Any experience with DTMF from FreeSwitch to Sonus with Vega ATAs?
Hello, Anyone has an experience with the above configuration? All sides are marked as RFC-2833 and the Sonus recognises most of the DTMF's twice. The Vega sends both RFC-2833 events and about 60msec of the DTMF tone (according to Vega's engineers it takes it about 60 msec to detect the DTMF tones). According to the RFC this is acceptable. The service provider I am working with says that it shouldn't be like that. Before I waste time in wars, has anybody had this issue and knows whether it is a matter of configuration at the Sonus side? Thanks! __Yehavi:
Sonus is known to do stupid crap with DTMF... Can you get a good pcap with a full end to end sip and rtp capture? K On 2/20/11 1:31 AM, "Yehavi Bourvine" <yehavi.bourvine@gmail.com> wrote:
Hello, Anyone has an experience with the above configuration? All sides are marked as RFC-2833 and the Sonus recognises most of the DTMF's twice. The Vega sends both RFC-2833 events and about 60msec of the DTMF tone (according to Vega's engineers it takes it about 60 msec to detect the DTMF tones). According to the RFC this is acceptable. The service provider I am working with says that it shouldn't be like that. Before I waste time in wars, has anybody had this issue and knows whether it is a matter of configuration at the Sonus side? Thanks! __Yehavi:
_______________________________________________ FreeSWITCH-users mailing list FreeSWITCH-users@lists.freeswitch.org http://lists.freeswitch.org/mailman/listinfo/freeswitch-users UNSUBSCRIBE:http://lists.freeswitch.org/mailman/options/freeswitch-users http://www.freeswitch.org
On 02/20/2011 03:31 PM, Yehavi Bourvine wrote:
Hello, Anyone has an experience with the above configuration? All sides are marked as RFC-2833 and the Sonus recognises most of the DTMF's twice. The Vega sends both RFC-2833 events and about 60msec of the DTMF tone (according to Vega's engineers it takes it about 60 msec to detect the DTMF tones). According to the RFC this is acceptable. The service provider I am working with says that it shouldn't be like that. Before I waste time in wars, has anybody had this issue and knows whether it is a matter of configuration at the Sonus side? Thanks! __Yehavi: Are you saying the Vega sends 60ms of DTMF as audio, then sends RFC2833 DTMF packets and mutes the audio? In what reading of the RFC is that considered appropriate action? In addition, a DTMF detector should detect is more like 40ms. 60ms is rather long.
Steve
Hello Steve, I uderstand from paragraph 3.2 in RFC-2833 that the sender may send some audio of the DTMF while sending the events, until it revognises that this is a DTMF tone. The 60mSec is what I've been told by Vega engineers. Thanks! __Yehavi: 2011/2/20 Steve Underwood <steveu@coppice.org>
On 02/20/2011 03:31 PM, Yehavi Bourvine wrote:
Hello, Anyone has an experience with the above configuration? All sides are marked as RFC-2833 and the Sonus recognises most of the DTMF's twice. The Vega sends both RFC-2833 events and about 60msec of the DTMF tone (according to Vega's engineers it takes it about 60 msec to detect the DTMF tones). According to the RFC this is acceptable. The service provider I am working with says that it shouldn't be like that. Before I waste time in wars, has anybody had this issue and knows whether it is a matter of configuration at the Sonus side? Thanks! __Yehavi: Are you saying the Vega sends 60ms of DTMF as audio, then sends RFC2833 DTMF packets and mutes the audio? In what reading of the RFC is that considered appropriate action? In addition, a DTMF detector should detect is more like 40ms. 60ms is rather long.
Steve
_______________________________________________ FreeSWITCH-users mailing list FreeSWITCH-users@lists.freeswitch.org http://lists.freeswitch.org/mailman/listinfo/freeswitch-users UNSUBSCRIBE:http://lists.freeswitch.org/mailman/options/freeswitch-users http://www.freeswitch.org
Hi Yehavi, Section 3.2 of RFC2833 is poorly worded, and causes a lot of issues in the real world. The first paragraph says you can choose to send both audio and events, but doesn't clarify how. Presumably RFC2198 should be used. Is your system using RFC2198 encoding? I doubt it. The second paragraph talks about tone onset getting through as audio. Valid DTMF can be as short as 45ms, so 60ms is far more than anything which might be termed onset. Because they didn't specify anything about what is tolerable as tone onset passing through the channel as audio we have chaos today. RFC2833 is an obsolete spec, and we should be talking about RFC4733 today. Section 2.5.1.3.1 of RFC4733 does clearly call for combined payloads to be sent as RFC2198 packets. RFC4733 is even vaguer about the onset issue, though. Steve On 02/20/2011 06:33 PM, Yehavi Bourvine wrote:
Hello Steve, I uderstand from paragraph 3.2 in RFC-2833 that the sender may send some audio of the DTMF while sending the events, until it revognises that this is a DTMF tone. The 60mSec is what I've been told by Vega engineers. Thanks! __Yehavi:
2011/2/20 Steve Underwood <steveu@coppice.org <mailto:steveu@coppice.org>>
On 02/20/2011 03:31 PM, Yehavi Bourvine wrote: > Hello, > Anyone has an experience with the above configuration? All sides are > marked as RFC-2833 and the Sonus recognises most of the DTMF's twice. > The Vega sends both RFC-2833 events and about 60msec of the DTMF tone > (according to Vega's engineers it takes it about 60 msec to detect the > DTMF tones). > According to the RFC this is acceptable. The service provider I am > working with says that it shouldn't be like that. Before I waste time > in wars, has anybody had this issue and knows whether it is a matter > of configuration at the Sonus side? > Thanks! __Yehavi: Are you saying the Vega sends 60ms of DTMF as audio, then sends RFC2833 DTMF packets and mutes the audio? In what reading of the RFC is that considered appropriate action? In addition, a DTMF detector should detect is more like 40ms. 60ms is rather long.
Steve
_______________________________________________ FreeSWITCH-users mailing list FreeSWITCH-users@lists.freeswitch.org <mailto:FreeSWITCH-users@lists.freeswitch.org> http://lists.freeswitch.org/mailman/listinfo/freeswitch-users UNSUBSCRIBE:http://lists.freeswitch.org/mailman/options/freeswitch-users http://www.freeswitch.org <http://www.freeswitch.org/>
_______________________________________________ FreeSWITCH-users mailing list FreeSWITCH-users@lists.freeswitch.org http://lists.freeswitch.org/mailman/listinfo/freeswitch-users UNSUBSCRIBE:http://lists.freeswitch.org/mailman/options/freeswitch-users http://www.freeswitch.org
Hello Steve, Thanks for clarifying it. __Yehavi: 2011/2/20 Steve Underwood <steveu@coppice.org>
Hi Yehavi,
Section 3.2 of RFC2833 is poorly worded, and causes a lot of issues in the real world. The first paragraph says you can choose to send both audio and events, but doesn't clarify how. Presumably RFC2198 should be used. Is your system using RFC2198 encoding? I doubt it. The second paragraph talks about tone onset getting through as audio. Valid DTMF can be as short as 45ms, so 60ms is far more than anything which might be termed onset. Because they didn't specify anything about what is tolerable as tone onset passing through the channel as audio we have chaos today.
RFC2833 is an obsolete spec, and we should be talking about RFC4733 today. Section 2.5.1.3.1 of RFC4733 does clearly call for combined payloads to be sent as RFC2198 packets. RFC4733 is even vaguer about the onset issue, though.
Steve
On 02/20/2011 06:33 PM, Yehavi Bourvine wrote:
Hello Steve, I uderstand from paragraph 3.2 in RFC-2833 that the sender may send some audio of the DTMF while sending the events, until it revognises that this is a DTMF tone. The 60mSec is what I've been told by Vega engineers. Thanks! __Yehavi:
2011/2/20 Steve Underwood <steveu@coppice.org <mailto:steveu@coppice.org
On 02/20/2011 03:31 PM, Yehavi Bourvine wrote: > Hello, > Anyone has an experience with the above configuration? All sides are > marked as RFC-2833 and the Sonus recognises most of the DTMF's twice. > The Vega sends both RFC-2833 events and about 60msec of the DTMF tone > (according to Vega's engineers it takes it about 60 msec to detect the > DTMF tones). > According to the RFC this is acceptable. The service provider I am > working with says that it shouldn't be like that. Before I waste time > in wars, has anybody had this issue and knows whether it is a matter > of configuration at the Sonus side? > Thanks! __Yehavi: Are you saying the Vega sends 60ms of DTMF as audio, then sends RFC2833 DTMF packets and mutes the audio? In what reading of the RFC is that considered appropriate action? In addition, a DTMF detector should detect is more like 40ms. 60ms is rather long.
Steve
_______________________________________________ FreeSWITCH-users mailing list FreeSWITCH-users@lists.freeswitch.org <mailto:FreeSWITCH-users@lists.freeswitch.org> http://lists.freeswitch.org/mailman/listinfo/freeswitch-users UNSUBSCRIBE: http://lists.freeswitch.org/mailman/options/freeswitch-users http://www.freeswitch.org <http://www.freeswitch.org/>
_______________________________________________ FreeSWITCH-users mailing list FreeSWITCH-users@lists.freeswitch.org 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@lists.freeswitch.org http://lists.freeswitch.org/mailman/listinfo/freeswitch-users UNSUBSCRIBE:http://lists.freeswitch.org/mailman/options/freeswitch-users http://www.freeswitch.org
participants (3)
-
Ken Rice
-
Steve Underwood
-
Yehavi Bourvine