Conference,
Extracting Acoustic Features of Singing Voice for Various Applications Related to MIR: A Review
(2013)
Abstract
Efficient and intelligent music information retrieval
(MIR) is a need of the 21st century. MIR addresses the problem
of querying and retrieving certain types of music from large
music data set. A singing voice is one of the key elements of
music. As most part of music is characterized by the performing
singer, analysis of singing voice reveals many characteristics
of a song. The unique qualities of a singer’s voice make it
relatively easy for carrying out numerous tasks in MIR. The
singing voice is completely characterized by its acoustic
features. Acoustic features like timbre, vibrato, pitch and
harmony describe the singing voice in the music and these
are discussed in the paper. There are many applications of
MIR which considers overall features of music, but the paper
presents a review of those applications of MIR concerned
directly to singing voice in the music. Also the paper lists the
feature extraction methods and identifies the suitable feature
appropriate for individual task of MIR.
