[LONG] RE: NTLK Re: Questions, questions...

From: Filmer, Paul E (pfilmer@nsf.gov)
Date: Sat Apr 29 2000 - 10:20:07 CDT

Warning! Warning! Fizzicist showing off!

This is called the "Nyquist limit" - think of it this way: say your sampling
rate was once every 24 hours (~125 milliHz!), and you sampled by looking out
your window. If you looked out at noon every day, you would conclude it was
light all the time. If you looked out at midnight, you would conclude it
was dark all the time. In fact, you might conclude that the sun doesn't seem
to move very much at all, no matter what time you choose to start your
sampling. You are missing anything that happens with a period shorter than
your sampling period - same happens for digitally recorded sound, and some
of the dropped data contains what out ears detect as the "difference"
between the various formats CD, MP3, DVD-A, and vinyl (analog).

It is interesting to note that this concept also applies to space as well as
time. Pixels on a screen represent samples of an image at regular intervals
over space, and so your screen resolution is subject to Nyquist's limit as
well - you need finer pixels to see finer detail - it seems obvious in this
case, because we deal with this phenomenon all the time.

An easy demonstration of how this becomes a problem is by selecting the font
Casual in 24 pt size - ugleee! You have to design fonts for each size of a
typeface so that they show up properly in each sampling frequency (in this
case, about 20 by 15 pixels/character). In a similar way, sounds at
particular frequencies will cause a problem for any sampling scheme, and the
cheap solution is simply to filter them out.

Paul F.

> -----Original Message-----
> From: gopi@sloth.org [SMTP:gopi@sloth.org]
> That's not "compression" in the digital data sense. And it is _not_
> similar to MD in any way at all. Since CDs are sampled at 44.1kHz, there
> is no possible way that you'll get a frequency higher than 22.5kHz.
> They're not "compressed" to remove this; it's simply a physical
> impossibility to record a higher sampling rate than that.
> gopi.
This NewtonTalk Message brought to you by:

EVOTE.COM, the ESPN of Politics on the
Internet. Visit EVOTE.COM for all the latest
news on Campaign 2000!

Visit http://www.evote.com today!


Need Subscribe/Unsubscribe Info?
visit http://www.planetnewton.com


This archive was generated by hypermail 2b29 : Thu Jun 01 2000 - 00:00:07 CDT