That's very much true. It's a major issue for interoperatability. During the
sip interop (SIPIT-12) event we observed that having the media attributes
match (such as vif size, silence suppression etc.,) at both ends required a
manual intervention several times (reconfiguring) due to the confusion with
the current SDP definitions/implementations.

-Satya
Telogy Networks
A Texas Instruments Company


-----Original Message-----
From: Christer Holmberg [mailto:[EMAIL PROTECTED] 
Sent: Tuesday, March 25, 2003 1:21 AM
To: Hearty, John
Cc: Attila Sipos; [EMAIL PROTECTED]; link.xie; Christian Stredicke;
[EMAIL PROTECTED]
Subject: Re: [Sip-implementors] SDP packet times for muliple codecs



Hi,

I would suggest taking this question to the mmusic list. This could affect
many codec related attributes, and not only ptime, so I think it is an
important issue.

Regards,

Christer (lmfchh)


"Hearty, John" wrote:

> Christer,
>
>  I believe Attila is correct.  To do it the way you suggest would mean 
> each ptime attribute would need to be position sensitive to match up 
> with the appropriate codec.  This behavior is not described anywhere 
> in RFC2327, and would need to be for implementations to be 
> interoperable.  There are also some mentions of ptime handling in 
> RFC3264, and they do not suggest this behaviour.  Those mentions 
> appear to indicate one ptime attribute associated with one stream.
>
>  This issue has been discussed a few times before on mmusic, it may be 
> more appropriate to continue the discussion there if needed.  I think 
> the bottom line is most implementations probabally associate one ptime 
> attribute to one m line.  The only way to get different ptimes for a 
> stream is to omit the attribute completely when specifying multiple 
> codecs with different default times, such as G711 & G723.  May have to 
> wait for sdpng to do this as desired.
>
> John Hearty
> Level3
>
> > -----Original Message-----
> > From: Christer Holmberg [mailto:[EMAIL PROTECTED]
> > Sent: Sunday, March 23, 2003 11:20 AM
> > To: Attila Sipos
> > Cc: [EMAIL PROTECTED]; link.xie; 'Christian Stredicke'; 
> > [EMAIL PROTECTED]
> > Subject: Re: [Sip-implementors] SDP packet times for muliple codecs
> >
> >
> >
> > Hi,
> >
> > Which "rules" are you refering to? RFC2327 does say ptime is a media 
> > attribute, but I can't find anywhere said that you can't specify 
> > separate values for specific codecs (using it together with rtpmap).
> > Please tell me if
> > I'm wrong.
> >
> > Regards,
> >
> > Christer Holmberg
> > Ericsson Finland
> >
> >
> > > I think the problem is that (according to the rules) the ptime 
> > > attribute is not related to a specific profile but to a specific 
> > > media description:
> > > > m=audio 10000 RTP/AVP 0 8
> > > > a=rtpmap: 0 PCMU/8000
> > > > a=ptime:20
> > > > a=rtpmap: 8 PCMA/8000
> > > > a=ptime:30
> > >
> > > So I guess, officially, you can only really have one ptime 
> > > parameter associated with each "m=" line. Personally I can't see 
> > > why we couldn't use multiple packet times in the way you suggested 
> > > but maybe it'll break some implementations. I don't know.
> > >
> > > As a recap, I remember the various other proposals to
> > handle the multiple
> > > profiles:
> > >   m=audio 10000 RTP/AVP 0 8
> > >   a=mptime:20 30
> > >   a=rtpmap: 0 PCMU/8000
> > >   a=rtpmap: 8 PCMA/8000
> > >
> > > or:
> > >   m=audio 10000 RTP/AVP 0 8
> > >   a=x-ptime:20 30
> > >   a=rtpmap: 0 PCMU/8000
> > >   a=rtpmap: 8 PCMA/8000
> > >
> > > I have another idea:  be explicit about the packet times used for 
> > > each profile (like with rtpmap):
> > >
> > >   m=audio 10000 RTP/AVP 0 8
> > >   a=rtpmap: 0 PCMU/8000
> > >   a=rtpmap: 8 PCMA/8000
> > >   a=ptimemap: 0 20
> > >   a=ptimemap: 8 30
> > >
> > > But, as you can see, it's less efficient that the mptime or 
> > > x-ptime parameters.
> > >
> > > Can anyone say how far away we are before all this "mulitple 
> > > packet time" stuff is standardised?
> > >
> > > Attila
> > >
> > > > -----Original Message-----
> > > > From: Christer Holmberg 
> > > > [mailto:[EMAIL PROTECTED]
> > > > Sent: 19 March 2003 22:43
> > > > To: [EMAIL PROTECTED]
> > > > Cc: link.xie; Attila Sipos; 'Christian Stredicke';
> > > > [EMAIL PROTECTED]
> > > > Subject: Re: [Sip-implementors] SDP packet times for
> > muliple codecs
> > > >
> > > >
> > > >
> > > > Hi,
> > > >
> > > > Why don't you use normal RTPMAP stuff, without any extensions 
> > > > and/or x- parameters?
> > > >
> > > > Your SDP would look like:
> > > >
> > > > m=audio 10000 RTP/AVP 0 8
> > > > a=rtpmap: 0 PCMU/8000
> > > > a=ptime:20
> > > > a=rtpmap: 8 PCMA/8000
> > > > a=ptime:30
> > > >
> > > > Or, have I missunderstood your issue?
> > > >
> > > > Regards,
> > > >
> > > > Christer Holmberg
> > > > Ericsson Finland
> > > >
> > > >
> > > >
> > > > [EMAIL PROTECTED] wrote:
> > > >
> > > > > Hi all,
> > > > >
> > > > > I think Link.xie is right that till the attribute is 
> > > > > registered 'X-mptime' should be the attribute name used.
> > > > >
> > > > > Though, we should move towards standardizing this attribute in 
> > > > > a draft.... because chances are other implementations will not 
> > > > > understand this unregistered attribute and may lead to 
> > > > > inter-op issues later. PacketCable seems to have taken a step 
> > > > > forward already in this regard.
> > > > >
> > > > > PS: Some more digging into the archive shows that we did
> > > > >     agree to bring out a draft on the 'mptime' attrib
> > > > >     for exactly the same reason.... Perhaps Kevin and
> > > > >     Colin could tell if one is already on its way :)
> > > > >     Else we can prepare one now ?
> > > > >
> > > > > (ref:
> > > > >
> > > > http://www1.ietf.org/mail-archive/working-groups/mmusic/curren
> > > > t/msg01030.ht
> > > > > ml)
> > > > >
> > > > > Cheers,
> > > > > Siddharth.
> > > > >
> > > > > --------------------------------------------------------
> > > > > Siddharth Toshniwal
> > > > > Hughes Software Systems
> > > > > Prestige Opal                  http://www.hssworld.com
> > > > > 146, Infantry Road          Ph (O): +91-80-2286390 (7094)
> > > > > Bangalore-560001, India           Mobile: +91-9845154068
> > > > > --------------------------------------------------------
> > > > >
> > > > > "link.xie" <[EMAIL PROTECTED]> on 03/12/2003 04:49:12 PM
> > > > >
> > > > > To:   "Attila Sipos" <[EMAIL PROTECTED]>, "'Christian
> > > > Stredicke'"
> > > > >       <[EMAIL PROTECTED]>
> > > > > cc:   [EMAIL PROTECTED] (bcc: Siddharth J
> > > > Toshniwal/HSSBLR)
> > > > >
> > > > > Subject:  Re: [Sip-implementors] SDP packet times for
> > muliple codecs
> > > > >
> > > > > I think X-mptime is better. Cause in rfc2327,  "Unregistered 
> > > > > attributes should begin with "X-" to
> > > > prevent inadvertent
> > > > > collision with registered attributes"
> > > > >
> > > > > ----- Original Message -----
> > > > > From: "Attila Sipos" <[EMAIL PROTECTED]>
> > > > > To: "'Christian Stredicke'" <[EMAIL PROTECTED]>; "'link.xie'" 
> > > > > <[EMAIL PROTECTED]>; "Attila Sipos" <[EMAIL PROTECTED]>
> > > > > Cc: <[EMAIL PROTECTED]>
> > > > > Sent: Wednesday, March 12, 2003 4:44 PM
> > > > > Subject: RE: [Sip-implementors] SDP packet times for
> > muliple codecs
> > > > >
> > > > > >
> > > > > > Hello all,
> > > > > >
> > > > > > Link Xie, thanks for your response:
> > > > > > > >     m=audio 10000 RTP/AVP 0 8
> > > > > > > >     a=X-ptime: 20 30
> > > > > >
> > > > > > Following Siddharth's link
> > > > > >
> > > > >
> > > > http://lists.cs.columbia.edu/pipermail/sip-implementors/2003-F
> > > > ebruary/00455
> > > > > 6
> > > > > > .html
> > > > > >
> > > > > > It seems that from this link the answer is to use "mptime":
> > > > > >
> > > > > > >So instead of redefining ptime, create a new "mptime"
> > > > parm that has a
> > > > > > >construct similar to the m line:
> > > > > > >
> > > > > > >m=audio 49170 RTP/AVP 0 4 18
> > > > > > >A=mptime:10 30 30
> > > > > > >
> > > > > >
> > > > > > The packetcable spec (from
> > > > http://www.packetcable.com/specifications/ )
> > > > > >
> > > > http://www.packetcable.com/downloads/specs/PKT-SP-EC-MGCP-I06-
> > > > 021127.pdf
> > > > > > also seem to use mptime:
> > > > > >
> > > > > > (taken from PacketCable(tm) Network-Based Call
> > Signaling Protocol
> > > > > > Specification)
> > > > > >
> > > > > > m=audio 3456 RTP/AVP 96 97 0
> > > > > > a=rtpmap:96 G726-32/8000
> > > > > > a=rtpmap:97 telephone-event/8000
> > > > > > a=mptime: 10 - 10
> > > > > >
> > > > > > Here, rfc2833 doesn't have a packet time so it's
> > shown with a '-'
> > > > > > (rfc2833 is supposed to have the same packet time as
> > the selected
> > > > > > audio codec).
> > > > > >
> > > > > >
> > > > > > Which is correct mptime or X-ptime?
> > > > > >
> > > > > >
> > > > > > Regards,
> > > > > >
> > > > > > Attila
> > > > > >
> > > > > >
> > > > > > > -----Original Message-----
> > > > > > > From: Christian Stredicke [mailto:[EMAIL PROTECTED]
> > > > > > > Sent: 12 March 2003 08:31
> > > > > > > To: 'link.xie'; 'Attila Sipos'
> > > > > > > Cc: [EMAIL PROTECTED]
> > > > > > > Subject: AW: [Sip-implementors] SDP packet times for
> > > > muliple codecs
> > > > > > >
> > > > > > >
> > > > > > > The "X-" is a bad idea. We were using this in another 
> > > > > > > scenario and many SDP parsers complained about "x-" 
> > > > > > > headers. Better use something without
> > > > > > > "x-" (e.g. "xptime").
> > > > > > >
> > > > > > > Christian
> > > > > > >
> > > > > > > > -----Urspr
> > > > >
> > > > >
> > > > --------------------------------------------------------------
> > > > ----------
> > > > >
> > > > > �ngliche Nachricht-----
> > > > > > > > Von: [EMAIL PROTECTED]
> > > > > > > [mailto:sip-implementors-
> > > > > > > > [EMAIL PROTECTED] Im Auftrag von link.xie
> > > > > > > > Gesendet: Mittwoch, 12. M�rz 2003 03:46
> > > > > > > > An: Attila Sipos; [EMAIL PROTECTED]
> > > > > > > > Betreff: Re: [Sip-implementors] SDP packet times for
> > > > muliple codecs
> > > > > > > >
> > > > > > > > Hi,Attila
> > > > > > > >     I think your suggestion is wrong. We also use
> > > > this solution
> > > > > > > before.
> > > > > > > > But we find that the meaning of 2 "m=" line repesent
> > > > 2 media stream.
> > > > > > > So
> > > > > > > > it's not suitalbe for this case.
> > > > > > > >     Our solution is introducing a new attribut 
> > > > > > > > "X-mptime"
> > > > > > > which means
> > > > > > > > multiple ptime
> > > > > > > >    and the SDP would like:
> > > > > > > >
> > > > > > > >     m=audio 10000 RTP/AVP 0 8      ----- 0 means g711u-law
> > > > > > > and 8 means
> > > > > > > > g711a-law
> > > > > > > >     a=X-ptime: 20 30                          -----20 for
> > > > > > > g711u 30 for
> > > > > > > > g711a
> > > > > > > >
> > > > > > > >     I think this can solve your problem
> > > > > > > >
> > > > > > > > Best Regard
> > > > > > > > Link.Xie
> > > > > > > > ----- Original Message -----
> > > > > > > > From: "Attila Sipos" <[EMAIL PROTECTED]>
> > > > > > > > To: <[EMAIL PROTECTED]>
> > > > > > > > Sent: Tuesday, March 11, 2003 7:33 PM
> > > > > > > > Subject: [Sip-implementors] SDP packet times for
> > > > muliple codecs
> > > > > > > >
> > > > > > > >
> > > > > > > > >
> > > > > > > > > Hello everyone,
> > > > > > > > >
> > > > > > > > > For multiple codecs we can have an SDP like this:
> > > > > > > > > m=audio 10000 RTP/AVP 0 8      ----- 0 means
> > > > g711u-law and 8 means
> > > > > > > > g711a-law
> > > > > > > > > a=ptime:20
> > > > > > > > > But this is only correct if the packet times match for 
> > > > > > > > > u-law and a-law.
> > > > > > > > >
> > > > > > > > >
> > > > > > > > > What if you have a system where, for some reason, the 
> > > > > > > > > codecs have different packet times?
> > > > > > > > >
> > > > > > > > > Can you do this?
> > > > > > > > > m=audio 10000 RTP/AVP 0 -             ---- g711-ulaw
> > > > > > > > > a=ptime:20
> > > > > > > > > m=audio 10000 RTP/AVP 8               ---- g711 a-law
> > > > > > > > > a=ptime:30
> > > > > > > > >
> > > > > > > > > Regards,
> > > > > > > > >
> > > > > > > > > Attila
> > > > > > > > >
> > > > > > > > > <http://www.vegastream.com>
> > > > > > > > > VegaStream : A World of difference for your Integrated
> > > > > > > Communications
> > > >
> >
> > _______________________________________________
> > Sip-implementors mailing list [EMAIL PROTECTED]
> > http://lists.cs.columbia.edu/mailman/listinfo/sip-implementors
> >

_______________________________________________
Sip-implementors mailing list
[EMAIL PROTECTED]
http://lists.cs.columbia.edu/mailman/listinfo/sip-implementors
_______________________________________________
Sip-implementors mailing list
[EMAIL PROTECTED]
http://lists.cs.columbia.edu/mailman/listinfo/sip-implementors

Reply via email to