A New Method of Onset and Offset Detection in Ensemble Singing
| Autor: | Freya Bailes, Sara D'Amario, Helena Daffern |
|---|---|
| Přispěvatelé: | D'Amario, Sara |
| Rok vydání: | 2018 |
| Předmět: |
Male
Offset (computer science) Phrase Time Factors Computer science Voice Quality Speech recognition Singing Pilot Projects Visual contact 030507 speech-language pathology & audiology 03 medical and health sciences Speech and Hearing 0302 clinical medicine Arts and Humanities (miscellaneous) Humans Phonation Overall performance 030223 otorhinolaryngology Interpersonal interaction Electrodiagnosis Signal Processing Computer-Assisted Acoustics LPN and LVN Legato Group Processes Female 0305 other medical science Algorithms |
| Zdroj: | Logopedics Phoniatrics Vocology |
| DOI: | 10.5281/zenodo.1241000 |
| Popis: | This paper presents a novel method combining electrolaryngography and acoustic analysis to detect the onset and offset of phonation as well as the beginning and ending of notes within a sung legato phrase, through the application of a peak-picking algorithm, TIMEX. The evaluation of the method applied to a set of singing duo recordings shows an overall performance of 78% within a tolerance window of 50 ms compared with manual annotations performed by three experts. Results seem very promising in light of the state-of-the-art techniques presented at MIREX in 2016 yielding an overall performance of around 60%. The new method was applied to a pilot study with two duets to analyse synchronization between singers during ensemble performances. Results from this investigation demonstrate bidirectional temporal adaptations between performers, and suggest that the precision and consistency of synchronization, and the tendency to precede or lag a co-performer might be affected by visual contact between singers and leader–follower relationships. The outcomes of this paper promise to be beneficial for future investigations of synchronization in singing ensembles. |
| Databáze: | OpenAIRE |
| Externí odkaz: |
|