Unlike fixed-pitch instruments such as the piano, human singing can stray from a target pitch by as much as a semitone while still being perceived as a single fixed note. This p...
The use of HMM (Hidden Markov Models) for speech recognition has been successful for various applications in the past decades. However, the use of continuous HMM (CHMM) for melody...
We present an approach to automatically retrieve and extract lyrics of arbitrary songs from the Internet. It is intended to provide easy and convenient access to lyrics for users,...
This paper describes a fast audio detection method for specific audio retrieval in the AV stream. The method is a histogram matching algorithm based on structural and perceptual f...
In this paper, we study the use of spectral patterns to represent the characteristics of the rhythm of an audio signal. A function representing the position of onsets over time is...