From: Tony H. (aamercado31@yahoo.com)
Date: Fri Feb 07 2003 - 04:16:08 GMT-3
HI
Not sure if i am getting into this kinda late...
I always had the impression to avoid rtp compression
unless you had total finish designing your qos but
congestions still exist on the network...then you
would need to use rtp compression. The rationale to
avoid it is that the overhead cost of cpu/memory is
unnessary (unless you still ran into
congestion)...also the extra config on both side can
lead to one way voice if somehow only one side is
config with rtp compression.
assuming g729 which is already compress..this should
be fine as additional compression will add on to (i
think) delay and contribute to a decrease in voice
qos.
on the next point, 729 should not be 30, i thought it
was 20 plus 4byte overhead=24 total.
Or am I totally off mark?
--- Mike Schlenger
<mschlenger@meridianitsolutions.com> wrote:
> ip rtp header-compression
>
> Investigate this command and see what it does for
> your application. It
> should take the IP/RTP/UDP header down from 40 bytes
> to 2-4 bytes.
>
> -Mike
>
> Michael Schlenger
> CCIE #7079
> Meridian IT/N2N Solutions
> mschlenger@meridianitsolutions.com
> 847.592.3912
>
>
> -----Original Message-----
> From: Lillemor Hamnqvist
> [mailto:lillemor@hamnqvist.org]
> Sent: Tuesday, February 04, 2003 2:04 PM
> To: ccielab@groupstudy.com
> Subject: OT: Voice compression
>
>
> Hi,
>
> I have a dedicated VoIP circuit, HDLC. What can I do
> to get the better
> compression? I'm using g729, but it looks like each
> call takes about 30kb.(I
> look at the load in the serial line). I would expect
> it to be lower. Is the
> load number in 'show interface serial' accurate ?
> What is your experience ?
> Thank you, /lillemor .
> .
>
This archive was generated by hypermail 2.1.4 : Sat Mar 01 2003 - 11:06:15 GMT-3