TWI414949B - A music video server and a method for setting favorite music video list - Google Patents

Abstract

A method for setting favorite music video list comprises steps of: determining whether an user is first login; calculating similarities of not played songs when the user is logout last time, if the user is not first login; generating a user favorite music video list according to the similarities of the not played songs and an user favorite degree to all songs; receiving user votes at different times; determining the user favorite degree to all the songs according to the user votes; determining whether the songs in the music list are all played; and calculating similarities of not played songs repeatly, if the songs in the music list are not all played. The present invention also provides a music video server.

Description

Music video server and its music video personalization setting method

The invention relates to a music video (MV) server and a music video personalization setting method thereof.

With the rapid development and popularization of digital technology, digital TV has become more and more popular among thousands of households, and users can select favorite programs according to their own preferences (for example, picking songs that they like to listen to). However, this situation requires the user to manually select among a large number of programs provided by the operator or the service provider, which not only takes a lot of time, but also causes inconvenience to the user.

Since the program selected by the user each time represents his personal preference, the personal preference is not saved, but disappears with each logout. Therefore, the user needs to re-select the next time to log in, resulting in time. waste.

A music video server is connected to a plurality of media servers and user equipments for generating a personalized music video playlist for the user, the media server for storing song content provided by an operator or a service provider, the user The end device is configured to play the songs, wherein the music video server comprises a receiving module, a storage module, a recording module, a judging module, a computing module, and a music video playlist generating module. The receiving module is configured to receive the identity identification message and the user voting message from the user equipment. The storage module is used to store the attributes of the song and its weight values. The recording module is used to record user voting messages. The judging module is used to identify the identity of the user equipment The other message determines whether the user logs in for the first time, and determines the user preference of each song based on the user voting message. The calculation module is configured to calculate the similarity of the song according to the song attribute and the weight value thereof, the weight value is a value corresponding to each attribute of the song, and the utility model may be divided into a preset weight value and a calculation weight value, wherein the preset weight value For the value preset by the music video server for each attribute, the calculated weight value is a value calculated from the unplayed song attribute calculated relative to the played song attribute. The music video playlist generation module is configured to generate a new music video playlist according to the user preference and similarity of the song.

A music video personalization setting method includes the following steps: receiving a user login message; determining whether the user logs in for the first time; if the user is not logging in for the first time, calculating the song play record and the song attribute and the weight value according to the last time the user left the user The similarity of the unplayed songs; generating a music video playlist of the user's preferences according to the similarity of the songs and the user's preference; playing the songs in the music video playlist one by one; receiving the voting messages of the users at different times; determining the users according to the voting messages Desirability; judge whether the song in the music video playlist is completely played; and if the song in the music video playlist is completely played, calculate the similarity of the unplayed song according to the song attribute and its weight value, the weight value is each song A value corresponding to an attribute, which may be divided into a preset weight value and a calculated weight value, wherein the preset weight value is a value preset by the music video server for each attribute, and the calculated weight value is an unplayed song attribute. The value obtained relative to the calculated song attribute.

The music video server and the music video personalization setting method of the present invention use the user voting mechanism to feed back the user's favorite message for the song, and at the same time, the music video server calculates the similarity according to the attribute of the song, and according to the calculated similarity degree. The user's preference is used to generate a personalized music video playlist that suits the user's preferences, saving the user's time for selecting the program and facilitating the user's use.

10‧‧‧Client equipment

21, 22‧‧‧Media Server

30‧‧‧Music video server

301‧‧‧ receiving module

302‧‧‧recording module

303‧‧‧ storage module

304‧‧‧Judgement module

305‧‧‧Computation Module

306‧‧‧Music video playlist generation module

307‧‧‧Reading module

1 is an application diagram of a personalized music video server of the present invention; FIG. 2 is an internal module diagram of the music video server of FIG. 1 according to the present invention; FIG. 3 is a diagram showing a music video personalization setting method of the present invention. FIG. 4 is a detailed flowchart of step S360 in FIG. 3 of the present invention; FIG. 5 is a detailed flowchart of step S380 in FIG. 3 of the present invention; FIG. A refinement flowchart of step S330; and FIG. 7 is another flowchart of the personalized music video personalization setting method of the present invention.

1 is a schematic diagram of an application environment of a music video (MV) server according to the present invention, including a client device 10, such as a set-top box (STB), a portable handheld device, and the like, and two media. The servers 21, 22 and the music video server 30. The media servers 21, 22 are used to store the content of the songs provided by the operator or the service provider. The client device 10 is connected to the media servers 21, 22 and the music video server 30, respectively, for playing the songs. The music video server 30 is also coupled to the media servers 21, 22 for generating a personalized music video playlist for the user. When the user wants to listen to the song, the music video server 30 generates a music video playlist, and the client device 10 downloads the corresponding song from the media servers 21, 22 for playback according to the list of songs. In the present embodiment, the number of the media servers 21 and 22 is two, and the other embodiments are not limited thereto.

2 is a block diagram of the interior of the media server 30 of FIG. 1 of the present invention. The media server 30 includes a receiving module 301, a recording module 302, a storage module 303, a determining module 304, a computing module 305, a music video playlist generating module 306, and a reading module 307.

The receiving module 301 is configured to receive the identity identification message and the user voting message from the client device 10.

The recording module 302 is used to record the voting message of the user. In this embodiment, the user voting message in the recording module 302 is coverable, and the user vote during the playback of the song can be defined as one round of voting. For the same song, the recording module 302 records the song. The last round of voting messages, that is, when the user makes a second round of voting on a certain song, the second round of voting will cover the first round of voting messages.

The storage module 303 is configured to store the attributes of the song and its weight values. Among them, the song attributes include song title, lead singer, music type (popular, classical, etc.), music rhythm (slow, fast), musical mood (sadness, happiness), plot (singing, dancing, etc.), language, lyrics, composition, and so on. In this embodiment, the attributes may be quantizable and non-quantizable respectively. For example, music mood and music rhythm are quantifiable attributes, which can use digital 1~5 to indicate the degree of gradual change of music emotion from sadness to happiness or music rhythm from slow to fast. Here, the numbers 1~5 are called quantized values. . The attributes listed above are not quantifiable except for musical emotions and musical rhythms. Meanwhile, in this embodiment, each song attribute has its own weight value, which can be divided into a preset weight value and a calculation weight value, wherein the preset weight value is preset by the music video server 30 for each attribute. The numerical value, for example, the preset weight of the title is 2, the preset weight of the lead singer is 20, the preset weight of the music type is 20, the preset weight of the musical mood is 15, the preset weight of the music rhythm is 15, and the plot is The preset weight is 15, the default weight of the language is 5, the preset weight of the lyrics is 3, and the preset weight of the composition is 5, and the preset weight can be adjusted according to the situation of the body. The calculated weight value is the value of the unplayed song attribute calculated relative to the played song attribute. In this embodiment, the weight values of all the attributes are added up to less than 100.

The judging module 304 is configured to determine the user preference of each song according to the user voting message, that is, determine whether each song belongs to a song that the user likes. At the same time, it is determined according to the identity identification message of the user equipment 10 whether the user logs in for the first time.

The calculation module 305 is configured to calculate the similarity of the song according to the song attribute in the storage module 303. among them The similarity is the degree of similarity of the unplayed song relative to the played song, which is only for the unplayed song, and the calculation method is obtained by adding all the attribute calculation weight values of the unplayed song, and the specific calculation method is as follows. This will be described in detail in Figure 6.

The music video playlist generation module 306 is configured to generate a new music video playlist based on the user preference and similarity of the song.

FIG. 3 is a flow chart showing a personalized music video personalization setting method according to the present invention. First, in step S310, the receiving module 301 receives the user login message. In step S320, the determining module 304 is configured to determine, according to the identity identification code of the user equipment 10, whether the user logs in for the first time, that is, whether the user uses the music video personalization setting service for the first time. In other embodiments of the invention, the music video server 30 may authenticate in other ways. If the user is logging in for the first time, step S321 is performed, and the music video playlist generating module 306 randomly generates a system music video playlist.

If the user is not the first time to log in, step S330 is executed, and the calculation module 305 calculates the similarity of the unplayed songs. In this case, the similarity of the songs that were not played when the user left the service last time is calculated. The similarity of the unplayed songs in this embodiment is calculated based on the attribute weight values of the played songs. For example, if the operator or the service provider provides a total of 100 songs, each music video playlist is included. There are only 10 songs. After listening to the first round, there are only 90 songs. If the songs in the first round of music video playlist are the first to the tenth, then the remaining 90 songs are the 11th to the 100th. first. When the songs in the first music video playlist are completely played, the calculation module 306 performs similarity calculations for the 11th to 100th songs with respect to the first to the 10th songs, respectively. That is to say, the 11th song will be similarly calculated with the first to the 10th songs respectively, that is, 10 similarity values are calculated, and then the maximum value among the 10 similarity values is taken as the 11th. The similarity of the first song, and so on, the similarity value of all the songs that are not played.

For the specific calculation method, please refer to FIG. 6. At step S610, the reading module 307 sequentially reads the attributes of the unplayed songs in the storage module 303. In step S620, the determining module 304 determines whether the read song attribute is a quantizable attribute. If it is a quantifiable attribute, step S621 is executed, and the calculation module 305 calculates a formula according to the quantizable attribute weight: (1 - (quantized value of the compared song - quantized value of the compared song) / quantized value interval) * this attribute is pre- Set the weight value and calculate the weight value for this quantifiable attribute. Here, the compared song is a played song, the comparison song is an unplayed song, and the quantized value interval is the minimum integer interval between the minimum quantized value and the maximum quantized value, for example: music mood and music rhythm system default quantization value It is 1~5. Since its minimum integer interval is 1, there are 4 intervals, and its quantization interval is 4.

If the attribute is a non-quantizable attribute, then step S630 is executed, and the determining module 304 determines whether the attribute of the unplayed song is consistent with the content of the attribute corresponding to the compared song. If the two are consistent, in step S640, the calculation module 305 obtains the preset weight value of the attribute. If the two are inconsistent, in step S631, the calculation module 305 obtains a weight value of 0 for this attribute. Next, in step S650, the calculation module 305 adds the weight calculation values of all the attributes of each of the above songs to obtain the similarity of the song.

For example, if the first song is a song that has already been played, its attributes and contents are: song name - Madrid unbelievable, lead singer - Jolin Tsai, music type - pop, music mood -5, music rhythm - 4, plot - go Sing, language-Chinese, lyrics-Huang Junlang, composition-Chen Mengqi; if the 11th song is an unplayed song, its attributes and contents are: song name - see me seventy-two, lead singer - Jolin Tsai, music type - Pop, music mood-3, music rhythm-5, plot-dance, language-Chinese, lyrics-Chen Zhenchuan, composition-Edward Chan/Charles Lee. The attribute weight value of the first song is a preset weight value. As mentioned above, the attribute weight value of the 11th song is a calculation weight value. That is to say, the 11th song attribute weight value is obtained by calculating with respect to the first song.

First, the judging module 304 first determines whether the content of the non-quantitative attributes of the two songs is consistent, that is, whether the song name, the lead singer, the music type, the plot, the language, the lyrics, and the composition content are the same, and secondly, The calculation module 305 calculates the weight value of the eleventh song quantizable attribute according to the quantifiable attribute weight calculation formula. For example, if the song name, plot, lyrics, and composition of the 11th song are different from those of the first song, the weight values of the attributes are all 0; the lead singer, music type, language of the 11th song are the same as the first song. If the first header is the same, the calculated weight value of these attributes and the weight value of the first corresponding attribute are the preset weight values of the attributes: 20, 20, and 5, respectively. For the quantified value of musical mood, the first is 5 and the eleventh is 3, then the calculated weight of the 11th musical emotion is: (1-(5-3)/4)*15=7.5. Correspondingly, the calculated weight of the 11th song music rhythm is: (1-(5-4)/4)=11.25. Finally, all the calculated weight values of the 11th song are added to obtain the similarity value, that is, the similarity of the 11th song = the name of the song (0) + the lead vocal (0) + the music type (20) + the musical mood (7.5) + music rhythm (11.25) + plot (0) + language (0) + lyrics (0) + composition (0) = 63.75.

After the similarity is calculated, the process returns to FIG. 3. In step S340, the music video play request generation module 306 generates a new music video playlist according to the song similarity and the user preference (refer to FIG. 5 in detail). The music video playlist is generated from two sets of songs: one is a song with a high degree of user preference (a favorite song) and a set of songs whose similarity is greater than a system preset value of 60, and the similarity is lower than a preset. For a set of songs with a value of 60, the music video playlist generation module 306 will randomly select the two sets according to the set ratio (eg, 8 for the first set and 2 for the second set). 10 songs as a new music video playlist. It should be noted that the reason for choosing among songs with low similarity is that since some users are limited to listening to a certain type of song, subjectively excludes the possibility of other songs being listened to, so that the user can contact the user more. Songs other than the favorite category, increase their preference for other songs. In the present embodiment, the number of selected songs in the two sets may be changed, and is not limited thereto.

In step S350, the client device 10 downloads the music video playlist generated in step S321 or step S340. In step S360, the client device 10 obtains related music video content from the media server 21 or 22 to play one by one according to the downloaded music video playlist.

In step S370, the receiving module 301 receives the user voting message at different times. That is to say, each song can be voted at any time during the playback process, for example, voting during playback, voting after the end of playback, and the like. In this embodiment, the voting message is divided into two types, one is a positive vote, and the other is a negative vote. Among them, the affirmative vote indicates that the user prefers a certain song. A negative indicates that the user does not like a song.

In step S380, the determining module 304 determines the user preference according to the voting message of the user for each song.

In step S390, the determining module 304 determines whether the songs in the music video playlist have been played. If not finished, returning to step S360, the client device 10 continues to play the songs in the list one by one. If the broadcast has been completed, step S330 is executed, and the calculation module 305 calculates the similarity of the unplayed songs.

FIG. 4 is a detailed flowchart of step S360 in FIG. 1 of the present invention. When the user is listening to a certain song, in step S410, the determination module 304 determines whether the user wants to leave halfway. If the user wants to leave midway, the entire process ends. If the user does not leave in the middle, step S420 is executed, and the determining module 304 determines whether the user votes according to whether the receiving module 301 receives the voting message. If the user has a vote, step S421 is performed, and after voting, the process returns to step S360 in FIG. In this embodiment, the user voting system at this time defaults to the vote voted by the user.

If the user does not vote, step S430 is executed, and the determination module 403 determines whether the currently played song is played. If the playback has been completed, step S440 is performed, and the user performs voting, that is, the receiving module 401 receives the user voting message. In this embodiment, the system defaults to the vote voted by the user at this time. Invoices, that is, only when the user likes a song, will they listen to it completely.

If the playback is not completed, step S431 is executed, and the determination module 304 continues to determine whether the user cuts off the playback during the song playback. If the playback of a certain song is not cut off, it returns to step S360 of FIG. If the playback of a certain song is cut off, step S440 is performed and the user votes. At this point, the system defaults to the user's vote, that is, the user does not like the current song, so it is cut off halfway.

In this embodiment, the user may not select a vote after a certain song is played. At this time, the system will default to the user's vote, that is, the user likes the song. Similarly, the user may not choose to vote when a song is cut off midway. At this time, the system defaults to the user's vote, that is, the user does not like the song.

As can be seen from FIG. 4, in the present embodiment, the user usually votes in three situations: one is to listen to the song in the middle of the vote, the other is to cut the song during the course of listening to the song, and the third is to vote after the song is played.

FIG. 5 is a detailed flowchart of step S380 in FIG. 3 of the present invention. In step S510, the determining module 304 determines whether the received voting message belongs to the vote or the vote. If the voting message is in favor of the vote, indicating that the user likes the song, step S511 is performed, and the recording module 302 adds 1 to the number of votes in favor, and the number of the votes is positive. During the whole process of listening to the song, the user casts more than once. In favor of the vote, the corresponding number of votes in favor has increased.

If the voting message is negative, indicating that the user does not like the song, step S512 is performed, and the recording module 302 adds 1 to the black number, and the negative number is a negative value. It is worth noting that since the user does not like a song, the song will be cut off, so the black will only appear once, but the ticket can be used multiple times.

In step S520, the determining module 306 determines according to the number of votes and the number of negative votes of a certain song. The user's preference for the song, that is, the song that the user likes or the song that he does not like. Specifically, in the embodiment, when the user cuts the song during the listening process, the record module 302 records that the user's vote for the song is a negative value of the ticket, and the positive value of the ticket is 0. I don't like this song; when the user has voted in favor during the listening process, but when I switch to the next song without listening, the record module 302 records that the user's vote for the song is a vote in favor. One vote is negative. In this case, the default user of the system likes this song, and the song is classified as such a song that the user likes. When the user voted for a number of votes during the listening process, the song is undoubtedly For the songs that users like.

It should be noted that in the present embodiment, in the process of calculating the similarity, it is possible that the following situation occurs: since each unplayed song is calculated similarly to the played song, as can be seen from FIG. 5, The songs that have been played include the songs that the user likes and the songs that the user does not like, that is, the voting of the played songs may be positive or negative, then, similar to the similarity, the similarity of an unplayed song The degree may also be positive or negative, so the choice of similarity should follow the following rules: Compare the maximum value of the positive value of the similarity with the maximum value of the absolute value of the negative value, which one is larger, Then, the user preference of the unplayed song is consistent with the largest one; if the two are as large, the default unplayed song is the song that the user likes. For example, if the similarity of the 11th song to the first to the 10th songs respectively has 10 values, there may be 6 positive values, 4 negative values, and firstly select the largest of the 6 positive values. The value, for example: 78, secondly selects the maximum of the four negative absolute values is -70, obviously, 78 is greater than the absolute value of -70, then the eleventh song belongs to the kind of song that the user likes. If the largest of the four negative absolute values is -80, obviously, 78 is less than the absolute value of -80, then the eleventh song belongs to the type of song that the user does not like. If the largest of the four negative absolute values is -78, obviously, 78 is equal to -78 absolute value, then the 11th song system defaults to the kind of song that the user likes. In this way, all the songs in the system will have their own user preferences after listening to each music video playlist, and the user preferences may be the same or different each time.

FIG. 7 is a variation of FIG. 3 of the present invention, which is substantially the same as FIG. 3, except that FIG. 7 further includes step S322. When the user is not logging in for the first time, the determining module 304 determines whether the song needs to be changed. Attribute weight. If the change is required, step S323 is executed, and the music video server 30 changes the song attribute weight value and stores it in the storage module 302, overwriting the original song attribute weight value. When the change is completed, the process returns to step S322. If no change is needed, step S330 is executed, and the calculation module 305 calculates the similarity of the unplayed songs. Here, there are many reasons for changing the weight of the song attribute. For example, after a period of user feedback, it is found that the user's preference for a certain song is not obvious, or some users prefer to listen to only certain lead singers. The songs of the lyrics, composers, and the like, or other reasons, such that the song attribute weights are modified such that the music video playlist generated by the music video playlist generation module 306 is more in line with the user's preference.

The music video server and music video personalization setting method of the present invention utilizes a user voting mechanism to feed back a user's favorite message for the song, and at the same time, the music video server 30 calculates the similarity according to the attribute of the song, and according to the calculated similarity. Degree and user preference to generate personalized music video playlists that match the user's preferences, saving users time to select programs and facilitating user access.

In summary, the present invention complies with the requirements of the invention patent and submits a patent application according to law. The above description is only the preferred embodiment of the present invention, and equivalent modifications or variations made by those skilled in the art of the present invention should be included in the following claims.

Claims (23)

A music video server, which is connected to a plurality of media servers and user equipments for generating a personalized music video playlist for the user, the media server for storing a plurality of songs provided by an operator or a service provider, the user The end device is configured to play the songs, wherein the music video server comprises: a receiving module, configured to receive a user identification message and a user voting message; and a storage module, configured to store attributes of the songs and weight values thereof a recording module for recording the user voting message; the determining module, configured to determine, according to the user identification message, whether the user associated with the user identification message is logged in for the first time, and determining, according to the user voting message, a user preference of the first song; a calculation module, configured to calculate a similarity of the song according to the attributes and weight values of the songs, the weight value is a value corresponding to each attribute of a song in the songs, and is divided into Presetting a weight value and calculating a weight value, wherein the preset weight value is pre-set for each attribute of a song in the songs a value obtained by calculating an unplayed song attribute relative to an attribute of a played song; and a music video playlist generating module for generating the user preference according to the song and the similarity New music video playlist. The music video server of claim 1, wherein the music video server further comprises a reading module for sequentially reading attributes of the songs. The music video server of claim 1, wherein the determining module is further configured to determine whether the songs in the music video playlist are all played. A music video server as described in claim 1, wherein the attributes of the songs include Quantizable attributes and non-quantizable attributes, wherein the quantifiable attributes include music rhythms and musical emotions, and non-quantitative attributes include music name, music type, plot, language, lyrics, and composition. The music video server of claim 1, wherein the similarity of the songs indicates the degree of similarity of the unplayed songs relative to the played songs, and the calculation manner is that the calculated weight values of all the attributes of the unplayed songs are Plus and get. A music video server as claimed in claim 4, wherein the quantizable attribute weight calculation formula = (1 - (quantized value of the compared song - quantized value of the compared song) / quantized value interval) * this attribute Preset weight value. The music video server according to claim 4, wherein the calculation rule of the non-quantizable attribute weight is as follows: when the content corresponding to the unplayed song attribute is consistent with the content of the corresponding attribute of the played song, the unplayed song The attribute is equal to its preset weight value; otherwise, the unplayed song attribute is equal to zero. A music video personalization setting method includes: receiving a user login message; determining whether the user logs in for the first time; if the user is not logging in for the first time, calculating the similarity of the unplayed song according to the attribute of the unplayed song and its weight value; Playing the similarity of the unplayed songs with the user's preference to generate a music video playlist of the user's preferences; playing the songs in the music video playlist one by one; receiving the voting message of the user at different times; determining the user's preference according to the voting message Determining whether the song in the music video playlist is completely played; and if the song in the music video playlist is completely played, according to the unplayed song relative to The song attribute in the music video playlist and the weight value thereof calculate the similarity of the unplayed song, and the weight value is a value corresponding to each attribute of a song in the unplayed song, and is divided into preset weight values. And calculating a weight value, wherein the preset weight value is a preset value of each attribute of a song in the unplayed song, and the calculated weight value is an attribute of a song in the unplayed song relative to The value obtained by calculating the properties of the played song. The music video personalization setting method of claim 8, wherein the step of determining whether the user logs in for the first time further comprises the step of randomly generating a system music video playlist when the user logs in for the first time. The music video personalization setting method of claim 8, wherein the step of whether the song in the music video playlist is completely played further comprises: if the song in the music video playlist is not completely played, continuing to perform the play one by one The steps for playing a song in a music video playlist. The music video personalization setting method of claim 8, further comprising the steps of: determining whether a weight value of the song attribute needs to be changed; and if the change is required, performing the step of changing the song attribute weight value. The music video personalization setting method of claim 8, wherein the step of calculating the similarity of the unplayed song further comprises: sequentially reading an attribute of the unplayed song and the corresponding content; determining the read attribute Whether it is a quantizable attribute; if it is a quantizable attribute, calculating a weight of the quantizable attribute according to a quantifiable attribute similarity calculation formula; if it is a non-quantizable attribute, the determining module determines the attribute corresponding to the unplayed song Whether the content is consistent with the content of the attribute corresponding to the compared song; if the two are consistent, the weight of the unplayable attribute of the unplayed song is equal to the preset weight value; If the two are inconsistent, the non-quantitative attribute of the unplayed song is equal to 0; and the weight values of the above attributes are respectively added to obtain the similarity. The music video personalization setting method according to claim 12, wherein the quantifiable attribute weight calculation formula = (1 - (quantized value of the compared song - quantized value of the compared song) / quantized value interval) * Item attribute preset weight value. The music video personalization setting method of claim 8, wherein the step of receiving the voting message of the user at different times further comprises: determining whether the user leaves in the middle of playing a certain song; if not leaving, determining the user Whether to vote; if the user does not vote, judge whether the song is played; if the song has been broadcast, receive the vote in favor; if the song is not finished, determine whether the user cuts off the song; and if The user cuts off the playback of this song and receives a vote against the vote. The music video personalization setting method of claim 14, wherein the step of voting by the user further comprises: if the user votes, receiving the voting message and continuing to perform the step of playing the songs in the music video playlist one by one. The music video personalization setting method of claim 14, wherein the step of determining whether the user cuts off the song further comprises: if the user does not cut off the playing of the song, continuing to perform the music playing one by one The steps for playing a song in the list. The music video personalization setting method of claim 8, wherein the step of determining the user preference according to the voting message further comprises: determining whether the voting is an affirmative vote; if the voting is a positive vote, the number of votes is plus 1; If the vote is negative, the number of votes is increased by one; and the user preference of the song is determined based on the number of votes in favor and the number of votes. A music video server for generating a user personalized media content playlist, wherein the music video server comprises: a receiving module, configured to receive a user identification message and associated with each media item in the first media item set a user voting message, wherein the first media item set includes one or more media items; and the storage module is configured to store an attribute of each media item in the second media item set and a weight value thereof, wherein the foregoing The second media item set includes one or more media items; the recording module is configured to record the user voting message; and the determining module is configured to determine, according to the user identification message, whether the user associated with the user identification message is Logging in for the first time, and determining the user preference of each media item in the first media item set according to the user voting message; the computing module is configured to compare each media item in the first media item set The attribute of the second media item set and its weight value to calculate each media in the first media item set The similarity between the target and the media item in the second media item set; wherein the weight value is a value corresponding to each attribute, and is divided into a preset weight value and a calculated weight value, wherein the preset weight value is each attribute a preset value, the calculated weight value is a value obtained by calculating an unplayed media item attribute in the first media item set relative to a played media item attribute in the second media item set; and a music video playlist generating mode A group for generating a new music video playlist based on the user preferences and similarities. The music video server of claim 1, wherein the determining module is further configured to determine whether the media items in the music video playlist are all played. A music video server as described in claim 1 wherein the properties of the media items are The quantifiable attribute includes a music rhythm and a musical mood, and the non-quantizable attributes include a music name, a music type, a plot, a language, a lyric, and a composition. The music video server of claim 18, wherein the similarity of the media items indicates the degree of similarity of the unplayed media items relative to the played media items, and the calculation manner is that all the attributes of the unplayed media items are The calculated weight values are added together. The music video server according to claim 20, wherein the quantizable attribute weight calculation formula = (1 - (quantized value of the compared media item - quantized value of the comparison media item) / quantized value interval) * Item attribute preset weight value. The music video server according to claim 20, wherein the calculation rule of the non-quantizable attribute weight is as follows: when the content corresponding to the unplayed media item attribute is consistent with the content of the corresponding attribute of the played media item, the The play media item attribute is equal to its preset weight value; otherwise, the unplayed media item attribute is equal to zero.