CN101587708A - Song emotion pressure analysis method and system - Google Patents
Song emotion pressure analysis method and system Download PDFInfo
- Publication number
- CN101587708A CN101587708A CNA200910087827XA CN200910087827A CN101587708A CN 101587708 A CN101587708 A CN 101587708A CN A200910087827X A CNA200910087827X A CN A200910087827XA CN 200910087827 A CN200910087827 A CN 200910087827A CN 101587708 A CN101587708 A CN 101587708A
- Authority
- CN
- China
- Prior art keywords
- lyrics
- song
- mood
- unit
- emotion
- 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
Images
Landscapes
- Information Retrieval, Db Structures And Fs Structures Therefor (AREA)
Abstract
The invention discloses a song emotion pressure analysis method comprising steps: S1, cutting the lyrics; S2, judging the structure of lyrics by analyzing the repetition, parataxis of the lyrics and the lyrics fragment sequence; S3, determining the average singing speed of the song according to the words of the lyrics and the singing time; S4, abstracting the emotion unit from the lyrics and judging the emotion pressure polar of the emotion unit; S5, calculating the lyrics emotion pressure index according to the structure of the lyrics, the average singing speed of the song and the emotion pressure polar of the emotion unit. Based on the method, the user can exactly position the required song, thereby improving the song searching hit rate and shortening the song searching time.
Description
Technical field
The present invention relates to the song analytical technology, particularly relating to a kind of is the song emotion pressure analysis method and the system of foundation with the lyrics.
Background technology
Current social grows with each passing day to the demand of song, listens song to turn to the online audiovisual in internet from PC.Popularizing gradually of 3G communication network must promote the song operation and expand to mobile phone from the internet.Be reply the demand, various intelligent song searches and commending system emerge in large numbers gradually.The song classification is the gordian technique of intelligent song search and recommendation, and target is to give song specific class label, to make things convenient for user search or system recommendation.Objective classification and intension classification are arranged in the song classification, and for example the region classification is the objective classification that sets according to the song creation place, and the mood classification is the intension classification that sets according to song emotion pressure.
It is bigger to realize that automatic song emotion is analyzed difficulty.In recent years, the song emotion classification at first occurs in Audio Signal Processing research, and people attempt to extract the audio frequency characteristics that may reflect mood by statistical means from sound signal, relends to help machine learning algorithm to realize the mood classification.This class research has nearly 20 years history, but can't obtain accurately to reflect the audio frequency characteristics of mood so far, and obtained effect is very limited, can't reach realistic scale.Analytically limitation is obvious at song emotion owing to sound signal, has occurred recently carrying out the research interest that song emotion is analyzed by the lyrics.
Summary of the invention
The purpose of this invention is to provide a kind of method and system of the emotional stress state of song being analyzed according to the lyrics.
For achieving the above object, a kind of song emotion pressure analysis method according to embodiment of the present invention is provided, said method comprising the steps of:
S1 carries out cutting to the lyrics;
S2 by analyzing repetition, parallelism and the lyrics fragment sequence of the lyrics, judges the structure of the lyrics;
S3, according to the number of words and the singing time of the lyrics, that determines song on average sings speed;
S4 extracts the mood unit from the lyrics, and judges the emotional stress polarity of mood unit;
S5 calculates lyrics emotional stress index according to the structure of the described lyrics, the emotional stress polarity of on average singing speed and mood unit of song.
Preferably, in step S1, the lyrics being carried out cutting comprise being statement with lyrics cutting earlier, is vocabulary with sentence segmentation again.
Preferably, in step S1, adopt during for vocabulary two-way maximum match strategy to judge the vocabulary border sentence segmentation.
Preferably, in step S2, analyze the repetition of the lyrics by analyzing lyrics file temporal information and character match.
Preferably, in step S2, mate the parallelism of analyzing the lyrics by analyzing lyrics fragments character.
Preferably, in step S2, analyze lyrics fragment sequence by analyzing the lyrics file time series.
Preferably, among the described step S4, from lyrics fragment, search the mood speech, in context, search qualifier and negative word afterwards, obtain the mood unit by utilizing mood speech dictionary.
Preferably, the emotional stress polarity judging step of mood unit comprises among the described step S4:
S4-1, the emotional stress polarity of extraction mood speech from mood speech dictionary;
S4-2 judges whether there is negative word in the mood unit, if then polarity is put anti-and forwarded step S4-3 to, if not, then directly forwards step S4-3 to;
S4-3 judges whether to exist qualifier, if, described emotional stress polarity is strengthened or slacken according to the degree of qualifier, if not, then finish.
Preferably, described emotional stress polarity comprises positive and passive.
Preferably, utilize following formula to calculate the emotional stress index of song among the described step S5:
Wherein, s is the emotional stress index, and its span is [0,16], wherein minimum value 0 show song pressure state for the lightest, maximal value 16 shows that the pressure state of song is for the heaviest; l
iBe certain lyrics, N is the sum of song emotion unit, Polar (u
i) be mood unit u
iPolarity, value be 1 ,-1}, corresponding { actively, passiveness }; Func (u
i) be mood unit u
iStrength factor, value be 1 ,-1}, corresponding { strengthen, weaken }; Struct (u
i) be mood unit u
iConstruction coefficient; Speed (u
i) be mood unit u
iSing fast coefficient, this coefficient is actually mood unit u
iFast s is sung in the present position
iFunction g (s
i), g (s
i) calculate according to following formula:
s
iFor singing speed, unit is word/second, and its span is [s
-, s
+] (s
-Be that the song overall situation croons speed most, s
+Be that the song overall situation is sung loudly speed most), g (s
i) span be [1,1].
For achieving the above object, a kind of song emotion pressure analytic system according to embodiment of the present invention also is provided, described system comprises:
Pretreatment module is used for the cutting lyrics;
The structure analysis module is connected with pretreatment module, is used for according to the repetition of the lyrics and the structure of the parallelism characteristic and the lyrics fragment sequence analysis lyrics;
Sing fast analysis module, be connected, be used for judging the performance speed of song according to lyrics number of words with pretreatment module;
Mood unit extraction module is connected with pretreatment module, is used for extracting the mood unit of the lyrics, and judges the pressure polarities of mood unit according to lyrics context;
The emotional stress computing module with the structure analysis module, sing fast analysis module and be connected with mood unit extraction module, is used for calculating the song emotion pressure index according to lyric structure, the distribution situation of singing speed, mood cell pressure polarity and mood unit.
Based on song emotion pressure analysis method provided by the present invention and system, can more effectively carry out the analysis of song intension, especially judge the emotional stress of song.As be applied to the situation of user search song or system recommendation song, then can make the user in the process of song search, keyword (for example singer's name) and song intension (for example emotional stress) classification according to song, can accurately locate needed song, comprise the user and do not know the new song of title, thereby improve the song search hit rate, shorten user's song searching time, constantly promote user experience.
Description of drawings
Fig. 1 is the schematic flow sheet according to the song emotion pressure analysis method of embodiment of the present invention;
Fig. 2 is the structural representation according to the song emotion pressure analytic system of embodiment of the present invention.
Embodiment
Below in conjunction with drawings and Examples, the specific embodiment of the present invention is described in further detail.Following examples are used to illustrate the present invention, but are not used for limiting the scope of the invention.
Fig. 1 shows the schematic flow sheet of the song emotion pressure analysis method that one embodiment of the invention provides.
In step S1, the lyrics are carried out cutting, be vocabulary for example with lyrics cutting, preferably, can be statement with lyrics cutting earlier, and then be vocabulary sentence segmentation.The lyrics of song can be the LYRIC file layouts (* .LRC) that comprises temporal information and lyrics information.The cutting strategy of statement is comparatively simple, with the capable statement that directly is considered as of the lyrics of LRC file.When segmenting words, take two-way maximum match strategy to judge the vocabulary border, used dictionary is the standard modern Chinese dictionary.Two-way maximum match is a kind of universal machine segmenting method, it simultaneously from forward (by left-to-right) and reverse (by the right side to a left side) entry Chinese character string to be analyzed and the general Chinese dictionary is mated; If find certain character string in dictionary, then the match is successful (identifying a speech); If two-wayly exist inconsistently, then judge in conjunction with statistical information.For example, " we will strengthen the democratic consciousness " is " we// reinforcements/democracy/consciousness " by cutting, rather than " we// reinforcement/people/idea/knowledge ".This is because two-way maximum match strategy can effectively be handled the segmenting words ambiguity.
In step S2,, judge the structure of the lyrics by analyzing repetition, parallelism and the lyrics fragment sequence of the lyrics.Lyrics fragment sequence is the vocabulary combination of having expressed certain emotion, is the embodiment of lyric structure relation, is the precedence that each fragment of the lyrics occurred according to the time.Lyrics fragment generally is a statement, is littler language particle sometimes, for example the part of sentence and even single vocabulary.Analyze the repetition of the lyrics by analyzing lyrics file temporal information and character match in the present embodiment.LRC form lyrics file explicit representation repetition lyrics fragments, for example what in the end the lyrics " it is good not very bad what in the end [03:52.00] [00:49.00] just calculates " expression lyrics fragment " is just calculated good not very bad " having repeated 2 times in the piece of writing lyrics entirely, promptly can directly extract repeated fragment from the lyrics by the LRC form lyrics.This embodiment also mates the parallelism of analyzing the lyrics by analyzing lyrics fragments character.For example following two lyrics constitute parallelism: [04:08.00] [01:04.00] really at all times not should, and [04:12.00] [01:08.00] is really not lovely." really at all times not should " and " really not lovely " formation " really XX just not XX " parallelism structure wherein.The present invention can find this parallelism structure by string matching." standard " parallelism sentence that exists length to differ sometimes, for example following two lyrics constitute " standard " parallelism: it is good not very bad what in the end [03:52.00] [00:49.00] just calculates, and what in the end [03:56.00] [00:52.00] adapts to these epoch.Wherein " what in the end just calculate good not very bad " and " what in the end adapting to these epoch " formation " standard " parallelism sentence.This " standard " parallelism phenomenon also uses in the present invention, and its confidence level is a little less than parallelism.
Find that the lyrics repeat and the parallelism structure is in order to analyze the lyric structure feature.This analysis method is a background with the full piece of writing of the lyrics, determines song structure according to the precedence that repeats and parallelism occurs.Present embodiment has mainly been considered following lyric structure:
(1)A-B-A-B
(2) A1-A2-B-A1 (or A2)-B
(3) A1-A2-B-A1 (or A2)
(4)A1-B-A2-B-C-B
(5)A-B-A-B-D
(6)A-B-B-D
(7)A-B-C-B-D
(8)B-A-B-D
Wherein A represents theme lyrics fragment, and Ai is the theme fragment of numbering, and B represents the refrain fragment, and C represents transition sentence (parenthesis), and the D representative finishes sentence (popular sentence).
Theme fragment (A) generally is the prelude part of song, is expressed as purpose with content, is mainly used to place mat, and for drawing the climax part, emotion is more flat usually, even inconsistent with the expressed emotion of climax part.The refrain part is the climax of song normally, all forms sharp contrast with the theme fragment on rhythm, intensity and tune.Refrain fragment (B) generally is the distillation of emotion, is the crucial touch part of the full lyrics.Wherein the composition of Shu Qinging is in the majority, and generality is very strong.The part of refrain simultaneously multiplicity is intended to strengthen mood than higher.Therefore, refrain fragment can reflect song emotion substantially.Transition fragment (C) is often referred to two sections transition lyrics between the tight adjacent refrain, and its purpose is in transition, thereby emotion expression service is flat.Finish (D) one or two lyrics being remembered easily of song latter end of sentence (popular sentence), its emotion expression service also relatively fully, near refrain.
Like this, the present invention is after obtaining lyric structure, and the refrain fragment that mood is abundant is carried out the weight enhancement process, and mood is more flat even carry out weight with inconsistent theme fragment of refrain fragment and transition fragment and reduce or ignore processing.Table 1 is the weight during the residing diverse location in mood unit in the different structure of setting among the present invention.
Table 1
| Lyrics clip types | Lyric structure | Weight |
| A | A-B-A-B | 0.4 |
| B | A-B-A-B | 1.2 |
| A1 | A1-A2-B-A1(A2)-B | 0.4 |
| A2 | A1-A2-B-A1(A2)-B | 0.5 |
| B | A1-A2-B-A1(A2)-B | 1.2 |
| A1 | A1-A2-B-A1(A2) | 0.4 |
| A2 | A1-A2-B-A1(A2) | 0.5 |
| B | A1-A2-B-A1(A2) | 1.2 |
| A1 | A1-B-A2-B-C-B | 0.5 |
| A2 | A1-B-A2-B-C-B | 0.5 |
| B | A1-B-A2-B-C-B | 1.2 |
| C | A1-B-A2-B-C-B | 0.2 |
| A | A-B-A-B-D | 0.4 |
| B | A-B-A-B-D | 1.2 |
| D | A-B-A-B-D | 1 |
| A | A-B-B-D | 0.4 |
| B | A-B-B-D | 1.2 |
| D | A-B-B-D | 1 |
| A | A-B-C-B-D | 0.4 |
| B | A-B-C-B-D | 1.2 |
| C | A-B-C-B-D | 0.2 |
| D | A-B-C-B-D | 1 |
| A | B-A-B-D | 0.4 |
| B | B-A-B-D | 1.2 |
| D | B-A-B-D | 1 |
In step S3, according to the number of words and the singing time of the lyrics, that determines song on average sings speed.Specifically, present embodiment with every capable content in the LRC file as a lyrics fragment, with the time in the lyrics square bracket be zero-time, the zero-time of next lyrics is the concluding time of the one's own profession lyrics with it, just can calculate the shared singing time of these lyrics fragment.The number of words of the statistics lyrics again divided by lyrics singing time, just can obtain the speed of singing of song earlier.For example the number of words of delegation's lyrics is 10 words, and singing zero-time is [03:52.00], and the concluding time is [03:56.00], and then the fragment holding time is 4 seconds, so the speed of singing of this fragment was 0.4 word/second.
Error may take place in above-mentioned singing time computing method on last lyrics of lyrics paragragh, but the lyrics that most of LRC file can musical portions are set to sky.Like this, error can be controlled in very little scope.
Mood speech dictionary is the dictionary of analyzing at lyrics mood, also can merge with existing emotion dictionary and adjust expansion to form.For example, the mood speech dictionary in the present embodiment has merged and has known net emotion dictionary (knowing that net emotion dictionary is the light version that provides on the internet).The emotion dictionary divides emotion keyword dictionary, qualifier dictionary and negative word dictionary three parts.Mood keyword lexicon file form is as follows:
[keyword] [polarity]
Polarity is divided actively (1) and passive (1) two classes.This document example fragment is as follows:
Think and act in one and the same way 1
The same outside and inside 1
The example 1
......
A tangled skein of jute-1
Pitch-dark-1
Mess-1
Qualifier lexicon file form is as follows:
[qualifier] [degree]
Degree is divided reinforcement (1) and is weakened (1) two classes.This document example fragment is as follows:
A little-1
A little-1
Slightly-1
......
Abundant 1
Unusual 1
Greatly 1
Negative word lexicon file form is simple, and every row is only deposited negative word.
In step S4, from the lyrics, extract the mood unit, and judge the emotional stress polarity of mood unit.The context of considering the mood unit does not exceed statement substantially, and the present invention is goal seeking mood unit with lyrics fragment.The mood unit is a meaning unit that comprises mood speech, negative word (optional) and qualifier (optional).The present invention at first by mood speech dictionary from lyrics sheet segment search mood speech, then in context (totally 8 speech) scope, search qualifier and negative word.So just can in contextual window, obtain a mood unit, be usually expressed as a phrase, for example from the lyrics " the too many query of [04:02.00] [00:58.00] is too many helpless paces up and down too much ", can obtain following three mood unit:
" too many query "
" helpless too much "
" pace up and down too much "
Wherein, " too much " is qualifier, and " query ", " helpless " and " pacing up and down " are the mood speech.
The polarity judging of mood unit is fairly simple.At first extract the mood polarity of mood speech from mood speech dictionary, for example the mood polarity of " query " is passive; See then whether the mood unit exists negative word, if exist then polarity to put instead; See whether there is qualifier at last, polarity strengthens or weakens if exist then.Because three vocabulary all belong to passive vocabulary, cut and all do not exist negative word to limit, and the qualifier " too much " that has degree to strengthen, so three mood unit are the passive polarity of enhancing.
In step S5, calculate lyrics emotional stress index according to the structure of the described lyrics, the emotional stress polarity of on average singing speed and mood unit of song.Make { u
i(value of i is 1～N) to represent the mood unit in the lyrics, and the emotional stress index definition of song is:
Wherein: s is the emotional stress index, and its span is [0,16], wherein minimum value 0 show song pressure state for the lightest, maximal value 16 shows that the pressure state of song is for the heaviest; l
iBe certain lyrics, N is the sum of song emotion unit, Polar (u
i) be mood unit u
iPolarity, value be 1 ,-1}, corresponding { actively, passiveness }; Func (u
i) be mood unit u
iStrength factor, value be 1 ,-1}, corresponding { strengthen, weaken }; Struct (u
i) be mood unit u
iConstruction coefficient; Speed (u
i) be mood unit u
iSing fast coefficient, this coefficient is actually mood unit u
iFast s is sung in the present position
iFunction g (s
i), g (s
i) calculate according to following formula:
s
iFor singing speed, unit is word/second, and its span is [s
-, s
+] (s
-Be that the song overall situation croons speed most, s
+Be that the song overall situation is sung loudly speed most), g (s
i) span be [1,1].
Song is followed 8 kinds of lyric structures that provide in the preamble generally speaking, and residing position, mood unit difference in the different structure is different to the song emotion pressure influence, such as being expressed as weight listed in the table 1.
Fig. 2 shows a kind of song emotion pressure analytic system of one embodiment of the invention, and it comprises: pretreatment module is used for the cutting lyrics; The structure analysis module is connected with pretreatment module, is used for according to the repetition of the lyrics and the structure of the parallelism characteristic and the lyrics fragment sequence analysis lyrics; Sing fast analysis module, be connected, be used for judging the performance speed of song according to lyrics number of words with pretreatment module; Mood unit extraction module is connected with pretreatment module, is used for extracting the mood unit of the lyrics, and judges the pressure polarities of mood unit according to lyrics context; The emotional stress computing module with the structure analysis module, sing fast analysis module and be connected with mood unit extraction module, is used for according to lyric structure, performance speed, the described position calculation song emotion pressure of mood cell pressure polarity and mood unit index.
Can more effectively carry out the analysis of song intension according to the described method and system of present embodiment, especially judge the emotional stress of song, and then exploitation song classification, retrieval and commending system, can be adopted by ISP, mobile phone value-added service provider so that provide accurate, flexible, personalized audition for the songs and recommendation service for its user.Its potential value advantage can reduce two aspects at least:
The first, because number of songs is numerous, and kind is numerous and diverse, and the audition for the songs service often becomes unhappy depressing because of service provides improper.Carry out song retrieval by keyword (for example vocabulary in song title, singer's name or the lyrics), need the inquiry to realize knowing on the one hand and use which keyword, on the other hand, the return results accuracy rate is low, and it is poor finally to cause audition for the songs to experience.Based on algorithm of the present invention, and, can develop the song search system that combines with intension based on keyword in conjunction with other song analytical technologies.The user not necessarily realizes knowing keyword, only need to specify keyword (for example singer's name) and song intension (for example emotional stress) classification of being understood, just can accurately locate needed song, comprise the user and do not know the new song of title, thereby improve the song hit rate, dwindle user's song searching time, constantly promote user experience.
The second, personalized in recent years song recommendations service more and more receives publicity.Based on algorithm of the present invention, and in conjunction with other song analytical technologies, can develop personalization, intelligent song recommendations system, according to collection of user's song and audition log acquisition user " my favorite " list of songs, make up personalized user song appreciation taste model, and serve as the new song of not knowing the recent creation of new singer of name according to the song, the especially user that meet user's taste from song storehouse coupling with this model, thereby constantly bring pleasantly surprisedly, constantly promote user experience to him.
The above only is a preferred implementation of the present invention; should be pointed out that for those skilled in the art, under the prerequisite that does not break away from the technology of the present invention principle; can also make some improvement and modification, these improve and modification also should be considered as protection scope of the present invention.
Claims (11)
1, a kind of song emotion pressure analysis method is characterized in that, said method comprising the steps of:
S1 carries out cutting to the lyrics;
S2 by analyzing repetition, parallelism and the lyrics fragment sequence of the lyrics, judges the structure of the lyrics;
S3, according to the number of words and the singing time of the lyrics, that determines song on average sings speed;
S4 extracts the mood unit from the lyrics, and judges the emotional stress polarity of mood unit;
S5 calculates lyrics emotional stress index according to the structure of the described lyrics, the emotional stress polarity of on average singing speed and mood unit of song.
2, song emotion pressure analysis method as claimed in claim 1 is characterized in that, in step S1, describedly the lyrics are carried out cutting comprises that with lyrics cutting be statement, and be vocabulary with sentence segmentation.
3, song emotion pressure analysis method as claimed in claim 2 is characterized in that, sentence segmentation is adopted during for vocabulary two-way maximum match strategy judge the vocabulary border.
4, song emotion pressure analysis method as claimed in claim 1 is characterized in that, in step S2, analyzes the repetition of the lyrics by analyzing lyrics file temporal information and character match.
5, as each described song emotion pressure analysis method among the claim 1-4, it is characterized in that, in step S2, mate the parallelism of analyzing the lyrics by analyzing lyrics fragments character.
6, song emotion pressure analysis method as claimed in claim 5 is characterized in that, in step S2, analyzes lyrics fragment sequence by analyzing the lyrics file time series.
7, song emotion pressure analysis method as claimed in claim 1 is characterized in that, among the described step S4, searches the mood speech by utilizing mood speech dictionary from lyrics fragment, searches qualifier and negative word afterwards in context, obtains the mood unit.
8, song emotion pressure analysis method as claimed in claim 7 is characterized in that, among the described step S4, the emotional stress polarity judging step of mood unit comprises:
S4-1, the emotional stress polarity of extraction mood speech from mood speech dictionary;
S4-2 judges whether there is negative word in the mood unit, if then polarity is put anti-and forwarded step S4-3 to, if not, then directly forwards step S4-3 to;
S4-3 judges whether to exist qualifier, if, described emotional stress polarity is strengthened or slacken according to the degree of qualifier, if not, then finish.
9, song emotion pressure analysis method as claimed in claim 8 is characterized in that, described emotional stress polarity comprises positive and passive.
10, song emotion pressure analysis method as claimed in claim 1 is characterized in that, utilizes following formula to calculate the emotional stress index of song among the described step S5:
Wherein, s is the emotional stress index, and its span is [0,16], wherein minimum value 0 show song pressure state for the lightest, maximal value 16 shows that the pressure state of song is for the heaviest; l
iBe certain lyrics, N is the sum of song emotion unit, Polar (u
i) be mood unit u
iPolarity, value be 1 ,-1}, corresponding { actively, passiveness }; Func (u
i) be mood unit u
iStrength factor, value be 1 ,-1}, corresponding { strengthen, weaken }; Struct (u
i) be mood unit u
iConstruction coefficient; Speed (u
i) be mood unit u
iSing fast coefficient, this coefficient is mood unit u
iFast s is sung in the present position
iFunction g (s
i), g (s
i) calculate according to following formula:
s
iFor singing speed, unit is word/second, and its span is [s
-, s
+], s wherein
-Be that the song overall situation croons speed most, s
+Be that the song overall situation is sung loudly speed most, g (s
i) span be [1,1].
11, a kind of song emotion pressure analytic system is characterized in that, described system comprises:
Pretreatment module is used for the cutting lyrics;
The structure analysis module is connected with pretreatment module, is used for according to the repetition of the lyrics and the structure of the parallelism characteristic and the lyrics fragment sequence analysis lyrics;
Sing fast analysis module, be connected, be used for judging the performance speed of song according to lyrics number of words with pretreatment module;
Mood unit extraction module is connected with pretreatment module, is used for extracting the mood unit of the lyrics, and judges the pressure polarities of mood unit according to lyrics context;
The emotional stress computing module with the structure analysis module, sing fast analysis module and be connected with mood unit extraction module, is used for calculating the song emotion pressure index according to lyric structure, performance speed, mood cell pressure polarity and distribution situation.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN200910087827XA CN101587708B (en) | 2009-06-26 | 2009-06-26 | Song emotion pressure analysis method and system |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN200910087827XA CN101587708B (en) | 2009-06-26 | 2009-06-26 | Song emotion pressure analysis method and system |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN101587708A true CN101587708A (en) | 2009-11-25 |
| CN101587708B CN101587708B (en) | 2012-05-23 |
Family
ID=41371927
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN200910087827XA Expired - Fee Related CN101587708B (en) | 2009-06-26 | 2009-06-26 | Song emotion pressure analysis method and system |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN101587708B (en) |
Cited By (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN102467939A (en) * | 2010-11-04 | 2012-05-23 | 北京彩云在线技术开发有限公司 | Song audio frequency cutting apparatus and method thereof |
| WO2016150275A1 (en) * | 2015-03-20 | 2016-09-29 | 广东欧珀移动通信有限公司 | Preset style song processing method and apparatus |
| CN107545883A (en) * | 2017-10-13 | 2018-01-05 | 广州酷狗计算机科技有限公司 | The method and apparatus for determining the rhythm speed grade of music |
| CN111159463A (en) * | 2019-11-26 | 2020-05-15 | 黑盒子科技（北京）有限公司 | Music emotion recognition method and system |
| CN112989105A (en) * | 2019-12-16 | 2021-06-18 | 黑盒子科技（北京）有限公司 | Music structure analysis method and system |
Family Cites Families (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| KR100772386B1 (en) * | 2005-12-10 | 2007-11-01 | 삼성전자주식회사 | Method of classifying music file and system thereof |
| CN101149950A (en) * | 2007-11-15 | 2008-03-26 | 北京中星微电子有限公司 | Media player for implementing classified playing and classified playing method |
-
2009
- 2009-06-26 CN CN200910087827XA patent/CN101587708B/en not_active Expired - Fee Related
Cited By (8)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN102467939A (en) * | 2010-11-04 | 2012-05-23 | 北京彩云在线技术开发有限公司 | Song audio frequency cutting apparatus and method thereof |
| CN102467939B (en) * | 2010-11-04 | 2014-08-13 | 北京彩云在线技术开发有限公司 | Song audio frequency cutting apparatus and method thereof |
| WO2016150275A1 (en) * | 2015-03-20 | 2016-09-29 | 广东欧珀移动通信有限公司 | Preset style song processing method and apparatus |
| US10031714B2 (en) | 2015-03-20 | 2018-07-24 | Guangdong Oppo Mobile Telecommunications Corp., Ltd. | Method and device for processing audio files |
| CN107545883A (en) * | 2017-10-13 | 2018-01-05 | 广州酷狗计算机科技有限公司 | The method and apparatus for determining the rhythm speed grade of music |
| CN111159463A (en) * | 2019-11-26 | 2020-05-15 | 黑盒子科技（北京）有限公司 | Music emotion recognition method and system |
| CN112989105A (en) * | 2019-12-16 | 2021-06-18 | 黑盒子科技（北京）有限公司 | Music structure analysis method and system |
| CN112989105B (en) * | 2019-12-16 | 2024-04-26 | 黑盒子科技（北京）有限公司 | Music structure analysis method and system |
Also Published As
| Publication number | Publication date |
|---|---|
| CN101587708B (en) | 2012-05-23 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN107240398B (en) | Intelligent voice interaction method and device | |
| CN101510222B (en) | Multilayer index voice document searching method | |
| CN103345922B (en) | A kind of large-length voice full-automatic segmentation method | |
| CN103164403B (en) | The generation method and system of video index data | |
| US8554755B2 (en) | Dynamic client interaction for search | |
| CN109637537B (en) | Method for automatically acquiring annotated data to optimize user-defined awakening model | |
| US8775174B2 (en) | Method for indexing multimedia information | |
| EP2245617B1 (en) | Computer-implemented method for interacting with a user via a speech-based user interface | |
| CN101587708B (en) | Song emotion pressure analysis method and system | |
| CN107316643A (en) | Voice interactive method and device | |
| CN103956169A (en) | Speech input method, device and system | |
| CN103605665A (en) | Keyword based evaluation expert intelligent search and recommendation method | |
| CN103106287B (en) | A kind of processing method and system of user search sentence | |
| CN107799116A (en) | More wheel interacting parallel semantic understanding method and apparatus | |
| CN107403619A (en) | A kind of sound control method and system applied to bicycle environment | |
| CN1535460A (en) | Hierarchichal language models | |
| CN102280106A (en) | VWS method and apparatus used for mobile communication terminal | |
| CN102880723A (en) | Searching method and system for identifying user retrieval intention | |
| CN108538294B (en) | Voice interaction method and device | |
| CN106407484A (en) | Video tag extraction method based on semantic association of barrages | |
| CN105374352A (en) | Voice activation method and system | |
| CN106933800A (en) | A kind of event sentence abstracting method of financial field | |
| CN101493812B (en) | Tone-character conversion method | |
| CN105551485A (en) | Audio file retrieval method and system | |
| CN105654947A (en) | Method and system for acquiring traffic information in traffic broadcast speech |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| C06 | Publication | ||
| PB01 | Publication | ||
| C10 | Entry into substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| C14 | Grant of patent or utility model | ||
| GR01 | Patent grant | ||
| CF01 | Termination of patent right due to non-payment of annual fee |
Granted publication date: 20120523 Termination date: 20150626 |
|
| EXPY | Termination of patent right or utility model |