CN107103915A

CN107103915A - A kind of audio data processing method and device

Info

Publication number: CN107103915A
Application number: CN201610090785.5A
Authority: CN
Inventors: 陈雪琪; 赵伟峰
Original assignee: Guangzhou Kugou Computer Technology Co Ltd
Current assignee: Guangzhou Kugou Computer Technology Co Ltd
Priority date: 2016-02-18
Filing date: 2016-02-18
Publication date: 2017-08-29

Abstract

The embodiment of the invention discloses a kind of audio data processing method and device, wherein method includes：Recorded audio data are obtained, and recognize each recording individual character in recorded audio data；Recorded audio data are split as multiple individual character voice datas according to each recording individual character；Lyrics individual character corresponding with each individual character voice data is obtained in lyrics file corresponding with recorded audio data；The corresponding standard original singer parameter of actual performance parameter and lyrics individual character of individual character voice data is compared, the parameter comparison information between individual character voice data standard original singer's parameter corresponding with lyrics individual character is determined；Parameter comparison information includes time registration and/or tone deviation；The corresponding type of error mark of individual character voice data is generated according to parameter comparison information, and shows that type of error is identified in the range areas of display lyrics individual character.Using the present invention, singing displaying content can be enriched, and be really given to the rational recommendation on improvement of user.

Description

A kind of audio data processing method and device

Technical field

The present invention relates to field of computer technology, more particularly to a kind of audio data processing method and device.

Background technology

Current most of singing softwares all possess marking function, and specific scoring process can be：Pass through contrast MIDI (the Musical Instrument Digital Interface, music of the spectrum waveform of recording and corresponding song Apparatus figure interface) curve degree of fitting, the performance level to the song gives a mark.Although by singing in antiphonal style Song, which is sung give a mark, can improve enthusiasm of the user to singing, but only be shown by a fraction, The general level that user sings can only be broadly informed, and user does not know the place for oneself singing mistake still, more Do not know how oneself should improve, it can be seen that, it is existing sing marking function displaying when content excessively It is single, it is impossible to be really given to the rational recommendation on improvement of user.

The content of the invention

The embodiment of the present invention provides a kind of audio data processing method and device, can enrich singing displaying content, And really it is given to the rational recommendation on improvement of user.

The embodiments of the invention provide a kind of audio data processing method, including：

Recorded audio data are obtained, and recognize each recording individual character in the recorded audio data；

The recorded audio data are split as multiple individual character voice datas according to each recording individual character；

Obtained in lyrics file corresponding with the recorded audio data and each individual character voice data pair The lyrics individual character answered；

By the actual performance parameter of the individual character voice data and the corresponding standard original singer parameter of the lyrics individual character It is compared, determines between individual character voice data standard original singer parameter corresponding with the lyrics individual character Parameter comparison information；The parameter comparison information includes time registration and/or tone deviation；

The corresponding type of error of the individual character voice data is generated according to the parameter comparison information to identify, and Show and the type of error mark is shown in the range areas of the lyrics individual character.

Correspondingly, the embodiment of the present invention additionally provides a kind of audio-frequency data processing device, including：

Identification module is obtained, for obtaining recorded audio data, and is recognized every in the recorded audio data Individual recording individual character；

Module is split, for the recorded audio data to be split as into multiple lists according to each recording individual character Word voice data；

Individual character acquisition module, for being obtained in lyrics file corresponding with the recorded audio data and each The corresponding lyrics individual character of the individual character voice data；

Compare determining module, for by the actual performance parameter and the lyrics individual character of the individual character voice data Corresponding standard original singer parameter is compared, and determines that the individual character voice data is corresponding with the lyrics individual character Parameter comparison information between standard original singer's parameter；The parameter comparison information includes time registration and/or sound Tuningout difference；

Display module is generated, it is corresponding for generating the individual character voice data according to the parameter comparison information Type of error is identified, and shows that the type of error is identified in the range areas for showing the lyrics individual character.

The embodiment of the present invention, can be with by each recording individual character in the recorded audio data accessed by identification Recorded audio data are split as multiple individual character voice datas according to each recording individual character, and in lyrics file Lyrics individual character corresponding with the individual character voice data in multiple individual character voice datas is obtained, further according to individual character audio Data it is corresponding it is actual sing parameter and the corresponding standard original singer parameter of lyrics individual character determine individual character voice data with Parameter comparison information between the corresponding standard original singer parameter of lyrics individual character, to generate individual character voice data correspondence Type of error mark, and display lyrics individual character range areas in show type of error mark, so as to Displaying content of singing is enriched, and can be as accurate as each problematic lyrics individual character and shows corresponding mistake Type identification, so as to really be given to the rational recommendation on improvement of user.

Brief description of the drawings

In order to illustrate more clearly about the embodiment of the present invention or technical scheme of the prior art, below will be to implementing The accompanying drawing used required in example or description of the prior art is briefly described, it should be apparent that, describe below In accompanying drawing be only some embodiments of the present invention, for those of ordinary skill in the art, do not paying On the premise of going out creative work, other accompanying drawings can also be obtained according to these accompanying drawings.

Fig. 1 is a kind of schematic flow sheet of audio data processing method provided in an embodiment of the present invention；

Fig. 1 a are a kind of surface charts of lyrics mark displaying provided in an embodiment of the present invention；

Fig. 2 is the schematic flow sheet of another audio data processing method provided in an embodiment of the present invention；

Fig. 3 is the schematic flow sheet of another audio data processing method provided in an embodiment of the present invention；

Fig. 4 is a kind of structural representation of audio-frequency data processing device provided in an embodiment of the present invention；

Fig. 5 is a kind of structural representation for comparing determining module provided in an embodiment of the present invention；

Fig. 6 is a kind of structural representation of individual character acquisition module provided in an embodiment of the present invention；

Fig. 7 is the structural representation of another audio-frequency data processing device provided in an embodiment of the present invention.

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 clear Chu, it is fully described by, it is clear that described embodiment is only a part of embodiment of the invention, rather than Whole embodiments.Based on the embodiment in the present invention, those of ordinary skill in the art are not making creation Property work under the premise of the every other embodiment that is obtained, belong to the scope of protection of the invention.

Fig. 1 is referred to, is a kind of schematic flow sheet of audio data processing method provided in an embodiment of the present invention, Methods described can include：

S101, obtains recorded audio data, and recognize each recording individual character in the recorded audio data；

Specifically, audio-frequency data processing device can obtain the recorded audio data of user by microphone, and Each recording individual character in the recorded audio data, the voice data are identified based on speech recognition algorithm Processing unit can also terminate according to the pronunciation initial time and pronunciation of each recording individual character identified Moment records each recording individual character and distinguishes corresponding actual singing time information, the actual singing time Information includes recording individual character initial time stamp and recording individual character ending timestamp, the recording individual character initial time Stamp is the pronunciation initial time of the recording individual character, and the recording individual character ending timestamp is the recording individual character Pronunciation finish time.For example, the recorded audio data that the audio-frequency data processing device is got are user A word " being all starlet over the sky " in a certain first song sung, then the audio-frequency data processing device can be recognized Go out it is each recording individual character respectively " full ", " my god ", " ", "Yes", " small ", " star ", " star ", the audio Data processing equipment can also be further according to the pronunciation initial time of each recording individual character identified Each recording individual character is recorded with pronunciation finish time and distinguishes corresponding actual singing time information, if knowing Do not go out the pronunciation initial time of " full " word for 1 point 2.37 seconds, pronunciation finish time be 1 point 3.11 seconds, i.e., " full " The complete tone period of word be from 1 point 3.11 seconds 2.37 seconds to 1 point, then can using 1 point 2.37 seconds be used as should Recording individual character initial time stamp in the corresponding actual singing time information of individual character of recording, and by 1 point 3.11 seconds It is used as the recording individual character ending timestamp in the corresponding actual singing time information of the recording individual character.

The recorded audio data are split as multiple individual character audio numbers by S102 according to each recording individual character According to；

Specifically, the audio-frequency data processing device identifies each recording individual character, and get described Each recording individual character is distinguished after corresponding actual singing time information, you can according to each recording individual character point Recording individual character initial time stamp and recording individual character ending timestamp in not corresponding actual singing time information, The recorded audio data are split as multiple individual character voice datas；Each individual character voice data is corresponded to One recording individual character, i.e., each individual character voice data is the corresponding recording individual character initial time stamp To the voice data in the corresponding recording individual character ending timestamp.If for example, the voice data is handled Device identify it is each recording individual character respectively " full ", " my god ", " ", "Yes", " small ", " star ", " star ", And get the corresponding actual singing time information of each recording individual character difference, then at the voice data Manage device and corresponding actual singing time information is further distinguished by the recording according to each recording individual character Voice data is split as multiple individual character voice datas, if the complete tone period of " full " word is from 1 point 2.37 Record in second 2.37 seconds being the corresponding actual singing time information of the recording individual character to 1 point of 3.11 seconds, i.e., 1 point Sound individual character initial time stamp, 1 point 3.11 seconds is the record in the corresponding actual singing time information of the recording individual character Sound individual character ending timestamp, then the audio-frequency data processing device can be according to the corresponding reality of recording individual character Singing time information is by 1 point of 2.37 seconds voice data to 1 point of 3.11 second period in the recorded audio data Split out, and regard the voice data splitted out as the corresponding individual character voice data of " full " word.

S103, is obtained and each individual character audio in lyrics file corresponding with the recorded audio data The corresponding lyrics individual character of data；

Specifically, the audio-frequency data processing device can be in the lyrics corresponding with recorded audio data text Each lyrics individual character is obtained in part and distinguishes corresponding standard original time information；The standard original time information Including lyrics individual character initial time stamp and lyrics individual character ending timestamp；The lyrics file can be QRC songs Each lyrics individual character in word file, the QRC lyrics files is corresponded to respectively time tag, when described Between label including lyrics individual character initial time and the duration of lyrics individual character；Wherein, each lyrics Corresponding standard original time information is to distinguish corresponding time mark according to each lyrics individual character to individual character respectively Label calculate what is obtained, for example, the corresponding time tags of lyrics individual character A include lyrics individual character A initial time T1 and the lyrics individual character A duration t2, then can calculate the corresponding lyrics individual character startings of lyrics individual character A Timestamp is t1, and the corresponding lyrics individual character ending timestamps of lyrics individual character A are t1+t2.

The audio-frequency data processing device further distinguishes corresponding recording individual character according to each recording individual character Initial time stamp and recording individual character ending timestamp, and each lyrics individual character distinguish corresponding lyrics list Word initial time stamp and lyrics individual character ending timestamp, obtain each recording individual character and each lyrics Time match result between individual character.Wherein, by by it is described it is each recording individual character it is corresponding recording individual character rise Beginning timestamp and recording individual character ending timestamp lyrics individual character initial time corresponding with each lyrics individual character Stamp and lyrics individual character ending timestamp are contrasted, and can generate each recording individual character and each song Time match result between word individual character, and by the time match result with record individual character recording individual character Initial time stamp and the recording individual character ending immediate lyrics individual character of timestamp are defined as and the recording individual character time The lyrics individual character that the match is successful., can when determining with the successful lyrics individual character of certain recording individual character time match So that the lyrics individual character is defined as into lyrics individual character corresponding with the recording individual character, that is, include the list of the recording individual character Word voice data is corresponding with the lyrics individual character, therefore, can be described in each according to the time match result Individual character voice data determines corresponding lyrics individual character respectively.

S104, the corresponding standard of the actual performance parameter and the lyrics individual character of the individual character voice data is former Sing parameter to be compared, determine individual character voice data standard original singer parameter corresponding with the lyrics individual character Between parameter comparison information；

Specifically, the actual parameter of singing includes the corresponding actual singing time letter of the individual character voice data Breath；The standard original singer parameter includes the corresponding standard original time information of the lyrics individual character；The audio Data processing equipment can be according to the record in the corresponding actual singing time information of the individual character voice data Sound individual character initial time stamp and recording individual character ending timestamp, and the corresponding standard of the lyrics individual character The lyrics individual character initial time stamp and lyrics individual character ending timestamp in original time information, are calculated Duration and non-coincidence duration are overlapped between the individual character voice data and the lyrics individual character, and calculates described Non-coincidence duration and the ratio overlapped between duration, to obtain the individual character voice data corresponding time Registration, and it regard the time registration as the parameter comparison information.For example, in individual character voice data The corresponding actual singing time information of recording individual character " love " be (01:02.37,01:03.02) (i.e. 1 point 2.37 seconds For the recording individual character initial time stamp, 1 point 3.02 seconds is the recording individual character ending timestamp), the list The standard original time information of the corresponding lyrics individual character " love " of word voice data is (01:02.39,01:03.21) (i.e. 1 point 2.39 seconds is lyrics individual character initial time stamp, and 1 point 3.21 seconds is lyrics individual character ending timestamp)； Wherein, (2.39-2.37) a length of when non-coincidence+(3.21-3.02)=0.21 second, a length of 3.02-2.39=0.63 during coincidence Second, therefore, the corresponding time registration of the individual character voice data can for Z=[(2.39-2.37)+ (3.21-3.02)]/(3.02-2.39).Wherein, Z symbol indicates the attribute of " fast " or " partially slow ", works as Z During for positive number, represent that user sings slowly, when Z is negative, represent that user sings soon；Z absolute value is big Small to indicate fast or partially slow degree, numerical value is bigger, and deviation is bigger.

Or, the actual parameter of singing includes the corresponding actual performance pitch value of the individual character voice data； The standard original singer parameter includes the corresponding standard original singer pitch value of the lyrics individual character；At the voice data Managing device can be corresponding with the lyrics individual character by the corresponding actual performance pitch value of the individual character voice data Standard original singer's pitch value is subtracted each other, to obtain the corresponding tone deviation of the individual character voice data, and will be described Tone deviation is used as the parameter comparison information；Wherein, the corresponding standard original singer tone of the lyrics individual character Value is according to the corresponding standard original time of the lyrics individual character in MIDI MIDI files What acquisition of information was arrived, if for example, the corresponding standard original time information of the lyrics individual character is (01:02.39, 01:03.21) (i.e. 1 point 2.39 seconds is lyrics individual character initial time stamp, and 1 point 3.21 seconds is that lyrics individual character ends up Timestamp), then it can obtain in 1 point of 2.39 seconds to 1 point period of 3.21 seconds in the MIDI files Standard original singer's pitch value, and the standard original singer pitch value got is corresponding as the lyrics individual character Standard original singer's pitch value.In another example, if actual pitch value of singing is Y, standard original singer pitch value is y, then may be used To calculate the corresponding tone deviation K=Y-y of the individual character voice data；K symbol indicate " higher " or The attribute of " relatively low ", when K is positive number, represents that user sings height；When K is negative, represent that user sings It is low；K order of magnitude indicates higher and relatively low degree, and numerical value is bigger, and deviation is bigger.

Or, the actual parameter of singing includes the corresponding actual singing time letter of the individual character voice data Breath and actual performance pitch value；When the standard original singer parameter includes the lyrics individual character corresponding standard original singer Between information and standard original singer's pitch value；Therefore, the audio-frequency data processing device can calculate described in the lump Time registration and the tone deviation, and using the time registration and the tone deviation as The parameter comparison information.

S105, generates the corresponding type of error of the individual character voice data according to the parameter comparison information and identifies, And the type of error mark is shown in the range areas for showing the lyrics individual character；

Specifically, the audio-frequency data processing device can be according to the time in the parameter comparison information Registration and/or the tone deviation generate the corresponding singing mass fraction of the individual character voice data, work as institute Singing mass fraction is stated to overlap including the time less than default score threshold, and the parameter comparison information When degree and tone deviation, it can further detect that time registration accounts for low point of principal element or tone is inclined Difference accounts for low point of principal element, if the time registration accounts for low point of principal element, and the time overlaps It is to sing soon to spend show, then can generate the corresponding type of error of the individual character voice data and be designated and " sing fast ", and show that type of error mark " singing soon " institute is right in the range areas for showing the lyrics individual character The icon answered；If the tone deviation accounts for low point of principal element, and the tone deviation show be Height is sung, then can generate the corresponding type of error of the individual character voice data and be designated " singing height ", and aobvious Show and show that the type of error identifies the icon corresponding to " singing height " in the range areas of the lyrics individual character.Can Choosing, can also be in display institute if the time registration and tone deviation all deviation ratios are larger State in the range areas of lyrics individual character while showing the corresponding type of error mark of the time registration and described The corresponding type of error mark of tone deviation.Or, when the singing mass fraction is less than default point Number threshold values, and the parameter comparison information include the time registration or tone deviation when, can be direct Corresponding type of error mark is generated according to the parameter comparison information.As can be seen here, the embodiment of the present invention can Corresponding type of error mark is shown to be accurate to each problematic lyrics individual character, so as to really be given to use The rational recommendation on improvement in family.

Further, it is a kind of lyrics mark displaying provided in an embodiment of the present invention then please also refer to Fig. 1 a Surface chart.The lyrics individual character shown in Fig. 1 a be respectively " I ", " thinking ", " band ", " you ", " returning ", " I ", " ", " outer ", " mother-in-law ", " family ", and Fig. 1 a are to represent different wrong classes by being differently directed for arrow Type is identified, and the upward type of error mark of arrow represents " singing height ", and the downward type of error mark of arrow is represented " singing low ", the type of error mark of arrow right represents " singing soon ", the type of error mark table of arrow to the left Show " singing slowly ".First, can according to the recorded audio data got, QRC files and MIDI files Height is sung when singing first " I " word to analyze user, sings low when singing " you " word, when singing " returning " word Sing height, sing low when singing " mother-in-law " word, sung soon when singing " family " word, then can " I " (first), " you ", " returning ", " mother-in-law ", the top of " family " these lyrics individual characters show corresponding type of error mark respectively, with Inform how user should improve when singing which word.

Fig. 2 is referred to again, is the flow signal of another audio data processing method provided in an embodiment of the present invention Figure, methods described can include：

S201, obtains recorded audio data, and recognize each recording individual character in the recorded audio data；

S202, corresponding actual singing time information is distinguished by the recording sound according to each recording individual character Frequency evidence is split as multiple individual character voice datas；

S203, it is right respectively in lyrics file corresponding with the recorded audio data to obtain each lyrics individual character The standard original time information answered；

S204, according to corresponding individual character initial time stamp and the recording individual character of recording of each recording individual character difference End up timestamp, and each lyrics individual character distinguishes corresponding lyrics individual character initial time stamp and lyrics list Word ending timestamp, obtains the time match knot between each recording individual character and each lyrics individual character Really；

Specifically, the audio-frequency data processing device is further corresponding according to each recording individual character difference Individual character initial time stamp of recording and recording individual character ending timestamp, and each lyrics individual character are corresponded to respectively Lyrics individual character initial time stamp and the lyrics individual character ending timestamp, obtain it is described it is each recording individual character with it is described Time match result between each lyrics individual character.Wherein, by by the corresponding record of each recording individual character Sound individual character initial time stamp and recording individual character ending timestamp lyrics individual character corresponding with each lyrics individual character Initial time stamp and lyrics individual character ending timestamp are contrasted, and can generate each recording individual character and institute State the time match result between each lyrics individual character, and by the time match result with recording individual character Recording individual character initial time stamp and the recording individual character ending immediate lyrics individual character of timestamp are defined as and the recording The successful lyrics individual character of individual character time match.

S205, according to the time match result and each recording individual character and each lyrics individual character it Between recognition sequence matching result, be that each individual character voice data determines corresponding lyrics individual character respectively；

Specifically, each recording individual character that the audio-frequency data processing device is identified by sequencing and elder generation The each lyrics individual character sequentially identified afterwards is contrasted, and can generate each recording individual character and described every Recognition sequence matching result between individual lyrics individual character, and by the content with the identical cis-position of recording individual character and identification Identical lyrics individual character be defined as with recording individual character order the lyrics individual character that the match is successful, if for example, recording Individual character " my god " be second word identified, and be similarly in the lyrics individual character of second position also for " my god ", then Can by lyrics individual character " my god " be defined as and recording individual character order the lyrics individual character that the match is successful.Therefore, when true Make with certain recording the successful lyrics individual character of individual character time match, and the lyrics individual character also be and the recording individual character During the order lyrics individual character that the match is successful, the lyrics individual character can be defined as to song corresponding with the recording individual character Word individual character, i.e. the individual character voice data comprising the recording individual character are corresponding with the lyrics individual character, therefore, according to institute State time match result and the recognition sequence matching result can be true for each individual character voice data difference Fixed corresponding lyrics individual character.

If obtaining 7 individual character voice datas for example, splitting, and identify respectively in individual character voice data A Pronounce for the pronunciation in " full " word, individual character voice data B for " my god " word, the pronunciation in individual character voice data C It is that pronunciation in "Yes" word, individual character voice data E is " small " for the pronunciation in " all " words, individual character voice data D Pronunciation in word, individual character voice data F is that the pronunciation in " star " word, individual character voice data G is " star " word, And it is relative with the actual audio temporal information of 7 individual character voice datas in the QRC lyrics files Identify 7 lyrics individual characters at the standard audio temporal information answered, respectively " full ", " my god ", " all ", "Yes", One in " small ", " star ", " star ", the then each individual character voice data that can be identified according to order, really The corresponding lyrics individual characters of order word voice data A are " full ", the corresponding lyrics individual characters of individual character voice data B are " my god ", the corresponding lyrics individual characters of individual character voice data C for " all ", the corresponding lyrics lists of individual character voice data D Word be "Yes", the corresponding lyrics individual characters of individual character voice data E be " small ", the corresponding songs of individual character voice data F Word individual character is that " star ", the corresponding lyrics individual characters of individual character voice data G are " star ", you can described every to determine Individual individual character voice data distinguishes corresponding lyrics individual character.

S206, according to the recording individual character in the corresponding actual singing time information of the individual character voice data Initial time stamp and recording individual character ending timestamp, and during the corresponding standard original singer of the lyrics individual character Between the lyrics individual character initial time stamp in information and lyrics individual character ending timestamp, calculate the list Duration and non-coincidence duration are overlapped between word voice data and the lyrics individual character；

S207, calculates the non-coincidence duration and the ratio overlapped between duration, to obtain the individual character The corresponding time registration of voice data, and it regard the time registration as the parameter comparison information；

For example, the corresponding actual singing time information of recording individual character " love " in individual character voice data is (01:02.37,01:03.02) (i.e. 1 point 2.37 seconds is the recording individual character initial time stamp, and 1 point 3.02 seconds is The recording individual character ending timestamp), the standard original singer of the corresponding lyrics individual character " love " of the individual character voice data Temporal information is (01:02.39,01:03.21) (i.e. 1 point 2.39 seconds is lyrics individual character initial time stamp, 1 point It is within 3.21 seconds lyrics individual character ending timestamp)；Wherein, (2.39-2.37) a length of when non-coincidence+(3.21-3.02) It is=0.21 second, a length of 3.02-2.39=0.63 seconds during coincidence, therefore, the individual character voice data corresponding time Registration can be Z=[(2.39-2.37)+(3.21-3.02)]/(3.02-2.39).Wherein, Z symbol mark The attribute of bright " fast " or " partially slow ", when Z is positive number, represents that user sings slowly, when Z is negative, Represent that user sings soon；Z order of magnitude indicates fast or partially slow degree, and numerical value is bigger, and deviation is bigger.

S208, generates the corresponding type of error of the individual character voice data according to the parameter comparison information and identifies, And the type of error mark is shown in the range areas for showing the lyrics individual character；

Specifically, the audio-frequency data processing device can be according to the time in the parameter comparison information Registration generates the corresponding singing mass fraction of the individual character voice data, when the singing mass fraction is less than Default score threshold, and the time registration be negative when, the individual character voice data pair can be generated The type of error answered is designated " singing soon ", and shows the mistake in the range areas for showing the lyrics individual character Miss the icon corresponding to type identification " singing soon "；As can be seen here, the embodiment of the present invention can be as accurate as each having The lyrics individual character of problem all shows corresponding type of error mark, is built so as to really be given to user and reasonably improve View.

Fig. 3 is referred to, is the flow signal of another audio data processing method provided in an embodiment of the present invention Figure, methods described can include：

S301, obtains recorded audio data, and recognize each recording individual character in the recorded audio data；

S302, corresponding actual singing time information is distinguished by the recording sound according to each recording individual character Frequency evidence is split as multiple individual character voice datas；

S303, it is right respectively in lyrics file corresponding with the recorded audio data to obtain each lyrics individual character The standard original time information answered；

S304, according to corresponding individual character initial time stamp and the recording individual character of recording of each recording individual character difference End up timestamp, and each lyrics individual character distinguishes corresponding lyrics individual character initial time stamp and lyrics list Word ending timestamp, obtains the time match knot between each recording individual character and each lyrics individual character Really；

S305, according to the time match result and each recording individual character and each lyrics individual character it Between recognition sequence matching result, be that each individual character voice data determines corresponding lyrics individual character respectively；

Wherein, S301-S305 steps may refer to the S201-S205 in above-mentioned Fig. 2 correspondences embodiment, here No longer repeated.

S306, the corresponding actual performance pitch value of the individual character voice data is corresponding with the lyrics individual character Standard original singer's pitch value is subtracted each other, to obtain the corresponding tone deviation of the individual character voice data, and will be described Tone deviation is used as the parameter comparison information；

Specifically, the audio-frequency data processing device can be by the corresponding actual performance of the individual character voice data Pitch value standard original singer pitch value corresponding with the lyrics individual character is subtracted each other, to obtain the individual character voice data Corresponding tone deviation, and it regard the tone deviation as the parameter comparison information；Wherein, it is described The corresponding standard original singer pitch value of lyrics individual character is according to described in MIDI MIDI files The corresponding standard original time acquisition of information of lyrics individual character is arrived, if for example, the corresponding mark of the lyrics individual character Quasi- original time information is (01:02.39,01:03.21) (i.e. 1 point 2.39 seconds is lyrics individual character initial time stamp, 1 point 3.21 seconds is lyrics individual character ending timestamp), then it can obtain 1 point 2.39 in the MIDI files Second to standard original singer's pitch value in 1 point of period of 3.21 seconds, and by the standard original singer tone got Value is used as the corresponding standard original singer pitch value of the lyrics individual character.In another example, if actual pitch value of singing is Y, Standard original singer pitch value is y, then can calculate the corresponding tone deviation of the individual character voice data K=Y-y；K symbol indicates the attribute of " higher " or " relatively low ", when K is positive number, represents that user sings height； When K is negative, represent that user sings low；K order of magnitude indicates higher and relatively low degree, Numerical value is bigger, and deviation is bigger.

S307, generates the corresponding type of error of the individual character voice data according to the parameter comparison information and identifies, And the type of error mark is shown in the range areas for showing the lyrics individual character.

Specifically, the audio-frequency data processing device can be according to the tone in the parameter comparison information Deviation generates the corresponding singing mass fraction of the individual character voice data, when the singing mass fraction is less than Default score threshold, and the tone deviation be negative when, the individual character voice data pair can be generated The type of error answered is designated " singing low ", and shows the mistake in the range areas for showing the lyrics individual character Miss the icon corresponding to type identification " singing low "；As can be seen here, the embodiment of the present invention can be as accurate as each having The lyrics individual character of problem all shows corresponding type of error mark, is built so as to really be given to user and reasonably improve View.

Optionally, the audio-frequency data processing device can also be counted in the lump when calculating the tone deviation Calculate such as the time registration in the S207 in above-mentioned Fig. 2 correspondence embodiments, you can with by the tone deviation With the time registration as the parameter comparison information, now, the audio-frequency data processing device can The individual character is generated with the time registration in the parameter comparison information and the tone deviation The corresponding singing mass fraction of voice data, when the singing mass fraction is less than default score threshold, Can further detect time registration account for low point of principal element or tone deviation account for low point it is main because Element, if the time registration accounts for low point of principal element, and it is to sing soon that the time registration, which shows, The corresponding type of error of the individual character voice data can then be generated to be designated " singing soon ", and showing the song The icon corresponding to the type of error mark " singing soon " is shown in the range areas of word individual character；If the sound Tuningout difference accounts for low point of principal element, and it is to sing height that the tone deviation, which shows, then can generate institute The corresponding type of error of individual character voice data is stated to be designated " singing height ", and in the scope for showing the lyrics individual character The icon corresponding to the type of error mark " singing height " is shown in region.Optionally, if the time overlaps Degree and tone deviation all deviation ratios are larger, then can also show the range areas of the lyrics individual character It is interior while showing that the corresponding type of error of the time registration identifies mistake corresponding with the tone deviation Type identification.

Fig. 4 is referred to, is a kind of structural representation of audio-frequency data processing device provided in an embodiment of the present invention, The audio-frequency data processing device 1 can include：Obtain identification module 10, split module 20, individual character acquisition Module 30, compare determining module 40, generation display module 50；

The acquisition identification module 10, for obtaining recorded audio data, and recognizes the recorded audio data In each recording individual character；

Specifically, the identification module 10 that obtains can obtain the recorded audio data of user by microphone, And each recording individual character in the recorded audio data is identified based on speech recognition algorithm, the acquisition is known Other module 10 can also terminate according to the pronunciation initial time and pronunciation of each recording individual character identified Moment records each recording individual character and distinguishes corresponding actual singing time information, the actual singing time Information includes recording individual character initial time stamp and recording individual character ending timestamp, the recording individual character initial time Stamp is the pronunciation initial time of the recording individual character, and the recording individual character ending timestamp is the recording individual character Pronunciation finish time.For example, the recorded audio data that the acquisition identification module 10 is got are sung for user A certain first song in a word " being all starlet over the sky ", then the acquisition identification module 10 can identify Each recording individual character respectively " full ", " my god ", " ", "Yes", " small ", " star ", " star ", it is described to obtain knowledge Other module 10 can also be further according to the pronunciation initial time and hair of each recording individual character identified Sound finish time records each recording individual character and distinguishes corresponding actual singing time information, if identifying The pronunciation initial time of " full " word be 1 point 2.37 seconds, pronunciation finish time be 1 point 3.11 seconds, i.e. " full " word Complete tone period be from 1 point 3.11 seconds 2.37 seconds to 1 point, then can be used as the record within 2.37 seconds using 1 point Recording individual character initial time stamp in the corresponding actual singing time information of sound individual character, and 1 point is made for 3.11 seconds For the recording individual character ending timestamp in the corresponding actual singing time information of the recording individual character.

The fractionation module 20, for being split as the recorded audio data according to each recording individual character Multiple individual character voice datas；

Specifically, the acquisition identification module 10 identifies each recording individual character, and get described every Individual recording individual character is distinguished after corresponding actual singing time information, and the fractionation module 20 specifically can be according to institute State the recording individual character initial time stamp in the corresponding actual singing time information of each recording individual character difference and recording The recorded audio data are split as multiple individual character voice datas by individual character ending timestamp；Each list Word voice data corresponds to a recording individual character, i.e., each individual character voice data is the corresponding record The voice data that sound individual character initial time stamp ends up in timestamp to the corresponding recording individual character.If for example, It is described acquisition identification module 10 identify it is each recording individual character respectively " full ", " my god ", " ", "Yes", " small ", " star ", " star ", and get the corresponding actual singing time information of each recording individual character difference, then it is described Split module 20 and corresponding actual singing time information is further distinguished by described according to each recording individual character Recorded audio data are split as multiple individual character voice datas, if the complete tone period of " full " word is from 1 point During 2.37 seconds to 1 point 3.11 seconds, i.e., 1 point 2.37 seconds are the corresponding actual singing time information of the recording individual character Recording individual character initial time stamp, 1 point 3.11 seconds be the corresponding actual singing time information of the recording individual character in Recording individual character ending timestamp, then it is described fractionation module 20 can actually be drilled according to the recording individual character is corresponding Temporal information is sung to tear 1 point of 2.37 seconds voice data to 1 point of 3.11 second period in the recorded audio data open Branch away, and regard the voice data splitted out as the corresponding individual character voice data of " full " word.

The individual character acquisition module 30, for being obtained in lyrics file corresponding with the recorded audio data Lyrics individual character corresponding with each individual character voice data；

Specifically, the individual character acquisition module 30 can be in lyrics file corresponding with the recorded audio data It is middle to obtain the corresponding standard original time information of each lyrics individual character difference；The standard original time packet Include lyrics individual character initial time stamp and lyrics individual character ending timestamp；The lyrics file can be the QRC lyrics Each lyrics individual character in file, the QRC lyrics files is corresponded to respectively time tag, the time Label includes the initial time of lyrics individual character and the duration of lyrics individual character；Wherein, each lyrics list Corresponding standard original time information is to distinguish corresponding time tag according to each lyrics individual character to word respectively Calculate what is obtained, for example, the corresponding time tags of lyrics individual character A include lyrics individual character A initial time t1 With lyrics individual character A duration t2, then when can calculate the starting of lyrics individual character A corresponding lyrics individual characters Between stamp be t1, lyrics individual character A corresponding lyrics individual characters ending timestamp is t1+t2.

It is single that the individual character acquisition module 30 further can distinguish corresponding recording according to each recording individual character Word initial time stamp and recording individual character ending timestamp, and each lyrics individual character distinguish the corresponding lyrics Individual character initial time stamp and lyrics individual character ending timestamp, obtain each recording individual character and each song Time match result between word individual character.Wherein, by by the corresponding recording individual character of each recording individual character When initial time stamp and corresponding with each lyrics individual character lyrics individual character starting of recording individual character ending timestamp Between stamp and the lyrics individual character ending timestamp contrasted, the individual character acquisition module 30 can generate it is described each The time match result recorded between individual character and each lyrics individual character, and by the time match result Recording individual character initial time stamp and the recording individual character ending immediate lyrics individual character of timestamp with individual character of recording is true It is set to and the successful lyrics individual character of recording individual character time match.When determine and certain recording individual character time match During successful lyrics individual character, the lyrics individual character can be defined as and the recording list by the individual character acquisition module 30 The corresponding lyrics individual character of word, i.e. the individual character voice data comprising the recording individual character are corresponding with the lyrics individual character, because This, the corresponding lyrics can be determined respectively according to the time match result for each individual character voice data Individual character.

It is described to compare determining module 40, for by the actual performance parameter of the individual character voice data and the song The corresponding standard original singer parameter of word individual character is compared, and determines the individual character voice data and the lyrics individual character Parameter comparison information between corresponding standard original singer parameter；The parameter comparison information includes time registration And/or tone deviation；

Specifically, when the actual parameter of singing includes the corresponding actual singing time of the individual character voice data Information, and the standard original singer parameter is when including the corresponding standard original time information of the lyrics individual character, please The structural representation for comparing determining module 40 in the lump referring to Fig. 5, it is described compare determining module 40 can With including：Duration calculation unit 401, calculating determining unit 402；

The duration calculation unit 401, for being believed according to the corresponding actual singing time of the individual character voice data The recording individual character initial time stamp and recording individual character ending timestamp in breath, and the lyrics list The lyrics individual character initial time stamp and lyrics individual character ending in the corresponding standard original time information of word Timestamp, calculates and overlaps duration and non-coincidence duration between the individual character voice data and the lyrics individual character；

The calculating determining unit 402, for calculating the non-coincidence duration and the ratio overlapped between duration Value, to obtain the corresponding time registration of the individual character voice data, and regard the time registration as institute State parameter comparison information.

For example, the corresponding actual singing time information of recording individual character " love " in individual character voice data is (01:02.37,01:03.02) (i.e. 1 point 2.37 seconds is the recording individual character initial time stamp, and 1 point 3.02 seconds is The recording individual character ending timestamp), the standard original singer of the corresponding lyrics individual character " love " of the individual character voice data Temporal information is (01:02.39,01:03.21) (i.e. 1 point 2.39 seconds is lyrics individual character initial time stamp, 1 point It is within 3.21 seconds lyrics individual character ending timestamp)；Wherein, what the duration calculation unit 401 was calculated is non-coincidence Shi Changwei (2.39-2.37)+(3.21-3.02)=0.21 second, it is a length of 3.02-2.39=0.63 seconds during coincidence, because This, the calculating determining unit 402 can calculate the corresponding time registration of the individual character voice data can Think Z=[(2.39-2.37)+(3.21-3.02)]/(3.02-2.39).Wherein, Z symbol is indicated " fast " Or the attribute of " partially slow ", when Z is positive number, represents that user sings slowly, when Z is negative, represent user Sing soon；Z order of magnitude indicates fast or partially slow degree, and numerical value is bigger, and deviation is bigger.

Or, pitch value is sung when the actual performance parameter includes the corresponding reality of the individual character voice data, And the standard original singer parameter is when including the corresponding standard original singer pitch value of the lyrics individual character；The comparison is true Cover half block 40 by the corresponding reality of the individual character voice data specifically for singing pitch value and the lyrics individual character Corresponding standard original singer pitch value is subtracted each other, to obtain the corresponding tone deviation of the individual character voice data, and It regard the tone deviation as the parameter comparison information；Wherein, the corresponding standard of the lyrics individual character is former It is former according to the corresponding standard of the lyrics individual character in MIDI MIDI files to sing pitch value Sing what temporal information was got, if for example, the corresponding standard original time information of the lyrics individual character is (01:02.39,01:03.21) (i.e. 1 point 2.39 seconds is lyrics individual character initial time stamp, and 1 point 3.21 seconds is song Word individual character end up timestamp), then can obtain in the MIDI files 1 point 3.21 seconds 2.39 seconds to 1 point Period in standard original singer's pitch value, and regard the standard original singer pitch value got as the lyrics The corresponding standard original singer pitch value of individual character.In another example, if actual pitch value of singing is Y, standard original singer's tone It is worth for y, then can calculates the corresponding tone deviation K=Y-y of the individual character voice data；K symbol mark The attribute of bright " higher " or " relatively low ", when K is positive number, represents that user sings height；When K is negative, Represent that user sings low；K order of magnitude indicates higher and relatively low degree, and numerical value is bigger, deviation It is bigger.

Or, the actual parameter of singing includes the corresponding actual singing time letter of the individual character voice data Breath and actual performance pitch value；When the standard original singer parameter includes the lyrics individual character corresponding standard original singer Between information and standard original singer's pitch value；Therefore, it is described when comparing determining module 40 and can calculate described in the lump Between registration and the tone deviation, and regard the time registration and the tone deviation as institute State parameter comparison information.

The generation display module 50, for generating the individual character voice data according to the parameter comparison information Corresponding type of error mark, and show the type of error in the range areas for showing the lyrics individual character Mark；

Specifically, the generation display module 50 can be according to the weight of the time in the parameter comparison information Right and/or described tone deviation generates the corresponding singing mass fraction of the individual character voice data, when described Singing mass fraction is less than default score threshold, and the parameter comparison information includes the time registration During with tone deviation, it can further detect that time registration accounts for low point of principal element or tone deviation Value accounts for low point of principal element, if the time registration accounts for low point of principal element, and the time registration What is shown is to sing soon, then can generate the corresponding type of error of the individual character voice data and be designated and " sing fast ", and show that type of error mark " singing soon " institute is right in the range areas for showing the lyrics individual character The icon answered；If the tone deviation accounts for low point of principal element, and the tone deviation show be Height is sung, then can generate the corresponding type of error of the individual character voice data and be designated " singing height ", and aobvious Show and show that the type of error identifies the icon corresponding to " singing height " in the range areas of the lyrics individual character.Can Choosing, can also be in display institute if the time registration and tone deviation all deviation ratios are larger State in the range areas of lyrics individual character while showing the corresponding type of error mark of the time registration and described The corresponding type of error mark of tone deviation.Or, when the singing mass fraction is less than default point Number threshold values, and the parameter comparison information include the time registration or tone deviation when, can be direct Corresponding type of error mark is generated according to the parameter comparison information.As can be seen here, the embodiment of the present invention can Corresponding type of error mark is shown to be accurate to each problematic lyrics individual character, so as to really be given to use The rational recommendation on improvement in family.

Further, then Fig. 6 is referred to, is the structural representation of individual character acquisition module 30, the individual character is obtained Modulus block 30 can include：Temporal information acquiring unit 301, matching result acquiring unit 302, individual character are true Order member 303；

The temporal information acquiring unit 301, in lyrics file corresponding with the recorded audio data Obtain each lyrics individual character and distinguish corresponding standard original time information；The standard original time information includes Lyrics individual character initial time stamp and lyrics individual character ending timestamp；

Specifically, the temporal information acquiring unit 301 can be in song corresponding with the recorded audio data Each lyrics individual character is obtained in word file and distinguishes corresponding standard original time information；The standard original time Information includes lyrics individual character initial time stamp and lyrics individual character ending timestamp；The lyrics file can be Each lyrics individual character in QRC lyrics files, the QRC lyrics files is corresponded to respectively time tag, The time tag includes the initial time of lyrics individual character and the duration of lyrics individual character；Wherein, it is described every Corresponding standard original time information is corresponding according to each lyrics individual character difference to individual lyrics individual character respectively Time tag calculates what is obtained, for example, the corresponding time tags of lyrics individual character A include rising for lyrics individual character A Time beginning t1 and the lyrics individual character A duration t2, then can calculate the corresponding lyrics lists of lyrics individual character A Word initial time stamp is t1, and the corresponding lyrics individual character ending timestamps of lyrics individual character A are t1+t2.

The matching result acquiring unit 302, it is single for distinguishing corresponding recording according to each recording individual character Word initial time stamp and recording individual character ending timestamp, and each lyrics individual character distinguish the corresponding lyrics Individual character initial time stamp and lyrics individual character ending timestamp, obtain each recording individual character and each song Time match result between word individual character；

Specifically, the matching result acquiring unit 302 is further right respectively according to each recording individual character The recording individual character initial time stamp and recording individual character ending timestamp answered, and each lyrics individual character difference Corresponding lyrics individual character initial time stamp and the lyrics individual character ending timestamp, obtain it is described it is each recording individual character with Time match result between each lyrics individual character.Wherein, by the way that each recording individual character is corresponding Recording individual character initial time stamp and the recording individual character ending timestamp lyrics corresponding with each lyrics individual character Individual character initial time stamp and lyrics individual character ending timestamp are contrasted, the matching result acquiring unit 302 The time match result between each recording individual character and each lyrics individual character can be generated, and by institute State recording individual character initial time stamp and recording individual character in time match result with individual character of recording and end up timestamp most Close lyrics individual character is defined as and the successful lyrics individual character of recording individual character time match.

The individual character determining unit 303, for according to the time match result and it is described it is each recording individual character with Recognition sequence matching result between each lyrics individual character, is that each individual character voice data difference is true Fixed corresponding lyrics individual character；

Specifically, each recording individual character that the individual character determining unit 303 is identified by sequencing and elder generation The each lyrics individual character sequentially identified afterwards is contrasted, and can generate each recording individual character and described every Recognition sequence matching result between individual lyrics individual character, and by the content with the identical cis-position of recording individual character and identification Identical lyrics individual character be defined as with recording individual character order the lyrics individual character that the match is successful, if for example, recording Individual character " my god " be second word identified, and be similarly in the lyrics individual character of second position also for " my god ", then The individual character determining unit 303 can by lyrics individual character " my god " be defined as with the recording individual character order the match is successful Lyrics individual character.Therefore, when determining and certain the recording successful lyrics individual character of individual character time match, and the song When word individual character is also with the recording individual character order lyrics individual character that the match is successful, the individual character determining unit 303 The lyrics individual character can be defined as to lyrics individual character corresponding with the recording individual character, i.e., comprising the recording individual character Individual character voice data is corresponding with the lyrics individual character, therefore, suitable according to the time match result and the identification Sequence matching result can determine corresponding lyrics individual character respectively for each individual character voice data.

Fig. 7 is referred to, is the structural representation of another audio-frequency data processing device provided in an embodiment of the present invention Figure.As shown in fig. 7, the audio-frequency data processing device 1000 can include：At least one processor 1001, Such as CPU, at least one network interface 1004, user interface 1003, memory 1005, at least one Communication bus 1002.Wherein, communication bus 1002 is used to realize the connection communication between these components.Wherein, User interface 1003 can include display screen (Display), keyboard (Keyboard), optional user interface 1003 Wireline interface, the wave point of standard can also be included.Network interface 1004 can optionally include standard Wireline interface, wave point (such as WI-FI interfaces).Memory 1005 can be high-speed RAM memory, Can also be non-labile memory (non-volatile memory), for example, at least one magnetic disk storage. Memory 1005 optionally can also be at least one storage device for being located remotely from aforementioned processor 1001. As shown in fig. 7, as operating system, net can be included in a kind of memory 1005 of computer-readable storage medium Network communication module, Subscriber Interface Module SIM and equipment control application program.

In the audio-frequency data processing device 1000 shown in Fig. 7, user interface 1003 is mainly used in as user The interface of recorded audio input is provided, the recorded audio data of user's output are obtained；And processor 1001 can be with The equipment control application program stored for calling in memory 1005, and specifically perform following steps：

In one embodiment, the processor 1001 is being performed the record according to each recording individual character When sound voice data is split as multiple individual character voice datas, following steps are specifically performed：

Corresponding actual singing time information is distinguished by the recorded audio data according to each recording individual character It is split as multiple individual character voice datas；The actual singing time information include recording individual character initial time stamp and Recording individual character ending timestamp；Each individual character voice data corresponds to a recording individual character.

In one embodiment, the processor 1001 is being performed in song corresponding with the recorded audio data When lyrics individual character corresponding with each individual character voice data is obtained in word file, following steps are specifically performed：

Each lyrics individual character is obtained in lyrics file corresponding with the recorded audio data and distinguishes corresponding mark Quasi- original time information；The standard original time information includes lyrics individual character initial time stamp and lyrics individual character End up timestamp；

When distinguishing corresponding recording individual character initial time stamp and recording individual character ending according to each recording individual character Between stab, and each lyrics individual character distinguishes corresponding lyrics individual character initial time stamp and the ending of lyrics individual character Timestamp, obtains the time match result between each recording individual character and each lyrics individual character；

According to the knowledge between the time match result and each recording individual character and each lyrics individual character Other order matching result, is that each individual character voice data determines corresponding lyrics individual character respectively.

In one embodiment, the actual parameter of singing includes that the individual character voice data is corresponding actually to drill Sing temporal information；The standard original singer parameter includes the corresponding standard original time information of the lyrics individual character；

The processor 1001 is being performed the actual performance parameter and the lyrics list of the individual character voice data The corresponding standard original singer parameter of word is compared, and determines that the individual character voice data is corresponding with the lyrics individual character Standard original singer's parameter between parameter comparison information when, specifically perform following steps：

When the recording individual character in the corresponding actual singing time information of the individual character voice data is originated Between stamp and the recording individual character ending timestamp, and the corresponding standard original time information of the lyrics individual character In the lyrics individual character initial time stamp and lyrics individual character ending timestamp, calculate the individual character audio Duration and non-coincidence duration are overlapped between data and the lyrics individual character；

The non-coincidence duration and the ratio overlapped between duration are calculated, to obtain the individual character audio number The parameter comparison information is used as according to corresponding time registration, and using the time registration.

In one embodiment, the actual parameter of singing includes that the individual character voice data is corresponding actually to drill Sing pitch value；The standard original singer parameter includes the corresponding standard original singer pitch value of the lyrics individual character；

The corresponding reality of the individual character voice data is sung into pitch value standard corresponding with the lyrics individual character former Pitch value is sung to subtract each other, to obtain the corresponding tone deviation of the individual character voice data, and the tone is inclined Difference is used as the parameter comparison information；

Wherein, the corresponding standard original singer pitch value of the lyrics individual character is in MIDI MIDI Arrived in file according to the corresponding standard original time acquisition of information of the lyrics individual character.

One of ordinary skill in the art will appreciate that all or part of flow in above-described embodiment method is realized, It can be by computer program to instruct the hardware of correlation to complete, described program can be stored in a calculating In machine read/write memory medium, the program is upon execution, it may include such as the flow of the embodiment of above-mentioned each method. Wherein, described storage medium can for magnetic disc, CD, read-only memory (Read-Only Memory, ) or random access memory (Random Access Memory, RAM) etc. ROM.

Above disclosure is only preferred embodiment of present invention, can not limit the present invention's with this certainly Interest field, therefore the equivalent variations made according to the claims in the present invention, still belong to the scope that the present invention is covered.

Claims

1. a kind of audio data processing method, it is characterised in that including：

2. the method as described in claim 1, it is characterised in that described to be incited somebody to action according to each recording individual character The recorded audio data are split as multiple individual character voice datas, specifically include：

3. method as claimed in claim 2, it is characterised in that it is described with the recorded audio data pair Lyrics individual character corresponding with each individual character voice data is obtained in the lyrics file answered, including：

4. method as claimed in claim 3, it is characterised in that the actual parameter of singing includes the list The corresponding actual singing time information of word voice data；The standard original singer parameter includes the lyrics individual character pair The standard original time information answered；

It is described by the actual performance parameter of the individual character voice data and the corresponding standard original singer of the lyrics individual character Parameter is compared, determine individual character voice data standard original singer parameter corresponding with the lyrics individual character it Between parameter comparison information, including：

5. method as claimed in claim 3, it is characterised in that the actual parameter of singing includes the list The corresponding actual performance pitch value of word voice data；The standard original singer parameter includes lyrics individual character correspondence Standard original singer's pitch value；

It is described by the actual performance parameter of the individual character voice data and the corresponding standard original singer of the lyrics individual character Parameter is compared, determine individual character voice data standard original singer parameter corresponding with the lyrics individual character it Between parameter comparison information, specifically include：

6. a kind of audio-frequency data processing device, it is characterised in that including：

7. device as claimed in claim 6, it is characterised in that

The fractionation module, specifically for distinguishing corresponding actual singing time according to each recording individual character The recorded audio data are split as multiple individual character voice datas by information；The actual singing time packet Include recording individual character initial time stamp and recording individual character ending timestamp；Each individual character voice data is corresponded to One recording individual character.

8. device as claimed in claim 7, it is characterised in that the individual character acquisition module includes：

Temporal information acquiring unit, for obtaining every in lyrics file corresponding with the recorded audio data Individual lyrics individual character distinguishes corresponding standard original time information；The standard original time information includes lyrics list Word initial time stamp and lyrics individual character ending timestamp；

Matching result acquiring unit, for according to the corresponding individual character starting of recording of each recording individual character difference Timestamp and recording individual character ending timestamp, and each lyrics individual character are distinguished corresponding lyrics individual character and risen Beginning timestamp and lyrics individual character ending timestamp, obtain each recording individual character and each lyrics individual character Between time match result；

Individual character determining unit, for according to the time match result and each recording individual character with it is described every Recognition sequence matching result between individual lyrics individual character, is that each individual character voice data determines correspondence respectively Lyrics individual character.

9. device as claimed in claim 8, it is characterised in that the actual parameter of singing includes the list The corresponding actual singing time information of word voice data；The standard original singer parameter includes the lyrics individual character pair The standard original time information answered；

It is described to compare determining module, including：

Duration calculation unit, for according in the corresponding actual singing time information of the individual character voice data The recording individual character initial time stamp and recording individual character ending timestamp, and lyrics individual character correspondence Standard original time information in the lyrics individual character initial time stamp and the lyrics individual character ending time Stamp, calculates and overlaps duration and non-coincidence duration between the individual character voice data and the lyrics individual character；

Determining unit is calculated, for calculating the non-coincidence duration and the ratio overlapped between duration, with The corresponding time registration of the individual character voice data is obtained, and regard the time registration as the parameter Comparative information.

10. device as claimed in claim 8, it is characterised in that the actual parameter of singing is including described The corresponding actual performance pitch value of individual character voice data；The standard original singer parameter includes the lyrics individual character pair The standard original singer's pitch value answered；

It is described to compare determining module, specifically for the corresponding reality of the individual character voice data is sung into pitch value Standard original singer pitch value corresponding with the lyrics individual character is subtracted each other, corresponding to obtain the individual character voice data Tone deviation, and it regard the tone deviation as the parameter comparison information；