Name: Estefania Cano Affiliation: Fraunhofer IDMT The system is based on a main melody extraction algorithm which extracts the most prominent voice in the recording. The melody extraction is followed by an F0 refinement stage that uses interpolation in the magnitude spectrogram to refine the location of the fundamental frequency. After having found an F0 value, the harmonic components of the tone are found based on a systematic search of spectral peaks around the calculated harmonic locations. After each tone is created, a binary spectral mask is created and the tone is filtered out from the original spectrogram. Solo and accompaniment tracks are re-synthesized using inverse FFT. RESULTS ( CPU: 2.66GHz) 1. test1__tamy-que_pena_tanto_faz__snip__mix.wav: 8.907 sec 2. test1__bearlin-roads__snip__mix.wav: 8.9006 sec 3. test2__glen_philips-the_spirit_of_shackleton__snip_163_185__mix.wav: 8.563 sec 4. test2__nine_inch_nails-the_good_soldier__snip_104_125__mix.wav: 9.547 sec 5. test2__shannon_hurley-sunrise__snip_62_85__mix.wav: 8.89 sec 6. dev2__another_dreamer-the_ones_we_love__snip_69_94__mix.wav: 9.938 sec 7. dev2__fort_minor-remember_the_name__snip_54_78__mix.wav: 9.094 sec 8. dev2__ultimate_nz_tour__snip_43_61__mix.wav: 7.5 sec REFERENCES 1. Estefanķa Cano and Corey Cheng, "Melody Line Detection and Source Separation in Classical Saxophone Recordings," in 12th International Conference on Digital Audio Effects (DAFx-09), Como, Italy, 2009, pp. 1-6. 2. Sascha Grollmisch, Estefanķa Cano, and Christian Dittmar, "Songs2See : Learn to play by playing," in AES 41st International Conference: Audio for Games, London, UK, 2011. 3. Estefanķa Cano, Gerald Schuller, and Christian Dittmar " Exploring Phase Information in Sound Source Separation Applications" in 13th International Conference on Digital Audio Effects (DAFx-10). Graz, Austria, 2010. 4. Estefanķa Cano, Christian Dittmar, and Gerald Schuller "Interaction of phase, magnitude and location of harmonic components in the perceived quality of extracted solo signals" in AES 42st International Conference: Semantic Audio, Ilmenau, Germany, 2011. -- M.Sc. Estefania Cano Ceron Semantic Music Technologies Fraunhofer IDMT Ehrenbergstr. 31 D-98693 Ilmenau, Germany email: cano@idmt.fraunhofer.de phone: +49 3677 467-110