A real string section has a different performance from each person, and I can't see a way to re-create that without recording a separate performance for each player. When I do that I think it's magical.
With a conventional string library, the natural variations amongst the musicians are already baked in. With DS that's not the case
That is exactly right and why I don't understand this weird attack on Dimension - it is a unique library in all sampling systems yet done - there is nothing else like it - and it became obvious to me the second I started using it the potential is huge, either using it alone or with the other VSL strings. It is truly like having each player of a string ensemble in your own room, playing exactly what you want them to play, with all their individual differences. And one may use separate tracks for each and every player, or just one track for all of them with humanize, so it has a lot of flexibility in how much you want to make of it.