In article , Justin
Ulysses Morse wrote:

Mike Rivers wrote:

I haven't given this any thought, so I'll just throw out the question
to ponder. Is it possible that time resolution between components of a
complex wave could be better with a higher sampler rate? Could the
phase relationship between the fundamental and, say, third harmonic of
a distorted guitar, be more accurately preserved at a higher sample
rate?


If my understanding is correct, and it's based mostly on what I read
here from I think Arny (but it makes perfect sense to me), then the
answer to your question is that yes, a higher sample rate (or a higher
bit rate for that matter) would improve that time resolution; but that
the resolution is already WAY higher than necessary. People assume
that the time-domain accuracy is one sample period; but it's really one
sample period divided by the quantization range. So the time domain
accuracy for CD audio would be 1/(44100*2^16) or 346 picoseconds.
That's about a decimal place or two off of what I vaguely remember Arny
saying, so I've probably messed it up. But you get the idea? I
probably shouldn't be paraphrasing from memory, so go back and read
Arny's post from yesterday in one of these threads.


ulysses

And you wouldn't believe it, but dither actually affects the time domain
as well. No kidding. Little known, but true - it's similarly helpful in
the time domain as it is for amplitude. Most of the science geeks say
that time resolution also becomes essentially infinite with proper
dither. So Mike's concern is probably something that doesn't need to be
worried about as some say the limit doesn't exist in a well designed
system, and even the competing camp says it's so small as to not make a
difference. In this case it doesn't matter who's right since it just
plain doesn't seem to matter. There's a few stragglers left over that
say it's possible that imaging can be affected when you have
multichannel in play, but even the significance of this deserves some
skepticism as it is far from proven or agreed upon, and the "infinite
time resolution in a properly implimented system" guys actually seem to
have the science to back it up. My vote is with them.

--
Jay Frigoletto
Mastersuite
Los Angeles
promastering.com