[GnomeMeeting-list] Re: [Xmeeting-ohphonex] Free Software/Open Source Telephony-Summit in Germany

Hi Craig and Damien,

Was there ever a follow-up or wrapup session writeup, or conclusions from this conference? In testing the openh323-based/pwlib-based application ohphoneX (which is otherwise regarded as a blessing by this community), a study by a serious but friendly commercial MCU vendor (Codian) found the following defects, as listed below.

I'd like to know whether these problems are in the openh323 or pwlib layers, and thus common to GnomeMeeting, for example, or specific to the otherwise very nice ohphoneX implementation.

Apologies to Mark Blake for quoting his e-mail without permission, but I think these are things that ought to be addressed by the ohphoneX/GM developers and user community and should not have to be worked around by vendors:

On Jan 18, 2005, at 12:03 PM, Mark Blake wrote:

Although all the endpoint manufacturers attempt to adhere to the the various standards the quality with which they manage to do so varies tremendously from manufacturer to manufacturer. For instance of “the big two” endpoint manufacturers we have seen something like 10x more bugs with one than the other. Typically these are real bugs that are not just a matter of interpretation. However there is almost always a work-around of some sort - so our code has quite a lot of stuff that attempt to detect various bugs in real time and implement work-arounds!

In the specific case of OhphoneX we managed to find the following major bugs / issues in only about 15 minutes of testing

1) Bad H.263 bitrate advertisment

2) Inability to cope with some H.261 fragmented packets

3) Incorrect timestamps on H.261 media

4) H.261 encoder only uses I-macroblocks

5) OhphoneX does not correctly issue fast update requests in all cases.

So (1) is why H.263 calls do not work

(2) is what was causing the corrupt video with H.261 but (3) would also be sufficient to cause major problems even if (2) were fixed.

(4) is not a bug as such but means that OhphoneX is only using the most basic form of H.261 encoding which means it will need roughly 10x the bandwidth for the same quality of video compared to a “full” H.261 encoder.

(5) means that OhphoneX has poor resilience to packet loss / bad network conditions.

(2) and (3) are worked around in the forthcoming MCU build John previously mentioned. This is sufficient for H.261 video to work ok.

On Jan 11, 2005, at 6:03 PM, Craig Southeren wrote:

To all,

  The German Unix Users Group is holding the second annual Free
Software/Open Source Telephony-Summit n Geilenkirchen, Germany this
weekend and early next week. Myself and many other developers and users
will be attending the event. If anyone is interested in attending,
please see the web page at:



 Craig Southeren      craigs postincrement com / craigs voxgratia org

 Phone:  +61 243654666      ICQ: #86852844
 Fax:    +61 243673140      MSN: craig_southeren hotmail com
 Mobile: +61 417231046   Jabber: craigs jabber voxgratia org

 Post Increment - Consulting & Services    http://www.postincrement.com
 Vox Gratia - The Open Source VoIP portal  http://www.voxgratia.org
Raving Of A Strange Mind - the VoIP blog http://www.southeren.com/blog

The SF.Net email is sponsored by: Beat the post-holiday blues
Get a FREE limited edition SourceForge.net t-shirt from ThinkGeek.
It's fun and FREE -- well, almost....http://www.thinkgeek.com/sfshirt
Xmeeting-ohphonex mailing list
Xmeeting-ohphonex lists sourceforge net

:  Alan Sill, Texas Tech University   Office: Science 117, MS 41051  :
:  e-mail: Alan Sill ttu edu    ph. 806-742-3730   fax 806-742-1182  :
:  CDF New Office Bldg 52-102   ph. 630-840-3701  cell 806-790-7462  :

[Date Prev][Date Next]   [Thread Prev][Thread Next]   [Thread Index] [Date Index] [Author Index]