This thesis examines the viability of audio descriptors within a synthe-sis context. It provides insight into acoustical modeling based on verbaldescriptions by quantifying the relationships between verbal attributes oftimbre and a set of audio descriptors. Various predictive models of verbalattribute magnitude estimation (VAME) are also tested. The results showthat is possible to create, classify and order sounds according to a verbaldescription. Finally, audio descriptive synthesis (AUDESSY) is introduced.This technique offers the possibility to synthesize and modulate sounds ac-cording to sonic morphologies, which are revealed by audio descriptors.