Brian,

The more I dig the sadder this situation becomes.
http://www.aes.org/sections/uk/meetings/0499.html

This was not a published paper but a meeting in the UK in 1999 with one Sony and one Philips guy. I would loved to have been at this event because I would have given them hell

Peter Eastty, Chief Consultant Engineer at Sony Brodcast Europe.

"In PCM coding,the sample rate is generally only about twice the required signal banndwidth (although DVD-Audio allows sample rates up to 192kHz)"

I can not believe what I'm reading here, this guy is questioning Shannon! Information theory states that the bandwidth MUST be at least twice the sampling rate to prevent aliasing.

Shannon's law (even I can understand this, and I just got O'Level math)

C=BW log2(1 + S/N)
BW = Bandwidth
S = Signal power
N = Noise power

Peter then goes on to say

"A DSD coded audio signal consists of a data stream at 2.8224Mbit/s"

Followed by "albeit with a bandwidth of 1.4MHz" Call me stupid but this guy has got it round the wrong way, the bandwidth is 2.8224Mbit/s and the information rate is 1.4112Mbit/s.

If Peter is to present a paper I need to be there to see him laughed off the stage.

Martin.