It's my understanding that, although the data for the 6th channel in DTS-ES(?) is in fact a discrete stream the 'center' (6th channel) info is ALSO matrixed into the surrounds for backward compatibility with older DTS decoders. This leaves the obvious problem that an ES decoder must remove the matrixed material from the surrounds, but the matrixing was done pre-compression, so the common signal still in the surrounds after compression will almost certainly not be the same as the 6th channel data after compression. This seems like a great chance to get some artifacting, given the right set of signals.

I can see the marketing advantage, and also the clear difference in trade-offs, but it would be hard for me to love either solution.

Part of good engineering is understanding and staying within design constraints. I wonder how DTS would sound at 450 kbps or what Dolby could cook up for 1.5 mbps.