The Research Of An Algorithm Based On Query By Humming

With growth of the society and the development of the science and technology, the various interactive media attach primary importance to people-oriented service more and more. The birth of Content Based Retrieval, transforms the technology of retrieval radically. One of the areas in the Content Based Retrieval is Query By Humming. Query By Humming's development is accelerated, and application prospect is very broad. This technology of retrieval searches the matching song by user's humming based on the melody of the song. Query By Humming has broad application prospect and commercial space in the searching of songs on the network, music producer, picking songs system of KTV room, authentication etc.In this dissertation, Query By Humming is analyzed in the three constituent parts: pretreatment, feature extraction, feature matching. The integration algorithm of Query By Humming is confirmed according to the pluses and minuses of varies methods.Two different methods are confirmed to manage two different signals, hum signal and music signal. This algorithm overcomes the shortcoming that the target of searching must be the specific sentence of the song. The user's hum is collected by the microphone which is connected with the computer. Then, the algorithm of Query By Humming can search the correct song for the user's hum.