BG> Russell posted these stats in USR_Modems today, and
BG> I've forwarded them to you just in case you missed them.

I do read all that echo now, tho I can be a bit behind in it.

BG> The frequency response plot is particularly interesting,
BG> and assuming as near as dammit to 0dB loss for the link,
BG> that almost dead flat -10dBm Tx level is actually 5dB less
BG> than the modem's reported Tx level, so I presume that it
BG> had automatically adjusted itself accordingly (the curve is
BG> plotted right at the beginning of the connect, as you know).

Thats the way its supposed to work. The V34 standard requires
that the modem puts standard test out, and the one at the other
end of the line then measures what it gets, and whats different to
the standard must be due to the line and so a measure of the line.

The modems use THAT measure of the line to set a variety of params that
they can change to get the best out of THAT measured line characteristic,
and the Tx level is one of the values that can be varied.

In fact you would expect the actual Tx value used to be dropped in the
situation where there is no loss at all between the modems. When you go thru
any phone line and a pair of line cards you will always get some loss, in fact
its not meant to deliver what the modem puts out for complicated reasons.
@EOT:

---