Apparatus and method for designating various segment classes

A method of designating various segment classes, to which temporal segments are assigned due to segment similarities, wherein the segment similarities are assigned, wherein the segments represent an audio piece that is structured into main parts repeatedly occurring in the audio piece, comprising:

• providing a representation of the audio piece, in which the segments of the audio piece are assigned to various segment classes, wherein one segment class each can be associated with a main part, and
• designating the segment classes
• by performing the main part selection for a first first-order candidate segment class and for a second second-order candidate segment class, wherein the second order is lower than the first order regarding a similarity of the segments of the second candidate segment class to segments of the audio piece,
• by examining whether a first-order candidate segment class has the temporally later segment in the audio signal, and
• by designating the first-order candidate segment class as refrain segment class, when it has the temporally later segment, and designating the second-order candidate segment class as stanza segment class.

A computer program for executing, when the computer program is executed on a computer, a method of designating various segment classes, to which temporal segments are assigned due to segment similarities, wherein the segment similarities are assigned, wherein the segments represent an audio piece that is structured into main parts repeatedly occurring in the audio piece, comprising:

• providing a representation of the audio piece, in which the segments of the audio piece are assigned to various segment classes, wherein one segment class each can be associated with a main part, and
• designating the segment classes
• by performing the main part selection for a first first-order candidate segment class and for a second second-order candidate segment class, wherein the second order is lower than the first order regarding a similarity of the segments of the second candidate segment class to segments of the audio piece,
• by examining whether a first-order candidate segment class has the temporally later segment in the audio signal, and
• by designating the first-order candidate segment class as refrain segment class, when it has the temporally later segment, and designating the second-order candidate segment class as stanza segment class.