This is an incredible experiment in machinima. It's interesting to see where the original cinematography is still effective, and where it fails. Also, it really shows the weaknesses in the facial animation of the characters - my biggest problem with HL2, in that it falls right in the "uncanny valley" for me.

Those two issues I think are hand in hand, though - I felt that during the more intense moments of Cruise's character's dialogue there needed to be a change in framing to highlight the intensity. This was not needed in the original film, however, because the subtleties of the performances increased the intention within the frame. These subtleties, of course, are lost in the translation.

To Randall's credit, though, the performances he DID get out of these animations are pretty awesome, and as detailed as I've ever seen in a machinima. I applaud his and his collaborator's efforts!