Note that I'm not saying there's a mismatch or that it flat out doesn't work! Jitter is much more subtle.
Lightpipe is a pretty jittery format, especially over longer runs, so it's best to clock both ends separately. In other words, lightpipe only sounds bad if something with converters syncs to the clock embedded in the datastream going through the lightpipe. If you ignore that clock and instead distribute word clock or digital black, it's a perfectly good format.
It's a good idea to use distributed clock in most - but not all - cases anyway, even when you're not using lightpipe. I tried to get those snobs on the Pro Audio list to explain why that is a few years ago, but they had their noses too far up their own you-know-whats and I got a private letter from the moderator saying that this was too basic a question for them. [:)] Maybe someone knows the answer?
What I do know is that the extra jitter in lightpipe is caused by refraction at the connectors. Also, some higher-end D/As have jitter attenuation built in (dual PLLs, usually), so they don't necessarily want to be clocked externally - which would make everything I'm saying invalid!
Bottom line, all I'm saying is that I'd rather run an audio cable than a lightpipe + a word clock cable to a speaker.