On Sun, 07 Nov 2004 09:27:06 +0000, Pooh Bear
wrote:


mick wrote:

I read somewhere (sorry, can't give a ref) that the error correction on
transports intended for audio is more lax than on those intended for data,
as your ears are incapable of detecting low error rates but are more
sensitive to the gaps caused by error correction. If that is so, then
using a data drive for audio may give a different sound, but not one that
is necessarily "better" as it will contain a different sort of inaccuracy!
I'm sure someone will be able to correct me on this if I'm wrong.

There's a lot of error correction capability on a CD. Hamming encoded IIRC.
Forget how many bits of error it can correct transparently. Philips / Sony
expected early CDs to have lots of errors so needed them to be correctable.

Bear in mind that it was expected that early CDs would *need* error correction.
I'm sure they are much better now.

I can't recall if the CD standard includes 'error concealment'. Anyone know ?

It does. Uncorrected errors may be concealed, i.e. the system takes a
'best guess' at what the missing sample(s) should have been, or
unconcealed 'mute' errors where the output is silenced - usually for
less than a millisecond. The general consensus is that with a standard
commercial CD you get one sub-millisecond concealed error about once
every five minutes, and somewhat less than one 'mute' error per disc.
Anyone here think that either of those will be audible? Otherwise, the
datastream is *perfect*.
--

Stewart Pinkerton | Music is Art - Audio is Engineering