IMHO, Rockwell's '56K standard' is vaporware.
28.8 is the real, demonstratable baud limit on analog lines. 33.6 is only
achieved by 'cheating', using the guard bands for data.
> Besides, if this turns out to be real then you are going to put all these
> devices on their own hunt group and charge extra for them anyway. Your
> users are not going to all wake up one day and go out and buy something
> that is certainly going to make the 28.8 modems look inexpensive by
> comparison all on the same day.
The PM-3 takes in calls via digital lines. A modem call from an analog
phone line comes in as 56K sampled data.
By Shannon's law(?), a local caller (one A->D conversion) actually can
push at most 28K through the pipe (1/2 the sampling rate). Slightly higher
rates can be achieved if the analog device matches the sampling rate at
the Central Office.
Calls from non-local users will often go through multiple A->D->A->D
conversions, and will lose some clarity, giving lower bandwidth.
56K uncompressed is technically impossible over a standard analog phone
line if any part of the PSTN path is digitized. It is possible on a
'dry pair', and MIGHT be possible if the route between the user and the
provider has no digital components, but the PM-3 requires digital input...
56K over analog isn't going to happen. Not for most vendors,
or for most users for that matter.