RE: [Asterisk-Users] iaxtel and jitterbuffer
On Sun, 29 Aug 2004, Kris Boutilier wrote: Is timestamp information calculated purely from the relative timestamps of each frame of the current incoming stream or is there some degree of RTC synchronization expected between the two endpoints? No sync is needed; its all relative. Similarly, are jitter calculations made seperately for each discrete channel (ie. the IAX level) or are they based on an aggregate of all channels between each pair of two endpoints (ie. the TCP/IP level)? De-jtter is done for each call independently. Steve ___ Asterisk-Users mailing list [EMAIL PROTECTED] http://lists.digium.com/mailman/listinfo/asterisk-users To UNSUBSCRIBE or update options visit: http://lists.digium.com/mailman/listinfo/asterisk-users
Re: [Asterisk-Users] iaxtel and jitterbuffer
On Sat, 28 Aug 2004, Michael George wrote: So even with X11 eliminated the sound is still bad to Digium. I tried another's 1700 number, and it sounded the same, so it's not something unique to digium and me. Would IAX/GSM be so sensitive to half-duplex that I cannot expect it to work with my ISP only giving me 1/2 duplex service? If you think that the jitter buffer isn't working right and should fix this, then please capture debug from the buffer and send over to me. To do that, in /etc/asterisk/logger.conf edit the debug line to be: debug = notice,warning,error,debug,verbose Then run asterisk like so: /usr/sbin/asterisk -vv -g -dd -c Then go iax2 debug at the CLI prompt. Do a test call, then send me the resulting /var/log/asterisk/debug file. THanks, Steve ___ Asterisk-Users mailing list [EMAIL PROTECTED] http://lists.digium.com/mailman/listinfo/asterisk-users To UNSUBSCRIBE or update options visit: http://lists.digium.com/mailman/listinfo/asterisk-users
Re: [Asterisk-Users] iaxtel and jitterbuffer
On Sat, 28 Aug 2004, Andrew Kohlsmith wrote: Please note that it seems impossible to disable jitter buffer between 20040806 CVS HEAD endpoints. The jitterbuffer numbers in iax2 show channels look live. The numbers look right (jitbuf 0ms) between 20040806 and RC1 (Nufone). I haven't upgraded since then. The numbers get reported still in the older version, but the buffer IS turned off. Steve ___ Asterisk-Users mailing list [EMAIL PROTECTED] http://lists.digium.com/mailman/listinfo/asterisk-users To UNSUBSCRIBE or update options visit: http://lists.digium.com/mailman/listinfo/asterisk-users
Re: [Asterisk-Users] iaxtel and jitterbuffer
On Sunday 29 August 2004 02:06, [EMAIL PROTECTED] wrote: On Sat, 28 Aug 2004, Andrew Kohlsmith wrote: Please note that it seems impossible to disable jitter buffer between 20040806 CVS HEAD endpoints. The jitterbuffer numbers in iax2 show channels look live. The numbers look right (jitbuf 0ms) between 20040806 and RC1 (Nufone). I haven't upgraded since then. The numbers get reported still in the older version, but the buffer IS turned off. Ok so the disparity between iax2 show channels between two 20040806 (looks live) and 20040806 and RC1 (shows 0s) is expected? Just making sure, as between the two 'new' versions it is live, but between the new and old, it looks dead, whereas your reply said the numbers are still reported in the older version and that's not what I'm seeing. :-) -A. ___ Asterisk-Users mailing list [EMAIL PROTECTED] http://lists.digium.com/mailman/listinfo/asterisk-users To UNSUBSCRIBE or update options visit: http://lists.digium.com/mailman/listinfo/asterisk-users
Re: [Asterisk-Users] iaxtel and jitterbuffer
On Sunday 29 August 2004 01:59, [EMAIL PROTECTED] wrote: If you think that the jitter buffer isn't working right and should fix this, then please capture debug from the buffer and send over to me. To do that, in /etc/asterisk/logger.conf edit the debug line to be: debug = notice,warning,error,debug,verbose Then run asterisk like so: /usr/sbin/asterisk -vv -g -dd -c Then go iax2 debug at the CLI prompt. Do a test call, then send me the resulting /var/log/asterisk/debug file. Is there any way to do this 'live'? I get it intermittently and capturing debug for days before the problem is manifest is probably not the best way to do it. I've tried leaving the debug line in and not invoking any kind of -d in the asterisk startup but the debug log still grows. I can't comment out the debug line in logger.conf because a logger reload or reload will NOT create the debug file, only a restart will. Ideally some way to create the debug file but write very litte to it until I connect with asterisk -rc or something would be best I imagine. Also, is are logs of problem conversations already in progress any use to you? You nailed down the dead audio after 65535ms problem but every now and again (very very rare) we will have a conversation where the incoming audio goes totally dead for about 2-4 seconds and then continues just fine. This occurs usually several minutes into the conversation, and I've never seen it occur twice in a conversation. Obviously this is next to impossible to catch. :-( I haven't heard a complaint about it since turning off jitter buffer to nufone. As always, thank you for your knowledge and input. -A. ___ Asterisk-Users mailing list [EMAIL PROTECTED] http://lists.digium.com/mailman/listinfo/asterisk-users To UNSUBSCRIBE or update options visit: http://lists.digium.com/mailman/listinfo/asterisk-users
Re: [Asterisk-Users] iaxtel and jitterbuffer
On Sunday 29 August 2004 01:59, [EMAIL PROTECTED] wrote: If you think that the jitter buffer isn't working right and should fix this, then please capture debug from the buffer and send over to me. I notice that the timing measurements are still showing wild values at times - here is a partial grab of an iax2 show channels: Lag Jitter JitBuf Format 00020ms 6291456ms ms ALAW 00012ms 6291440ms ms ALAW 00017ms 0004ms ms ALAW 00012ms 286523393ms ms ALAW 00012ms 0025ms ms ALAW -978714621ms 6293280ms ms ALAW ___ Asterisk-Users mailing list [EMAIL PROTECTED] http://lists.digium.com/mailman/listinfo/asterisk-users To UNSUBSCRIBE or update options visit: http://lists.digium.com/mailman/listinfo/asterisk-users
Re: [Asterisk-Users] iaxtel and jitterbuffer
Those wild times especially occur before any audio is sent. (e.g. while ringing or pre ringing). At 17:10 29/08/2004, you wrote: On Sunday 29 August 2004 01:59, [EMAIL PROTECTED] wrote: If you think that the jitter buffer isn't working right and should fix this, then please capture debug from the buffer and send over to me. I notice that the timing measurements are still showing wild values at times - here is a partial grab of an iax2 show channels: Lag Jitter JitBuf Format 00020ms 6291456ms ms ALAW 00012ms 6291440ms ms ALAW 00017ms 0004ms ms ALAW 00012ms 286523393ms ms ALAW 00012ms 0025ms ms ALAW -978714621ms 6293280ms ms ALAW ___ Asterisk-Users mailing list [EMAIL PROTECTED] http://lists.digium.com/mailman/listinfo/asterisk-users To UNSUBSCRIBE or update options visit: http://lists.digium.com/mailman/listinfo/asterisk-users ___ Asterisk-Users mailing list [EMAIL PROTECTED] http://lists.digium.com/mailman/listinfo/asterisk-users To UNSUBSCRIBE or update options visit: http://lists.digium.com/mailman/listinfo/asterisk-users
Re: [Asterisk-Users] iaxtel and jitterbuffer
On Sat, Aug 28, 2004 at 11:31:48PM -0400, Andrew Kohlsmith wrote: On Saturday 28 August 2004 23:01, Michael George wrote: It's a PII 266 (okay, not the fatest system) with 192MB RAM. X is not running and the Framebuffer has been turned off in /boot/grum/menu.lst. I have disabled all the servers except for sshd. I have the latest source from CVS HEAD as of about 30min ago. Should be fine. I ran * on a P90 for a while; it did everything I needed except iLBC. :-) Okay, that's a good assurance. Unfortunately, I have discovered that either the HDD or the ide controller in that system is bad because it cannot stay up overnight. When I stress it with a YaST update, it will die much more reliably. Until I can address that issue, I will have to work on my main system. I'll just have to take it down to init 3 and stop many of the other server processes that will still be running. There is no Zap card in this sytem. The only phone on it is a SIP phone. With it I dial in to digium's 1-700 number. The audio is better, but still choppy and unacceptible. Is your SIP phone doing any kind of silence suppression? It must be turned off because asterisk takes its timing from the RTP stream and if the phone's not transmitting frames continuously you'll get shitty audio. Good suggestion and I have double checked it. I am and was not doing that. I think I'd read about it in a Granstream-* page Note that latest CVS HEAD looks like they're making provisions for self-timing but without a stable clock source it's unlikely to help you. There are ztdummy modules which use the RTC or certain brands of USB controller to provide adequate timing but ideally you want some kind of Zaptel hardware in there providing a 1000Hz interrupt. Hmm, I thought that the timing source was only needed for trunking. I don't have on on the little box, but I do have a TDM400 (which seems to have faults, also, but Digium suggested moving the FXO to socket 4, we'll see if that helps) in the main box, so that should be all set for a timing source. Also -- make sure your uplink is acceptable. First test: make sure there is nothing plugged into your upstream except for your asterisk box and the phone. Some routers are known to play silly bugger with your packets which naturally wreaks havoc with asterisk. :-) The only things on the net when I run the next test will be my main server. Since I have to test on that with X turned off, I don't even need the SIP phone active. In case it might be relevant (there are SO many pieces to this puzzle that I want to mention all I can think of in case they ring a trouble-bell in someone's mind...) my router is a Netgear FVS318 acting as a NAT to my ISP. So even with X11 eliminated the sound is still bad to Digium. I tried another's 1700 number, and it sounded the same, so it's not something unique to digium and me. Perhaps something to do with your upstream or connection to IAXtel. That's why I'm recommending having nothing but asterisk and the phone on the connection, at least until we nail down what the poor audio's being caused by. That's possible. I've checked with my ISP and he said that the connection is surely half-duplex, but you say that you have 1/2 also and it works fine for you, so that's not it. I'm also inquiring about other filters they might have in place. I've heard them mention before that they had some cool router software that could detect traffic patterns usually associated with software and music piracy and then throttle that traffic into a small part of The Pipe. I haven't yet heard back, and I'm hoping that isn't the case. However, if it *is*, a VPN between offices might help. IAXtel would be shot, though. Hoever, if that *is* the case, I can probably convince them to tell their software to leave me alone on a couple specified ports. Would IAX/GSM be so sensitive to half-duplex that I cannot expect it to work with my ISP only giving me 1/2 duplex service? It has nothing to do with IAX or GSM. Stop blaming them. My upstream is half duplex as well (pretty much anyone on DSL or cable is on a half duplex connection whether they realize it or not). There are many, many people using asterisk every day for long distance and in environments where audio quality is crucial. Let's stop blaming asterisk and take a good hard look at what's happenning, shall we? My apologies. I'm not trying to blame anyone, I love * and except for a couple glitches that we're working on (with all your gracious help), I'm very impressed. My one glitch may be with the hardware, so that's a separate issue, but the other is trying to figure out this issue with IAX/GSM. When I ask about sensitivity, I don't mean to be accusatory. IAX is open and freely available and GSM is freely usable, and I'm glad. Sometimes OSS has its limitations and I am willing to work with them. So I do not intend any condescention(sp?),
Re: [Asterisk-Users] iaxtel and jitterbuffer
At 17:10 29/08/2004, you wrote: I notice that the timing measurements are still showing wild values at times - here is a partial grab of an iax2 show channels: Lag Jitter JitBuf Format 00020ms 6291456ms ms ALAW 00012ms 6291440ms ms ALAW 00017ms 0004ms ms ALAW 00012ms 286523393ms ms ALAW 00012ms 0025ms ms ALAW -978714621ms 6293280ms ms ALAW Those wild times especially occur before any audio is sent. (e.g. while ringing or pre ringing). That maybe true, but the examples above appeared to be established calls! Linus ___ Asterisk-Users mailing list [EMAIL PROTECTED] http://lists.digium.com/mailman/listinfo/asterisk-users To UNSUBSCRIBE or update options visit: http://lists.digium.com/mailman/listinfo/asterisk-users
Re: [Asterisk-Users] iaxtel and jitterbuffer
On Sun, Aug 29, 2004 at 07:59:20AM +0200, [EMAIL PROTECTED] wrote: On Sat, 28 Aug 2004, Michael George wrote: So even with X11 eliminated the sound is still bad to Digium. I tried another's 1700 number, and it sounded the same, so it's not something unique to digium and me. Would IAX/GSM be so sensitive to half-duplex that I cannot expect it to work with my ISP only giving me 1/2 duplex service? If you think that the jitter buffer isn't working right and should fix this, then please capture debug from the buffer and send over to me. I'm not sure what the problem is. What I am hearing does sound like the descriptions I've read w.r.t. the jitter buffer, but making jitter buffer changes haven't really changed the effect. That gives 2 possibilities: 1. That the jitter buffer isn't working and it *should* fix the problem. 2. That the problem is completely independent of the JB so there is nothing the JB can do to fix it. To do that, in /etc/asterisk/logger.conf edit the debug line to be: debug = notice,warning,error,debug,verbose Then run asterisk like so: /usr/sbin/asterisk -vv -g -dd -c Then go iax2 debug at the CLI prompt. Do a test call, then send me the resulting /var/log/asterisk/debug file. I will do that. Hopefully that will help us isolate the problem and perhaps eliminate the jitterbuffer from the equasion. :) I will try to run this test today and report back my findings. Also, on Thursday I will be going into the main office. I will take my little * box and try the IAXtel test there. That should help determine if it's my local office net connection that is the problem. Thank you! -- -M There are 10 kinds of people in this world: Those who can count in binary and those who cannot. ___ Asterisk-Users mailing list [EMAIL PROTECTED] http://lists.digium.com/mailman/listinfo/asterisk-users To UNSUBSCRIBE or update options visit: http://lists.digium.com/mailman/listinfo/asterisk-users
Re: [Asterisk-Users] iaxtel and jitterbuffer
On Sun, 29 Aug 2004, Andrew Kohlsmith wrote: Also, is are logs of problem conversations already in progress any use to you? You nailed down the dead audio after 65535ms problem but every now and again (very very rare) we will have a conversation where the incoming audio goes totally dead for about 2-4 seconds and then continues just fine. This occurs usually several minutes into the conversation, and I've never seen it occur twice in a conversation. Logs of parts of a call are fine. The jitter buffer makes all its decisions about dejittering based on the timestamps of incoming frames. There a fundamental expectation that the sending side is correctly stamping each frame - 20msec, 40msec etc etc. The problem is that the sending side doesn't always do that. Sometimes for one reason or another the stamps jump. The receiver has no way of telling that the sender mangled the timestamps, and assumes that the packets with the new stamps have been delayed, or arrived early, or whatever. Either way, the jitter buffer does its thing and unknowingly makes things worse. Unfortunately, this is why you can still be better off without it - but the problem really needs to be fixed by fixing the timestamp generation on the sender. Steve ___ Asterisk-Users mailing list [EMAIL PROTECTED] http://lists.digium.com/mailman/listinfo/asterisk-users To UNSUBSCRIBE or update options visit: http://lists.digium.com/mailman/listinfo/asterisk-users
Re: [Asterisk-Users] iaxtel and jitterbuffer
On Sun, 29 Aug 2004, joachim wrote: Those wild times especially occur before any audio is sent. (e.g. while ringing or pre ringing). Yeah - because the sender does weird things to the timestamps it generates. This is the problem that needs to be resolved; the jitter buffer just shows up the issue. Steve ___ Asterisk-Users mailing list [EMAIL PROTECTED] http://lists.digium.com/mailman/listinfo/asterisk-users To UNSUBSCRIBE or update options visit: http://lists.digium.com/mailman/listinfo/asterisk-users
Re: [Asterisk-Users] iaxtel and jitterbuffer
On Sunday 29 August 2004 15:52, [EMAIL PROTECTED] wrote: The jitter buffer makes all its decisions about dejittering based on the timestamps of incoming frames. There a fundamental expectation that the sending side is correctly stamping each frame - 20msec, 40msec etc etc. Right, this makes sense. :-) The problem is that the sending side doesn't always do that. Sometimes for one reason or another the stamps jump. The receiver has no way of telling that the sender mangled the timestamps, and assumes that the packets with the new stamps have been delayed, or arrived early, or whatever. Either way, the jitter buffer does its thing and unknowingly makes things worse. Unfortunately, this is why you can still be better off without it - but the problem really needs to be fixed by fixing the timestamp generation on the sender. Hmm... I think next CVS update I'm gonna add a bit of code in chan_iax2 that tries to verify that timestamps aren't getting sent incorrectly. Fun fun fun. :-) -A. ___ Asterisk-Users mailing list [EMAIL PROTECTED] http://lists.digium.com/mailman/listinfo/asterisk-users To UNSUBSCRIBE or update options visit: http://lists.digium.com/mailman/listinfo/asterisk-users
Re: [Asterisk-Users] iaxtel and jitterbuffer
On Sun, 29 Aug 2004, Andrew Kohlsmith wrote: Hmm... I think next CVS update I'm gonna add a bit of code in chan_iax2 that tries to verify that timestamps aren't getting sent incorrectly. Fun fun fun. :-) Its not that the generation is broken. Its that various optimisations and things have been added over time. The result is that sometimes the source of the timestamps changes - and suddenly. Like - we're playing locally generated Playback() audio down the line, then the dialplan rings another IAX2/ address. Then the other end answers. First the timestamps come from the Playback, then the ring generator, then from the remote IAX2/ system... So the discontinuities get in. There is also effort in the sending IAX2code to lock the timestamps to exact intervals (20msec), but sometimes it gives up and lets it jump to get back into step... Steve ___ Asterisk-Users mailing list [EMAIL PROTECTED] http://lists.digium.com/mailman/listinfo/asterisk-users To UNSUBSCRIBE or update options visit: http://lists.digium.com/mailman/listinfo/asterisk-users
Re: [Asterisk-Users] iaxtel and jitterbuffer
On Sat, Aug 28, 2004 at 11:31:48PM -0400, Andrew Kohlsmith wrote: On Saturday 28 August 2004 23:01, Michael George wrote: It has nothing to do with IAX or GSM. Stop blaming them. My upstream is half duplex as well (pretty much anyone on DSL or cable is on a half duplex connection whether they realize it or not). There are many, many people using asterisk every day for long distance and in environments where audio quality is crucial. Let's stop blaming asterisk and take a good hard look at what's happenning, shall we? Someone suggested that perhaps the machine is too slow. If someone who uses IAX2 between offices wouldn't mind, could you please indicate how heavy a system you are using for Zap -- IAX/GSM -- VOIP. Perhaps I am underestimating the HP required for the voice coding... -- -M There are 10 kinds of people in this world: Those who can count in binary and those who cannot. ___ Asterisk-Users mailing list [EMAIL PROTECTED] http://lists.digium.com/mailman/listinfo/asterisk-users To UNSUBSCRIBE or update options visit: http://lists.digium.com/mailman/listinfo/asterisk-users
RE: [Asterisk-Users] iaxtel and jitterbuffer
Is timestamp information calculated purely from the relative timestamps of
each frame of the current incoming stream or is there some degree of RTC
synchronization expected between the two endpoints?
Similarly, are jitter calculations made seperately for each discrete channel
(ie. the IAX level) or are they based on an aggregate of all channels
between each pair of two endpoints (ie. the TCP/IP level)?
k.
-Original Message-
From: [EMAIL PROTECTED] [mailto:[EMAIL PROTECTED]
Sent: August 29, 2004 12:53 PM
To: Asterisk Users Mailing List - Non-Commercial Discussion
Subject: Re: [Asterisk-Users] iaxtel and jitterbuffer
{clip}
The jitter buffer makes all its decisions about dejittering based on the
timestamps of incoming frames. There a fundamental expectation that the
sending side is correctly stamping each frame - 20msec, 40msec etc etc.
The problem is that the sending side doesn't always do that. Sometimes
for one reason or another the stamps jump. The receiver has no way of
telling that the sender mangled the timestamps, and assumes that the
packets with the new stamps have been delayed, or arrived early, or
whatever. Either way, the jitter buffer does its thing and unknowingly
makes things worse.
Unfortunately, this is why you can still be better off without it - but
the problem really needs to be fixed by fixing the timestamp generation on
the sender.
Steve
___
Asterisk-Users mailing list
[EMAIL PROTECTED]
http://lists.digium.com/mailman/listinfo/asterisk-users
To UNSUBSCRIBE or update options visit:
http://lists.digium.com/mailman/listinfo/asterisk-users
Re: [Asterisk-Users] iaxtel and jitterbuffer
Does this also effect 1.0-RC2? I am having a similar issue at a
customer site over a frame relay network that is having occasional
choppy sound over a fairly open line, with the jitter buffer enabled,
as well as trunk=yes enabled.
Thanks!
Brian
On Fri, 27 Aug 2004 12:47:05 -0700, Kris Boutilier
[EMAIL PROTECTED] wrote:
Had this problem earlier this week - ensure 'trunk=no' in iax.conf if you're
using fairly current CVS code. There is something not right w/the trunking
that causes choppy sound. See the wiki for more info.
Kris Boutilier
Information Systems Coordinator
Sunshine Coast Regional District
-Original Message-
From: Michael George [mailto:[EMAIL PROTECTED]
Sent: August 27, 2004 11:58 AM
To: [EMAIL PROTECTED]
Subject: [Asterisk-Users] iaxtel and jitterbuffer
I am trying to work out IAX -- IAX communications with my * box. I have a
registration with iaxtel and I thought I would start there for my learning.
I am able to call the number for Digium's support line (700-428-6000), but
the
sound is horribly chopping. Some reading revealed the jitterbuffer
settings,
so I enabled them in iax.conf. I have the following now:
{clip}
___
Asterisk-Users mailing list
[EMAIL PROTECTED]
http://lists.digium.com/mailman/listinfo/asterisk-users
To UNSUBSCRIBE or update options visit:
http://lists.digium.com/mailman/listinfo/asterisk-users
___
Asterisk-Users mailing list
[EMAIL PROTECTED]
http://lists.digium.com/mailman/listinfo/asterisk-users
To UNSUBSCRIBE or update options visit:
http://lists.digium.com/mailman/listinfo/asterisk-users
Re: [Asterisk-Users] iaxtel and jitterbuffer
Had this problem earlier this week - ensure 'trunk=no' in iax.conf if you're using fairly current CVS code. There is something not right w/the trunking that causes choppy sound. See the wiki for more info. I am using current CVS code and I have trunk=no. Still sounds crappy. I need to check with my ISP and make sure they aren't throttling in that range. I'm only getting about 4.5Kbps of throughput... Any available codecs that can use that level of bandwidth? I do a lot of work with companies throughout the US on network performance and we _frequently_ run into routers, switches, servers, etc, that are allowed to auto-negotiate their half vs full duplex nic interfaces. About 50% of the time, systems will get it wrong as there are no standards as to how the negotiation should be done. A recent case this past week indicated that data flow between two servers on the same layer-2 network was around 400 kbps when it should have been able to sustain at least 80 meg. You might double check each of your ethernet interfaces to ensure duplex settings are correct. If not at full duplex all the way through, you'll run into the strangeness you're seeing under varying traffic loads. ___ Asterisk-Users mailing list [EMAIL PROTECTED] http://lists.digium.com/mailman/listinfo/asterisk-users To UNSUBSCRIBE or update options visit: http://lists.digium.com/mailman/listinfo/asterisk-users
Re: [Asterisk-Users] iaxtel and jitterbuffer
On Aug 28, 2004, at 7:39 AM, Rich Adamson wrote: I do a lot of work with companies throughout the US on network performance and we _frequently_ run into routers, switches, servers, etc, that are allowed to auto-negotiate their half vs full duplex nic interfaces. About 50% of the time, systems will get it wrong as there are no standards as to how the negotiation should be done. No standard? Huh? You mean besides 'NWay', which is part of 802.3? http://www.scyld.com/NWay.html I've certainly seen problems, particularly with older Cisco switches and routers, but newer hardware seems to be pretty good. In fact, autonegotiation is *required* with GigE; you aren't even allowed to disable it according to the specs. Of course, that's sort of moot, because 1000/half isn't even slightly useful due to its 640-byte minimum packet size. At my previous employer, we were having tons of duplex problems. They mostly boiled down to forced duplex problems, where someone would force one end of a link, but leave the other end to autonegotiate. With most of Cisco's hardware, forcing 100/full *completely* disables autonegotiation. IMHO, it should still participate in autonegotiation, but only advertise the 100/full ability. Instead, Cisco tells the other end I don't negotiate. So, if you set one end to 100/full and fail to force the other end, then it will try to negotiate, fail, and fall back to 100/half, because that's the only reasonable thing to when negotiation fails. At this point, one end is 100/full and the other is 100/half, and you're about to have trouble. The really fun thing with this sort of link is that it works just fine with low traffic levels--a normal ping won't show problems, but it'll break when you actually try to use it for anything non-trivial. Using larger ping packets helps: ping -s 1 totally fails if the duplex is broken anywhere along the link. With newer IOS and CatOS builds, you can get around this by leaving CDP enabled; CDP v2 shares duplex information, and it'll log duplex mis-matches when both ends of the link use Cisco hardware. I wrote a small CDP listener for Linux boxes and did the same thing, logging duplex mis-matches. With 700 servers over 2 years, the only mismatches we ever found were caused by forced 100/full on the switches. One easy fix that we found, at least for IOS switches, was to set the speed to auto but force the duplex. That apparently leaves NWay negotiation running but only advertises full duplex as an option. Since nothing *ever* uses NWay to negotiate the speed of the link, this has the same result as forcing 100/full, but it fails in the right direction if you only force one end of the link. Of course, knowing Cisco, this only applies for every third model of switch running even-numbered IOS builds. Scott ___ Asterisk-Users mailing list [EMAIL PROTECTED] http://lists.digium.com/mailman/listinfo/asterisk-users To UNSUBSCRIBE or update options visit: http://lists.digium.com/mailman/listinfo/asterisk-users
Re: [Asterisk-Users] iaxtel and jitterbuffer
On Sat, Aug 28, 2004 at 08:39:50AM -0600, Rich Adamson wrote: I do a lot of work with companies throughout the US on network performance and we _frequently_ run into routers, switches, servers, etc, that are allowed to auto-negotiate their half vs full duplex nic interfaces. About 50% of the time, systems will get it wrong as there are no standards as to how the negotiation should be done. A recent case this past week indicated that data flow between two servers on the same layer-2 network was around 400 kbps when it should have been able to sustain at least 80 meg. You might double check each of your ethernet interfaces to ensure duplex settings are correct. If not at full duplex all the way through, you'll run into the strangeness you're seeing under varying traffic loads. My ISP has a half-duplex connection between me and the world-at-large. It doesn't seem like that should be a problem, though, because we've been running VOIP with Multitech proprietary hardware for over two years now with little trouble and excellent voice quality. That was using a 9.6KBps codec. The difference between that and what I'm getting from IAX/GSM is profound, with GSM being intolerably poor quality. As a test, I ran two internal * machines with IAX/GSM between them. A conversation would consume from 7-10KBps, varying. Then I would call Digium's iaxtel number and I could see traffic from 4.5-8KBps and the voice was all choppy. I called another person's system (knowing they had IVR, of course) and the audio was also choppy, but when it got through the message and sent ring tones, they sounded fine. Then another voice message and it was choppy again. So I tried digium again. This time I could see the bandwidth being consumed, but I heard nothing on the line at all. So I tried calling my own iaxtel number. I could see my badwidth usage jump to about 10KBps, as I would expect and * told me that it was playing out the appropriate audo to the incoming caller. I heard nothing, however. Does this perhaps give any further indication of what might be wrong? I have in my [general] section: bandwidth=low disallow=all allow=gsm jitterbuffer=yes dropcount=10 maxjitterbuffer=500 maxexcessbuffer=100 minexcessbuffer=10 jittershrinkrate=1 trunk=no register = me:[EMAIL PROTECTED] tos=lowdelay I'm working towards a client install of IAX which will be used for inter-office VOIP, but I need to get this issue worked out or it's not deployable. -- -M There are 10 kinds of people in this world: Those who can count in binary and those who cannot. ___ Asterisk-Users mailing list [EMAIL PROTECTED] http://lists.digium.com/mailman/listinfo/asterisk-users To UNSUBSCRIBE or update options visit: http://lists.digium.com/mailman/listinfo/asterisk-users
Re: [Asterisk-Users] iaxtel and jitterbuffer
On Sat, Aug 28, 2004 at 03:00:26PM -0400, Michael George wrote: On Sat, Aug 28, 2004 at 08:39:50AM -0600, Rich Adamson wrote: I do a lot of work with companies throughout the US on network performance and we _frequently_ run into routers, switches, servers, etc, that are allowed to auto-negotiate their half vs full duplex nic interfaces. About 50% of the time, systems will get it wrong as there are no standards as to how the negotiation should be done. A recent case this past week indicated that data flow between two servers on the same layer-2 network was around 400 kbps when it should have been able to sustain at least 80 meg. You might double check each of your ethernet interfaces to ensure duplex settings are correct. If not at full duplex all the way through, you'll run into the strangeness you're seeing under varying traffic loads. I just saw a page on the wiki that mentions that running X11 or a VESA frame buffer can cause jittery sound. I only have this problem with IAX2, but that might be cause when I use Zap -- Zap or Zap -- SIP there is no en/decoding involved. I am running * on my main home server, which does run X and other software. Perhaps that's the problem? Maybe if I juiced it up with more RAM, might that help? It's at .5GB now, but I can easily take it to 1GB. Or, maybe a 900MHz Athlon still can't handle the coding with X11 running? -- -M There are 10 kinds of people in this world: Those who can count in binary and those who cannot. ___ Asterisk-Users mailing list [EMAIL PROTECTED] http://lists.digium.com/mailman/listinfo/asterisk-users To UNSUBSCRIBE or update options visit: http://lists.digium.com/mailman/listinfo/asterisk-users
Re: [Asterisk-Users] iaxtel and jitterbuffer
On Sat, 28 Aug 2004 15:24:01 -0400, Michael George wrote: On Sat, Aug 28, 2004 at 03:00:26PM -0400, Michael George wrote: On Sat, Aug 28, 2004 at 08:39:50AM -0600, Rich Adamson wrote: I do a lot of work with companies throughout the US on network performance and we _frequently_ run into routers, switches, servers, etc, that are allowed to auto-negotiate their half vs full duplex nic interfaces. About 50% of the time, systems will get it wrong as there are no standards as to how the negotiation should be done. A recent case this past week indicated that data flow between two servers on the same layer-2 network was around 400 kbps when it should have been able to sustain at least 80 meg. You might double check each of your ethernet interfaces to ensure duplex settings are correct. If not at full duplex all the way through, you'll run into the strangeness you're seeing under varying traffic loads. I just saw a page on the wiki that mentions that running X11 or a VESA frame buffer can cause jittery sound. I only have this problem with IAX2, but that might be cause when I use Zap -- Zap or Zap -- SIP there is no en/decoding involved. I am running * on my main home server, which does run X and other software. Perhaps that's the problem? Maybe if I juiced it up with more RAM, might that help? It's at .5GB now, but I can easily take it to 1GB. Or, maybe a 900MHz Athlon still can't handle the coding with X11 running? My understand, admittedly limited, is that the windowing system (X or other) generates a lot of interupts. This can burden the system that is also engaged in real-time tasks such as rpt for voip. That said, my Asterisk server is is an AMD2500+ with 512 MB ram. I did install the Gnome desktop with Fedora Core 1, but I don't do anything else on the server. It runs headless. I just ssh in to tweak * as needed. Michael Michael -- Michael Graves [EMAIL PROTECTED] Sr. Product Specialist www.pixelpower.com Pixel Power Inc. [EMAIL PROTECTED] o713-861-4005 o800-905-6412 c713-201-1262 An ounce of pretention is worth a pound of manure. ** Tag(s) inserted by Bandit Tagger98 - http://www.gbar.dtu.dk/~c918704 ___ Asterisk-Users mailing list [EMAIL PROTECTED] http://lists.digium.com/mailman/listinfo/asterisk-users To UNSUBSCRIBE or update options visit: http://lists.digium.com/mailman/listinfo/asterisk-users
Re: [Asterisk-Users] iaxtel and jitterbuffer
On Saturday 28 August 2004 15:00, Michael George wrote: The difference between that and what I'm getting from IAX/GSM is profound, with GSM being intolerably poor quality. That's odd; every single voice call coming in and out of the company I work for is using the GSM codec with asterisk and IAX2... even the music on hold is passable. I have in my [general] section: bandwidth=low get rid of it; you're giving codecs below. disallow=all allow=gsm jitterbuffer=yes dropcount=10 maxjitterbuffer=500 maxexcessbuffer=100 minexcessbuffer=10 jittershrinkrate=1 My jitter settings are similar. max 500, maxexcess 100, minexcess 50, dropcount=2 (10, are you *insane*?!), jittershrink of 1. I'd slow down the shrink even more if I could, as even at 1 it's still noticeable. Please note that it seems impossible to disable jitter buffer between 20040806 CVS HEAD endpoints. The jitterbuffer numbers in iax2 show channels look live. The numbers look right (jitbuf 0ms) between 20040806 and RC1 (Nufone). I haven't upgraded since then. trunk=no I found 20040806 CVS HEAD to have odd little problems with trunking too. Trunking between 20040806 and RC1 (again, with nufone) work fine. I can't trunk to VPC at all or they can't hear me (I can hear them). Just to make clear: I have completely disabled the jitter buffer between myself and Nufone and the call quality has gone up slightly. I wasn't expecting this. Regards, Andrew ___ Asterisk-Users mailing list [EMAIL PROTECTED] http://lists.digium.com/mailman/listinfo/asterisk-users To UNSUBSCRIBE or update options visit: http://lists.digium.com/mailman/listinfo/asterisk-users
Re: [Asterisk-Users] iaxtel and jitterbuffer
On Saturday 28 August 2004 15:24, Michael George wrote: I just saw a page on the wiki that mentions that running X11 or a VESA frame buffer can cause jittery sound. I only have this problem with IAX2, but that might be cause when I use Zap -- Zap or Zap -- SIP there is no en/decoding involved. Asterisk is an application requiring hard realtime performance. Pretty much any telephony application is. Running *anything* in addition to asterisk is just asking for trouble. Actually I would be curious to see if asterisk performs better in a soft-realtime environment (i.e. what's actually easily possible with commodity PC hardware). -A. ___ Asterisk-Users mailing list [EMAIL PROTECTED] http://lists.digium.com/mailman/listinfo/asterisk-users To UNSUBSCRIBE or update options visit: http://lists.digium.com/mailman/listinfo/asterisk-users
Re: [Asterisk-Users] iaxtel and jitterbuffer
On Sat, Aug 28, 2004 at 05:08:30PM -0400, Andrew Kohlsmith wrote: On Saturday 28 August 2004 15:24, Michael George wrote: I just saw a page on the wiki that mentions that running X11 or a VESA frame buffer can cause jittery sound. I only have this problem with IAX2, but that might be cause when I use Zap -- Zap or Zap -- SIP there is no en/decoding involved. Asterisk is an application requiring hard realtime performance. Pretty much any telephony application is. Running *anything* in addition to asterisk is just asking for trouble. Since X11 and other daemons might be a problem on my main * server, I pulled out my little testbed and fired it up. It's a PII 266 (okay, not the fatest system) with 192MB RAM. X is not running and the Framebuffer has been turned off in /boot/grum/menu.lst. I have disabled all the servers except for sshd. I have the latest source from CVS HEAD as of about 30min ago. There is no Zap card in this sytem. The only phone on it is a SIP phone. With it I dial in to digium's 1-700 number. The audio is better, but still choppy and unacceptible. Looking at the * hardware recommendations page, this is by no means near the smallest recorded setup, so teh system shouldn't be underpowered. So even with X11 eliminated the sound is still bad to Digium. I tried another's 1700 number, and it sounded the same, so it's not something unique to digium and me. Would IAX/GSM be so sensitive to half-duplex that I cannot expect it to work with my ISP only giving me 1/2 duplex service? -- -M There are 10 kinds of people in this world: Those who can count in binary and those who cannot. ___ Asterisk-Users mailing list [EMAIL PROTECTED] http://lists.digium.com/mailman/listinfo/asterisk-users To UNSUBSCRIBE or update options visit: http://lists.digium.com/mailman/listinfo/asterisk-users
Re: [Asterisk-Users] iaxtel and jitterbuffer
On Saturday 28 August 2004 23:01, Michael George wrote: It's a PII 266 (okay, not the fatest system) with 192MB RAM. X is not running and the Framebuffer has been turned off in /boot/grum/menu.lst. I have disabled all the servers except for sshd. I have the latest source from CVS HEAD as of about 30min ago. Should be fine. I ran * on a P90 for a while; it did everything I needed except iLBC. :-) There is no Zap card in this sytem. The only phone on it is a SIP phone. With it I dial in to digium's 1-700 number. The audio is better, but still choppy and unacceptible. Is your SIP phone doing any kind of silence suppression? It must be turned off because asterisk takes its timing from the RTP stream and if the phone's not transmitting frames continuously you'll get shitty audio. Note that latest CVS HEAD looks like they're making provisions for self-timing but without a stable clock source it's unlikely to help you. There are ztdummy modules which use the RTC or certain brands of USB controller to provide adequate timing but ideally you want some kind of Zaptel hardware in there providing a 1000Hz interrupt. Also -- make sure your uplink is acceptable. First test: make sure there is nothing plugged into your upstream except for your asterisk box and the phone. Some routers are known to play silly bugger with your packets which naturally wreaks havoc with asterisk. :-) So even with X11 eliminated the sound is still bad to Digium. I tried another's 1700 number, and it sounded the same, so it's not something unique to digium and me. Perhaps something to do with your upstream or connection to IAXtel. That's why I'm recommending having nothing but asterisk and the phone on the connection, at least until we nail down what the poor audio's being caused by. Would IAX/GSM be so sensitive to half-duplex that I cannot expect it to work with my ISP only giving me 1/2 duplex service? It has nothing to do with IAX or GSM. Stop blaming them. My upstream is half duplex as well (pretty much anyone on DSL or cable is on a half duplex connection whether they realize it or not). There are many, many people using asterisk every day for long distance and in environments where audio quality is crucial. Let's stop blaming asterisk and take a good hard look at what's happenning, shall we? -A. ___ Asterisk-Users mailing list [EMAIL PROTECTED] http://lists.digium.com/mailman/listinfo/asterisk-users To UNSUBSCRIBE or update options visit: http://lists.digium.com/mailman/listinfo/asterisk-users
RE: [Asterisk-Users] iaxtel and jitterbuffer
Had this problem earlier this week - ensure 'trunk=no' in iax.conf if you're
using fairly current CVS code. There is something not right w/the trunking
that causes choppy sound. See the wiki for more info.
Kris Boutilier
Information Systems Coordinator
Sunshine Coast Regional District
-Original Message-
From: Michael George [mailto:[EMAIL PROTECTED]
Sent: August 27, 2004 11:58 AM
To: [EMAIL PROTECTED]
Subject: [Asterisk-Users] iaxtel and jitterbuffer
I am trying to work out IAX -- IAX communications with my * box. I have a
registration with iaxtel and I thought I would start there for my learning.
I am able to call the number for Digium's support line (700-428-6000), but
the
sound is horribly chopping. Some reading revealed the jitterbuffer
settings,
so I enabled them in iax.conf. I have the following now:
{clip}
___
Asterisk-Users mailing list
[EMAIL PROTECTED]
http://lists.digium.com/mailman/listinfo/asterisk-users
To UNSUBSCRIBE or update options visit:
http://lists.digium.com/mailman/listinfo/asterisk-users
Re: [Asterisk-Users] iaxtel and jitterbuffer
On Fri, Aug 27, 2004 at 12:47:05PM -0700, Kris Boutilier wrote: Had this problem earlier this week - ensure 'trunk=no' in iax.conf if you're using fairly current CVS code. There is something not right w/the trunking that causes choppy sound. See the wiki for more info. I am using current CVS code and I have trunk=no. Still sounds crappy. I need to check with my ISP and make sure they aren't throttling in that range. I'm only getting about 4.5Kbps of throughput... Any available codecs that can use that level of bandwidth? I'll have to check out the speex codec... -- -M There are 10 kinds of people in this world: Those who can count in binary and those who cannot. ___ Asterisk-Users mailing list [EMAIL PROTECTED] http://lists.digium.com/mailman/listinfo/asterisk-users To UNSUBSCRIBE or update options visit: http://lists.digium.com/mailman/listinfo/asterisk-users
