We present a novel algorithm for automatic music synthesis based on a physical model of the electric bass guitar. The main objective of this work is to integrate of 11 typical bass guitar playing techniques into the physical model. We focus on the 5 plucking styles finger- style (FS), picked (PK), muted (MU), slap-thumb (ST), and slap- pluck (SP) and the 6 expression styles normal (NO), vibrato (VI), bending (BE), slide (SL), harmonics (HA), and dead-note (DN).
For that purpose, all playing techniques are analyzed in terms of their physical influence on the sound that is produced on the instrument. Then, different parts of the model are designed to mimic these influences during the sound synthesis process.
The sound quality of the algorithm was evaluated using a MUSHRA listening test in the context of an audio coding application. The results of the test revealed a higher acceptance of the algorithms sound quality in comparison to other low bit rate coders. In another listening test a threshold of the audibility of frequency modulations in electric bass signals was determined with the help of the synthesis algorithm. The test was performed to find a detection threshold of modulation effects for analyzing and parametrization algorithms.