Audio Information Retrieval using Sinusoidal Modeling based Features

Abstract

In this paper, we propose an approach for searching a speech query word in an audio data base. In this approach, we explore Sinusoidal modeling based features such as Amplitude, Frequency and Phase of the speech signal. Three independent systems are built using these features. Further, the Majority voting logic is used to arrive at a conclusion to locate (time stamp) the query word in the reference utterances. The studies are performed on the TIMIT database. The results show that, Sinusoidal based features can be used for speech processing in place of conventional approaches.