Abstract
A method has been developed for estimating the parameters of virtual musical instrument synthesizers to obtain isolated instrument sounds without distortion and noise. First, a number of instrument sounds are generated from randomly generated parameters of a synthesizer. Lowlevel acoustic features and their delta features are extracted for each time frame and accumulated into statistics. Multiple linear regression is used to model the relationship between the acoustic features and instrument parameters. Experimental evaluations showed that the proposed method estimated parameters with a best case error of 0.004 and signal-to-distortion ratio of 17.35 dB, and reduced noise to smaller distortions in several cases.
| Original language | English |
|---|---|
| Title of host publication | Proceedings - 40th International Computer Music Conference, ICMC 2014 and 11th Sound and Music Computing Conference, SMC 2014 - Music Technology Meets Philosophy: From Digital Echos to Virtual Ethos |
| Publisher | National and Kapodistrian University of Athens |
| Pages | 1426-1431 |
| Number of pages | 6 |
| ISBN (Print) | 9789604661374 |
| Publication status | Published - 2014 |
| Externally published | Yes |
| Event | 40th International Computer Music Conference, ICMC 2014, Joint with the 11th Sound and Music Computing Conference, SMC 2014 - Music Technology Meets Philosophy: From Digital Echos to Virtual Ethos - Athens Duration: 2014 Sept 14 → 2014 Sept 20 |
Other
| Other | 40th International Computer Music Conference, ICMC 2014, Joint with the 11th Sound and Music Computing Conference, SMC 2014 - Music Technology Meets Philosophy: From Digital Echos to Virtual Ethos |
|---|---|
| City | Athens |
| Period | 14/9/14 → 14/9/20 |
ASJC Scopus subject areas
- Music
- Media Technology
- Computer Science Applications