From what I've heard, sample libraries of acoustic instruments are a system of presets and directed constraints (inability to replicate complex dynamics, articu

From what I've heard, sample libraries of acoustic instruments are a system of presets and directed constraints (inability to replicate complex dynamics, articulation, and pitch bends). I haven't tried physical modeling, and don't know if it manages to escape this issue, or ends up too complex to learn/perform or stuck in an uncanny valley. (This is from the perspective of orchestral-inspired sound design rather than more synthetic compositions.)