I'll continue the spirit of topposting...

When NPR went on the birds I think they were using EBU compression. That is
essentially a floating-point encoding of the waveform with 10 or 11 bits of
precision sent in one millisecond blocks with a scale factor [exponent] for
each block. There were 31 kilosamples per second??? Didn't NPR change to a
method - unknown to me - which claimed a much lower the bit rate? I'd think
that EBU, with its low sampling rate and only 10-bit encoding, would have
real artifacts. This is certainly destructive encoding since you can only
approximately recreate the input, more nearly so than MP3, I'll grant.

One Dalet system I know of uses MPEG 2 encoding with a 256 (?) kilobit per
second rate.

Bob C.



"Bob Haberkost" wrote in message
...
There are two kinds of codecs....destructive and (you might figure)
non-destructive.

..WAV files are encoded with a non-destructive codecs...basically a
1-for-1 ratio of
bits to the waveform's voltage at the instant of sampling. The digital
encoding for
NPR is, at most, companded, but this, like Dolby, can be undone....thus,
non-destructive encoding.

MP3s and methods like IBOC or, from what I've heard, Canadian DAB radio,
have
encoding artifacts due to the effort used to eliminate redundant (and I
guess the use
of that word is a judgement call, considering what IBOC concludes is
redundant) leave
out too much, and thus when the encoded audio is reconstituted, it's
missing stuff.
Sortof like freeze-dried ice cream.

And from what I know about the one production audio system, Dalet, this
implementation uses .wav files for storage. So it's about as good as the
source, and
thus quite suitable for air. Even these jukeboxes you speak of use a
codecs which
result in larger files than MP3s at the same sample rate, but the audio's
quite
unaffected by the process. MP3s and other internet and broadcast
streaming methods
need to cut down the data rate to levels that can't possibly allow true
high fidelity
or even a facsimile of it.
--
For direct replies, take out the contents between the hyphens. -Really!-


"R J Carpenter" wrote in message
...

"Rich Wood" wrote in message
...

I have both Sirius and XM. I agree that the quality depends on the
streams. Classical and Jazz seem to be the highest quality. For analog
I listen to WFCR, Amherst, MA. It's NPR, Classical and Jazz. Very
lightly processed.

And how does NPR get to them? Isn't there a little bird up there
chirping
digits to all the NPR affiliates?

I presume they do their music locally from CDs - but haven't the CDs
been
compressed onto some automation system's hard disk...?

Bob C.