Addition to the characteristics:
# Can be implemented (in file formats) by any audio coding scheme
(lossless PCM or lossy compressed schemes, different sample rates or
bit depths).
There's been discussion about head room and bitdepth ... is this
statement a bit too lose? ... or plain wrong?
Etienne