In article , Chris Hornbeck
wrote:

On Sun, 16 Nov 2003 22:57:17 GMT, Jay - atldigi
wrote:

The myth is the dynamic equivalent to the argument that 4 samples on a
20kHz sine wave will render it more accurately than 2, and 8 samples
even more so. That's not true either.

But 4 samples will render it more accurately in *time* than two.
Alternatively, a smaller quantization step will also. Quantization
could be said to "jitter" the conversion in either time or
amplitude.

This is a contention of some, but when the system is viewed as a whole
including proper dither, time resolution essentially becomes infinite.
People Bob Stuart and Tom Holman have suggested that the time issue is
significant for imaging if you are dealing with two or more channels.
Other digital audio heavy hitters and PhDs point out the seldom
understood fact that the right dither has an effect on the time
resolution as well as the preventing of truncation distortion we know
and love it for. So it is not agreed upon that 4 samples render it more
accurately in time. In fact, the science seems to be against it. It's
too bad JJ formerly from AT&T isn't around anymore to offer up all the
empirical data for us. I wouldn't mind Dick Piecrce or Dave Collins
making an appearance either.

None of this means that higher sample rates and more bits don't sound
better. It's just that the explanations often given are flawed, and the
requirements demanded are often excessive. 64kHz 20 bit sampling is
probably the minimum necessary. 96/24 offers a margin of safety. More
than that may actually cause more problems than it solves. As stated
before - upsamling for non linear processes, oversampling, and signal
processing with double precision are not the same subject and can be
beneficial.

--
Jay Frigoletto
Mastersuite
Los Angeles
promastering.com