Dear TREGers,

Many, many thanks to all those who replied to my recent queries about
trans-hybrid loss. Your answers helped us to "stick to our guns" with a
customer, who now also fully agrees with us, and has apparently solved some
software difficulties that were causing poor near-end echo cancelation. It seems
that we had already done just about all we could to ensure that the THL for our
design would be no worse than others.

I have some questions relating to the ability of (half-duplex) group 3 facsimile
modems (V.17/V.29/V.27ter) and (full-duplex) data modems (V.34/V.32bis, etc.) to
withstand short drop-outs. I suppose that any of you who develop modem DSP
software, or who work with those who do, may be able to offer advice on this
subject.

Can any of you provide a definitive answer as to how these modems will behave in
the event of the speech connection between two such modems being occasionally
lost for 1 millisecond? By "occasionally", I mean once per call, or possibly
once every few minutes, at a random time after the speech connection has been
established between the two subscribers' equipment. I am only referring to a
speech drop-out, with the DC connection between each modem and its telephone
exchange being unaffected. During the drop-out period, there is silence.

If you feel that a 1 millisecond break could cause serious problems, how short
could the interruption be without causing the modems to drop the connection, or
to cause more than a few seconds' delay before resynchronization and data
channel recovery occur?

What is/are the mechanism(s) that causes the connection to fail in the case of
such drop-outs?

Clearly, data are lost during such an interruption, but the question is whether
the connection will recover after such an interruption, and if so, how long it
would take for the modems to resynchonize, and how far-reaching the
repercussions might be. Our interest is in how standard quality modems on the
market in the described circumstances, rather than how some unusually
well-designed modem happens to work.

>From our own experience with facsimile modems, the modems are normally able to
resynchronize and in most cases the impact will be retransmission of V.21
handshaking or a block of image data at V.17/V.29/V.27ter. However, I have not
done an exhaustive study on the abilities of the various fax modems (and their
controlling software) to cope with an impairment of the type described.

We have no direct experience with V.34 and other data modems, except that we
have found that if someone picks up a telephone on the same extension to which
one of our V.34 modems is connected, it is virtually guaranteed that the
connection will rapidly be dropped altogether.

Best regards,

David Drori

-----------------------------------------------------------
Novarex Enterprises Ltd.,
POB 2833,
306-6, Mevo Hapartisanim Street,
Jerusalem 91028,
Israel.
Tel: +972 2 5810995
Fax: +972 2 5813750
-----------------------------------------------------------

Reply via email to