"Brenda Ann" wrote in message
.. .

"Brenda Ann" wrote in message
.. .

" David Eduardo wrote:
In fact, the Census Bureau fairly conclusively showed that the Census
could be done more accurately by a sample than a census... the problem
is the constitution requires, specifically, a census.


Say WHAT?????

I don't care how good your stats happen to be, there is no way in hell
that a sample of less than 100% can be as accurate, let alone MORE
accurate, than a sample of 100%. That's blowing smoke into anal crevices.
Period. That's the same warped logic that tries to convince people that
a 16Kb digital stream sounds as good as a 15KHz analog signal. There's
just not enough samples there to get an accurate representation of the
original analog. And I don't care what anyone says, there's no way that
digital will ever be 'as good as analog', let alone better, because to
get a perfect representation of the original analog waveform (especially
a complex waveform) you would have to have an infinite number of samples.



BTW, it doesn't matter that the human ear can make up for much of the loss
of proper waveform from digital. That's just working to the least common
denominator, and eventually you end up losing more information than the
human ear can make up for, since the lowest common denominator tends to
keep getting lower with each generation of a technology.

Per Mr. Orban... even including a link...

"Don't rely on the measurements you may be familiar with to evaluate ANY
bit reduction system incorporating a psychoacoustic model. The whole point
of
using the model is to throw away stuff that people can't hear anyway. Such
stuff is likely to appear on spectrum analyzers and the like, but
interpreting what your measurements mean (if anything) is non-trivial, to
say
the least.

There is an accepted objective method of measuring the performance of such
codecs (an ITU standard called PEAQ) but it isn't perfect, and, of course,
it
too must incorporate a psychoacoustic model.

Here's a bibliography on objective measurement of bit reduction systems
incorporating psychoacoustic models. Be warned -- this is heavy reading.

http://www.opticom.de/technology/literature.html