Sample Rate of a Digital Audio Delivery Medium (i.e., CD)

On Tuesday, November 1, 2016 at 4:20:29 PM UTC-7, Randy Yates wrote:

(snip)

> Agreed. The problem is that some people may reasonably argue that the
> current CD standard is too limited in resolution and/or sample rate. As
> Tim can attest, some people can hear up to 24 kHz. Also I think in some
> special borderline cases >16 bits resolution could be useful. So an
> updated audio standard may be reasonable (for some value of "reasonable"
> ha ha).

I am not so convinced, but it isn't easy to test.

If you have a 24kHz source that isn't perfect, anywhere along 
the way, you might get some other components, maybe 16kHz or 12kHz,
that were audible.

But more to the point, even if you can hear 24kHz alone, you won't
hear it, at normal levels, mixed in with other musical signals.

I have done (just) a few experiments, filtering symphony orchestra
recordings through low-pass filters.  There really isn't all that
much up there.

The popular 20Hz to 20kHz for audio equipment means that it should
be nice and flat over the range that you actually hear. 

The highest key on an 88 key piano is 4186Hz. There could be
some second or third harmonic around 8kHz or 12kHz, but still
a long way from 20kHz.  And most of the music is much lower
than that.

Seems to me that the only reason for higher sampling rates
is to get better results in the lower (but still high)
frequencies.  But well designed resampling digital filters
should be able to do everything we need.

Note that when CDs were first designed, the electronics
for fancy digital filters weren't around.  Even more, they
used gas lasers, as semiconductor lasers weren't ready.

I am not sure if they knew that diode lasers would come
along, and make for affordable CD players, though I suppose
mass production of gas lasers would have brought the costs
down enough.