Content-based Music Retrieval from Acoustic Input (CBMR).ppt

Content-based Music Retrieval from Acoustic Input (CBMR) Outline What is CBMR? Methods Signal processing Similarity comparison Experiment results Demo Future work What is CBMR? CBMR : Content-based Music Retrieval Traditional database query : Text-based or SQL-based Our goal : Music retrieval by singing/humming Related Work Query by humming by Ghias,Loga and Chamberlin in 1995 Autocorrelation pitch detection 183 songs in database MELDEX system by New Zealand Digital Library Project in 1996 Gold/Rabiner Algorithm (800 songs) Sing ‘la’ or ‘ta’ when transposition Karaoke song recognizer by J.F. Wang in 1997 Novel pitch detection 50 songs in database Flowchart Original Wave Input Single Frame Pitch Tracking Range E2 - C6 82 Hz - 1047 Hz ( - ) Method Auto-correlation Auto-correlation without Clipping Center Clipping Auto-correlation with Clipping Pitch Contour Signal Process Remove violent point short notes Down sampling smoothing Frequency to semitone Semitone : A music scale based on A440 Pitch Contour (After Smoothing) Mid-level Representation Mid-level Representation without Rest Similarity Comparison Goal Find the most similar Midi file Challenge Tempo variance Dynamic time warping (DTW) Tune variance Key transposition Compare by DTW Dynamic Time Warping (DTW) DTW (cont.) Example of DTW Key Transposition Mean sift Binary search in the searching area O( N) -- O (log N) Example of Key Transposition Score Function m : length of match string n : length of input string e : DTW distance A = 0.8 B = 0.6 Experiment Environment 290 wave files Wave length : 5 - 8 sec Wave format : PCM, 11025Hz, 8bits, Mono Environment Celeron 450 with 128Mb RAM under Matlab 5.3 Database 493 midi files Experiment Result (Histogram) Experiment Result (Pie) Experiment Result (Pie) - With Rest How to Accelerate? Branch and bound O(N) - O(lnN) Triangle inequality d(a,b) + d(b,c) ≧ d(a,c) Hierarchical 2 phase 3/32 sec 2/32 sec Experiment Result (Pie) - 3/32 sec Experiment Result