Quote:
Originally Posted by Roc Mixwell  Actually, it is, but you don't need to understand any of it to hear a change, with ALL KNOWN VARIABLES ACCOUNTED FOR. Anyway, I'm not going to carry on here. I just don't think you're going to be able to prove anything in either camp without being in the same room with each other. And even then results are inconclusive between testers. I've seen it happen, but best believe, I'm not going to degrade my experience no matter what. |
I have a [day job] client who is into racing sailboats. One of the members of his team is a brain surgeon. People are understandably fascinated.
When folks ask about it, he tells them, "Brain surgery? Well... you know... it's not
rocket science."
FWIW, Roc, I'm not trying to
prove anything
. What folks prefer is what they prefer. (At least when they actually
know they prefer it. See rant on double blind testing.) I'm just trying to make sure that folks have the generally accepted
facts on the increased jitter likely to result from slaving a converter to an external clock source.*
(Of course, setting
one master clock source and slaving all other clocks to it
is absolutely necessary when yoking multiple converters together. Still, it
can be better, in some respects, even there, to use the
internal clock of one of the converters as the master; since, in that circumstance, at least
one converter will not have internal jitter increased by attempting to maintain sync with an outside clock source.)
*BTW, it's not that difficult to get confused about what a certain external clock maker says about their jitter correction technology. If one is less than
rabbinical in his approach to some of that company's representatives' comments, one can get the false impression that they are claiming their jitter correction technology can somehow clean up the bad internal clocking of converters slaved to it. However, all a dedicated wordclock generator can really do is put out accurate, pristine clock signal, nothing more.
It
cannot pass any positive voodoo down the clock wire. A given dedicated clock may well be able to
clean up incoming clock signal, removing jitter and making it a better clock signal -- and when necessary that ability should be
golden -- but that juju is
internal. What it passes out to slaved clients is, at best, simply good, clean clock signal. If those slaved converters have bad clock circuits that can't keep their
own internal time it's
very likely that they will
also do a bad job of the much harder task of keeping sync with an outside clock signal. No matter how clean that external clock is.