February 2009 - FreeSWITCH-users - FreeSWITCH Mailing Lists

Re: [Freeswitch-users] Voicemail prompts and playback speed
by Michael S Collins 07 Feb '09

07 Feb '09

On Feb 6, 2009, at 6:27 PM, Michael Jerris <mike(a)jerris.com> wrote: > Not sure what you mean by playback speed. All the prompts for > voicemail are defined in the phrase macros in the configuration. Check out conf/lang/en/vm/sounds.xml -MC

1 0

Voicemail prompts and playback speed
by Maxim Karp 07 Feb '09

07 Feb '09

Hello all, Can anyone please let me know how I might be able to configure the voice mail prompts and their playback speed? Thanks, Maxim.

2 1

Separate NICs for Performance
by Gabriel Kuri 07 Feb '09

07 Feb '09

Hey Folks: For a FS box that's generally handling higher amounts of inbound/outbound call traffic (say 500 - 700 calls) and registrations (30 - 50 per second), is it recommended to split off the signaling and media traffic onto separate NICs for performance reasons? Or is it better to split all the traffic for the phones and outside traffic to carriers into separate profiles and split those profiles onto separate NICs? or perhaps none of the above and there's still a better solution I didn't mention ;) ? Thanks, Gabe

3 5

Re: [Freeswitch-users] Transcoding G723
by Mitul Limbani 06 Feb '09

06 Feb '09

BODY { font-family:Arial, Helvetica, sans-serif;font-size:12px; }Hello guys, Sorry it was my mistake, i re-read the entire proposal and it looks like their specialized 1U Hardware with custom CPU can handle close to 1500 simultaneous G729 encoding n transmission, also this hardware offers from any codec to any codec transcoding path. Also this box easily works over FreeSWITCH, Asterisk, RTPProxy, OpenSIPS are their claims. and their pricing are First year £16,000 Second year £13,000/annum recurring This is found slightly on higher side though, Thanks & Regards, Mitul Limbani, Founder & CEO, Enterux Solutions, The Enterprise Linux Company (TM), www.enterux.com +91-9820332422 On Sat 07/02/09 02:00 , Michael Giagnocavo mgg(a)giagnocavo.net sent: $22K would buy quite a few machines with many core Xeons. I just don't see how it'd be effective at that price. Not to mention a yearly figure. The only thing to take into consideration would be the G729 licenses. But in bulk, the price should be pretty effective, even figuring in hardware. (Not to mention what things like Larrabee will mean for encoding -- 32 1.5GHz "Pentium 4 x64" cores, each with 4 threads?) And... if you're running Asterisk, um, isn't 5000 channels a wee bit over what you can handle anyways? ;) -Michael -----Original Message----- From: freeswitch-users-bounces(a)lists.freeswitch.org [freeswitch-users-bounces(a)lists.freeswitch.org [1]] On Behalf Of Mitul Limbani Sent: Friday, February 06, 2009 12:28 PM To: freeswitch-users(a)lists.freeswitch.org [2] Subject: Re: [Freeswitch-users] Transcoding G723 Hi, I came across a company who was selling hardware which could do like 5000 G729 conversion simultanoeusly, I was like this sounds cool, they have support for asterisk, I haven't enquirer yet how they do this, but anyone wishes to buy it cost £15000/year for the hrdware + support Any one on the list requires such high throughput calls can connect with me so that we can really test if it can work with freeswitch. Regards, Mitul Limbani, Founder & CEO, Enterux Solutions Pvt Ltd, The Enterprise Linux Company(r), http://www.enterux.com/ On 06-Feb-09, at 21:58, Shelby Ramsey wrote: > Steve, > > You definitely have a better grasp on this topic than me. But I > think it's a tough sell on the host based processing ... when you > look at products like what audio codes can do on a card (3 DS-3's > worth of transcoding) ... but I have had a couple of soft switch > vendors claim though that they could do 2,000 calls per host (but I > seriously doubt it). > > I agree that producing a card that does 120 channels is pretty > worthless ... but having something that could do say a 1000 would be > very helpful. > > G723 is a pretty big deal internationally ... I'm even seeing crazy > requests like AMR from folks trying to originate VoIP off of mobile > devices in Europe. > > But I agree just being able to g729 --> ulaw or ulaw --> g729 would > be a great first step (host based or otherwise). > > SDR > > _______________________________________________ > Freeswitch-users mailing list > Freeswitch-users(a)lists.freeswitch.org [4] > http://lists.freeswitch.org/mailman/listinfo/freeswitch-users > UNSUBSCRIBE:http://lists.freeswitch.org/mailman/options/freeswitch-users > http://www.freeswitch.org _______________________________________________ Freeswitch-users mailing list Freeswitch-users(a)lists.freeswitch.org [5] http://lists.freeswitch.org/mailman/listinfo/freeswitch-users UNSUBSCRIBE:http://lists.freeswitch.org/mailman/options/freeswitch-users http://www.freeswitch.org _______________________________________________ Freeswitch-users mailing list Freeswitch-users(a)lists.freeswitch.org [6] http://lists.freeswitch.org/mailman/listinfo/freeswitch-users UNSUBSCRIBE:http://lists.freeswitch.org/mailman/options/freeswitch-users http://www.freeswitch.org ------------------------- Msg sent via Enterux Enterprise Email Server : http://www.enterux.com/ Links: ------ [1] mailto:freeswitch-users-bounces@lists.freeswitch.org [2] mailto:freeswitch-users@lists.freeswitch.org [3] mailto:sicfslist@gmail.com [4] mailto:Freeswitch-users@lists.freeswitch.org [5] mailto:Freeswitch-users@lists.freeswitch.org [6] mailto:Freeswitch-users@lists.freeswitch.org

2 1

Re: [Freeswitch-users] mod_shout delay in trunk
by freeswitch-users＠digitaldan.com 06 Feb '09

06 Feb '09

For me it is. For what I'm using it for I can tolerate around a second or two delay. I have the icecast server setup to only buffer 1K for their on-connect burst as well as my flash/flex player to only buffer 1k (yes I might as well not buffer at all, which I may end up doing). In 1.0.2 this worked very well. Is this buffer configurable? If not, where is it being set? Thanks Dan- ----- Original Message ----- From: "Brian West" <brian(a)freeswitch.org> To: freeswitch-users(a)lists.freeswitch.org Sent: Friday, February 6, 2009 12:47:53 PM GMT -07:00 US/Canada Mountain Subject: Re: [Freeswitch-users] mod_shout delay in trunk Let me clarify.. yes this is normal file buffering was added so we wouldn't thrash your hard drive with tiny bits of data when recording calls so now it buffers and writes larger chunks to disk. This is why you have this delay which is 100% normal.... is realtime a critical thing? It is shout cast so you know it doesn't have to be realtime.. in fact some clients will buffer a little bit anyway and add to it. /b On Feb 6, 2009, at 1:43 PM, freeswitch-users(a)digitaldan.com wrote: I have, do you know what would have changed between 1.0.2 and trunk that would cause the buffer to change? Also if its not in mod_shout.c (which I copied from 1.0.2 to trunk for testing with no luck), where else would fs be buffering? One thing I have noticed is that in 1.0.2 as soon as the dial plan hits my record statement I see mod_shout logging that it has connected to the icecast server, in trunk it takes about 5 seconds to see the same log mesage. Below is my current svn info Path: . _______________________________________________ Freeswitch-users mailing list Freeswitch-users(a)lists.freeswitch.org http://lists.freeswitch.org/mailman/listinfo/freeswitch-users UNSUBSCRIBE:http://lists.freeswitch.org/mailman/options/freeswitch-users http://www.freeswitch.org

2 1

Re: [Freeswitch-users] mod_shout delay in trunk
by freeswitch-users＠digitaldan.com 06 Feb '09

06 Feb '09

I have, do you know what would have changed between 1.0.2 and trunk that would cause the buffer to change? Also if its not in mod_shout.c (which I copied from 1.0.2 to trunk for testing with no luck), where else would fs be buffering? One thing I have noticed is that in 1.0.2 as soon as the dial plan hits my record statement I see mod_shout logging that it has connected to the icecast server, in trunk it takes about 5 seconds to see the same log mesage. Below is my current svn info Path: . URL: http://svn.freeswitch.org/svn/freeswitch/trunk Repository Root: http://svn.freeswitch.org/svn Repository UUID: d0543943-73ff-0310-b7d9-9358b9ac24b2 Revision: 11669 Node Kind: directory Schedule: normal Last Changed Author: brian Last Changed Rev: 11669 Last Changed Date: 2009-02-06 11:29:51 -0700 (Fri, 06 Feb 2009) ----- Original Message ----- From: "Brian West" <brian(a)freeswitch.org> To: freeswitch-users(a)lists.freeswitch.org Sent: Friday, February 6, 2009 12:11:54 PM GMT -07:00 US/Canada Mountain Subject: Re: [Freeswitch-users] mod_shout delay in trunk Yes this will be normal due to buffering. Have you tested svn trunk? /b On Feb 6, 2009, at 12:59 PM, Dan wrote: Hi, With the 1.0.2 release i was able to to stream a call using mod_shout to an icecast server with only a 1 or two second delay to clients. With the current trunk that delay is now 8 to 10 seconds. I thought it might have been a change to mod_shout.c. I tried tweaking a few outbound buffer sizes with no luck so I just copied the 1.0.2 version of mod_shout.c over, compiled and reinstalled the module, restarted fs and still the delay is 8 to 10 seconds. I'm a little stumped. I currently have both versions installed (trunk and 1.0.2) for testing. Both are streaming to the same icecast server. My current svn revision is 11669, the calls are coming in via sip using g.711 ulaw and it looks like lame/mod_shout is streaming it as a 16kbs, 8khz mono mp3 stream. I'm using a flash/flex applet I wrote to consume the icecast stream, although I have used totem to listen to the stream as well. Any thoughts _______________________________________________ Freeswitch-users mailing list Freeswitch-users(a)lists.freeswitch.org http://lists.freeswitch.org/mailman/listinfo/freeswitch-users UNSUBSCRIBE:http://lists.freeswitch.org/mailman/options/freeswitch-users http://www.freeswitch.org

2 1

mod_shout delay in trunk
by Dan 06 Feb '09

06 Feb '09

Hi, With the 1.0.2 release i was able to to stream a call using mod_shout to an icecast server with only a 1 or two second delay to clients. With the current trunk that delay is now 8 to 10 seconds. I thought it might have been a change to mod_shout.c. I tried tweaking a few outbound buffer sizes with no luck so I just copied the 1.0.2 version of mod_shout.c over, compiled and reinstalled the module, restarted fs and still the delay is 8 to 10 seconds. I'm a little stumped. I currently have both versions installed (trunk and 1.0.2) for testing. Both are streaming to the same icecast server. My current svn revision is 11669, the calls are coming in via sip using g.711 ulaw and it looks like lame/mod_shout is streaming it as a 16kbs, 8khz mono mp3 stream. I'm using a flash/flex applet I wrote to consume the icecast stream, although I have used totem to listen to the stream as well. Any thoughts

2 1

Caller ID not being passed
by Nik Middleton 06 Feb '09

06 Feb '09

Try as I might, I cannot seem to get caller ID passed to the external sip gateway This GW happily processes caller id from Asterisk If tried adding param name="caller-id-in-from" value="true" in gw definition, and even <action application="set" data="effective_caller_id_number=07539600000"/> in the dial plan to no avail Can anyone shed some light on this? Regards, This is a sip debug sent 1382 bytes to udp/[XXX.XXX.XXX.XXX]:5060 at 21:45:17.404609: INVITE sip:07539000000@mygw.net SIP/2.0 Via: SIP/2.0/UDP 87.238.75.206:5080;rport;branch=z9hG4bKm2UveeUvtNc5D Max-Forwards: 69 From: "Extension 1000" <sip:me@mygw.net;transport=udp>;tag=j6m5U7g4F3XNa To: <sip:07539600000@gk.magrathea.net> Call-ID: 1e8c5d1d-6e71-122c-fdba-001a4b0a67ca CSeq: 110816823 INVITE Contact: <sip:me@myextIP:5080;transport=udp> Expires: 600 User-Agent: FreeSWITCH-mod_sofia/1.0.2-exported Allow: INVITE, ACK, BYE, CANCEL, OPTIONS, PRACK, MESSAGE, SUBSCRIBE, NOTIFY, REFER, UPDATE, REGISTER, INFO Supported: timer, precondition, path, replaces Allow-Events: talk, refer Proxy-Authorization: Digest username="me", realm="mygw.net", nonce="498b5f19b089db1bf9d13b2c83f45407048ede9b", algorithm=MD5, uri="sip:07539600000@mygw.net", response="03dbbeef1387a11c7100965dcfd01052" Min-SE: 120 Content-Type: application/sdp Content-Disposition: session Content-Length: 311 P-Key-Flags: keys="3"

2 9

Freeswitch freezes on increasing call traffic
by shehzad p 06 Feb '09

06 Feb '09

Hi all, Yesterday my Freeswitch server faced a problem when call traffic increased to more than 100. When I start Freeswitch, it works fine and then after some time (approximately 15 to 20 minutes) it stops functioning (means no call is being processed, no CLI command is working and it just freezes) until I restart the freeswitch. I am using Freeswitch 1.0.1. Debug (gdb) trace as on wiki page http://wiki.freeswitch.org/wiki/Debugging_Freeswitch#fs_debug.sh is attached http://www.nabble.com/file/p21701744/fs_debgu.txt fs_debgu.txt -- View this message in context: http://www.nabble.com/Freeswitch-freezes-on-increasing-call-traffic-tp21701… Sent from the Freeswitch-users mailing list archive at Nabble.com.

6 27

Freeswitch hangup cause code issue
by Ankit Gandhi 06 Feb '09

06 Feb '09

Hello, I want my caller to detect hangup cause as 34 so that he can try next provider according to lcr. Here is my setup. Caller -> switch (fs) -> Terminator. Now when terminator sends "503 Service Unavailable", I want to override this cause, so that the caller gets hangup case 34. (According to the terminator he is sending hangup cause 34 from his side, but in freeswitch we are getting hangup cause 41 for that call and the same hangup cause on caller side). When I tried asterisk as caller, I get hangup cause 34 in that case. But when I tried freeswitch as caller, then we are getting hangup cause 41, the same as we are getting in switch (fs). >From the switch, I tried one of this condition through javascript to override the hangup cause before sending to caller: -> session.execute("respond","503"); -> session.execute("hangup","NORMAL_CIRCUIT_CONGESTION"); -> session.execute("hangup","34"); -> session.hangup(34); In all the above cases, asterisk properly detects the hangup cause 34, but freeswitch does not detect that. It detects the same hangup cause 41 for the call. Other callers also get the same hangup cause 41 for such calls. How can I override this cause, so that the caller gets hangup cause 34 in such cases? Here is the sip trace, on the caller side returned through switch (fs). ss = switch cc = caller -------------------------------------------------------------------- U ss.ss.ss.ss:5060 -> cc.cc.cc.cc:5080 SIP/2.0 503 Service Unavailable. Via: SIP/2.0/UDP cc.cc.cc.cc:5080;rport=5080;branch=z9hG4bK8HvDU4Z88KU5m;received=122.169.29.122. From: "654321" <sip:654321@cc.cc.cc.cc>;tag=8p7aF18aN0mSj. To: <sip:123456@ss.ss.ss.ss>;tag=Nj8ypjDyvKUXe. Call-ID: fd1ab774-6ec5-122c-9fac-001cc086141d. CSeq: 110835048 INVITE. User-Agent: FreeSWITCH-mod_sofia/1.0.trunk-hacked. Accept: application/sdp. Allow: INVITE, ACK, BYE, CANCEL, OPTIONS, PRACK, MESSAGE, SUBSCRIBE, NOTIFY, REFER, UPDATE, REGISTER, INFO, PUBLISH. Supported: timer, precondition, path, replaces. Allow-Events: talk, presence, dialog, call-info, sla, include-session-description, presence.winfo, message-summary, refer. Reason: Q.850;cause=34;text="NORMAL_CIRCUIT_CONGESTION". Content-Length: 0. ----------------------------------------------------------------------- -- View this message in context: http://www.nabble.com/Freeswitch-hangup-cause-code-issue-tp21868781p2186878… Sent from the Freeswitch-users mailing list archive at Nabble.com.

2 1