CN106487531A - A kind of voice automatic record method with automatic error correction function - Google Patents
A kind of voice automatic record method with automatic error correction function Download PDFInfo
- Publication number
- CN106487531A CN106487531A CN201510529137.0A CN201510529137A CN106487531A CN 106487531 A CN106487531 A CN 106487531A CN 201510529137 A CN201510529137 A CN 201510529137A CN 106487531 A CN106487531 A CN 106487531A
- Authority
- CN
- China
- Prior art keywords
- word message
- voice
- voice signal
- conversion
- word
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Pending
Links
Landscapes
- Machine Translation (AREA)
Abstract
The invention discloses a kind of voice automatic record method with automatic error correction function, including：Pre-enter speech roster；Obtain voice signal；Voice signal conversion is identified as by corresponding Word message by voice conversion software, and stored, when identifying conversion voice signal first, choose and to the name corresponding to this voice signal from list, and extract the tamber characteristic of this voice messaging, the name chosen is associated with tamber characteristic simultaneously；When carrying out the conversion identification of next voice signal, judging whether its tone color is identical with the tone color formerly extracted, if identical, the name being associated with this tone color is shown in beginning of the sentence, if differing, entering the operation repeating to identifying conversion voice signal first；Word message is ranked up in a document show, in addition, also having the function that the Word message to sequencing display is corrected.This relatively existing professional equipment and professional, the cost of implementation of the present invention is lower, is substantially all and can realize in various occasions.
Description
Technical field
The present invention relates to signal processing technology field, more particularly, it is that a kind of voice with automatic error correction function is remembered automatically
Recording method.
Background technology
In meeting scene, it is all typically now the voice data to record meeting scene by the way of recording, or by scene
Scribe artificial input record is carried out by special recording equipment.Although both modes are not asked in realization
Topic, but, if adopting the former method, cannot timely on-the-spot meeting record manuscript, need after the meeting by manually to receive
The mode listening session recording is recorded；If using the method for the latter, needing to buy special recording equipment, and need specially
The typist of industry just enables the synchronous recording at meeting scene, and high cost is it is impossible to popularize, only in some official's meeting occasions
Just can use.
Therefore, how quickly to realize the speech of meeting scene spokesman is automatically recorded, and do not need the record of specialty
Personnel and the equipment of specialty, have just become a great problem of the art.
Content of the invention
In view of the above problems, the present invention provides a kind of voice automatic record method with automatic error correction function, for realizing not
Automatically the speech content at meeting scene is recorded in the case of needing professional.Its concrete technical scheme is：
A kind of voice automatic record method with automatic error correction function, including：Pre-enter speech roster；Obtain voice letter
Number；The conversion of described voice signal is identified as by corresponding Word message by voice conversion software, and is stored, wherein, in head
During secondary identification conversion voice signal, choose and to the name corresponding to this voice signal from the speech roster pre-entering, and
Extract the tamber characteristic of this voice messaging, the name chosen is associated with described tamber characteristic simultaneously；Carrying out next language
During message number conversion identification, judge whether its tone color is identical with the tone color of formerly described extraction, if identical, will associate with this tone color
Name show in beginning of the sentence, if differing, enter repeat to identify first conversion voice signal operation；Described word is believed
Breath is ranked up showing in a document, and wherein, described Word message includes correct Word message and wrong Word message；To document
In wrong Word message be marked, and the voice by the described mistake Word message of labelling and corresponding described wrong Word message
Signal is associated linking；When clicking on described mistake Word message, using described voice conversion software pair and described mistake word letter
The voice signal of breath association link is recognized for, and it is aobvious that in a document the secondary Word message identifying is carried out with editable
Show.
Preferably, the method obtaining described voice signal includes：Spoken sounds are converted into by described voice signal by mike.
Preferably, the method for sequencing display Word message includes in a document：Come to described literary composition according to every section of continuous voice signal
Word information carries out segmentation sequencing display.
Preferably, described according to the method that every section of continuous voice signal to carry out segmentation sequencing display to described Word message it is：
When described voice conversion software completes previous voice signal conversion identification, start timing；Arrive in described voice conversion software receipt
During current speech signal, stop timing, and be calculated the time difference between current speech signal and previous voice signal；Judge
Whether the described time be more than default time difference, if so, the Word message that current speech signal conversion identification obtains is carried out point
Section sequencing display；If it is not, the Word message that then current speech signal conversion identification obtains carries out arranged in sequence showing.
Hinge structure, the present invention does not need the typewriting apparatuss turning it is not required that the typing personnel of specialty, can be achieved with to meeting
The automatic conversion of view live speeches content, thus obtaining the Word message corresponding with described speech content, and records, with
When in the Word message showing, and can also have the function that the Word message to sequencing display is corrected, clearly know
Every, road words are who says.This relatively existing professional equipment and professional, the cost of implementation of the present invention is lower, and function is more preferable,
In hgher efficiency.
Brief description
For the scheme being illustrated more clearly that in the embodiment of the present invention, below will be attached to use required described in specific embodiment
Figure be briefly described it should be apparent that, drawings in the following description are only some embodiments of the present invention, for this area
For technical staff, on the premise of not paying creative work, other accompanying drawings can also be obtained according to these accompanying drawings.
A kind of flowchart of voice automatic record method with automatic error correction function that Fig. 1 provides for the present invention.
Specific embodiment
Below in conjunction with the accompanying drawing in the embodiment of the present invention, the technical scheme in the embodiment of the present invention is carried out clearly and completely
Description is it is clear that described embodiment is only a part of embodiment of the present invention, rather than whole embodiments.Based on this
Inventive embodiment, all other enforcement that those of ordinary skill in the art are obtained on the premise of not making creative work
Example, broadly falls into the scope of protection of the invention.
The optimized integration of the present invention is existing speech software.It is all more ripe that existing speech recognition technology and dress change technology,
Phonitic entry method or speech recognition technology are had on the mobile terminals such as mobile phone, as long as can be achieved with setting for terminal by voice
For being controlled or operating, as long as or for terminal unit speech, the content that this just can be talked by the software on terminal unit
It is automatically recognized as word, and shown.
Under the background of above prior art, in the present embodiment, give a kind of voice side of record automatically with automatic error correction function
Method, below will be described in detail to the method.
See Fig. 1, give a kind of flowchart of the voice automatic record method with automatic error correction function, methods described bag
Include following steps：
Step S1, pre-enters speech roster.
Step S2, obtains voice signal.
In being embodied as, the speech content of spokesman in meeting scene can be passed through by Mike by equipment such as mike and audio amplifiers
Wind is changing into the signal of telecommunication, and plays back in sound-box device.Or directly obtained in the speech of spokesman by mike
Hold, that is, directly by required voice signal in spoken sounds conversion cost embodiment.In this and prior art, speak against mobile phone,
Come to mobile phone transmission voice signal to be a reason with this.
Step S3, described voice signal is changed into corresponding Word message, and is stored, and wherein, identifies conversion first
During voice signal, choose and to the name corresponding to this voice signal from the speech roster pre-entering, and extract this voice
The name chosen is associated with described tamber characteristic by the tamber characteristic of information simultaneously.
In being embodied as, because in meeting, everyone is not continuous it is entirely possible to be a people one time of speech, deposit
In this situation alternately, then shown by not can know that in such scheme, Word message out is to be said by whom, because
Gone back in this present embodiment method being overcome.
First, pre-enter speech roster；
Then, when identifying conversion voice signal first, choose from the speech roster pre-entering and to this voice signal institute
Corresponding name, and extract the tamber characteristic of this voice messaging, the name chosen is associated with described tamber characteristic simultaneously；
Then, when carrying out the identification of next voice signal, judge whether its tone color is identical with the tone color of formerly described extraction, if phase
Same then the name being associated with this tone color is shown in beginning of the sentence；If differing, entering and repeating to convert voice signal to identifying first
Operation.Rapidly and efficiently can know that by said method whom the spokesman of every words is in a document.
For example, there is " Zhang San ", " Li Si " and " king five " in the list of typing in advance, after meeting starts, voice is changed
Identification software is when carrying out voice signal identification first, if the first voice signal is the speech content of speech " Zhang San ", that
Choose Zhang San from list, afterwards, if " Zhang San " talks always, then from the beginning of second voice signal, then meeting
Automatically labelling " Zhang San " before the Word message identifying.In addition, if from the beginning of second or the 3rd voice signal
After be other people speeches, then by according to the same to the speech recognition labelling of Zhang San, repeat no more here.
In being embodied as, software can be converted by voice and convert voice signals into corresponding Word message.By this side
Formula will identify that the Word message coming saves, and reaches the purpose of record.
Step S4, when carrying out the conversion identification of next voice signal, judges whether its tone color is identical with the tone color of formerly described extraction,
If identical, the name being associated with this tone color is shown in beginning of the sentence, if differing, entering and repeating to convert voice letter to identifying first
Number operation.
Step S5, described Word message is ranked up in a document show, wherein, described Word message includes correct word letter
Breath and wrong Word message.
Step S6, is marked to the wrong Word message in document, and will be described with corresponding for the described mistake Word message of labelling
The voice signal of mistake Word message is associated linking；
Step S7, when clicking on described mistake Word message, is associated with described mistake Word message using described voice conversion software pair
The voice signal of link is recognized for, and carries out editable to the secondary Word message identifying in a document and show.
In being embodied as, absolutely can not correctly identify the speech content in voice signal due to reason switching software,
Or can there is a certain proportion of wrong word, in order to overcome this problem further, present invention also offers implementation below.
When in a document Word message is ranked up with display, described Word message includes correct Word message and wrong word letter
Breath.Therefore, the present invention is marked using to the dislocation Word message of sequencing display in a document, in being embodied as, permissible
Carry out underlined in red labelling, or be changed font color being marked, be marked also or by way of annotation.
Meanwhile, the wrong Word message of this labelling is associated linking with the voice signal of corresponding described mistake Word message, when
When clicking on wrong Word message, the voice signal of described for correspondence mistake Word message is recognized for, and right in a document
The secondary Word message identifying carries out editable and shows.So, in being embodied as it is possible in being shown by editable
Wrong Word message is carried out with corrigendum editor, to obtain the Word message corrected, and is replaced with the Word message of described corrigendum described
Mistake Word message.
For example, have a voice signal A, the content said in A (hereinafter referred content B) be " weather of today is very good, I
Go together to stroll in the park ", voice conversion software after voice signal A is identified with conversion, (the hereafter letter of the content that obtains
Claim content C) be " just, we go to close public member's plate the weather of today together ", then can see, wherein " just " and
" closing public member's plate " is the Word message (hereinafter referred wrong content D) of mistake, therefore, when being ranked up display in a document,
Wrong content D " just " and " closing public member's plate " can be marked.Now, can be by manually coming to mistake at meeting scene
Content D is corrected, and the method for corrigendum is exactly to click on the wrong content D being marked in document, because wrong content D closes
Connection is linked to voice messaging A, then voice messaging A is recognized for change by starting voice conversion knowledge software, and in literary composition
Carry out editable and show in shelves, such as, be shown as " very good, whole good, true, pin, earn, demonstrate,prove ... ", from editable content
In have correct word, then " very good " can be clicked directly on and selected, then afterwards " very good " will replace mistake in
Hold D " just ", if correctly not corresponding to word in editable content, be such as shown that " whole good, true, pin,
Earn, demonstrate,prove ... ", then can first click on "true", then next automatically can show the word with "true" pairing again,
As " good, bold and unconstrained, number ... ", at this point it is possible to reselection " good ", the corrigendum of wrong Word message is completed with this.
In being embodied as, because the purpose of meeting is to record the speech content in meeting in a document, accordingly, it would be desirable to
The Word message identifying is ranked up show, that is to say that typesetting shows.
Pass through said method can automatically record the speech content at meeting scene well, without by professional
Typewriting apparatuss and the typist of specialty, can be carried out in any place.Relatively existing method, cost of the present invention is lower, reliable
Property is higher.
Further, in being embodied as, the speech content of on-the-spot meeting can also be carried out with live forwarding in real time, specifically real
Applying method may be referred to implementation below.
Individual in being embodied as, after identification obtains correct Word message, can described Word message will be sent by network
Live display in real time is carried out to website.Specifically can be achieved in that, the document showing described Word message will be used for even first
It is connected in a website, whenever having Word message to be identified sequencing display in a document after dress changes, Word message is detected,
The described Word message of institute does not comprise mistake Word message, then sends Word message and is shown to website.So meeting it
Outer other people just can watch Word message by refreshing this website.
In addition, if detect identified conversion after Word message includes mistake Word message when, then can be artificial
After it is corrected, by manually by corrigendum after Word message send to website, website after receiving this Word message,
Before the same, will show to receiving Word message.
In being embodied as, the Word message after identification can also be sent in the social software to mobile terminal, come with this
Forwarded in real time in certain circle or in scope to conference content with live.For example, it is possible to Word message is passed through wechat
Software is sent in group automatically.At this point it is possible to be achieved in, first log into wechat account；Then needs will be chosen
The group being transmitted or good friend, then now by document associations of described display Word message to the group that chosen or good
In the transmission backstage of friend, when there being Word message to be identified sequencing display in a document after dress changes, first Word message is detected,
If described Word message does not comprise mistake Word message, then sent Word message to group or good friend；If described literary composition
Word information includes wrong Word message, then after artificial corrigendum, by manually sending it in group or good friend.Or
Person periodically can also detected to shown Word message, once detect in shown Word message not including
Dislocation Word message, then be just automatically sent in group or good friend.
In being embodied as, when automatically being sent to Word message, dependence to be the detection to Word message to identify whether
Comprise mistake Word message, wherein detect that the standard of wrong Word message is just to detect whether that shown Word message includes labelling,
If Word message has description of symbols, if it is wrong Word message it is to be understood that by manually coming wrong Word message is carried out
After corrigendum, can cancel to original to labelling, and then be detected, thus realizing automatically sending.
Further, in a document sequencing display Word message when, the sequence to Word message can be according to every section of continuous language
Being ranked up, concrete methods of realizing is message number：When described voice conversion software completes previous voice signal conversion identification,
Start timing；When described voice conversion software receipt is to current speech signal, stops timing, and be calculated current speech letter
Time difference number and previous voice signal between；Judge whether the described time is more than default time difference, if so, to current language
The Word message that message conversion identification obtains carries out segmentation sequencing display；If it is not, then current speech signal conversion identification obtains
Word message carries out arranged in sequence and shows.
The principle of preceding method is that the people of general speech can make a short pause after finishing one section of word and put off until some time later the second word, then
Can also talk about according to every to carry out compartment for one section and show in the form that Word message is ranked up, pass through between every words
Compartment shows.For example, when first dress changes just to be identified to voice signal, current speech signal and previous voice signal are judged
Time of origin difference, if time difference is more than 0.5 second, then the Word message swapping out with regard to the identified dress of current speech signal
Carry out compartment to show.
Certainly it is to be understood that specific interval time can formerly be arranged, not necessarily 0.5 second, due to everyone
Word speed inconsistent, therefore it provides formerly interval time setting, more preferable word-information display effect can be reached.
From the point of view of to sum up, the present invention does not need the typewriting apparatuss turning it is not required that the typing personnel of specialty, can be achieved with existing to meeting
The automatic conversion of field speech content, thus obtaining the Word message corresponding with described speech content, and records.This is relatively
Existing professional equipment and professional, the cost of implementation of the present invention is lower, is substantially all and can realize in various occasions.
Above-mentioned the specific embodiment only principle of the illustrative present invention and its effect, not for the restriction present invention.Any it is familiar with this skill
The personage of art all can carry out modifications and changes without prejudice under the spirit and the scope of the present invention to above-described embodiment.Therefore, such as
Those of ordinary skill in the art completed under without departing from disclosed spirit and technological thought all etc.
Effect modifications and changes, must be covered by the claim of the present invention.
Claims (4)
1. a kind of voice automatic record method with automatic error correction function is it is characterised in that include：
Pre-enter speech roster；
Obtain voice signal；
The conversion of described voice signal is identified as by corresponding Word message by voice conversion software, and is stored, wherein,
First during identification conversion voice signal, choose and to the name corresponding to this voice signal from the speech roster pre-entering
Word, and extract the tamber characteristic of this voice messaging, the name chosen is associated with described tamber characteristic simultaneously；
When carrying out the conversion identification of next voice signal, judge whether its tone color is identical with the tone color of formerly described extraction, if phase
Same then the name being associated with this tone color is shown in beginning of the sentence, if differing, entering and repeating to identifying conversion voice signal first
Operation；
Described Word message is ranked up in a document show, wherein, described Word message includes correct Word message and mistake
Word message；
Wrong Word message in document is marked, and by the described mistake Word message of labelling and corresponding described wrong word
The voice signal of information is associated linking；
The language of software pair and described mistake Word message association link when clicking on described mistake Word message, is converted using described voice
Message number is recognized for, and carries out editable to the secondary Word message identifying in a document and show.
2. the voice automatic record method with automatic error correction function according to claim 1 is it is characterised in that obtain institute's predicate
The method of message number includes：Spoken sounds are converted into by described voice signal by mike.
3. the voice automatic record method with automatic error correction function according to claim 1 is it is characterised in that arrange in a document
Sequence shows that the method for Word message includes：Described Word message is carried out segmentation sequence according to every section of continuous voice signal aobvious
Show.
4. the voice automatic record method with automatic error correction function according to claim 3 it is characterised in that described according to every
The method that the continuous voice signal of section to carry out segmentation sequencing display to described Word message is：
When described voice conversion software completes previous voice signal conversion identification, start timing；
When described voice conversion software receipt is to current speech signal, stopping timing, and it is calculated current speech signal and front
Time difference between one voice signal；
Judge whether the described time is more than default time difference, the word letter if so, current speech signal conversion identification being obtained
Breath carries out segmentation sequencing display；If it is not, the Word message that then current speech signal conversion identification obtains carries out arranged in sequence showing.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201510529137.0A CN106487531A (en) | 2015-08-26 | 2015-08-26 | A kind of voice automatic record method with automatic error correction function |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201510529137.0A CN106487531A (en) | 2015-08-26 | 2015-08-26 | A kind of voice automatic record method with automatic error correction function |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN106487531A true CN106487531A (en) | 2017-03-08 |
Family
ID=58234501
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201510529137.0A Pending CN106487531A (en) | 2015-08-26 | 2015-08-26 | A kind of voice automatic record method with automatic error correction function |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN106487531A (en) |
Cited By (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN111048093A (en) * | 2018-10-12 | 2020-04-21 | 深圳海翼智新科技有限公司 | Conference sound box, conference recording method, device, system and computer storage medium |
| CN111816172A (en) * | 2019-04-10 | 2020-10-23 | 阿里巴巴集团控股有限公司 | Voice response method and device |
| CN116229975A (en) * | 2023-03-17 | 2023-06-06 | 杭州盈禾嘉田科技有限公司 | System and method for voice reporting of field diseases and insect pests in intelligent interaction scene |
-
2015
- 2015-08-26 CN CN201510529137.0A patent/CN106487531A/en active Pending
Cited By (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN111048093A (en) * | 2018-10-12 | 2020-04-21 | 深圳海翼智新科技有限公司 | Conference sound box, conference recording method, device, system and computer storage medium |
| CN111816172A (en) * | 2019-04-10 | 2020-10-23 | 阿里巴巴集团控股有限公司 | Voice response method and device |
| CN116229975A (en) * | 2023-03-17 | 2023-06-06 | 杭州盈禾嘉田科技有限公司 | System and method for voice reporting of field diseases and insect pests in intelligent interaction scene |
| CN116229975B (en) * | 2023-03-17 | 2023-08-18 | 杭州盈禾嘉田科技有限公司 | System and method for voice reporting of field diseases and insect pests in intelligent interaction scene |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN106486113A (en) | A kind of minutes method | |
| CN106486116A (en) | A kind of online generation method of on-the-spot meeting summary | |
| US10885318B2 (en) | Performing artificial intelligence sign language translation services in a video relay service environment | |
| US8611507B2 (en) | Systems and methods for intelligent call transcription | |
| US9218808B2 (en) | Computer-implemented system and method for reducing voice transcription error | |
| US9571638B1 (en) | Segment-based queueing for audio captioning | |
| EP3258392A1 (en) | Systems and methods for building contextual highlights for conferencing systems | |
| US8484040B2 (en) | Social analysis in multi-participant meetings | |
| US10574827B1 (en) | Method and apparatus of processing user data of a multi-speaker conference call | |
| US8391445B2 (en) | Caller identification using voice recognition | |
| CN105100360A (en) | Communication auxiliary method and device for voice communication | |
| US20180293996A1 (en) | Electronic Communication Platform | |
| US20150237214A1 (en) | Method and apparatus for providing ambient social telephony | |
| CN106487531A (en) | A kind of voice automatic record method with automatic error correction function | |
| CN106357932A (en) | Call information recording method and mobile terminal | |
| CN109688276A (en) | A kind of incoming call filter system and method based on artificial intelligence technology | |
| CN104618615B (en) | A kind of TeleConference Bridge meeting summary method for pushing based on short message | |
| CN109587664A (en) | A kind of voice dialing system of edge calculations in conjunction with cloud computing | |
| CN108320761B (en) | Audio recording method, intelligent recording device and computer readable storage medium | |
| CN106487532A (en) | A kind of voice automatic record method | |
| CN107659370A (en) | University student classroom total management system and method | |
| EP2913822B1 (en) | Speaker recognition | |
| CN104580772A (en) | Teleconference system conference summary pushing method based on audio | |
| CN106486123A (en) | A kind of conference content online record method | |
| US20230247131A1 (en) | Presentation of communications |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| C06 | Publication | ||
| PB01 | Publication | ||
| WD01 | Invention patent application deemed withdrawn after publication |
Application publication date: 20170308 |
|
| WD01 | Invention patent application deemed withdrawn after publication |