provides a convenient environment to escape this passive, one-dimensionality.

For example, in the case of a foreign language lecture voice file, a separate dictionary or reference material was used when an unknown word, context, or sentence appeared.

However, the present invention ameliorates this inconvenience, allowing the help of the time to be provided at various stages (eg, word level, context level and sentence level) through a "help" input at that time location.

A voice annotation player (program type or web service) is provided for the voice annotation function.

In addition, an online or offline voice annotation editing program is provided to conveniently edit voice annotation subtitles.

While the present invention has been described in connection with what is presently considered to be practical exemplary embodiments, it is to be understood that the invention is not limited to the disclosed embodiments, but, on the contrary, And such variations and modifications are intended to fall within the scope of the appended claims.

100: Server 110: Disconnected Read Module
120: dictation data storage module 130: dictation multi-faceted evaluation module
140: data recovery module 150: joint recording data partitioning module
160: recording results integration module 200: client
210: playback control unit 220: multi-sided user raw data collection unit
230: dictation data upload unit 240: sink and annotation editor unit
250: recording request posting unit 260: recording result verification unit
270: voice annotation playback unit

Claims (7)

A dictation data search module for searching for a voice file of a level desired by the user and searching for the voice file matching a keyword or a request presented by the user;
Analyze the value of the time length of the non-voice interval for the speech file to obtain a probability distribution map according to the time length, and store the data by dividing it into sentences, idioms, and words, respectively, according to the obtained probability, and reproducing the dictation of the user. A disconnect reading module for providing a voice file and said stored data;
Receive and store dictation data of the user for each voice file from the client below, and include items of dictated text in the stored dictation data by default, and items of a time interval from the time of reproduction until the dictation time until dictation Dictate evaluation module for analyzing the dictation ability of the user by evaluating by each evaluation item if at least one item of the repetition number of times of repetition and sentence length dictated at one time is further included and
A server comprising a data recovery module for storing the dictation data as a temporary data after a predetermined time interval or a predetermined amount of input and enabling recovery at an emergency end; And
At the request of the user, the playback control unit for playing the speech file provided by the cut-out diary module by cutting the requested sentence, idiom or word, and repeatedly playing a specific section of the speech file or moving back and forth to the section. Wow,
A multi-faceted user raw data collection unit for collecting the dictation text input by the user corresponding to the reproduced voice file for each multi-faceted evaluation item and transmitting the dictation text to the server as the dictation data;
And a client including a dictation data upload unit for uploading the voice file and text for dictation. delete The method of claim 1, wherein the server,
Joint recording data splitting module for automatic coordination between workers for transcript collaboration
And further comprising a recording result integration module for integrating the collaboration recording results. The method of claim 1, wherein the client,
A dictation system for online listening and dictation, further comprising a recording request posting section selected by the recording author and the recording request and payment cost. The method of claim 1, wherein the client,
Listening dictation system online, characterized in that it further comprises a recording result verification unit for providing an environment for verifying the recording result. The method of claim 1, wherein the client,
And a voice annotation reproducing unit which provides a description related to the point in time when a predetermined key is pressed while listening. The method of claim 1, wherein the client,
And a sync and annotation editor for sync editing and annotating the data of the voice file and the text.

Images