Richard Rauch wrote:
> On Wed, Nov 24, 2004 at 06:47:57PM +0100, Timo Sch?ler wrote:
>  [...]
>> cdrecord should complain on 48kHz .wav files... (it did for me, as i 
>> transitioned audio from DAT [48kHz] to cd-r...)
> 
> So what about dear old analog audio cassette tape?  Is there a
> theoretical maximum quality of a recording, played back on inexpensive
> consumer players?

There certainly exists a maximum quality level to an analog recording.  This 
is characterised mostly by the frequency response of the tape and playback 
equipment, and the signal-to-noise ratio.  Your typical audio tape has a S/N 
ratio of about 55 dB, and a frequency response that varies by more than 5 dB 
over the audible range.

> (I can readily record up to 48Kbps, but if recorded
> at that level, am I just retaining imperfections of the medium with
> higher fidelity?

That's about right.

> Or, if the tape is in good shape, is there always
> room for better sampling, due to the analog nature of the tape?)

Analog signals don't have infinite resolution.  This is a popular 
misconception that people have with digital versus analog recordings.  Most 
people learn something about how CD's work, and they compare a digital 
"stairstep" signal used in many examples versus the smooth curves of a sine 
wave and (correctly) decide that the two aren't comparible.

The crucial factor missing in the popular understanding of digital audio is 
that "stairstep" signal produced by the DAC's *must* be put through a 
so-called brickwall or reconstruction filter at the Nyquist frequency, which 
is half of the sampling rate.  In plain English, the "stairstep" signal with 
square edges contains very high freqencies which go away when you filter them 
properly.

-- 
-Chuck

PS: Apologies for drifting a bit off-topic.  Happy turkey day, all.  :-)