Method and device for generating music playlist

1. A method for generating a music playlist, the method comprising:

classifying a plurality of songs into first songs and second songs, the first songs being sample songs with mood vectors, and the second songs being new songs with no mood vectors;

comparing physical attributes of the first songs to physical attributes of each second song; determining which first song of the first songs has physical attributes most similar to the physical attributes of each second song;

assigning the mood vector of the determined first song having the most similar physical attributes to each second song; and

generating a music playlist containing songs, all with mood vectors, by combining the second songs with mood vectors assigned thereto and the first songs,

wherein the classifying of the plurality of songs comprises:

inputting a sound source file corresponding to a first song;

analyzing the inputted sound source file;

calculating physical attributes of the sound source file, including at least one of: scale, average energy, harmonics, rhythm, tempo, articulation, intensity, and tonality;

normalizing each of the physical attributes to the same scale;

calculating a first positional value representing a position on a mood plane including a plurality of moods where a meta-mood attribute corresponding to the mood vector of the first song is located using at least one of the normalized physical attributes;

calculating a second positional value representing the position on the mood plane using at least one of the physical attributes other than those used to calculate the first positional value; and

assigning the meta-mood attribute to the sound source file by positioning the sound source file with the first positional value and the second positional value on the mood plane.

2. The method of claim 1 , wherein the physical attributes are mel-frequency cepstral coefficient (MFCC) vectors.

3. The method of claim 1 , wherein:

the calculating of the first positional value comprises:

calculating a sum of scale, average energy, rhythm, and harmonics, among the normalized physical attributes; and

setting the calculated sum as a distance from a center of the mood plane for the first positional value,

the calculating of the second positional value comprises:

calculating a sum of tempo, articulation, intensity, and tonality, among the normalized physical attributes, and

setting the calculated sum as an angle of rotation about an X-axis of the mood plane for the second positional value, and

the assigning of the meta-mood attribute comprises:

assigning the meta-mood attribute to the sound source file by positioning the sound source file according to the distance and the angle on the mood plane.

4. The method of claim 3 , wherein the assigning of the mood vector of the first song comprises:

inputting a new sound source file corresponding to a second song of the second songs;

analyzing the inputted new sound source file;

calculating physical attributes of the new sound source file, including at least one of: scale, average energy, rhythm, harmonics, tempo, articulation, intensity, and tonality;

normalizing each of the physical attributes to the same scale;

comparing the normalized physical attributes of the new sound source file to the physical attributes of the sound source files for the first songs; and

setting the mood vector of the second song based on results of the comparison.

5. A device for generating a music playlist, the device comprising:

a controller including a memory configured to store program instructions and a processor configured to execute the stored program instructions, wherein the controller controls operation of:

a receiver that receives first songs and second songs, the first songs being sample songs with mood vectors, and the second songs being new songs with no mood vectors;

an analyzer that analyzes the first songs and the second songs received by the receiver, respectively, calculates physical attributes of the first songs and the second songs, and calculates mood vectors of the first songs based on the physical attributes of the first songs;

a comparator that compares the physical attributes of each second song to the physical attributes of the first songs in order to determine which first song of the first songs has physical attributes most similar to physical attributes of each second song and assigns the mood vector of the determined first song to each second song; and

a generator that generates a music playlist containing songs, all with mood vectors, by combining the second songs with mood vectors assigned thereto and the first songs,

wherein the analyzer:

analyzes a sound source file corresponding to a first song;

calculates physical attributes of the sound source file, including at least one of: scale, average energy, harmonics, rhythm, tempo, articulation, intensity, and tonality;

normalizes each of the physical attributes to the same scale;

calculates a first positional value representing a position on a mood plane including a plurality of moods where a meta-mood attribute corresponding to the mood vector of the first song is located using at least one of the normalized physical attributes;

calculates a second positional value representing the position on the mood plane using at least one of the physical attributes other than those used to calculate the first positional value; and

assigns the meta-mood attribute to the sound source file by positioning the sound source file with the first positional value and the second positional value on the mood plane.

6. The device of claim 5 , wherein the physical attributes are mel-frequency cepstral coefficient (MFCC) vectors.

7. The device of claim 5 , wherein the analyzer:

calculates a sum of scale, average energy, rhythm, and harmonics, among the normalized physical attributes;

sets the calculated sum as a distance from a center of the mood plane where an emotional attribute corresponding to the mood vector of the first song is located for a first positional value;

calculates a sum of tempo, articulation, intensity, and tonality, among the normalized physical attributes;

sets the calculated sum as an angle of rotation about the X-axis of the mood plane for a second positional value; and

assigns the emotional attribute to the sound source file by positioning the sound source file according to the distance and the angle on the mood plane.

8. The device of claim 7 , wherein the comparator:

analyzes a new sound source file corresponding to a second song of the second songs;

calculates physical attributes of the new sound source file, including at least one of: scale, average energy, harmonics, rhythm, tempo, articulation, intensity, and tonality;

normalizes each of the physical attributes to the same scale;

compares the normalized physical attributes of the new sound source file to the physical attributes of the sound source files for the first songs; and

sets the mood vector of the second song based on results of the comparison.