Acoustic Data Analysis from Multi-Sensor Capture in Rare Singing
Résumé
This paper deals with new capturing technologies to safeguard and transmit endangered intangible cultural heritage including Corsican multipart singing technique. The described work, part of the European FP7 i-Treasures project, aims at increasing our knowledge on rare singing techniques. This paper includes (i) a presentation of our light hyper-helmet with 5 non-invasive sensors (microphone, camera, ultrasound sensor, piezoelectric sensor, electroglottograph), (ii) the data acquisition process and software modules for visualization and data analysis, (iii) a case study on acoustic analysis of voice quality for the UNESCO labelled traditional Cantu in Paghjella. We have identified specific features for this singing style, such as changes in vocal quality, especially concerning the energy in the speaking and singing formant frequency region, a nasal vibration that seems to occur during singing, as well as laryngeal mechanism characteristics. These capturing and analysis technologies will contribute to define relevant features for a future educational platform.