Okay, I may be mistaken, but here goes....
DC offset is when a sound wave is not centred on the reference voltage of a signal. It's easy to see when you look at the waveform of a sound, it won't be centred (vertically)
****
<PRE>
no offset DC offset
_____________________________________________
.
. _
. / \ _
. / \ / \
.__/_____\_____________/___\_________________
. \ / _ _/_ _ _\ _ _ _ /_ _ _
. \ / \ /
. \_/ \ /
. \_/
.____________________________________________
</PRE>
****** (copy and paste the above mess into windows notepad, it'll make more sense)
It's a problem because it lowers dynamic range of the recording: the wave form can't get as loud because on side reaches it's maximum magnitude before the other, while the other side doesn't use its entire range. In a sample where there is no offset, both sides will reach their respective limits at around the magnitude, so there isn't any wasted range.
Generally, it's not useful to have DC offset, I don't know of anycase where it's useful.
Hope this was useful (and accurate!!)
William Underwood
[This message has been edited by cwillu (edited 11-24-1999).]