Dear Simon, H323plus members,

Thanks for your reply.

One thing that I am not understand is why it happened at the packets which are sent by polycom site, but the bug was fixed at our site(h323plus) system?

Thanks. On Wed, Feb 23, 2011 at 7:25 PM, Simon Horne s.horne@packetizer.com wrote:

Kc

Yes you are correct there was a bug with the timestamp on the marker packet being incorrect, that should have been fixed in the last release.

Update to the latest release.

That should also help resolve some issues with packets being discarded.

Simon

*From:* h323plus-bounces@lists.packetizer.com [mailto: h323plus-bounces@lists.packetizer.com] *On Behalf Of *kc ooi *Sent:* 23 February 2011 19:54 *To:* h323plus@lists.packetizer.com" *Subject:* [h323plus] RTP Packet Issue

Dear H323plus Team members,

Recently, I have found some issue when I am doing H.323 testing with polycom system (eg. PVX 8.0.2) and checked the packet traffic using wireshark.

I use RTP stream analysis of wireshark for viewing the result, and I notice that every packet(which is final slice of a frame, contain Marker=1) has a status of "Incorrect timestamp".

1. Why it behave like this?

2. Does it affect the receiving site during a conference meeting?

3. Is there any timeout for RTP packet handling(like dropping packet) in

h323plus? Because I found out that it cause large packet loss during the conference compare to other h.323

system on the same situation(ie. same executing environment).

Btw, I have done some enhancement on MyOutput display according to the suggestion of "Simon" to one of the h323 member, to avoid dropping packet.

Thanks.

Dear Simon, Jan, and H323plus member,

So how it will affect the system during conference as I am not manage to notice any abnormal behavior during display? Thanks. On Thu, Feb 24, 2011 at 5:07 AM, Jan Willamowius jan@willamowius.de wrote:

Hi,

using the latest H323Plus, I see the same timestamp error on video packets with a marker in Wireshark, but I see it in both directions (Polycom PVX to H323Plus and reverse).

Regards, Jan

Simon Horne wrote:

...

Kc

Yes you are correct there was a bug with the timestamp on the marker

packet

...

being incorrect, that should have been fixed in the last release.

Update to the latest release.

That should also help resolve some issues with packets being discarded.

Simon

From: h323plus-bounces@lists.packetizer.com [mailto:h323plus-bounces@lists.packetizer.com] On Behalf Of kc ooi Sent: 23 February 2011 19:54 To: h323plus@lists.packetizer.com" Subject: [h323plus] RTP Packet Issue

Dear H323plus Team members,

Recently, I have found some issue when I am doing H.323 testing with polycom system (eg. PVX 8.0.2) and checked the packet traffic using wireshark.

I use RTP stream analysis of wireshark for viewing the result, and I

notice

...

that every packet(which is final slice of a frame, contain Marker=1) has

a

...

status of "Incorrect timestamp".

1. Why it behave like this?

2. Does it affect the receiving site during a conference meeting?

3. Is there any timeout for RTP packet handling(like dropping packet) in

h323plus? Because I found out that it cause large packet loss during the conference compare to other h.323

system on the same situation(ie. same executing environment).

Btw, I have done some enhancement on MyOutput display according to the suggestion of "Simon" to one of the h323 member, to avoid dropping

packet.

...

Thanks.

-- Jan Willamowius, jan@willamowius.de, http://www.gnugk.org/

Simon,

I haven't seen any negative effects either and since Polycom generates similar packets there is a good chance that most endpoints will accept the RTP as is.

But Wireshark is usually pretty good at analyzing network packets and if somebody would be able to say why it is complaining, we should fix it.

BTW: Wireshark also complains about the length of H323Plus' RTCP packets and claims they are 4 bytes short. I think there might be some mandatory padding missing, but I'm not sure.

Regards, Jan

Simon Horne wrote:

Guys

I'm a little confused. I'm not seeing the problem on h323plus CVS. Marker is definitely on the last packet for a given timestamp.

Anyway the issue had been there for quite some time without causing any issue.

The fix was added only when I was looking for something unrelated in wireshark.. J

Simon

From: h323plus-bounces@lists.packetizer.com [mailto:h323plus-bounces@lists.packetizer.com] On Behalf Of kc ooi Sent: 24 February 2011 11:56 To: Jan Willamowius Cc: h323plus@lists.packetizer.com Subject: Re: [h323plus] RTP Packet Issue

Dear Simon, Jan, and H323plus member,

So how it will affect the system during conference as I am not manage to notice any abnormal behavior during display?

Thanks.

On Thu, Feb 24, 2011 at 5:07 AM, Jan Willamowius jan@willamowius.de wrote:

Hi,

using the latest H323Plus, I see the same timestamp error on video packets with a marker in Wireshark, but I see it in both directions (Polycom PVX to H323Plus and reverse).

Regards, Jan

Simon Horne wrote:

...

Kc

Yes you are correct there was a bug with the timestamp on the marker

packet

...

being incorrect, that should have been fixed in the last release.

Update to the latest release.

That should also help resolve some issues with packets being discarded.

Simon

From: h323plus-bounces@lists.packetizer.com [mailto:h323plus-bounces@lists.packetizer.com] On Behalf Of kc ooi Sent: 23 February 2011 19:54 To: h323plus@lists.packetizer.com" Subject: [h323plus] RTP Packet Issue

Dear H323plus Team members,

Recently, I have found some issue when I am doing H.323 testing with polycom system (eg. PVX 8.0.2) and checked the packet traffic using wireshark.

I use RTP stream analysis of wireshark for viewing the result, and I

notice

...

that every packet(which is final slice of a frame, contain Marker=1) has a status of "Incorrect timestamp".

1. Why it behave like this?

2. Does it affect the receiving site during a conference meeting?

3. Is there any timeout for RTP packet handling(like dropping packet) in

h323plus? Because I found out that it cause large packet loss during the conference compare to other h.323

system on the same situation(ie. same executing environment).

Btw, I have done some enhancement on MyOutput display according to the suggestion of "Simon" to one of the h323 member, to avoid dropping packet.

Thanks.

-- Jan Willamowius, jan@willamowius.de, http://www.gnugk.org/

KC

It looks like some multiplexing where 2 different RTP streams are sent over the same connection. No idea why. PVX sends audio tones embedded in the audio channel but I don't know what would be sent on the video channel.

PVX does some strange things. Last week I committed a number of custom patches to get H.239 interworking fixed as PVX does it's own strange logic.

Simon

From: h323plus-bounces@lists.packetizer.com [mailto:h323plus-bounces@lists.packetizer.com] On Behalf Of kc ooi Sent: 02 March 2011 17:42 To: h323plus@lists.packetizer.com" Subject: Re: [h323plus] RTP Packet Issue

Dear H323plus Team members,

Recently, I have done H.323 testing (simple conference without presentation channel opened) and captured the video packets using wireshark.

I found that when both endpoints are polycom PVX system, there are extra video packets (ie. coded slice data partition A, coded slice of an IDR picture, and subsequent ..,) with difference with dynamic payload type value (eg. actual video payload type is 109, it has payload type 99) tranferring to each other site during conference.

It will not happened if either site using other H.323 system(not polycom system).

What is the purpose of this packet as It seems like the received site can form even a complete picture?

Thanks.