Use of Locality Sensitive Hashing in Query by Humming Method of Audio Signal

Authors:

Sonal Goel (Manav Rachna College of Engineering, Faridabad)

Mehak Malik (Manav Rachna College of Engineering, Faridabad)

Charu Pathak (Manav Rachna University, Faridabad)

Keywords: Music, Information retrieval, Database query processing, Audio Systems

Abstract:

This paper is based on Locality Sensitive Hashing (LSH) which proposes a query by humming method. This method constructs an index of melodic fragments by extracting pitch vectors. This method automatically deciphers a song query into notes and then concentrates the pitch vectors as similar to the index construction. This method searches for similar fragments in the database to obtain a directory of candidate melodies for each query pitch vector. This is performed efficiently by using LSH. In our experiments, the method achieved mean reciprocal rank of 0.885 for 2797 queries when searching
from database of 6030 Musical Instrumental Digital Interface (MIDI) melodies. MIDI allows multiple instruments to be played from a single controller, which makes stage setups much more portable.

References:

[1] R. Typke, Music Retrieval based on Melodic Similarity, Ph.D. thesis, Universiteit Utrecht, 2007

[2] K. Lemstr¨om, String Matching Techniques for Music Retrieval, Ph.D. thesis, University of Helsinki, 2000.

[3] C. Meek and W. Birmingham, “Applications of binary classification and adaptive boosting to the query-by-humming problem,” in Proc. 3rd International Conference on Music InformatioRetrieval, 2002.

[4] J.-S. R. Jang, C.-L. Hsu, and H.-R. Lee, “Continuous HMM and its enhancement for singing/humming query retrieval,” in Proc. 6th International Conference on Music Information Retrieval, 2005.

[5] J.-S. R. Jang and M.-Y. Gao, “A query-by-singing system based on dynamic programming,” in Proc. International Workshop on Intelligent Systems Resolutions, 2000.

[6] A. Duda, A. N¨urnberger, and S. Stober, “Towards query by N humming/singing on audio databases,” in Proc. 7th International Conference on Music Information Retrieval, 2007.

[7] X. Wu, M. Li, J. Yang, and Y. Yan, “A top-down approach to melody match in pitch countour for query by humming,” in Proc. International Conference of Chinese Spoken Language Processing, 2006.

[8] A. Andoni and P. Indyk, “Near-optimal hashing algorithms for approximate nearest neighbor in high dimensions,” in 47th Annual IEEE Symposium on Foundations of Computer Science (FOCS’06), 2006, pp. 459–468.

[9] M. Datar, N. Immorlica, P. Indyk, and V. Mirrokni, “Locality sensitive hashing scheme based on p-stable
distributions,” in Proc. ACM Symposium on Computational Geometry, 2004, pp. 253–262.

[10] M. Casey and M. Slaney, “Fast recognition of remixed music audio,” in Proc. 2007 IEEE International Conference on Acoustics, Speech, and Signal Processing, 2007.

[11] M. Covell and S. Baluja, “Known-audio detection using Waveprint: Spectrogram fingerprinting by wavelet hashing,” in Proc. 2007 IEEE International Conference on Acoustics, Speech, and Signal Processing, 2007.

[12] M. Ryyn¨anen and A. Klapuri, “Transcription of the singing melody in polyphonic music,” in Proc. 7th
International Conference on Music Information Retrieval, 2006.

[13] Sundberg J (1977) The Acoustics of the Singing Voice, Scientific American offprints, vol 356. W.H Freeman, New York Ghias A, Logan J, Chamberlin D, Smith BC (1995)

[14] Query by humming: Musical information retrieval in an audio database. In: Proceedings of the Third ACM International Conference on Mul-timedia, MULTIMEDIA ’95. ACM, New York, pp 231–236

[15] Nagavi TC, Bhajantri NU (2012) An extensive analysis of query by singing/humming system through query proportion. Int J Multimed Appl 4(6). doi:10.5121/ ijma.2012.4606

[16] Weinstein E (2005) Query by humming: a survey. Tech. rep., NYU and Google