| TAZ-TFG-2020-3844 |
Harmonic Change Detection from Musical Audio
Ramoneda Franco, Pedro
Bernardes de Almeida, Gilberto (dir.)
Tardos Solano, Juan Domingo (ponente)
Universidad de Zaragoza,
EINA,
1997
Departamento de Informática e Ingeniería de Sistemas, Área de Ingeniería de Sistemas y Automática
Graduado en Ingeniería Informática
Resumen: In this dissertation, we advance an enhanced method for computing Harte et al.’s [31] Harmonic Change Detection Function (HCDF). HCDF aims to detect harmonic transitions in musical audio signals. HCDF is crucial both for the chord recognition in Music Information Retrieval (MIR) and a wide range of creative applications. In light of recent advances in harmonic description and transformation, we depart from the original architecture of Harte et al.’s HCDF, to revisit each one of its component blocks, which are evaluated using an exhaustive grid search aimed to identify optimal parameters across four large style-specific musical datasets. Our results show that the newly proposed methods and parameter optimization improve the detection of harmonic changes, by 5.57% (f-score) with respect to previous methods. Furthermore, while guaranteeing recall values at > 99%, our method improves precision by 6.28%. Aiming to leverage novel strategies for real-time harmonic-content audio processing, the optimized HCDF is made available for Javascript and the MAX and Pure Data multimedia programming environments. Moreover, all the data as well as the Python code used to generate them, are made available.
Tipo de Trabajo Académico: Trabajo Fin de Grado
Notas: El proyecto ha sido realizado en el Sound And Music Computing Lab FEUP Universidade do Porto and INESCTEC
Enlace permanente:
El registro pertenece a las siguientes colecciones:
Trabajos académicos > Trabajos Académicos por Centro > Escuela de Ingeniería y Arquitectura
Trabajos académicos > Trabajos fin de grado