news:cfaj3n$e18$1@news5.jaring.my

Three words: AES preprint 4826.

More specifically:

AES Preprint 4826 discusses this on page 23. It says that the threshold of
audibility with a 20 kHz tone is 10 nanoseconds rms, and at 4 kHz the
threshold of audiblity is 100 nanaseconds rms.

However, at 4 kHz, the distortion products from 100 nanoseconds of jitter
are about one fifth the
amplitude as they are with  20 KHz and 100 nanoseconds of jitter.

IOW, the distortion products from 4 KHz and 100 ns of jitter are about the
same size as the ones from 20 KHz and 20 ns of jitter. This accounts for
about half of the observed diffference in perceptual threshold.

We can see that the perceptiion of jitter is approximately controlled by the
size of the garbage that the jitter creates, as it is reduced by masking in
the ear. The paper mentions masking by the musical program material.

Extrapolating this to lower frequencies suggests that at 20 Hz, 20,000 ns (2
uS) is the threshold of audibility for jitter.

We can conclude that the worst case is for musical content at 20 KHz, and at
that point the threshold of audibility is about 20 nanaseconds or 20,000 ps.

The natural basis for a jitter spec in ppm is the clock period, If the clock
frequency is 44 KHz, then the natural basis for a ppm-based spec is about 22
microseconds. 5 times 10 to the minus 6, times 22 times 10 to the minus 6 is
110 times 10 to the minus twelfth.  In 1/1000's the sequence goes seconds,
milliseconds, microseconds, nanoseconds, picoseconds.   5 ppm at 44 KHz  is
something like 110 picoseconds (ps).

Calculations equate that with approximately 110 picseconds (ps).

Per AES preprint 4826, the audibility of jitter varies with the frequency of
the signal being modulated by the jitter. But, their audibility figures are
given in nanoseconds which are 1000 times greater than picoseconds. The
paper suggests strongly that the ear is most sensitive to jitter as applied
to 20 KHz signals. At this point the threshold of audiblity is about 20,000
ps which is far less than 110 ps. IOW, 110 ps jitter is nit.

Let's look at this situation critically.

My statement of the situation:

A sighted, non-time-synched, non-level-matched listening *test*  wth no
other on-site technical tests backing it, is positive for audible
differences.

Somehow, the conclusion is reached that *any* and *all* perceived
differences are due to differences in jitter specs that in fact, may or not
be representative of the equipment being compared, AND is probably inaudible
by a factor of  100 or more.

What's wrong with this picture, logically-speaking?

With all due respect I'd say, just about everything is wrong with this
picture.

With all due respect I'd say that the  listening tests described have about
6 times more uncontrolled variables than it would take to make assigning any
technical cause to its outcome a pure and pristine, nearly science-free
wild-ass guess.