Yes pitch-shifting is the best way to do it and why VSL did not sample 2nd versions of any of the instruments except for the ones that are variations in playing quality or instrument, like the second flute or lyric piccolo.
BTW you not only can do this pitch-shifting approach with divisi, but you really should if you want the sound to be accurate. In other words, simply using a solo clarinet for two clarinets when you only need one line, and then splitting it into two solo instruments when you need two, is a computerized cheat that you can probably get away with sometimes but is not (as you noted) like the real thing. Also, you might better use the ensemble clarinets, or other ensemble instruments if you are going from 3 part divisi to unisons with 3 players, and have them all on separate tracks so that the stereo image size is adjusted properly for the audible size of three players vs. a solo.
I have been a little obsessed with this divisi accuracy subject, as I have been doing some stuff with divis strings and orchestral horns, which very often go from unison to 2 notes to 4 notes. This absolutely requires extra steps like the pitch shifting, since there are no 2 player horn samples (and they are not really needed all that much).
In strings it is much more complicated, and doing a Debussy piece accurately could easily drive someone over the edge as the numbers of the Appassionata, Orchestral and Chamber ensembles do not necessarily match what you need in divisi. Though here, you can indeed cheat successfully as going from for example Appassionata strings unison to Orchestral 2 note divisi works very well. If you have 4 notes, you will certainly need the Chamber strings as the sound begins to be too thick. If you have even more (as in Debussy's Images) you would need to bring in a combination of Solo and Chamber parts. I had divisi in one piece I did (Apotheosis) that was 16 note in violas-violins and used only solos panned to approximate the stereo spread of the Appassionata unison violas-violins.
