CN106409294A - Method and apparatus for preventing voice command misidentification - Google Patents
Method and apparatus for preventing voice command misidentification Download PDFInfo
- Publication number
- CN106409294A CN106409294A CN201610909229.6A CN201610909229A CN106409294A CN 106409294 A CN106409294 A CN 106409294A CN 201610909229 A CN201610909229 A CN 201610909229A CN 106409294 A CN106409294 A CN 106409294A
- Authority
- CN
- China
- Prior art keywords
- lyrics
- time
- false triggering
- chinese
- phrase
- 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.)
- Granted
Links
- 238000000034 method Methods 0.000 title claims abstract description 49
- 238000004321 preservation Methods 0.000 claims description 22
- 150000001875 compounds Chemical class 0.000 claims description 13
- 238000012790 confirmation Methods 0.000 claims description 5
- 239000003550 marker Substances 0.000 claims description 3
- 238000004364 calculation method Methods 0.000 abstract description 2
- 230000007547 defect Effects 0.000 abstract description 2
- 230000001960 triggered effect Effects 0.000 abstract 2
- 230000000630 rising effect Effects 0.000 abstract 1
- 238000002054 transplantation Methods 0.000 abstract 1
- 230000006870 function Effects 0.000 description 6
- 230000008569 process Effects 0.000 description 6
- 230000007812 deficiency Effects 0.000 description 5
- 230000008878 coupling Effects 0.000 description 4
- 238000010168 coupling process Methods 0.000 description 4
- 238000005859 coupling reaction Methods 0.000 description 4
- 238000002372 labelling Methods 0.000 description 3
- 230000008859 change Effects 0.000 description 2
- 239000000203 mixture Substances 0.000 description 2
- 230000002123 temporal effect Effects 0.000 description 2
- 230000009471 action Effects 0.000 description 1
- 230000009286 beneficial effect Effects 0.000 description 1
- 238000012937 correction Methods 0.000 description 1
- 238000011161 development Methods 0.000 description 1
- 238000005516 engineering process Methods 0.000 description 1
- 238000009434 installation Methods 0.000 description 1
- 238000012986 modification Methods 0.000 description 1
- 230000004048 modification Effects 0.000 description 1
- 238000012545 processing Methods 0.000 description 1
Classifications
-
- G—PHYSICS
- G10—MUSICAL INSTRUMENTS; ACOUSTICS
- G10L—SPEECH ANALYSIS TECHNIQUES OR SPEECH SYNTHESIS; SPEECH RECOGNITION; SPEECH OR VOICE PROCESSING TECHNIQUES; SPEECH OR AUDIO CODING OR DECODING
- G10L15/00—Speech recognition
- G10L15/22—Procedures used during a speech recognition process, e.g. man-machine dialogue
-
- G—PHYSICS
- G10—MUSICAL INSTRUMENTS; ACOUSTICS
- G10L—SPEECH ANALYSIS TECHNIQUES OR SPEECH SYNTHESIS; SPEECH RECOGNITION; SPEECH OR VOICE PROCESSING TECHNIQUES; SPEECH OR AUDIO CODING OR DECODING
- G10L15/00—Speech recognition
- G10L15/26—Speech to text systems
Landscapes
- Engineering & Computer Science (AREA)
- Computational Linguistics (AREA)
- Health & Medical Sciences (AREA)
- Audiology, Speech & Language Pathology (AREA)
- Human Computer Interaction (AREA)
- Physics & Mathematics (AREA)
- Acoustics & Sound (AREA)
- Multimedia (AREA)
- Electrophonic Musical Instruments (AREA)
- Reverberation, Karaoke And Other Acoustics (AREA)
Abstract
The embodiments of the invention disclose a method and apparatus for preventing voice command misidentification. The method comprises the following steps: obtaining a lyrics file matching a song to be played; searching for a phrase which can be wrongly triggered easily in the lyrics file; according to the lyrics file, calculating first initial time and first finishing time of play of the phrase which can be wrongly triggered easily; and playing the song to be played, closing a voice identification module when it is the first initial time, and starting the voice identification module when it is the first finishing time. According to the technical scheme provided by the embodiments of the invention, the technical defects of increased processor calculation burden, rising equipment power consumption and difficult voice identification algorithm transplantation which are caused by increased complexity of a voice identification algorithm for reducing the probability of voice wrong trigger in the prior art are overcome, and voice misidentification caused by song play can also be reliably reduced without improving the complexity of the voice identification algorithm.
Description
Technical field
The present embodiments relate to data processing technique, more particularly, to a kind of method preventing voice command misrecognition and dress
Put.
Background technology
With scientific and technical continuous development, and people, to quality of the life constantly higher pursuit, more and more set
Get everything ready and have voice control function.
In automobile, household electrical appliance and mobile phone, great majority all can install voice control function, and these equipment can also simultaneously
Play the audio files such as song, cross-talk, open Voice command work(when playing audio file it is possible to can open by mistake in these equipment
Can be so that equipment does the action making mistake.Pass through in prior art to improve the complexity of speech recognition algorithm, optimize speech recognition
Algorithm, reduces the probability of false triggering.
The defect of prior art is:The computation burden that improve CPU leads to equipment power dissipation to rise, and speech recognition algorithm moves
Plant difficult, and cannot fundamentally avoid the probability of false triggering.
Content of the invention
In view of this, embodiments provide a kind of method and apparatus preventing voice command misrecognition, to optimize
Existing reduce voice false triggering probabilistic technique it is achieved that the complexity of speech recognition algorithm need not be improved it is also possible to reliable
Reduce and misidentified due to the voice that song lyrics lead to.
In a first aspect, embodiments provide a kind of prevent voice command misrecognition method, including:
Obtain the lyrics file mated with song to be played;
Search the easy false triggering phrase in described lyrics file, wherein, described easy false triggering phrase and default Voice command
The language of order is same or like;
According to described lyrics file, at the end of calculating the first initial time and first that described easy false triggering phrase is play
Between;
Play described song to be played, close sound identification module when reaching described first initial time, reach institute
State and during the first end time, start described sound identification module.
In the above-mentioned methods it is preferred that described play described song to be played, when reaching described first initial time
Close sound identification module, after starting described sound identification module when reaching described first end time, also include:
Preserve described easy false triggering phrase and corresponding described first initial time and described first end time;
Described the first initial time and the first end that according to described lyrics file, the described easy false triggering phrase of calculating is play
Before time, also include:
Confirm that described lyrics file does not have the described easy false triggering phrase of preservation;
The described song to be played of described broadcasting, closes sound identification module when reaching described first initial time, arrives
Reach before starting described sound identification module during described first end time, also include:
If there is the described easy false triggering phrase of preservation, read easy false triggering phrase pair described in described song to be played
Described first initial time answered and described first end time.
In the above-mentioned methods it is preferred that described lyrics file is Chinese lyrics file;
Described default voice control command is Chinese speech control command;
The described easy false triggering phrase searched in described lyrics file includes:
The pronunciation attribute of all Chinese lyrics in traversal described Chinese lyrics file, wherein, described pronunciation attribute at least wraps
Include tone, initial consonant and simple or compound vowel of a Chinese syllable;
If each Chinese character in described all Chinese one of lyrics Chinese characters or a Chinese phrase all with the described Chinese
In language voice control command, the described pronunciation attribute of the Chinese character of correspondence position is identical, then confirm this Chinese character or Chinese phrase with described
The language of Chinese speech control command is identical;
If each Chinese character in described all Chinese one of lyrics Chinese characters or a Chinese phrase all with the described Chinese
In the described pronunciation attribute of the Chinese character of correspondence position in language voice control command at least a kind of different and at least two kinds identical,
Then confirm that this Chinese character or Chinese phrase are close with the language of described Chinese speech control command;
One Chinese character or one Chinese phrase are labeled as easy false triggering phrase.
In the above-mentioned methods it is preferred that described according to described lyrics file, calculate what described easy false triggering phrase was play
First initial time and the first end time include:
According to described lyrics file, obtain the second initial time that the lyrics sentence belonging to described easy false triggering phrase is play
With the second end time;
According to described second initial time, described second end time, described lyrics sentence comprise lyrics unit number,
Described easy false triggering phrase comprises the number and described easy false triggering phrase of the lyrics unit position in described lyrics sentence, meter
Calculate described first initial time and described first end time that described easy false triggering phrase is play.
In the above-mentioned methods it is preferred that described according to described lyrics file, calculate what described easy false triggering phrase was play
First initial time and the first end time include:
According to described lyrics file, obtain the lyrics sentence broadcasting belonging to described easy false triggering phrase described second initiates
Time and described second end time;
Obtain the first compensation time of described first initial time and the second compensation time of described first end time;
According to when described second initial time, described second end time, described first compensation time, described second compensation
Between, the described lyrics sentence number that comprises lyrics unit, the described easy false triggering phrase number that comprises lyrics unit and described easily
Position in described lyrics sentence for the false triggering phrase, calculate described first initial time that described easy false triggering phrase plays and
Described first end time.
In the above-mentioned methods it is preferred that also including:
After described playback of songs to be played terminates, statistics causes and does not cause described default voice control command misrecognition
Described easy false triggering phrase;
Revise described easy false triggering phrase corresponding described first compensation time and described second compensation time, recalculate
Described first initial time and described first end time that described easy false triggering phrase is play.
In second aspect, embodiments provide a kind of device preventing voice command misrecognition, including:
Lyrics file acquisition module, for obtaining the lyrics file mated with song to be played;
Easily false triggering phrase searching modul is for searching the easy false triggering phrase in described lyrics file, wherein, described easy
False triggering phrase is same or like with the language of default voice control command;
Reproduction time computing module, for according to described lyrics file, calculating the first of described easy false triggering phrase broadcasting
Initial time and the first end time;
Sound identification module control module, for playing described song to be played, when reaching described first initial time
Close sound identification module, start described sound identification module when reaching described first end time.
It is preferred that after described sound identification module control module, also including in said apparatus:
Reproduction time preserving module, for preserving described easy false triggering phrase and corresponding described first initial time and institute
Stated for the first end time;
Before described reproduction time computing module, also include:
Easily false triggering phrase confirms module, for confirming that described lyrics file do not have the described easy false triggering word of preservation
Group;
Before described sound identification module control module, also include:
Reproduction time read module, if for the described easy false triggering phrase that there is preservation, reads described song to be played
Easy corresponding described first initial time of false triggering phrase and described first end time described in song.
It is preferred that described lyrics file is Chinese lyrics file in said apparatus;
Described default voice control command is Chinese speech control command;
Described easy false triggering phrase searching modul includes:
Lyrics Traversal Unit, for traveling through the pronunciation attribute of all Chinese lyrics in described Chinese lyrics file, wherein, institute
State pronunciation attribute and at least include tone, initial consonant and simple or compound vowel of a Chinese syllable;
Same words group acknowledge unit, if in one of described all Chinese lyrics Chinese character or a Chinese phrase
Each Chinese character all identical with the described pronunciation attribute of the Chinese character of correspondence position in described Chinese speech control command, then confirm should
Chinese character or Chinese phrase are identical with the language of described Chinese speech control command;
Close phrase confirmation unit, if in one of described all Chinese lyrics Chinese character or a Chinese phrase
Each Chinese character all at least a kind of with the described pronunciation attribute of the Chinese character of correspondence position in described Chinese speech control command
Different and at least two kinds identical, then confirm that this Chinese character or Chinese phrase are close with the language of described Chinese speech control command;
Easily false triggering phrase marker unit, for being labeled as easy false touch by one Chinese character or one Chinese phrase
Send out phrase.
It is preferred that described reproduction time computing module includes in said apparatus:
Second reproduction time determining unit, for according to described lyrics file, obtaining belonging to described easy false triggering phrase
The second initial time and the second end time that lyrics sentence is play;
First reproduction time computing unit, for according to described second initial time, described second end time, described song
The number that word sentence comprises lyrics unit, described easy false triggering phrase comprise the number of lyrics unit and described easy false triggering phrase
Position in described lyrics sentence, calculates described first initial time and described first knot that described easy false triggering phrase is play
The bundle time.
It is preferred that described reproduction time computing module includes in said apparatus:
Second reproduction time determining unit, for according to described lyrics file, obtaining belonging to described easy false triggering phrase
Described second initial time and described second end time that lyrics sentence is play;
Compensate time acquisition unit, for obtaining the first compensation time of described first initial time and described first end
The second compensation time of time;
Second reproduction time computing unit, for according to described second initial time, described second end time, described
One compensation time, described second compensation time, described lyrics sentence comprise the number of lyrics unit, described easy false triggering phrase bag
Position in described lyrics sentence for the number and described easy false triggering phrase containing lyrics unit, calculates described easy false triggering phrase
Described first initial time play and described first end time.
It is preferred that also including in said apparatus:
Misrecognition statistical module, after terminating for described playback of songs to be played, statistics causes and does not cause described default
The described easy false triggering phrase of voice control command misrecognition;
Compensate time complexity curve module, for revising described easy false triggering phrase corresponding described first compensation time and described
Second compensation time, at the end of recalculating described first initial time and described first that described easy false triggering phrase is play
Between.
The method and apparatus preventing voice command misrecognition provided in an embodiment of the present invention, by first obtaining and song to be played
The lyrics file of bent coupling, then looks up the easy false triggering phrase in lyrics file, calculates the first of easy false triggering phrase broadcasting
Initial time and the first end time, finally play song to be played, close speech recognition mould when reaching the first initial time
Block, starts sound identification module when reaching for the first end time, overcomes in prior art to reduce voice false triggering
Probability, and then increase the complexity of speech recognition algorithm, lead to processor computation burden to increase, equipment power dissipation rises and voice
The difficult technological deficiency of recognizer transplanting is it is achieved that the complexity of speech recognition algorithm need not be improved it is also possible to reliably subtract
Few voice leading to due to playing song misidentifies.
Brief description
Fig. 1 is a kind of flow chart of method preventing voice misrecognition that the embodiment of the present invention one provides;
Fig. 2 is a kind of flow chart of method preventing voice misrecognition that the embodiment of the present invention two provides;
Fig. 3 is a kind of flow chart of method preventing voice misrecognition that the embodiment of the present invention three provides;
Fig. 4 is a kind of structure chart of device preventing voice misrecognition that the embodiment of the present invention four provides.
Specific embodiment
In order that the object, technical solutions and advantages of the present invention are clearer, the concrete reality to the present invention below in conjunction with the accompanying drawings
Apply example to be described in further detail.It is understood that specific embodiment described herein is used only for explaining the present invention,
Rather than limitation of the invention.
It also should be noted that, for the ease of description, illustrate only in accompanying drawing part related to the present invention rather than
Full content.It should be mentioned that some exemplary embodiments are described before exemplary embodiment is discussed in greater detail
Become the process described as flow chart or method.Although operations (or step) are described as the process of order by flow chart,
It is that many of which operation can be implemented concurrently, concomitantly or simultaneously.Additionally, the order of operations can be by again
Arrange.Described process can be terminated when its operations are completed, it is also possible to have the additional step being not included in accompanying drawing.
Described process can correspond to method, function, code, subroutine, subprogram etc..
Embodiment one
A kind of flow chart of method preventing voice command misrecognition that Fig. 1 provides for the embodiment of the present invention one, this enforcement
The method of example can be by preventing voice misrecognition device from executing, and this device can be realized by way of hardware and/or software, and
Typically can be integrated in and there is voice command control function and can play in the equipment of audio file, for example:Mobile phone, automobile etc..
The method of the present embodiment specifically includes:
The lyrics file that step 110, acquisition are mated with song to be played.
In general, in the prior art, there is setting of voice command control function and audio frequency document display function simultaneously
Standby, get play audio file order when, be all directly to play out, without before broadcasting, reduced due to
Play the operation of the voice command misrecognition that song leads to.
In the present embodiment, before playing song to be played, the lyrics file mated with song to be played can first be obtained.Its
In, the mode of acquisition can be specifically to obtain from locally stored lyrics file it is also possible to by internet from being stored with
Obtain in the server of required lyrics file, the present embodiment is not limited to this.
Wherein, lyrics file specifically refers to comprise initial time of all lyrics of song to be played and every lyrics etc.
The file of information.Can be typically:The file of suffix entitled .LRC .SNC and .KRC etc..
In addition, the described lyrics in this programme are not limited to the lyrics of the song of singer's performance, also include other audio frequency
The content misidentifying may be led in file, for example, read aloud, give a lecture etc. with sound broadcasting as display mode, reading aloud original text simultaneously or drill
The content of multimedia that lecture notes can be embodied in the way of lyrics file.
Step 120, the easy false triggering phrase searched in lyrics file, wherein, easy false triggering phrase and default Voice command
The language of order is same or like.
In the present embodiment, easy false triggering phrase specifically refers to cause the word of default voice control command false triggering
Group, that is, the phrase that pronunciation is same or like with the pronunciation of the language of default voice control command.Wherein, default Voice command
Order specifically refers to prestore, and can be used to implement voice-operated language.
Due to may there is easy false triggering phrase in the lyrics, lead to song during playing, cause Voice command
Order misrecognition, does the operation making mistake, and therefore, in the present embodiment, after obtaining the lyrics file of song to be played, first can
Search whether the lyrics in this lyrics file include easy false triggering phrase.
Wherein, the concrete mode searching easy false triggering phrase can be that all lyrics traveling through in lyrics file are searched and pre-
If voice control command in each Chinese character tone, initial consonant and simple or compound vowel of a Chinese syllable all same phrase as easy false triggering phrase,
Can be that all lyrics lookups traveling through in lyrics file are equal with the phonetic symbol of each English word in default voice control command
Identical phrase is as modes such as easy false triggering phrases.Wherein, all lyrics in traversal lyrics file search easy false triggering word
The mode of group can be to search just for a default voice control command to correspond to therewith while traveling through all lyrics
Easy false triggering phrase that is to say, that how many default voice control command will travel through how many all over all lyrics, also may be used
To be only to travel through all lyrics, search easy false triggering while traversal compared with all of default voice control command
Phrase will be entered with all of default voice control command when any one word or word in the lyrics that is to say, that often traversing
Row contrast.
It is further to note that being only illustrated for two kinds of situations of Chinese and english to the lyrics above, work as the lyrics
During for language beyond Chinese and english, the method for the present embodiment is equally applicable, because either any language, all to should have
The distinctive phone set of this language, and the pronunciation of all single word of this language or single word be all by this language distinctive one or
Multiple phonemes are constituted, when searching easy false triggering phrase it is possible to be used the distinctive phoneme of this language as the basis of comparison, when
When certain word or certain phrase are same or like with the phoneme of each word or each word in default voice control command, then judge
This word or this phrase are easy false triggering phrase.
Step 130, according to lyrics file, at the end of calculating the first initial time and first that easy false triggering phrase is play
Between.
In the present embodiment, the first initial time specifically refers to the time that easy false triggering phrase commences play out, the first end
Time specifically refers to the time that easy false triggering phrase terminates to play, and wherein, the first initial time and the first end time are all phases
The concrete time calculated for the initial reproduction time of song to be played, the initial reproduction time of song to be played can be remembered
Record as 0 point of 0 second 0 millisecond or 0 point of time format such as 0 second.
In the present embodiment, due to this easy false triggering phrase when searching easy false triggering phrase, can be known in song simultaneously
Particular location in word is (for example:The the 3rd to the 6th word in the 5th lyrics), and every can be known according to lyrics file
The temporal informations such as the initial reproduction time of the lyrics, therefore, according to particular location in the lyrics for the easy false triggering phrase and every
The temporal informations such as the initial reproduction time of the lyrics can relatively easily calculate that easy false triggering phrase plays first initial when
Between and the first end time.
Further, due to it cannot be guaranteed that each lyrics are all typically will not be detailed at the uniform velocity broadcasting, and lyrics file
Carefully record the initial reproduction time of each word or word, typically only record the initial reproduction time of every lyrics, so, calculating
When the first initial time and the first end time, if all words in each lyrics of acquiescence or word are all at the uniform velocity broadcastings,
Result of calculation may have certain error with the actual initial reproduction time of easy false triggering phrase and end reproduction time, therefore,
Calculated first initial time and the first end time slightly can be adjusted, make them be more nearly easy false triggering phrase
Actual initial reproduction time and end reproduction time.Wherein, the mode of adjustment can be the setting compensation time, will be calculated
The first initial time deduct this compensation time, the first end time add this compensation time, the first initial time and first knot
The compensation time of bundle time can identical it is also possible to differ, the present embodiment is not limited to this.
Step 140, play song to be played, close sound identification module when reaching the first initial time, reach the
Start sound identification module during one end time.
In the present embodiment, complete to calculate the first initial time and the first end time that easy false triggering phrase is play
Afterwards, commence play out song to be played, when playing this song, when reaching the first initial time, may turn off speech recognition mould
Block, causes voice to misidentify with the broadcasting preventing easy false triggering phrase, leads to do the operation making mistake, at the end of reaching first
Between when may turn on sound identification module, to carry out the identification of voice control command in real time.
The method preventing voice command misrecognition provided in an embodiment of the present invention, is mated with song to be played by first obtaining
Lyrics file, then look up the easy false triggering phrase in lyrics file, when calculate that easy false triggering phrase plays first is initial
Between and the first end time, finally play song to be played, reach the first initial time when close sound identification module, arrive
Reach startup sound identification module during the first end time, overcome in prior art to reduce the probability of voice false triggering, enter
And increase the complexity of speech recognition algorithm, lead to processor computation burden to increase, equipment power dissipation rises and speech recognition is calculated
The difficult technological deficiency of method transplanting it is achieved that need not improve speech recognition algorithm complexity it is also possible to reliably reduce due to
Play the voice misrecognition that song leads to.
Embodiment two
Fig. 2 is a kind of flow chart of method preventing voice command misrecognition that the embodiment of the present invention two provides.This enforcement
Example is optimized based on above-described embodiment, in the present embodiment, lyrics file is optimized for Chinese lyrics file;
Default voice control command is optimized for Chinese speech control command;
Accordingly, it is optimized for searching the easy false triggering phrase in lyrics file:In the Chinese lyrics file of traversal all in
The pronunciation attribute of the civilian lyrics, wherein, pronunciation attribute at least includes tone, initial consonant and simple or compound vowel of a Chinese syllable;If in all Chinese lyrics
Each Chinese character in individual Chinese character or a Chinese phrase is all belonged to the pronunciation of the Chinese character of correspondence position in Chinese speech control command
Property is identical, then confirm that this Chinese character or Chinese phrase are identical with the language of Chinese speech control command;If in all Chinese lyrics
A Chinese character or a Chinese phrase in each Chinese character all with Chinese speech control command the Chinese character of correspondence position send out
In sound attribute at least a kind of different and at least two kinds identical, then confirm that this Chinese character or Chinese phrase control life with Chinese speech
The language of order is close;One Chinese character or a Chinese phrase are labeled as easy false triggering phrase.
Further, the first initial time and the first end that easy false triggering phrase is play will be calculated according to lyrics file
Time-optimized it is:According to lyrics file, obtain the second initial time that the lyrics sentence belonging to easy false triggering phrase plays and the
Two end times;The number of lyrics unit, easy false triggering are comprised according to the second initial time, the second end time, lyrics sentence
Phrase comprises the number and easy false triggering phrase of the lyrics unit position in lyrics sentence, calculates what easy false triggering phrase was play
First initial time and the first end time.
Further, playing song to be played, closing sound identification module when reaching the first initial time, reaching
After starting sound identification module during the first end time, can also include:Preserve easy false triggering phrase and corresponding the first
Time beginning and the first end time.
Accordingly, at the end of the first initial time and first play according to lyrics file, the easy false triggering phrase of calculating
Between before, can also include:Confirm that lyrics file does not have the easy false triggering phrase of preservation.
Accordingly, play song to be played, closing sound identification module when reaching the first initial time, reach the
Before starting sound identification module during one end time, can also include:If there is the easy false triggering phrase of preservation, reading is treated
Play easy corresponding first initial time of false triggering phrase and the first end time in song.
Accordingly, the method for the present embodiment specifically includes:
The Chinese lyrics file that step 201, acquisition are mated with song to be played.
In the present embodiment, the lyrics of song to be played are Chinese, and the lyrics file of coupling is Chinese lyrics file.
Step 202, the Chinese lyrics file of judgement whether there is the easy false triggering phrase of preservation, if not existing, execute
Step 203, if exist, execution step 209.
In the present embodiment, if play before song to be played, then will there is the easy false triggering word of preservation
Group and corresponding first initial time and the first end time, now, it is no need for searching the easy false triggering in the Chinese lyrics again
Phrase and calculate the first initial time and the first end time that easy false triggering phrase is play, can directly invoke and preserve before
Related content.
The pronunciation attribute of all Chinese lyrics in the Chinese lyrics file of step 203, traversal, wherein, pronunciation attribute at least wraps
Include tone, initial consonant and simple or compound vowel of a Chinese syllable.
In the present embodiment, if Chinese lyrics file does not have the easy false triggering phrase of preservation, be from Chinese song
Easy false triggering phrase is searched, the concrete mode searching easy false triggering phrase is each of the Chinese lyrics of traversal Chinese character in word
Pronunciation attribute, according to the matching degree of this Chinese character and the pronunciation attribute of the language of Chinese speech control command, judges that this Chinese character is
No for or whether belong to easy false triggering phrase.
Wherein, pronunciation attribute specifically refers to the property set being made up of tone, initial consonant and simple or compound vowel of a Chinese syllable etc. to the related attribute that pronounces
Close.In the Chinese lyrics pronunciation attribute of Chinese character specifically can obtain from Chinese lyrics file it is also possible to by internet from
Download in server, the present embodiment is not limited to this.
The language identical Chinese character of step 204, confirmation and Chinese speech control command or Chinese phrase, specifically, if
Each Chinese character in one of all Chinese lyrics Chinese character or a Chinese phrase is all corresponding with Chinese speech control command
The pronunciation attribute of the Chinese character of position is identical, then confirm that this Chinese character or Chinese phrase are identical with the language of Chinese speech control command.
In a specific example, when the Chinese speech control command being contrasted is "ON", in traversal Chinese song
During word file, find have in Chinese lyrics file a Chinese character pronunciation attribute identical with the pronunciation attribute of "ON" word, then general
Easy false triggering phrase confirmed as in this Chinese character in Chinese lyrics file.
In a specific example, when the Chinese speech control command being contrasted is " increase volume ", in traversal
During Chinese lyrics file, find the pronunciation attribute phase of the pronunciation attribute having a Chinese character A in Chinese lyrics file and " increasing " word
With, then continue to judge the pronunciation attribute of Chinese character A Chinese character B below whether with " plus " word is identical, if differing then it is assumed that by Chinese character
The phrase of A and Chinese character B composition is not the easy false triggering phrase of corresponding " increase volume ";If identical, continue to judge after Chinese character B
Chinese character C pronunciation attribute whether identical with the pronunciation attribute of " sound " word, if differing then it is assumed that being made up of Chinese character A, B and C
Phrase is not the easy false triggering phrase of corresponding " increase volume ";If identical, continue to judge the pronunciation of Chinese character C Chinese character D below
Whether attribute is identical with the pronunciation attribute of " measuring " word, if differing then it is assumed that the phrase being made up of Chinese character A, B, C and D is not right
Answer the easy false triggering phrase of " increase volume ";If identical then it is assumed that being corresponding " to increase sound by the phrase that Chinese character A, B, C and D form
The easy false triggering phrase of amount ".
Step 205, the confirmation Chinese character close with the language of Chinese speech control command or Chinese phrase, specifically, if
Each Chinese character in one of all Chinese lyrics Chinese character or a Chinese phrase is all corresponding with Chinese speech control command
In the pronunciation attribute of the Chinese character of position at least a kind of different and at least two kinds identical, then confirm this Chinese character or Chinese phrase with
The language of Chinese speech control command is close.
In the present embodiment, when searching the Chinese character close with the language of Chinese speech control command or Chinese phrase, only
Will in pronunciation attribute and the pronunciation attribute of Chinese character of correspondence position in Chinese speech control command of this Chinese character or Chinese phrase extremely
Rare a kind of different and at least two kinds identical, being considered as this Chinese character or Chinese phrase is easy false triggering phrase.
For example, Chinese speech control command is " turning off the light ", has two Chinese character A and B of next-door neighbour, Chinese character in the Chinese lyrics
The tone harmony parent phase of the tone of A and initial consonant and "Off" word with and Chinese character A simple or compound vowel of a Chinese syllable different with the simple or compound vowel of a Chinese syllable of "Off" word, Chinese character B's
The initial consonant of initial consonant and simple or compound vowel of a Chinese syllable and " lamp " word and simple or compound vowel of a Chinese syllable is identical and the tone of Chinese character B and the tone of " lamp " word different then it is assumed that by the Chinese
The phrase of word A and B composition is the easy false triggering phrase of corresponding " turning off the light ".
Accordingly, judge whether one of Chinese lyrics Chinese character or a Chinese phrase are the concrete of easy false triggering phrase
Method with step 204 for two examples identical, will not be described in detail herein, be this Chinese character or Chinese phrase in this step
At least a kind of in pronunciation attribute should be different from the Chinese character of correspondence position in corresponding Chinese speech control command and at least two
Plant and identical with the Chinese character of correspondence position in corresponding Chinese speech control command should just meet the bar being judged to easy false triggering phrase
Part.
Step 206, a Chinese character or a Chinese phrase are labeled as easy false triggering phrase.
In the present embodiment, after determining all easy false triggering phrase in the Chinese lyrics, need to these easy false triggerings
Phrase is marked, and labeling method can be directly directly labeled in Chinese lyrics file or by easily
The relevant information of false triggering phrase is stored in the discernible file of another one, and the present embodiment is not limited to this.
Step 207, the second initial time play according to lyrics file, the lyrics sentence belonging to the easy false triggering phrase of acquisition
With the second end time.
In the present embodiment, the second initial time that lyrics sentence is play and the second end time specifically refer to this lyrics language
Initial time and end time that sentence is play, wherein, the second initial time and the second end time are relative to song to be played
The concrete time that bent initial reproduction time calculates.
In the present embodiment, while finding easy false triggering phrase, easy false triggering phrase institute can correspondingly be recorded
The lyrics sentence belonging to, and its position in this lyrics sentence.Accordingly, if during labelling easy false triggering phrase, it is direct
Chinese lyrics file is directly labeled, then be easy to know lyrics sentence belonging to easy false triggering phrase and at this
Position in lyrics sentence;If during labelling easy false triggering phrase, it is that in addition the relevant information of easy false triggering phrase is stored in
In one discernible file, then record in this document simultaneously lyrics sentence belonging to easy false triggering phrase and
Position in this lyrics sentence.
It will be appreciated by persons skilled in the art that the initial of each lyrics broadcasting typically all can be recorded in lyrics file
Time, also can record the end time that the time span of each lyrics broadcasting or each lyrics are play, therefore, when easy simultaneously
In the case that lyrics sentence belonging to false triggering phrase has determined, can relatively easily be obtained according to lyrics file or calculate
Go out the second initial time and the second end time that the lyrics sentence belonging to easy false triggering phrase is play.But, LRC lyrics file
Only record the broadcasting initial time of every lyrics and do not record the end time of every lyrics or the time span of broadcasting, therefore,
When the type of the lyrics file of song to be played coupling is LRC file, then the broadcasting initial time giving tacit consent to next lyrics is
The broadcasting end time of upper lyrics.
Step 208, the first initial time calculating easy false triggering phrase broadcasting and the first end time, specifically, according to
The number that second initial time, the second end time, lyrics sentence comprise lyrics unit, easy false triggering phrase comprise lyrics unit
Position in lyrics sentence of number and easy false triggering phrase, calculate the first initial time that easy false triggering phrase plays and the
One end time.
In the present embodiment, lyrics unit specifically refers to form the ultimate unit of the lyrics, for example:The song of Chinese lyrics file
Word unit is Chinese character, the lyrics unit of English lyrics file is English word etc..
It is assumed that each lyrics are all at the uniform velocity to sing in a specific example, the easy false triggering phrase place lyrics
The initial reproduction time of sentence is t1, and end reproduction time is t2, and this lyrics sentence has 10 Chinese characters, this easy false triggering phrase
The 3rd, 4,5 words positioned at this lyrics sentence, then, the first initial time T1 of this easy false triggering phrase and the first end time
The computing formula of T2 is respectively:
T1=t1+2 [(t2-t1)/10], T2=t1+5 [(t2-t1)/10].
Easy corresponding first initial time of false triggering phrase and the first end time in step 209, reading song to be played.
In the present embodiment, if Chinese lyrics file has the easy false triggering phrase of preservation, need not again search
Easily false triggering phrase, only need to directly invoke the easy false triggering phrase of this preservation.
Step 210, play song to be played, close sound identification module when reaching the first initial time, reach the
Start sound identification module during one end time.
Step 211, the easy false triggering phrase of preservation and corresponding first initial time and the first end time.
In the present embodiment, after playback of songs to be played finishes, need to preserve easy false triggering phrase and corresponding the first
Time beginning and the first end time, directly invoke during song so that next time plays a song.
The method preventing voice command misrecognition provided in an embodiment of the present invention, is mated with song to be played by first obtaining
Lyrics file, be, the no easy false triggering phrase that there is preservation to carry out respectively directly reading easy false triggering according to lyrics file
Phrase and its corresponding first initial time and the operation of the first end time, and by travel through the lyrics in all Chinese characters send out
Sound attribute determines easy false triggering phrase, obtains the second initial time and second that the lyrics sentence belonging to easy false triggering phrase is play
End time, and calculate the first initial time and the operation of the first end time that easy false triggering phrase is play, then play and treat
Play song simultaneously to close in good time, open sound identification module, finally preserve easy false triggering phrase and corresponding first initial time and the
One end time, overcome in order to reduce the probability of voice false triggering in prior art, and then increase answering of speech recognition algorithm
Miscellaneous degree, leads to processor computation burden to increase, equipment power dissipation rises and speech recognition algorithm transplants difficult technological deficiency, real
Show the complexity that need not improve speech recognition algorithm and known it is also possible to reliably reduce the voice leading to due to playing song by mistake
Not.
Embodiment three
Fig. 3 is a kind of flow chart of method preventing voice command misrecognition that the embodiment of the present invention three provides.This enforcement
Example is optimized based on above-described embodiment, in the present embodiment, will calculate easy false triggering phrase and play according to lyrics file
The first initial time and the first end time be optimized for:According to lyrics file, obtain the lyrics language belonging to easy false triggering phrase
The second initial time and the second end time that sentence is play;At the end of the first compensation time and first obtaining the first initial time
Between second compensation the time;According to the second initial time, the second end time, the first compensation time, the second compensation time, the lyrics
Number that sentence comprises lyrics unit, easy false triggering phrase comprise the number of lyrics unit and easy false triggering phrase in lyrics sentence
In position, calculate the first initial time and the first end time that easy false triggering phrase is play.
Further, can also include:After playback of songs to be played terminates, statistics causes and does not cause default Voice command
The easy false triggering phrase of order misrecognition;Revise easy false triggering phrase corresponding first compensation time and the second compensation time, weight
New the first initial time calculating easy false triggering phrase broadcasting and the first end time.
Accordingly, the method for the present embodiment specifically includes:
The lyrics file that step 301, acquisition are mated with song to be played.
Step 302, the Chinese lyrics file of judgement whether there is the easy false triggering phrase of preservation, if not existing, execute
Step 303, if exist, execution step 307.
Step 303, the easy false triggering phrase searched in lyrics file, wherein, easy false triggering phrase and default Voice command
The language of order is same or like.
Step 304, the second initial time play according to lyrics file, the lyrics sentence belonging to the easy false triggering phrase of acquisition
With the second end time.
The second compensation time of step 305, the first compensation time of acquisition the first initial time and the first end time.
According to the explanation in embodiment one, if it is assumed that song to be played is at the uniform velocity broadcasting completely, then according to
The first initial time that the easy false triggering phrase that this situation is calculated is play and the first end time may be with practical situations
There is error, therefore, in order that the first initial time of easy false triggering phrase broadcasting and the first end time are more accurate, this enforcement
The first compensation time and the second compensation time is increased, with the first initial time and first that easy false triggering phrase is play in example
End time is modified.
Wherein, the first compensation time is specifically for adjusting the first initial time, and second compensates the time specifically for adjustment the
One end time, first compensate the time and second compensate the time can identical it is also possible to different, the present embodiment is not limited to this
System.First compensation time and second compensation the time concrete numerical value can be empirical value (for example:1 second etc.), also can close
Arbitrarily arrange in the range of reason.
Further, since LRC lyrics file only records the broadcasting initial time of every lyrics and does not record the knot of every lyrics
Bundle time or the time span of broadcasting, therefore, when the type of the lyrics file of song to be played coupling is LRC file, and work as
When easily having musical background between the lyrics sentence at false triggering phrase place and next lyrics sentence, if giving tacit consent to next lyrics
Play the broadcasting end time that initial time is upper lyrics, then the broadcasting end time of a lyrics acquiescence and reality on this
The broadcasting end time on border differs, and therefore, the broadcasting initial time according to next lyrics of acquiescence is upper lyrics
The required time point that the broadcasting end time calculates has error with actual time point, and the introducing compensating the time can reduce very
To eliminating this error.
Step 306, the first initial time calculating easy false triggering phrase broadcasting and the first end time, specifically, according to
Second initial time, the second end time, the first compensation time, the second compensation time, lyrics sentence comprise the individual of lyrics unit
Several, easy false triggering phrase comprises the number and easy false triggering phrase of the lyrics unit position in lyrics sentence, calculates easy false touch
Send out the first initial time and the first end time that phrase is play.
In the present embodiment, calculate the first initial time T1 that easy false triggering phrase is play ' and the first end time T2 '
Method is:Calculate the first initial time T1 and the first end that easy false triggering phrase does not consider the broadcasting during compensation time first
Time T2, concrete steps may refer to illustrating in step 208, when then compensating time T ' and second compensation according to first
Between T " calculate T1 ' and T2 ', formula is:
T1'=T1-T', T2'=T2+T ", wherein, T ' and T " it is positive number.
Easy corresponding first initial time of false triggering phrase and the first end time in step 307, reading song to be played.
Step 308, play song to be played, close sound identification module when reaching the first initial time, reach the
Start sound identification module during one end time.
After step 309, playback of songs to be played terminate, statistics causes and does not cause default voice control command misrecognition
Easily false triggering phrase.
In the present embodiment, after playback of songs to be played terminates, to causing and default voice control command can not caused by mistake
The easy false triggering phrase of identification is counted.
Step 310, the easy false triggering phrase of correction corresponding first compensate time and the second compensation time, recalculate easy mistake
The first initial time and the first end time that trigger word group is play.
In the present embodiment, when easy false triggering phrase causes default Voice command in the playing process of song to be played
The misrecognition of order, then it is assumed that corresponding first initial time of this easy false triggering phrase and the first end time are inaccurate, needs
It is adjusted.The method of adjustment can be specifically accordingly to increase the first compensation time and the second compensation time, and both increase
Time quantum can identical it is also possible to different, for example the first compensation time and the second compensation time can be increased by 10% simultaneously, so
Recalculate the first of this easy false triggering phrase broadcasting using the first compensation time after increasing and the second compensation time afterwards to initiate
Time and the first end time.
In the present embodiment, when easy false triggering phrase does not cause default voice in the playing process of song to be played
The misrecognition of control command, then can accordingly reduce the first compensation time and second compensation the time numerical value, both reduce when
The area of a room can identical it is also possible to different, for example the first compensation time and the second compensation time can be reduced 5% simultaneously, then
Using the first compensation time after reducing and the second compensation time recalculate that this easy false triggering phrase plays first initial when
Between and the first end time.
Step 311, the easy false triggering phrase of preservation and corresponding first initial time and the first end time.
The method preventing voice command misrecognition provided in an embodiment of the present invention, is mated with song to be played by first obtaining
Lyrics file, be, the no easy false triggering phrase that there is preservation to carry out respectively directly reading easy false triggering according to lyrics file
Corresponding first initial time of phrase and the operation of the first end time, and belonged to by traveling through the pronunciation of all Chinese characters in the lyrics
Property determine easy false triggering phrase, obtain easy false triggering phrase belonging to lyrics sentence play the second initial time and second end
Time and the first compensation time and second compensate with strength, and calculate the first initial time and first that easy false triggering phrase is play
The operation of end time, then plays song to be played and closes in good time, opens sound identification module, play and count easy false touch after terminating
Send out whether phrase causes the situation of speech recognition mistake and correspondingly the first initial time and the first end time are modified,
After preserve easy false triggering phrase and corresponding first initial time and the first end time, overcome in prior art to reduce
The probability of voice false triggering, and then increase the complexity of speech recognition algorithm, lead to the increase of processor computation burden, equipment power dissipation
Rise and speech recognition algorithm transplant difficult technological deficiency it is achieved that the complexity of speech recognition algorithm need not be improved,
Can reliably reduce due to playing the voice misrecognition that song leads to, and open language to greatest extent while playing song
Sound identification module.
Example IV
Fig. 4 is a kind of structure chart of device preventing voice command misrecognition that the embodiment of the present invention four provides.As Fig. 4 institute
Show, described device includes:Lyrics file acquisition module 101, easy false triggering phrase searching modul 102, reproduction time computing module
103 and sound identification module control module 104.Wherein:
Lyrics file acquisition module 101, for obtaining the lyrics file mated with song to be played;
Easily false triggering phrase searching modul 102, for searching the easy false triggering phrase in lyrics file, wherein, easy false touch
Send out phrase same or like with the language of default voice control command;
Reproduction time computing module 103, when initiateing for according to lyrics file, calculating the first of easy false triggering phrase broadcasting
Between and the first end time;
Sound identification module control module 104, for playing song to be played, closes language when reaching the first initial time
Sound identification module, starts sound identification module when reaching for the first end time.
The device preventing voice command misrecognition provided in an embodiment of the present invention, is mated with song to be played by first obtaining
Lyrics file, then look up the easy false triggering phrase in lyrics file, when calculate that easy false triggering phrase plays first is initial
Between and the first end time, finally play song to be played, reach the first initial time when close sound identification module, arrive
Reach startup sound identification module during the first end time, overcome in prior art to reduce the probability of voice false triggering, enter
And increase the complexity of speech recognition algorithm, lead to processor computation burden to increase, equipment power dissipation rises and speech recognition is calculated
The difficult technological deficiency of method transplanting it is achieved that need not improve speech recognition algorithm complexity it is also possible to reliably reduce due to
Play the voice misrecognition that song leads to.
On the basis of the various embodiments described above, after sound identification module control module, can also include:
Reproduction time preserving module, at the end of preserving easy false triggering phrase and corresponding first initial time and first
Between;
Before reproduction time computing module, can also include:
Easily false triggering phrase confirms module, for confirming that lyrics file do not have the easy false triggering phrase of preservation;
Before sound identification module control module, can also include:
Reproduction time read module, if for the easy false triggering phrase that there is preservation, reads in song to be played and easily misses
Triggering corresponding first initial time of phrase and the first end time.
On the basis of the various embodiments described above, lyrics file can be Chinese lyrics file;
Default voice control command can be Chinese speech control command;
Easily false triggering phrase searching modul can include:
Lyrics Traversal Unit, for traveling through the pronunciation attribute of all Chinese lyrics in Chinese lyrics file, wherein, pronunciation belongs to
Property at least includes tone, initial consonant and simple or compound vowel of a Chinese syllable;
Same words group acknowledge unit, if for each in one of all Chinese lyrics Chinese character or a Chinese phrase
Individual Chinese character is all identical with the pronunciation attribute of the Chinese character of correspondence position in Chinese speech control command, then confirm this Chinese character or Chinese words
Group is identical with the language of Chinese speech control command;
Close phrase confirmation unit, if for each in one of all Chinese lyrics Chinese character or a Chinese phrase
Individual Chinese character is all at least a kind of different from the pronunciation attribute of the Chinese character of correspondence position in Chinese speech control command and at least
Two kinds identical, then confirm that this Chinese character or institute's Chinese phrase are close with the language of Chinese speech control command;
Easily false triggering phrase marker unit, for being labeled as easy false triggering phrase by a Chinese character or a Chinese phrase.
On the basis of the various embodiments described above, reproduction time computing module can include:
Second reproduction time determining unit, for according to lyrics file, obtaining the lyrics sentence belonging to easy false triggering phrase
The second initial time play and the second end time;
First reproduction time computing unit, for comprising to sing according to the second initial time, the second end time, lyrics sentence
The number of word unit, easy false triggering phrase comprise the number and easy false triggering phrase of the lyrics unit position in lyrics sentence,
Calculate the first initial time and the first end time that easy false triggering phrase is play.
On the basis of the various embodiments described above, reproduction time computing module can include:
Second reproduction time determining unit, for according to lyrics file, obtaining the lyrics sentence belonging to easy false triggering phrase
The second initial time play and the second end time;
Compensate time acquisition unit, for obtaining the of the first compensation time of the first initial time and the first end time
Two compensation times;
Second reproduction time computing unit, for according to the second initial time, the second end time, first compensate the time,
The number that second compensation time, lyrics sentence comprise lyrics unit, easy false triggering phrase comprise the number of lyrics unit and easy mistake
Triggering position in lyrics sentence for the phrase, calculates the first initial time and the first end time that easy false triggering phrase is play.
On the basis of the various embodiments described above, can also include:
Misrecognition statistical module, after terminating for playback of songs to be played, statistics causes and does not cause default Voice command
The easy false triggering phrase of order misrecognition;
Compensate time complexity curve module, when compensating for revising the easy false triggering phrase corresponding first compensation time and second
Between, recalculate the first initial time and the first end time that easy false triggering phrase is play.
The device of voice command misrecognition that what the embodiment of the present invention was provided prevent can be used for executing the present invention arbitrarily to be implemented
The method preventing voice command misrecognition of example offer, possesses corresponding functional module, realizes identical beneficial effect.
Obviously, it will be understood by those skilled in the art that each module of the above-mentioned present invention or each step can be by as above
Described server implementation.Alternatively, the embodiment of the present invention can be realized with the executable program of computer installation, thus can
To be executed by processor with being stored in storage device, described program can be stored in a kind of computer-readable storage
In medium, storage medium mentioned above can be read only memory, disk or CD etc.;Or they are fabricated to respectively each
Individual integrated circuit modules, or the multiple modules in them or step are fabricated to single integrated circuit module to realize.So,
The present invention is not restricted to the combination of any specific hardware and software.
The foregoing is only the preferred embodiments of the present invention, be not limited to the present invention, for those skilled in the art
For, the present invention can have various change and change.All any modifications made within spirit and principles of the present invention, equivalent
Replace, improve etc., should be included within the scope of the present invention.
Claims (12)
1. a kind of method preventing voice command misrecognition is it is characterised in that include:
Obtain the lyrics file mated with song to be played;
Search the easy false triggering phrase in described lyrics file, wherein, described easy false triggering phrase and default voice control command
Language same or like;
According to described lyrics file, calculate the first initial time and the first end time that described easy false triggering phrase is play;
Play described song to be played, close sound identification module when reaching described first initial time, reach described the
Start described sound identification module during one end time.
2. method according to claim 1 is it is characterised in that the described song to be played of described broadcasting, reaches described the
During one initial time close sound identification module, reach described first end time when start described sound identification module it
Afterwards, also include:
Preserve described easy false triggering phrase and corresponding described first initial time and described first end time;
Described the first initial time and the first end time that according to described lyrics file, the described easy false triggering phrase of calculating is play
Before, also include:
Confirm that described lyrics file does not have the described easy false triggering phrase of preservation;
The described song to be played of described broadcasting, closes sound identification module when reaching described first initial time, reaches institute
State before starting described sound identification module during the first end time, also include:
If there is the described easy false triggering phrase of preservation, read easily false triggering phrase described in described song to be played corresponding
Described first initial time and described first end time.
3. method according to claim 1 is it is characterised in that described lyrics file is Chinese lyrics file;
Described default voice control command is Chinese speech control command;
The described easy false triggering phrase searched in described lyrics file includes:
The pronunciation attribute of all Chinese lyrics in traversal described Chinese lyrics file, wherein, described pronunciation attribute at least includes sound
Tune, initial consonant and simple or compound vowel of a Chinese syllable;
If each Chinese character in described all Chinese one of lyrics Chinese characters or a Chinese phrase all with described Chinese language
In sound control command, the described pronunciation attribute of the Chinese character of correspondence position is identical, then confirm this Chinese character or Chinese phrase and described Chinese
The language of voice control command is identical;
If each Chinese character in described all Chinese one of lyrics Chinese characters or a Chinese phrase all with described Chinese language
In the described pronunciation attribute of the Chinese character of correspondence position in sound control command at least a kind of different and at least two kinds identical, then really
Recognize this Chinese character or Chinese phrase close with the language of described Chinese speech control command;
One Chinese character or one Chinese phrase are labeled as easy false triggering phrase.
4. the method according to any one of claim 1-3 it is characterised in that described according to described lyrics file, calculate institute
State the first initial time of easy false triggering phrase broadcasting and the first end time included:
According to described lyrics file, obtain the second initial time that the lyrics sentence belonging to described easy false triggering phrase plays and the
Two end times;
The number of lyrics unit, described is comprised according to described second initial time, described second end time, described lyrics sentence
Easily false triggering phrase comprises the number and described easy false triggering phrase of the lyrics unit position in described lyrics sentence, calculates institute
State described first initial time and described first end time that easy false triggering phrase is play.
5. the method according to any one of claim 1-3 it is characterised in that described according to described lyrics file, calculate institute
State the first initial time of easy false triggering phrase broadcasting and the first end time included:
According to described lyrics file, obtain described second initial time that the lyrics sentence belonging to described easy false triggering phrase is play
With described second end time;
Obtain the first compensation time of described first initial time and the second compensation time of described first end time;
According to described second initial time, described second end time, described first compensate the time, described second compensate the time,
The number that described lyrics sentence comprises lyrics unit, described easy false triggering phrase comprise the number of lyrics unit and described easy false touch
Send out position in described lyrics sentence for the phrase, calculate described first initial time that described easy false triggering phrase plays and described
First end time.
6. method according to claim 5 is it is characterised in that also include:
After described playback of songs to be played terminates, statistics causes and does not cause the described of described default voice control command misrecognition
Easily false triggering phrase;
Revise described easy false triggering phrase corresponding described first compensation time and described second compensation time, recalculate described
Described first initial time and described first end time that easily false triggering phrase is play.
7. a kind of device preventing voice command misrecognition is it is characterised in that include:
Lyrics file acquisition module, for obtaining the lyrics file mated with song to be played;
Easily false triggering phrase searching modul, for searching the easy false triggering phrase in described lyrics file, wherein, described easy false touch
Send out phrase same or like with the language of default voice control command;
Reproduction time computing module, for according to described lyrics file, calculate described easy false triggering phrase broadcasting first initiates
Time and the first end time;
Sound identification module control module, for playing described song to be played, closes when reaching described first initial time
Sound identification module, starts described sound identification module when reaching described first end time.
8. device according to claim 7 is it is characterised in that after described sound identification module control module, also include:
Reproduction time preserving module, for preserving described easy false triggering phrase and corresponding described first initial time and described
One end time;
Before described reproduction time computing module, also include:
Easily false triggering phrase confirms module, for confirming that described lyrics file do not have the described easy false triggering phrase of preservation;
Before described sound identification module control module, also include:
Reproduction time read module, if for the described easy false triggering phrase that there is preservation, reads in described song to be played
Described corresponding described first initial time of easy false triggering phrase and described first end time.
9. device according to claim 7 is it is characterised in that described lyrics file is Chinese lyrics file;
Described default voice control command is Chinese speech control command;
Described easy false triggering phrase searching modul includes:
Lyrics Traversal Unit, for traveling through the pronunciation attribute of all Chinese lyrics in described Chinese lyrics file, wherein, described
Sound attribute at least includes tone, initial consonant and simple or compound vowel of a Chinese syllable;
Same words group acknowledge unit, if for each in one of described all Chinese lyrics Chinese character or a Chinese phrase
Individual Chinese character is all identical with the described pronunciation attribute of the Chinese character of correspondence position in described Chinese speech control command, then confirm this Chinese character
Or Chinese phrase is identical with the language of described Chinese speech control command;
Close phrase confirmation unit, if for each in one of described all Chinese lyrics Chinese character or a Chinese phrase
Individual Chinese character is all at least a kind of different from the described pronunciation attribute of the Chinese character of correspondence position in described Chinese speech control command
And at least two kinds identical, then confirm that this Chinese character or Chinese phrase are close with the language of described Chinese speech control command;
Easily false triggering phrase marker unit, for being labeled as easy false triggering word by one Chinese character or one Chinese phrase
Group.
10. the device according to any one of claim 7-9 is it is characterised in that described reproduction time computing module includes:
Second reproduction time determining unit, for according to described lyrics file, obtaining the lyrics belonging to described easy false triggering phrase
The second initial time and the second end time that sentence is play;
First reproduction time computing unit, for according to described second initial time, described second end time, described lyrics language
Number that the sentence number that comprises lyrics unit, described easy false triggering phrase comprise lyrics unit and described easy false triggering phrase are in institute
State the position in lyrics sentence, at the end of calculating described first initial time and described first that described easy false triggering phrase is play
Between.
11. devices according to any one of claim 7-9 are it is characterised in that described reproduction time computing module includes:
Second reproduction time determining unit, for according to described lyrics file, obtaining the lyrics belonging to described easy false triggering phrase
Described second initial time and described second end time that sentence is play;
Compensate time acquisition unit, for obtaining the first compensation time of described first initial time and described first end time
Second compensation the time;
Second reproduction time computing unit, for according to described second initial time, described second end time, described first benefit
Repay the time, number that described second compensation time, described lyrics sentence comprise lyrics unit, described easy false triggering phrase comprise to sing
The number of the word unit and described easy false triggering phrase position in described lyrics sentence, calculates described easy false triggering phrase and plays
Described first initial time and described first end time.
12. devices according to claim 11 are it is characterised in that also include:
Misrecognition statistical module, after terminating for described playback of songs to be played, statistics causes and does not cause described default voice
The described easy false triggering phrase of control command misrecognition;
Compensate time complexity curve module, for revising described easy false triggering phrase corresponding described first compensation time and described second
The compensation time, recalculate described first initial time and described first end time that described easy false triggering phrase is play.
Priority Applications (2)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201610909229.6A CN106409294B (en) | 2016-10-18 | 2016-10-18 | The method and apparatus for preventing voice command from misidentifying |
PCT/CN2016/113279 WO2018072327A1 (en) | 2016-10-18 | 2016-12-29 | Method and device for preventing misrecognition of voice command |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201610909229.6A CN106409294B (en) | 2016-10-18 | 2016-10-18 | The method and apparatus for preventing voice command from misidentifying |
Publications (2)
Publication Number | Publication Date |
---|---|
CN106409294A true CN106409294A (en) | 2017-02-15 |
CN106409294B CN106409294B (en) | 2019-07-16 |
Family
ID=58013014
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN201610909229.6A Active CN106409294B (en) | 2016-10-18 | 2016-10-18 | The method and apparatus for preventing voice command from misidentifying |
Country Status (2)
Country | Link |
---|---|
CN (1) | CN106409294B (en) |
WO (1) | WO2018072327A1 (en) |
Cited By (5)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN108231076A (en) * | 2018-01-04 | 2018-06-29 | 广州视源电子科技股份有限公司 | A kind of sound control method, device, equipment and storage medium |
CN110827792A (en) * | 2019-11-15 | 2020-02-21 | 广州视源电子科技股份有限公司 | Voice broadcasting method and device |
CN110970027A (en) * | 2019-12-25 | 2020-04-07 | 上海博泰悦臻电子设备制造有限公司 | Voice recognition method, device, computer storage medium and system |
CN111433737A (en) * | 2017-12-04 | 2020-07-17 | 三星电子株式会社 | Electronic device and control method thereof |
CN116884399A (en) * | 2023-09-06 | 2023-10-13 | 深圳市友杰智新科技有限公司 | Method, device, equipment and medium for reducing voice misrecognition |
Families Citing this family (1)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN112509567B (en) * | 2020-12-25 | 2024-05-10 | 阿波罗智联(北京)科技有限公司 | Method, apparatus, device, storage medium and program product for processing voice data |
Citations (13)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN101290767A (en) * | 2007-04-20 | 2008-10-22 | 华硕电脑股份有限公司 | Portable computer with speech recognition function and processing method therefor |
CN101383150A (en) * | 2008-08-19 | 2009-03-11 | 南京师范大学 | Control method of speech soft switch and its application in geographic information system |
CN101634987A (en) * | 2008-07-21 | 2010-01-27 | 上海天统电子科技有限公司 | Multimedia player |
CN101998138A (en) * | 2009-08-25 | 2011-03-30 | 北京达鸣慧科技有限公司 | Television channel monitoring system and real-time monitoring method thereof |
CN102006373A (en) * | 2010-11-24 | 2011-04-06 | 深圳市子栋科技有限公司 | Vehicle-mounted service system and method based on voice command control |
CN102118886A (en) * | 2010-01-04 | 2011-07-06 | 中国移动通信集团公司 | Recognition method of voice information and equipment |
CN102208184A (en) * | 2010-03-31 | 2011-10-05 | 索尼公司 | Information processing device, information processing method, and program |
CN102236686A (en) * | 2010-05-07 | 2011-11-09 | 盛乐信息技术(上海)有限公司 | Voice sectional song search method |
CN102280106A (en) * | 2010-06-12 | 2011-12-14 | 三星电子株式会社 | VWS method and apparatus used for mobile communication terminal |
CN102332265A (en) * | 2011-06-20 | 2012-01-25 | 浙江吉利汽车研究院有限公司 | Method for improving voice recognition rate of automobile voice control system |
CN103151038A (en) * | 2011-12-06 | 2013-06-12 | 张国鸿 | Method of achieving voice recognition control in electronic products |
US20140214416A1 (en) * | 2013-01-30 | 2014-07-31 | Tencent Technology (Shenzhen) Company Limited | Method and system for recognizing speech commands |
US20150088525A1 (en) * | 2013-09-24 | 2015-03-26 | Tencent Technology (Shenzhen) Co., Ltd. | Method and apparatus for controlling applications and operations on a terminal |
Family Cites Families (5)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
US6243683B1 (en) * | 1998-12-29 | 2001-06-05 | Intel Corporation | Video control of speech recognition |
DE10058786A1 (en) * | 2000-11-27 | 2002-06-13 | Philips Corp Intellectual Pty | Method for controlling a device having an acoustic output device |
US8738382B1 (en) * | 2005-12-16 | 2014-05-27 | Nvidia Corporation | Audio feedback time shift filter system and method |
CN101753871A (en) * | 2008-11-28 | 2010-06-23 | 康佳集团股份有限公司 | Voice remote control TV system |
CN102945672B (en) * | 2012-09-29 | 2013-10-16 | 深圳市国华识别科技开发有限公司 | Voice control system for multimedia equipment, and voice control method |
-
2016
- 2016-10-18 CN CN201610909229.6A patent/CN106409294B/en active Active
- 2016-12-29 WO PCT/CN2016/113279 patent/WO2018072327A1/en active Application Filing
Patent Citations (13)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN101290767A (en) * | 2007-04-20 | 2008-10-22 | 华硕电脑股份有限公司 | Portable computer with speech recognition function and processing method therefor |
CN101634987A (en) * | 2008-07-21 | 2010-01-27 | 上海天统电子科技有限公司 | Multimedia player |
CN101383150A (en) * | 2008-08-19 | 2009-03-11 | 南京师范大学 | Control method of speech soft switch and its application in geographic information system |
CN101998138A (en) * | 2009-08-25 | 2011-03-30 | 北京达鸣慧科技有限公司 | Television channel monitoring system and real-time monitoring method thereof |
CN102118886A (en) * | 2010-01-04 | 2011-07-06 | 中国移动通信集团公司 | Recognition method of voice information and equipment |
CN102208184A (en) * | 2010-03-31 | 2011-10-05 | 索尼公司 | Information processing device, information processing method, and program |
CN102236686A (en) * | 2010-05-07 | 2011-11-09 | 盛乐信息技术(上海)有限公司 | Voice sectional song search method |
CN102280106A (en) * | 2010-06-12 | 2011-12-14 | 三星电子株式会社 | VWS method and apparatus used for mobile communication terminal |
CN102006373A (en) * | 2010-11-24 | 2011-04-06 | 深圳市子栋科技有限公司 | Vehicle-mounted service system and method based on voice command control |
CN102332265A (en) * | 2011-06-20 | 2012-01-25 | 浙江吉利汽车研究院有限公司 | Method for improving voice recognition rate of automobile voice control system |
CN103151038A (en) * | 2011-12-06 | 2013-06-12 | 张国鸿 | Method of achieving voice recognition control in electronic products |
US20140214416A1 (en) * | 2013-01-30 | 2014-07-31 | Tencent Technology (Shenzhen) Company Limited | Method and system for recognizing speech commands |
US20150088525A1 (en) * | 2013-09-24 | 2015-03-26 | Tencent Technology (Shenzhen) Co., Ltd. | Method and apparatus for controlling applications and operations on a terminal |
Cited By (8)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN111433737A (en) * | 2017-12-04 | 2020-07-17 | 三星电子株式会社 | Electronic device and control method thereof |
CN108231076A (en) * | 2018-01-04 | 2018-06-29 | 广州视源电子科技股份有限公司 | A kind of sound control method, device, equipment and storage medium |
CN110827792A (en) * | 2019-11-15 | 2020-02-21 | 广州视源电子科技股份有限公司 | Voice broadcasting method and device |
CN110827792B (en) * | 2019-11-15 | 2022-06-03 | 广州视源电子科技股份有限公司 | Voice broadcasting method and device |
CN110970027A (en) * | 2019-12-25 | 2020-04-07 | 上海博泰悦臻电子设备制造有限公司 | Voice recognition method, device, computer storage medium and system |
CN110970027B (en) * | 2019-12-25 | 2023-07-25 | 博泰车联网科技(上海)股份有限公司 | Voice recognition method, device, computer storage medium and system |
CN116884399A (en) * | 2023-09-06 | 2023-10-13 | 深圳市友杰智新科技有限公司 | Method, device, equipment and medium for reducing voice misrecognition |
CN116884399B (en) * | 2023-09-06 | 2023-12-08 | 深圳市友杰智新科技有限公司 | Method, device, equipment and medium for reducing voice misrecognition |
Also Published As
Publication number | Publication date |
---|---|
CN106409294B (en) | 2019-07-16 |
WO2018072327A1 (en) | 2018-04-26 |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
CN106409294A (en) | Method and apparatus for preventing voice command misidentification | |
CN110148427B (en) | Audio processing method, device, system, storage medium, terminal and server | |
US11398236B2 (en) | Intent-specific automatic speech recognition result generation | |
US20220156039A1 (en) | Voice Control of Computing Devices | |
US10884701B2 (en) | Voice enabling applications | |
US8200490B2 (en) | Method and apparatus for searching multimedia data using speech recognition in mobile device | |
US8527272B2 (en) | Method and apparatus for aligning texts | |
CN106683677B (en) | Voice recognition method and device | |
CN111710333B (en) | Method and system for generating speech transcription | |
CN109635270B (en) | Bidirectional probabilistic natural language rewrite and selection | |
US10917758B1 (en) | Voice-based messaging | |
KR102390940B1 (en) | Context biasing for speech recognition | |
US20150100314A1 (en) | Multiple web-based content category searching in mobile search application | |
US20110054900A1 (en) | Hybrid command and control between resident and remote speech recognition facilities in a mobile voice-to-speech application | |
US20110060587A1 (en) | Command and control utilizing ancillary information in a mobile voice-to-speech application | |
CN111429903B (en) | Audio signal identification method, device, system, equipment and readable medium | |
US10366690B1 (en) | Speech recognition entity resolution | |
US10783876B1 (en) | Speech processing using contextual data | |
KR20020027382A (en) | Voice commands depend on semantics of content information | |
JP3911178B2 (en) | Speech recognition dictionary creation device and speech recognition dictionary creation method, speech recognition device, portable terminal, speech recognition system, speech recognition dictionary creation program, and program recording medium | |
US11582174B1 (en) | Messaging content data storage | |
Brazier et al. | On-line audio-to-lyrics alignment based on a reference performance | |
CN113536029A (en) | Method and device for aligning audio and text, electronic equipment and storage medium | |
JP4794429B2 (en) | Reading information generation device, reading information generation method, reading information generation program, and speech synthesizer | |
JP2008116650A (en) | Reading information creating apparatus, reading information creating method, reading information creating program and voice synthesizer |
Legal Events
Date | Code | Title | Description |
---|---|---|---|
C06 | Publication | ||
PB01 | Publication | ||
C10 | Entry into substantive examination | ||
SE01 | Entry into force of request for substantive examination | ||
GR01 | Patent grant | ||
GR01 | Patent grant |