System and methods for providing automatic classification of media entities according to sonic properties
First Claim
1. A method for automatically classifying spectral properties of audio data, comprising:
- applying input audio data (1) to a critical band filtering process to form first output data and (2) to an entropy calculation process to form second output data;
applying the first output data to a first derivative process to form third output data; and
inputting said first, second and third output data to an averaging process to form a spectral feature vector representing the input audio data.
2 Assignments
0 Petitions
Accused Products
Abstract
In connection with a classification system for classifying media entities that merges perceptual classification techniques and digital signal processing classification techniques for improved classification of media entities, a system and methods are provided for automatically classifying and characterizing sonic properties of media entities. Such a system and methods may be useful for the indexing of a database or other storage collection of media entities, such as media entities that are audio files, or have portions that are audio files. The methods also help to determine media entities that have similar sonic properties by utilizing classification chain techniques that test distances between media entities in terms of their properties. For example, a neighborhood of songs may be determined within which each song has similar sonic properties.
115 Citations
35 Claims
-
1. A method for automatically classifying spectral properties of audio data, comprising:
-
applying input audio data (1) to a critical band filtering process to form first output data and (2) to an entropy calculation process to form second output data;
applying the first output data to a first derivative process to form third output data; and
inputting said first, second and third output data to an averaging process to form a spectral feature vector representing the input audio data. - View Dependent Claims (2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16, 17, 18, 19, 20)
-
-
21. A method of classifying data according to spectral properties of the data, comprising:
-
assigning to each media entity of a plurality of media entities in a data set to at least one spectral properties class;
processing each media entity of said data set to extract at least one spectral properties characteristic based on digital signal processing of each media entity;
generating a plurality of spectral properties vectors for said plurality of media entities, wherein each spectral properties vector includes said at least one spectral properties class and at least one spectral properties characteristic based on digital signal processing; and
forming a classification chain based upon said plurality of feature vectors. - View Dependent Claims (22, 23, 24, 25, 26, 27, 28, 29)
-
-
30. A computing system, comprising:
-
a computing device including;
a classification chain data structure stored thereon having a plurality of classification vectors, wherein each vector includes data representative of a spectral properties class as classified by humans and spectral properties characteristics as determined by digital signal processing; and
processing means for comparing an unclassified media entity to the classification chain data structure to determine an estimate of the spectral properties class of the unclassified media entity. - View Dependent Claims (31, 32)
-
-
33. A classification chain data structure utilized in connection with the classification of spectral properties of new unclassified media entities, comprising:
a plurality of classification vectors, wherein each vector includes;
spectral properties data as classified by humans; and
spectral properties data determined by digital signal processing techniques.
-
34. A method for classifying data according to its spectral properties, comprising:
- a construction phase, comprising;
classifying by human experts a representative set of sounds according to their spectral perceptual qualities;
assigning each entry in the representative set at least one value;
reducing the results of the construction phase to a set of numbers called the characteristic vector of each sound; and
storing the characteristic vector in a classification chain for later calculations; and
a classification phase, comprising;
presenting an unclassified sound for classification;
calculating the characteristic vector of the unknown sound; and
presenting the characteristic vector to the classification chain, which returns an estimate of the spectral properties of the unknown sound. - View Dependent Claims (35)
- a construction phase, comprising;
Specification