Thanks for the feedback everyone.
The sounds are from Sibelius piping its "mecanico" MIDI into NotePerformer. The advantage of NotePerformer is that it will take board-stiff expressionless notation score MIDI (dynamics, bowing techniques, etc.) and turn it into something that sounds like it was performed by humans by interpreting the dynamics, slurs, articulations, etc, and even introduce good-sounding timing discrepancies. The sounds themselves are physical models and not as realistic as the high-end orchestral samples, but the end result is more musical than if I had sent the same MIDI into VSL (which I have) or other high-end libraries (which I have as well). The tone would be richer, but it would sound very stiff.
I should take the time to play each part in, but that would easily make it 15 times the amount of work.