Content-based retrieval of MP3 songs based on query by singing

• Published in: 2004 IEEE International Conference on Acoustics, Speech, and Signal Processing Vol. 5; pp. V - 929
• Main Authors: LIE, Wen-Nung, SU, Chen-Kang
• Format: Conference Proceeding
• Language: English
• Published: Piscataway, N.J IEEE 2004
• Subjects: Acoustic noise; Applied sciences; Content based retrieval; Content management; Digital audio players; Discrete cosine transforms; Exact sciences and technology; Information retrieval; Information theory; Information, signal and communications theory; Interconnected networks; Internet; Miscellaneous; Music information retrieval; Networks and services in france and abroad; Signal and communications theory; Signal processing; Signal representation. Spectral analysis; Signal, noise; Support vector machine classification; Support vector machines; Telecommunications; Telecommunications and information theory; Teleprocessing networks. Isdn; Signal compression; Multistage method; Spectral method; Audio signal processing; Query; Acoustic signal; Warping; Information retrieval; K means algorithm; Support vector machine; Acoustic signal processing; Multimedia communication; Signal classification; Singing voice; Audio signal; Database; Signal processing; Content-based retrieval; Internet; Discrete cosine transforms; Content analysis
• ISBN: 9780780384842; 0780384849
• ISSN: 1520-6149; 2379-190X
• DOI: 10.1109/ICASSP.2004.1327264

With the growth of multimedia in the Internet, content analysis of multimedia plays an important role for humanistic management. We investigate the content-based retrieval of MP3 songs based on the interface of query by singing. MDCT (modified DCT) spectral coefficients are directly used to represent the tonic characteristics of a short-term sound. This spectral profile is used for detailed matching between two audio segments. Perceptual features are also computed from MDCT coefficients for audio classification. Two pre-stages based on SVM and k-means classifications are used to remove incorrect (or noisy) segment candidates and to speed up the subsequent matching process. On the other hand, exponential key-scaling schemes and time-warping techniques are developed to overcome key difference and tempo variation between different singers. Experiments show that the retrieval probability of our design can achieve up to 76% among the top 5 out of a total of 114 excerpts in the database.