Errors / ambiguities / problems found in specs at last week's int erop

Paul Long plong at SMITHMICRO.COM
Wed Mar 1 13:51:24 EST 2000


Charles,

Maybe we should use G.723.1 for guidance. There is a single G.723.1
capability, although there are two rates (and two distinct encodings, unlike
G.729/G.729a) which are signaled in-band. All decoders are required to
support both rates. The only place in H.245 where the rates are
distinguished is in the RequestMode message.

It seems to me, however, that the decision to encode G.729 versus G.729a
should always be local, since the major impact, complexity and therefore CPU
usage, is local. Otherwise, the remote EP would always want to receive the
higher-quality audio of G.729.

Paul Long
Smith Micro Software, Inc.

-----Original Message-----
From: Hutton, Charles [mailto:charles.hutton at ATTWS.COM]
Sent: Wednesday, March 01, 2000 12:24 PM
To: ITU-SG16 at MAILBAG.INTEL.COM
Subject: Re: Errors / ambiguities / problems found in specs at last
week's int erop


Paul:

729 and 729A do indeed use the same encoder. Only the decoder is
different.
Performance of 729A is poorer, particularly with respect to tandeming.

I can see that it may sometimes be beneficial to control whether the 729
or
729A algorithm (if a choice is available) is used by the decoder in
order to
prevent significant degradation of audio quality. If that is a real
scenario, then it becomes more than "a local issue".

Perhaps others can comment on the "realness" of this possibility.


Charles (Chuck) Hutton
Strategic Architecture Engineering
Wireless Local Technology Group
AT&T Wireless Services
PO Box 97059
Redmond, WA 98073 (USPS only)
9461 Willows Road
Redmond, WA 98052 (FEDEX/UPS)
425-702-2938 Voice
425-702-2518 FAX
charles.hutton at attws.com



More information about the sg16-avd mailing list