Automatic Accompaniment Music Generation and Singing Skill Evaluation for Vocal Melodies

  • K.L. Jayaratne

Abstract

Music composition may be one of the most difficult
tasks for non-musicians. To allow non-musicians to get a taste
of music creation, we propose a system for generating accompaniment
music for vocal melodies with singing skill evaluation
to encourage the user to improve the skill of singing. In the
proposed system, a pitch detection module first extracts the
pitch classes from the vocal melody using a time domain autocorrelation
method. According to these pitch classes, a singing
skill evaluation module classifies the skill of the vocal melody
as good or poor based on the pitch and the tempo. The pitch
based singing skill classifier uses two features of pitch interval
accuracy and classifies the singing skill using a trained Support
Vector Machine. The tempo based singing skill classifier uses a
vibrato suppression based onset detection technique to estimate
the tempo of the vocal melody and classifies the singing skill. A
Hidden Markov Model with learned probabilities then constructs
the best acceptable chord progression for the vocal melody by
applying the Viterbi algorithm. The system then generates the
accompaniment music using the constructed chord progression
and combines it with the vocal melody to obtain the final outcome.
Finally, we present the results from the first and the second study
demonstrating that the performance of singing skill classifiers is
fairly good. We present results from the third study showing that
our system is able to construct acceptable chord progressions for
vocal melodies. The results from the final study shows that the
users are enthusiastic about the final outcome of our system.

Published
2018-08-29
How to Cite
JAYARATNE, K.L.. Automatic Accompaniment Music Generation and Singing Skill Evaluation for Vocal Melodies. GSTF Journal on Computing (JoC), [S.l.], v. 6, n. 2, p. 12, aug. 2018. ISSN 2010-2283. Available at: <http://dl6.globalstf.org/index.php/joc/article/view/1751>. Date accessed: 17 dec. 2018.