[RESOLVED]Loud noises in meetme conferences


I’m having a bit of a trouble with meetme conferences. People are reporting loud noises on the conference rooms. I recorded a audio sample of what can be heard when entering.


I also noticed sip channels stuck in the room where hit happens, although there were no real person there and the rtptimeout is set to 2 minutes. The loud noise only stops when I kick those ‘ghost’ channels.

My system has a peculiarity, I have two boxes, one acting like a normal PBX and the other only for conferences, with sip trunking linking the two.

My diaplan is something like this

[code]Box 1 (PBX)

exten => _1XX,n,HangUp()

Box 2 (meetme server)

exten => _1XX,1,MeetMe(${EXTEN},d)
exten => _1XX,n,HangUp()


System info:
Asterisk versions

I suggest the following:

  1. Try reducing the timeout. Does it kick out the ghost user earlier and improves results ?

  2. If possible try to upgrade you conference to confbridge-10. I think it is much more reliable and provides higher quality. Also it does not require external clock synchronization. Take a look at the following documentation https://wiki.asterisk.org/wiki/display/AST/ConfBridge+10

  3. Try using a noise reduction software like PBXMate.

Thank you valer77,

Let’s see:

1- The rtptimeout has no effect whatsoever, I guess because of the loud and constant noise generated being sent to all the channels.

2 - Interesting, I’ll look for that!

3 - The noise is extreme, It’s not ambient. Sometimes it sounds like an endless loop of a fraction of what someone said mixed with very loud radio static.

Any other ideas (while I’m reading about confbridge)?

Are you using IP phones? If so, which type.

I don’t have an m4a codec on my PC, so I’m not going to listen, but we discovered that Cisco phones don’t like it when the timestamp jumps but the synchronisation source doesn’t change. Setup was via CUCM, but the RTP stream was direct. The result was buzzing at the frame rate.

Unfortunately, we are using an old version of Asterisk, and the developer mailing list is silently discarding everything I post to it.

Incidentally, if you are recording strange noises you really should use a lossless codec.

Drop an e-mail to Kevin Fleming, he should be able to help you with that.


Drop an e-mail to Kevin Fleming, he should be able to help you with that.[/quote]

I was trying to follow this up with Bryan Johns, but my last email (Feb 17th) didn’t get a reply. I’ve tried forwarding to my best guess at the email address for Kevin, and it hasn’t bounced yet.

There is only one lock in all of app_meetme (version 1.2) and that is on the linked list that holds the configurations for each conference. When something needs to be changed in a single conference, such as adding another channel, all of the conferences get locked. This is not that big of a deal when you have one 1 or 2 conferences but when you have 10 to 15 it becomes a somewhat of a problem and when you have 150 with channels being added and removed all the time it is a big problem. Sure using a single lock does a great job of helping to prevent deadlocks, but it would be far better to use thread safe programming techniques which do not need a lock.

I know this post is really old, but I recently had to go back to this matter and I did find out the root of the problem.

It turns out the users where transfering the conference room to itself by mistake(usually when calling people outside the company). This usually isn’t possible when using local meetme of confbridge, but when you use a remote conference server like I do the transfer is like any other.

All I have to do is block users from transfer a conference room to itself.