In this talk I will present our ongoing research on automatic transcription, melodic similarity and comparative analysis of a cappella and accompanied (guitar) flamenco singing recordings. Our transcription system is based on estimating the fundamental frequency (f0) of the singing voice. For this step, several algorithms for monophonic f0 estimation are considered to analyze a cappella singing. For polyphonic recordings, we evaluate the use of a state-of-the-art method for predominant f0 estimation vs the combination of source separation and monophonic f0 estimation. Once the f0 envelop is estimated, it is combined with other low-level features to perform note segmentation and labelling following an iterative strategy. In order to evaluate our approach, we contrast the computed transcription with manual annotations for different flamenco styles. These transcriptions are used to compute melodic similarity, and similarity matrices are then used for style classification. Finally, f0 and note descriptors are used to detect significant melodic patterns. During my presentation, I will introduce to the particularities of flamenco singing and discuss the difficulties found for its automatic characterization and for gathering human annotations and similarity judgements. I will analyze the influence of the different steps of the algorithm, state the main limitations of our approach and discuss the challenges for future studies.
Emilia Gómez is a researcher at the Music Technology Group (MTG) and assistant professor at the Department of Information and Communication Technologies (DTIC), Universitat Pompeu Fabra (UPF) in Barcelona, Spain. She graduated as a Telecommunication Engineer at Universidad de Sevilla and then received a DEA ATIAM at IRCAM. In 2006, she completed her PhD in Computer Science and Digital Communication at UPF, on the topic of Tonal Description of Music Audio Signals. She has also been visiting researcher at Royal Institute of Technology (KTH) in Stockholm and CIRMMT (Centre for Interdisciplinary Research in Music Media and Technology) McGill University in Montreal. Her research interests are related to melodic and tonal description of music audio signals, computer-assisted music analysis and computational ethnomusicology. She is very active in the field of Sound and Music Computing, acting as a reviewer of a number of journals and taking part in the organization of diverse scientific events. She has have been involved in several research projects funded by public institutions in Spain, Europe and Canada.