KR102343336B1 - Video Authoring System and Method - Google Patents

Abstract

The present invention relates to a system for producing a video work using speech recognition and speech synthesis, comprising: an audio extraction module 100 for extracting audio data from a video image; a speech recognition module 200 for recognizing the audio data extracted from the audio extraction module 100 and converting it into text to generate an explanatory audio script; a script editing module 300 for editing the explanatory audio script generated through the voice recognition module 200; a speech synthesis module 400 for recognizing the finally completed explanatory audio script through the script editing module 300 and synthesizing the speech to generate audio material of the work; a video extraction module 500 for extracting a video file from a video image to generate a work video material; and a work creation module 600 for generating a moving picture work by mixing the work audio material and the work video material.

Description

Video Authoring System and Method}

The present invention relates to a new concept of a video authoring system and method that can conveniently and quickly produce or edit a video work by utilizing speech recognition technology, speech synthesis technology, automatic translation technology, audio and video editing technology, and the like.

In general, video works such as video lecture materials are produced by using a recording device such as a camera to directly record lecture contents and edit them. The shooting process can be repeated several to tens of times until the user who is not accustomed to shooting can express the desired content well. There is a problem that the process is cumbersome.

Among the existing technologies for generating lecture materials, there is Patent Publication No. 10-2017-0029725, which is a lecture video using video synthesis, a presentation video, or a video to be transmitted to the other party in a video conference (hereinafter collectively referred to as 'lecture'). A device for generating an image (referred to as 'image'), and a camera image receiving module that receives a captured image (hereinafter referred to as 'camera image') of a lecturer or video conference (hereinafter referred to as 'instructor') or audience from a shooting camera ; an editing interface providing module for providing an editing screen for configuring a screen of a lecture video in 2D or 3D to a user; From the user, with respect to a plurality of images (hereinafter referred to as 'composition images') including the camera image and presentation material screen image, a setting (hereinafter '' a setting input receiving module for receiving input through the editing screen using a mouse, a keyboard, or a wireless pointing device, etc.; a lecture image generating module for generating a lecture image configured according to the received screen layout setting; When a specific condition is satisfied, a lecture video adjustment module for changing the configuration of the lecture video according to the satisfied condition; and a controller for controlling each module of the apparatus for generating a lecture image using image synthesis to perform a series of processing related to generating a lecture image using image synthesis.

If such a lecture video generating device is used, (a) materials such as PPT and video images to be used for the presentation, and the video captured by the lecturer, presenter, video conference (hereinafter referred to as 'instructor') or the audience from a plurality of cameras (hereinafter referred to as 'camera image') and receiving one or more presentation material screen images; (b) an editing screen for composing lecture video screens to be shown to the other lecturer in 2D or 3D space to the lecturer or user (hereinafter collectively referred to as 'user') who uses the lecture image generating device; and presentation; It provides a composite scene of inputs from cameras, such as data and presenters, and sets it on a plurality of 2D layers or 3D space (hereinafter referred to as 'screen layout setting') in order to make the composition images into a beautiful composite screen with various scenes. , setting in advance the synthetic scenes to be used while giving a lecture in real time; (c) the user selects a lecture video screen in real time and gives a lecture, in which the user or the lecturer alone manipulates the lecture material by simply using a mouse, a touch screen, or a pointer such as a wireless pointing device to switch lecture materials and change the screen layout It allows transitions, scene transitions, and presentation material changes at the same time, and provides a real-time lecture by pointing the pointer and switching presentation materials in real time while viewing the necessary composite scene by simply operating with a pointer such as a mouse, screen touch, or wireless pointing device. step; (d) The lecturer gives a lecture, and the video produced in real time, such as synthetic scenes and lecture material screens, is recorded, and the output is used as an input to the video conference program, so that the output video is transmitted to the video conference program like a single camera. recording and transmitting using the remote presentation conferencing method to be made available for use; and (e) when a specific condition is met, changing the configuration of the lecture video according to the satisfied condition can be implemented, but voice recognition technology, speech synthesis technology, automatic It does not provide a concept for conveniently and quickly creating or editing a moving image by utilizing translation technology, audio and video editing technology, etc.

[Prior art literature]

Publication No. 10-2017-0029725

Registered Patent No. 10-1292563

It is an object of the present invention to provide a new concept of a video authoring system and method capable of conveniently and quickly producing or editing a video work by utilizing speech recognition technology, speech synthesis technology, automatic translation technology, audio and video editing technology, etc. do it with

The technical configuration of the present invention created to achieve the above object is as follows.

The present invention relates to a system for producing a video work using speech recognition and speech synthesis, comprising: an audio extraction module 100 for extracting audio data from a video image; a speech recognition module 200 for recognizing the audio data extracted from the audio extraction module 100 and converting it into text to generate an explanatory audio script; a script editing module 300 for editing (adding/correcting/deleting) an explanatory audio script generated through the voice recognition module 200; a speech synthesis module 400 for recognizing the finally completed explanatory audio script through the script editing module 300 and synthesizing the speech to generate audio material of the work; a video extraction module 500 for extracting a video file from a video image to generate a work video material; and a work creation module 600 for generating a moving picture work by mixing the work audio material and the work video material.

In addition, the present invention relates to a method for producing a video work using a video authoring system, comprising: an audio data extraction process of extracting audio data from a video image previously produced through the audio extraction module 100; A descriptive audio script generation process in which the speech recognition module 200 recognizes the voice of the audio data extracted from the audio extraction module 100 and converts it into text to generate a descriptive audio script; a work audio material generation process in which the speech synthesis module 400 recognizes the finally completed explanatory audio script through the script editing module 300 and synthesizes the speech to generate the work audio material; a work video material creation process in which the video extraction module 500 extracts a video file from a pre-produced video image to generate a work video material; and a moving picture work creation process in which the work creation module 600 creates a moving picture work while mixing or editing the work audio material and the work video material.

Technical effects according to the configuration of the present invention are as follows.

First, using voice recognition technology to create explanatory audio scripts, edit (add/correct/delete) the generated explanatory audio scripts as needed, and then use the speech synthesis technology to create audio material for the work. It can be mixed with video material to create a video work with a quality suitable for your requirements.

Second, since the multilingual translation module 800 for translating the text included in the explanatory audio script into multiple languages is provided, it is possible to produce an image work translated into various languages.

Third, in the process of extracting the audio data included in the video file, the purpose can be specified by separating it into explanatory audio and background effect audio, and the speech recognition module 200 recognizes only the content specified as explanatory audio and converts it into text for explanation The audio script is generated, and the voice synthesis module 400 recognizes the explanatory audio script for the content designated as the explanatory audio and synthesizes the voice to generate the work audio material, and the content designated as the background effect audio is the voice recognition module 200 ) and the speech synthesis module 400 can be used as it is as an audio material of the work.

Fourth, it is easy to modify the work by modifying and supplementing the explanatory audio script (authoring script) once produced, and making a new work by cutting and collecting script blocks between several explanatory audio scripts (authoring script), changing the order, etc. It is easy, and it is possible to make various and detailed adjustments to the work by changing the voice or gender of the voice synthesis or adjusting the pronunciation speed in the explanatory audio script (authoring script) created once.

1 is a block diagram showing the overall configuration of the present invention.
2 is a flowchart illustrating a video work production process according to the present invention.

Hereinafter, specific embodiments of the present invention will be described in more detail with reference to the accompanying drawings.

The present invention relates to a system for producing a video work using voice recognition and voice synthesis, and as shown in FIG. 1 , an input module 700, an audio extraction module 100, a voice recognition module 200, and a script editing module 300 , a voice synthesis module 400 , a video extraction module 500 , a work creation module 600 , a multilingual translation module 800 , a block recognition module 900 , and the like.

These components can be composed of various hardware devices and programs (including web applications) that are built-in and perform pre-set arithmetic operations, and are divided based on functions. A mouse, a digital camera, a monitor, and a main body having a built-in memory and CPU may be configured.

The audio extraction module 100 serves to extract audio data from a video image.

The video image includes an image, explanatory audio, and background effect audio. The audio extraction module 100 can separate and extract audio data from these video images. For this operation, the audio extraction module 100 includes ffmpeg or Moviepy Commercially available video editing programs such as .

In the process of extracting audio data, the audio extraction module 100 may designate a use by dividing the audio data into explanatory audio and background effect audio.

The voice recognition module 200 recognizes only the content designated as explanatory audio and converts it into text to generate an explanatory audio script, and the speech synthesis module 400 recognizes the explanatory audio script for the content designated as explanatory audio and performs speech synthesis. This will create a work audio material.

As can be seen in the flowchart of FIG. 2 , the content designated as background effect audio in the audio data extraction process is used as an audio material of the work as it is without going through the voice recognition module 200 and the voice synthesis module 400 . That is, it is used after being mixed with the explanatory audio that has passed through the speech synthesis module 400 .

The block recognition module 900 distinguishes and recognizes the time zone (slot) in which the description content designated as the description audio exists among the audio data extracted from the audio extraction module 100 and the time zone (slot) in which the description content does not exist, and the description audio is If there is no explanatory audio for a preset period of time, it is divided into another script block.

The function of the block recognition module 900 can be selected as needed. When the block recognition module 900 is operated, the audio data extracted from the audio extraction module 100 is converted into a plurality of scripts by the block recognition module 900 . Blocks are automatically divided, and the voice recognition module 200 may generate a descriptive audio script for each divided script block.

When an explanation audio script is created for each script block, the script editing process, multilingual translation process, and work audio material creation process can be performed for each divided script block, and it is matched with the audio material of the work created for each divided script block While the work creation module 600 mixes or edits the work video material to be used, it is possible to finally create a moving image work.

This script block is the basic unit of work that creates the final work. That is, each script block is listed in order in the final work to compose one moving picture, but the work process of translating and synthesizing each script block and combining it with a video is an independent process that is not related to each other.

Therefore, the work time for creating the final work is shortened in the following two ways.

1) Parallel work: In a batch job that regenerates works from start to finish, each script block creates works by block at the same time as parallel work, and when all works by block are created, they are combined into a single video.

2) Background work: Usually, the editing work sequentially creates and edits script blocks. While editing the next block, the previously edited block is created for each block in the background process. Usually, the block work creation time is shorter than the editing work time, so it is finished before moving on to the next block edit, but even if it is not completed, it is possible to proceed with the creation of several block works at the same time. After editing the last block and creating the last block asset, all you need to do is assemble all the block assets into the final asset.

Usually, it takes the longest time to create a work for each script block, and if this operation is processed as a parallel operation, the operation time is reduced by about 1/5 in the case of 50 blocks of 20 minutes video, and if processed as a background operation, the script block in the background of the editing operation Because you create an asset, the creation process takes very little time.

In the process of creating a final work by combining works by script block, parallel work or background work is not possible, so work is done sequentially separately.

The voice recognition module 200 recognizes the voice of the audio data extracted from the audio extraction module 100 and converts it into text to generate an explanatory audio script, extracting the explanatory audio (voice) from the video image and The process of generating an audio script by recognizing a voice can also be confirmed in the flowchart of FIG. 2 .

Such a speech recognition module 200 may perform a speech recognition (STT, Speech To Text) function by calling and using an API (Application Program Interface) provided by google, amazon, or naver. It can be found at the URL of

google:

https://cloud.google.com/speech-to-text/docs/apis?hl=en

amazon:

https://aws.amazon.com/en/transcribe/

naver:

https://www.ncloud.com/product/aiService/csr

The script editing module 300 performs a function of editing (adding/correcting/deleting) an explanatory audio script generated through the voice recognition module 200, and a general text editor function including a memo pad can be used.

In other words, you can add information necessary for authoring. Use commas or periods to indicate the rest and end positions of the work description, or adjust the rest and end times by adding or subtracting the number of commas or periods. Background music or background music can be added during this break.

In addition, the misrecognized text can be freely corrected, and necessary explanations can be added.

That is, as can be seen from the flowchart shown in FIG. 2 , the explanatory audio script can be edited, corrected, or updated, and can be newly typed or copied and reused.

The voice synthesis module 400 recognizes the finally completed explanatory audio script through the script editing module 300 and synthesizes the voice to perform the process of generating audio material of the work, as can be seen in the flowchart of FIG. 2 , such voice synthesis The module 400 may perform a speech synthesis (TTS, Text To Speech) function by calling and using an API (Application Program Interface) provided by google, amazon, or naver. The specific usage method is at the following URL. can be checked

google:

https://cloud.google.com/text-to-speech/docs/reference/rest

amazon:

https://aws.amazon.com/en/polly/

naver:

https://www.ncloud.com/product/aiService/css

The video extraction module 500 extracts a video file from a video image as shown in the flowchart of FIG. 2 to generate a work video material. For this task, the video extraction module 500 includes an existing Use the editing function of a commercial video editing program of , and designate the start and end positions of the part to be used in the original video image.

The video extraction module 500 may extract some or all of an image, text, or PPT (presentation) material as can be seen from the flowchart of FIG. 2 as well as an existing video image to generate a work video material.

In other words, you can use photos, screen captures, and camera shots made in various applications as a source, create a video with text (characters) to be displayed on the screen, and video containing PPT screen switching/animation functions, etc. can also be used as a video source.

As can be seen from the flowchart of FIG. 2, the work creation module 600 performs the process of creating a moving picture work while mixing or editing the work audio material and the work video material. An editing function of a video editing program may be used.

The input module 700 may be a device such as a microphone for inputting voice or a keyboard or mouse for inputting text. is recognized and converted into text to generate an explanatory audio script, and audio material of the work is generated through the speech synthesis module 400 .

When a mouse or keyboard (or touch screen) is selected as the input module 700 , and text is inputted through this, an explanatory audio script is directly generated without going through the voice recognition module 200 , and the voice synthesis module 400 is used. Through this, the work audio material is created.

As a method of entering text, you can directly type the content or use a text manuscript of an existing document (a method of cutting or copying and pasting a text portion of an existing document).

The multilingual translation module 800 performs a process of translating the text included in the explanatory audio script into multiple languages, as can be seen in the flowchart of FIG. 2 .

The multilingual translation module 800 may utilize Google's Translater API, etc., and the speech synthesis module 400 recognizes the explanatory audio script translated into multiple languages through the multilingual translation module 800 and performs speech synthesis in multiple languages. This will create a work audio material.

That is, by using the explanatory audio script written in the Korean language, it is possible to easily create an audio material of a work composed of multiple languages.

The method for producing an audio material of a work according to the present invention includes the following process.

(1) Audio data extraction process

As a process of extracting audio data from a video image produced in advance through the audio extraction module 100 , a commercial video editing program such as ffmpeg or Moviepy may be embedded in the audio extraction module 100 .

Alternatively, if you speak into the microphone during the audio script editing process, you can create or edit an audio script with the STT function.

The video image includes an image, explanatory audio, and background effect audio, and the audio extraction module 100 can designate a use by separating it into explanatory audio and background effect audio in the process of extracting audio data.

In the audio data extraction process, a plurality of script blocks can be divided and specified as needed. The block recognition module 900 is a time period ( slot) and a non-existent time slot (slot) are recognized, and if the content designated as explanatory audio exists and the content designated as explanatory audio does not exist for a preset period of time, it is divided into another script block. The subsequent process may be performed for each divided script block.

(2) Description audio script creation process

The voice recognition module 200 is a process of recognizing the voice of the audio data extracted from the audio extraction module 100 and converting it into text to generate an explanatory audio script, the voice recognition module 200 is google, amazon, or Speech recognition (STT, Speech To Text) function can be performed by calling and using the API (Application Program Interface) provided by naver.

The voice recognition module 200 recognizes only the content designated as the explanatory audio and converts it into text to generate the explanatory audio script, and the content designated as the background effect audio in the audio data extraction process is the voice recognition module 200 and the voice synthesis module ( 400) and is used as an audio material for the work as it is.

The description audio script is a method of directly inputting a voice through the input module 700 such as a microphone, and the voice recognition module 200 recognizes the voice and converts it into text to generate the explanation audio script, or a mouse or keyboard (or touch screen). ), when text is input through the input module 700 , a method of directly generating an explanatory audio script without going through the voice recognition module 200 may be utilized.

In the case where each script block is divided and designated, the voice recognition module 200 generates the explanatory audio script for each divided script block in this explanatory audio script generation process.

(2-1) Script editing process

As a process in which the script editing module 300 edits (adds/corrects/deletes) the explanatory audio script, a general text editor function including a notepad may be used, or a specialized editor tailored to the purpose of the explanatory audio script may be used.

In other words, you can add information necessary for authoring. Use commas or periods to indicate the rest and end positions of the work description, or adjust the rest and end times by adding or subtracting the number of commas or periods. Background music or background music can be added during this break. In addition, the misrecognized text can be freely corrected, and necessary explanations can be added.

This script editing process is performed for each divided script block.

An important function of the present invention is to edit an existing moving picture work. For example, if you change some explanatory audio or some video of an existing moving picture work, you can change it with little effort by using the present invention.

In particular, in the case of lecture video works, there are many cases where editing is mainly focused on the audio of explanations.

In the existing method, editing is done after re-shooting or re-recording, but the present invention automatically divides the video work into blocks with the block recognition function and then edits only the description audio of the block to be edited with the editing method of the present invention. Editable with little effort and time.

Of course, in this case, there may be some recognition error in the description audio that automatically recognizes the block not to be edited. In this case, there are two choices.

(1) How to use only the original audio by putting the original audio as a background sound and not including the explanatory audio

(2) How to fix all recognition errors of recognized explanatory audio scripts

In the case of method (1), the explanation audio of the modified block is a machine synthesized sound, and the original audio is different because the voice of the original author. For lecture writings, where the accuracy of content delivery is more important, this is an option.

In the case of method (2), it takes a lot of initial effort and time, but in the case of a work that will be used for a long time, such as a lecture video, it is possible to modify the work created once with very little effort after the second revision.

According to my experience in making lecture works, it takes quite a bit of effort to make the lecture video I wrote into an audio script for the first time when the accuracy of lecture content is important, such as in mathematics or engineering, but once the audio script is created, the lecture content can be changed. The process of listening and editing the lecture video (by increasing the speed) is repeated several times in order to increase the accuracy and the understanding of the listener.

(2-2) Multilingual Translation Course

The multilingual translation module 800 is a process of translating the text included in the description audio script into multiple languages, and the multilingual translation module 800 may utilize Google's Translater API and the like.

When a multilingual translation process is added, the speech synthesis module 400 recognizes the explanatory audio script translated into multiple languages through the multilingual translation module 800 and synthesizes the speech in multiple languages to generate audio material of the work or multilingual subtitles corresponding thereto. do.

This multilingual translation process is performed for each divided script block.

(3) The process of creating audio material of the work

The voice synthesis module 400 recognizes the finally completed explanatory audio script through the script editing module 300 and synthesizes the voice to generate audio material of the work. This voice synthesis module 400 is google, amazon, or The voice synthesis (TTS, Text To Speech) function can be performed by calling and using the API (Application Program Interface) provided by naver.

In the case of being divided and designated for each script block, the speech synthesis module 400 performs speech synthesis for each divided script block to generate audio material of the work.

(4) The process of creating a work video material

The video extraction module 500 is a process of extracting video data from a pre-produced video image to generate a work video material. and designates the start and end positions of the image to be used in the original video image.

The video extraction module 500 may extract some or all of images, texts, or PPT (presentation) materials as well as existing video images to generate copyrighted video materials, photos or screen captures produced by various applications. , camera photos, etc. can be used as a source, a video can be made with text (characters) to be displayed on the screen, and a video including a screen change/animation function of PPT can be used as a video source.

(5) The process of creating a video work

The work creation module 600 is a process of creating a video work while mixing or editing the work audio material and the work video material.

For this task, the editing function of an existing commercial video editing program such as ffmpeg or Moviepy can be used.

When the script block is divided and designated, the work creation module 600 mixes or edits the work audio material generated for each divided script block and the work video material matching it to create a moving picture work.

(6) The entire process for creating a moving picture work can dramatically improve the processing speed through parallel processing of data. Because processing is possible, multiple script blocks can be processed in parallel and at the same time.

When the work for each script block is completed, it is only necessary to collect the asset elements for each script block to create the final video work.

For example, if one work consists of a first block, a second block, and a third block, the work of the second block and the third block may be performed simultaneously in parallel before the work on the first block is completed, , when the work for each block is completed, the work efficiency can be maximized by collecting the work elements of the first block, the second block, and the third block and performing the editing work to create the final moving picture work, and this parallel processing process has already been commercialized You can use the technical elements that are already in place.

As described above, specific embodiments of the present invention have been described with reference to the accompanying drawings, but the protection scope of the present invention is not necessarily limited only to these embodiments, and various design changes within the scope that do not change the technical gist of the present invention, It is made clear that addition or deletion of known technology, simple numerical limitation, etc. fall within the protection scope of the present invention.

100: audio extraction module
200: voice recognition module
300: script editing module
400: speech synthesis module
500: video extraction module
600: work creation module
700: input module
800: multilingual translation module
900: block recognition module

Claims (10)

It relates to a system for producing a video work using speech recognition and speech synthesis,
an audio extraction module 100 for extracting audio data from a video image;
a speech recognition module 200 for recognizing the audio data extracted from the audio extraction module 100 and converting it into text to generate an explanatory audio script;
a script editing module 300 for editing the explanatory audio script generated through the voice recognition module 200; and,
a speech synthesis module 400 for recognizing the finally completed explanatory audio script through the script editing module 300 and synthesizing the speech to generate audio material of the work;
includes,
In the process of extracting audio data, the audio extraction module 100 separates the audio data into explanatory audio and background effect audio to designate a use,
Among the audio data extracted from the audio extraction module 100, a time zone (slot) in which the content designated as explanatory audio exists and a time slot (slot) in which the content designated as explanatory audio does not exist is distinguished and recognized, and explanation audio is present for a preset time a block recognition module 900 for dividing audio into other script blocks when there is no audio;
This includes more,
The audio data extracted from the audio extraction module 100 is automatically designated and divided into a plurality of script blocks by the block recognition module 900,
The voice recognition module 200 is a video authoring system, characterized in that it can generate a description audio script for each divided script block. In claim 1,
an input module 700 for inputting voice or text;
This includes more,
When a voice is input through the input module 700, the voice recognition module 200 recognizes the voice and converts it into text to generate an explanatory audio script,
A video authoring system, characterized in that when text is input through the input module (700), the audio script is directly generated without going through the voice recognition module (200). In claim 1,
a video extraction module 500 capable of generating a work video material by extracting a video file from a video image, or extracting some or all of an image, text, or PPT material to generate a work video material; and,
a work creation module 600 for creating a moving picture work by mixing the work audio material and the work video material;
Video authoring system, characterized in that it is further included. In claim 1,
The voice recognition module 200 recognizes only the content designated as explanatory audio and converts it into text to generate an explanatory audio script,
The voice synthesis module 400 recognizes the explanatory audio script for the content designated as explanatory audio and synthesizes the voice to generate audio material of the work,
A video authoring system, characterized in that the content designated as the background effect audio is mixed as an audio material of the work as it is without going through the voice recognition module 200 and the voice synthesis module 400 and used. In claim 1,
a multilingual translation module 800 for translating text included in the description audio script into multiple languages;
This includes more
The speech synthesis module 400,
A video authoring system, characterized in that by recognizing an explanatory audio script translated into multiple languages through the multilingual translation module (800) and synthesizing speech in multiple languages, the work audio material and corresponding multilingual subtitles are generated. It relates to a video authoring method using a video authoring system,
an audio data extraction process of extracting audio data from a video image produced in advance through the audio extraction module 100;
A descriptive audio script generation process in which the speech recognition module 200 recognizes the voice of the audio data extracted from the audio extraction module 100 and converts it into text to generate a descriptive audio script;
A work audio data generation process in which the speech synthesis module 400 recognizes the finally completed explanatory audio script through the script editing module 300 and synthesizes the speech to generate the work audio data;
A work video material creation process in which the video extraction module 500 extracts a video file from a pre-produced video image to generate a work video material, or extracts some or all of an image, text, or PPT material to create a work video material ; and,
a video work creation process in which the work creation module 600 creates a video work while mixing or editing the work audio material and the work video material;
including,
The audio data extraction process includes:
In the process of extracting audio data, it is divided into explanatory audio and background effect audio to specify the purpose,
Among the audio data extracted from the audio extraction module 100, the time zone (slot) in which the content designated as the description audio exists and the time zone (slot) in which the content designated as the description audio does not exist is distinguished and recognized, and the description audio exists for a preset time while the description audio exists If does not exist, separate it into another script block,
The process of generating the audio script described above is,
The voice recognition module 200 generates an explanatory audio script for each divided script block,
The process of creating the audio material of the work,
The speech synthesis module 400 performs speech synthesis for each divided script block to generate audio material of the work,
The work creation process is
A video authoring method, characterized in that the work creation module 600 creates a video work while mixing or editing the work audio material generated for each divided script block and the work video material matching it. In claim 6,
The audio data extraction process includes:
Contents designated as background effect audio are mixed and used as audio material of the work without going through the speech recognition module 200 and the speech synthesis module 400 in the process of creating the audio material of the work,
The process of generating the audio script described above is,
A method of authoring a video, characterized in that only the content specified as the description audio is recognized and converted into text to generate the description audio script. In claim 6,
a multilingual translation process of translating text included in the explanatory audio script generated in the explanatory audio script generating process into multiple languages;
is added,
When the multilingual translation process is added, the speech synthesis module 400 recognizes the explanatory audio script translated into multiple languages through the multilingual translation module 800 and synthesizes speech in multiple languages to generate audio material of the work or multilingual subtitles corresponding thereto A video authoring method, characterized in that
delete delete

Images