Since VoI is
basically what's been done for a long time at the CO,
only doing it at (or at least closer to) the phone, [...]. A
ploss-free VoI call _should_ be able to run 56K.
I'd suspect only 28.8K or
perhaps 33.6K, but not beyond that. As I
understand it, 56K is a function of the telco 64K sampling and if
you're doing 56K answer, it's done with the assistance of the local
CO--i.e., no consumer-grade modem will operate at 56 in answer mode.
Perhaps, but the reason you can't run at 56K is then not the fault of
the VoI call, but rather because you can't run 56K between two CPE
devices at all.
And I don't think 56K is done with the assistance of the local CO; I
think it's done by running a T1 (or moral equivalent) into a modem-bank
device at the ISP. Certainly that's what the ISP I work at did back
when they still ran a modem bank, though I don't recall whether that
particular bank did 56K. After all, once the end-user's CO digitizes
the signal, you can ship it halfway around the world and the only
effect will be increased RTT. (Telco circuits do not just drop little
chunks of data the way the Internet does; they're circuit-switched
rather than packet-switched, after all.)
/~\ The ASCII Mouse
\ / Ribbon Campaign
X Against HTML mouse at
rodents-montreal.org
/ \ Email! 7D C8 61 52 5D E7 2D 39 4E F1 31 3E E8 B3 27 4B