Humming Search Humming Information Processing Method

Music Information Retrieval system via Humming involves music theory, the signal processing, the pattern recognition and so on related domain comprehensive discipline research. It has the very vital significance in music database administration, Internet music retrieval as well as life entertainment.the researches of effective humming query processing methods and exact melodic character information and representation have not been examined maturely.This article mainly revolves to humming query processing in the retrieval via Humming following several aspects to launch the research work:1) After studying the musical theoretic, the musical conceptions, the methods to describe the musical features and the theoretic of musical signals analyzing, some signal processing conceptions and algorithms and signal features about music tone extracting is presented.2) After studying each kind of the humming query processing method to hum the information present domestic and foreign, the analysis is made to each method's applicable scope and the limitation.3) Proposed one kind of improvement pronunciation enhancement technology to realize speech enhancement and noise reduction based on the spectral subtraction4) Above existing humming query processing, proposed one method about the melody feature extraction on the time domain, including audio segmentation based on short-term energy5) Proposed that one kind algorithm of pitch detection based on autocorrelation, this algorithm extract signal voiced sound section by autocorrelation function.6) the structure of the midi file is analyzed in detail,and the melody information is extracted from the midi music database.7) An intermediate form at for melody representation has been proposed.Unifying the above algorithm and the fundamental research, designed the featureextraction experiment in the humming retrieval, and performs in matlab, the experimental result had proven this kind of information processing method via humming is accurate effective.The study work of this thesis provides the preparation for the developing of content-based musical retrieval research and pushes forward the study on melody retrieval.