Thread: [OpenSIPStack] Silence Detection
Brought to you by:
joegenbaclor
From: Whit T. <de...@wh...> - 2007-07-21 18:55:22
|
Hey Ilian, Thanks for exposing these methods in ATLSIP! Has there been any more work on the adaptive Silence Detection? The threshold of 3 is pretty stable, but I have end-users who complain that all calls are extremely 'one-sided' meaning that once one side of the call begins talking, the other person can't easily interrupt them. Logically that makes sense, but the end users I have love to interrupt each other so its rather annoying for them to wait till the other person is finished. Has anyone else deployed ATLSIP to real world end users? If so, what has your experience been like? I'd love to see what kind of settings (Silence, CN etc) others are using in the field. Regards, Whit -----Original Message----- From: ope...@li... [mailto:ope...@li...] On Behalf Of Ilian Jeri C. Pinzon Sent: Friday, June 29, 2007 6:58 AM To: ope...@li... Subject: Re: [OpenSIPStack] Problem with OnOutgoingCallConnected The attachment was not sent. Probably filtered by the server. Anyway, here it is: =========================================================================== Hi all, I have exposed the setting of silence detection mode and audio jitter delay in ATLSIP and SoftPhoneInterface. Here are the methods: DisableSilenceDetection() - Disables silence detection. Disables CNG as well. EnableFixedSilenceDetection( ULONG threshold ) - Enables fixed silence detection. Any sound level below the threshold is treated as silence (and CN is generated as a result). Don't use too high threshold values or you'll only hear comfort noise. Try threshold=3 as suggested by Whit in another thread. EnableFixedSilenceDetectionEx( ULONG threshold , ULONG signalDeadband, ULONG silenceDeadband ) - An extended version of the previous method. Don't tinker with this unless you know what you're doing. For reference on how signalDeadband and silenceDeadband are used, look in OpalSilenceDetector::ReceivedPacket(). EnableAdaptiveSilenceDetection( ULONG adaptivePeriod ) - Enables an adaptive silence detection. Supposedly this enables the threshold to *adapt* to the current sound level every adaptivePeriod milliseconds. However, its silence detection doesn't seem to be very effective (at least in my machine). I'll look into this further to see what's wrong. This mode with adaptivePeriod=4800 is the default mode for ATLSIP. EnableAdaptiveSilenceDetectionEx( ULONG adaptivePeriod, ULONG signalDeadband, ULONG silenceDeadband ) - An extended version of the previous method. Don't tinker with this unless you know what you're doing. For reference on how signalDeadband and silenceDeadband are used, look in OpalSilenceDetector::ReceivedPacket(). SetAudioJitterDelay( ULONG minDelay, ULONG maxDelay ) - Sets audio jitter delay settings. Regards, Ilian Ilian Jeri C. Pinzon wrote: > ATLSIP's default silence detection is still faulty. You can try > setting the silence detection to fixed mode. Somebody said in another > thread that a threshold of 3 will do the trick. You may have to tweak > this value though. > > Attached is the thread containing the functions to call for setting > the silence detection mode. > > - Ilian > > tomach wrote: >> Hello! >> >> Ok thank you very much! Till now everythign works correct :) Now some >> more advanced test...about quality of voice etc... >> >> It started to work correctly (From Tag was ok) after I build it as >> debug and reregistered again :) Now everythign works correct:) >> >> I still have like last question, do you have comfort noise generator? >> Because now its totally silence if somoebdy do not talk at all... and >> it could be uncomcofortable for subscirbers... >> >> Best Regards, >> Tom >> |
From: Ilian J. C. P. <ip...@so...> - 2007-07-23 10:46:05
|
Whit Thiele wrote: > Hey Ilian, > Hi Whit, > Thanks for exposing these methods in ATLSIP! > > Has there been any more work on the adaptive Silence Detection? The > threshold of 3 is pretty stable, but I have end-users who complain that all > calls are extremely 'one-sided' meaning that once one side of the call > begins talking, the other person can't easily interrupt them. Logically that > makes sense, but the end users I have love to interrupt each other so its > rather annoying for them to wait till the other person is finished. > I'm sorry I haven't had time to dig into the adaptive silence detection bug further. I'm currently busy with some other aspect of the softphone. But I promise I'll look into this closely in the near future. Btw last time I checked, OpalSilenceDetector::ReceivedPacket(..) was the culprit. You might want to take a look in that function if you have time. Regards, Ilian > Has anyone else deployed ATLSIP to real world end users? > > If so, what has your experience been like? I'd love to see what kind of > settings (Silence, CN etc) others are using in the field. > > > Regards, > Whit > > > > > > -----Original Message----- > From: ope...@li... > [mailto:ope...@li...] On Behalf Of Ilian > Jeri C. Pinzon > Sent: Friday, June 29, 2007 6:58 AM > To: ope...@li... > Subject: Re: [OpenSIPStack] Problem with OnOutgoingCallConnected > > The attachment was not sent. Probably filtered by the server. Anyway, here > it is: > > =========================================================================== > > Hi all, > > I have exposed the setting of silence detection mode and audio jitter > delay in ATLSIP and SoftPhoneInterface. > > Here are the methods: > > DisableSilenceDetection() > - Disables silence detection. Disables CNG as well. > > EnableFixedSilenceDetection( ULONG threshold ) > - Enables fixed silence detection. Any sound level below the threshold > is treated as silence (and CN is generated as a result). Don't use too > high threshold values or you'll only hear comfort noise. Try threshold=3 > as suggested by Whit in another thread. > > EnableFixedSilenceDetectionEx( ULONG threshold , ULONG signalDeadband, > ULONG silenceDeadband ) > - An extended version of the previous method. Don't tinker with this > unless you know what you're doing. For reference on how signalDeadband > and silenceDeadband are used, look in OpalSilenceDetector::ReceivedPacket(). > > EnableAdaptiveSilenceDetection( ULONG adaptivePeriod ) > - Enables an adaptive silence detection. Supposedly this enables the > threshold to *adapt* to the current sound level every adaptivePeriod > milliseconds. However, its silence detection doesn't seem to be very > effective (at least in my machine). I'll look into this further to see > what's wrong. This mode with adaptivePeriod=4800 is the default mode for > ATLSIP. > > EnableAdaptiveSilenceDetectionEx( ULONG adaptivePeriod, ULONG > signalDeadband, ULONG silenceDeadband ) > - An extended version of the previous method. Don't tinker with this > unless you know what you're doing. For reference on how signalDeadband > and silenceDeadband are used, look in OpalSilenceDetector::ReceivedPacket(). > > SetAudioJitterDelay( ULONG minDelay, ULONG maxDelay ) > - Sets audio jitter delay settings. > > > Regards, > Ilian > > > > Ilian Jeri C. Pinzon wrote: > >> ATLSIP's default silence detection is still faulty. You can try >> setting the silence detection to fixed mode. Somebody said in another >> thread that a threshold of 3 will do the trick. You may have to tweak >> this value though. >> >> Attached is the thread containing the functions to call for setting >> the silence detection mode. >> >> - Ilian >> >> tomach wrote: >> >>> Hello! >>> >>> Ok thank you very much! Till now everythign works correct :) Now some >>> more advanced test...about quality of voice etc... >>> >>> It started to work correctly (From Tag was ok) after I build it as >>> debug and reregistered again :) Now everythign works correct:) >>> >>> I still have like last question, do you have comfort noise generator? >>> Because now its totally silence if somoebdy do not talk at all... and >>> it could be uncomcofortable for subscirbers... >>> >>> Best Regards, >>> Tom >>> >>> > > > > ------------------------------------------------------------------------- > This SF.net email is sponsored by: Splunk Inc. > Still grepping through log files to find problems? Stop. > Now Search log events and configuration files using AJAX and a browser. > Download your FREE copy of Splunk now >> http://get.splunk.com/ > _______________________________________________ > opensipstack-devel mailing list > ope...@li... > https://lists.sourceforge.net/lists/listinfo/opensipstack-devel > > > |
From: tomach <to...@dg...> - 2007-07-27 11:29:57
|
Hello! About ATLsip Is it possible to obtain what codec is using during talk? (from the list of audiocodecs0-5)...or the only way is to analyze sip messages coming on SIPMessage event? BR, TOm |
From: Ilian J. C. P. <ip...@so...> - 2007-07-30 02:51:50
|
Hi Tom, This isn't exposed to ATLSIP yet. So as of the moment reading the SIP message is your only workaround. I'll put this in my list of TODOs. Btw, just curious. Why do you need to know the codec? Regards, Ilian tomach wrote: > Hello! > > About ATLsip > Is it possible to obtain what codec is using during talk? (from the list of audiocodecs0-5)...or the only way is to analyze sip messages coming on SIPMessage event? > > BR, > TOm > > ------------------------------------------------------------------------- > This SF.net email is sponsored by: Splunk Inc. > Still grepping through log files to find problems? Stop. > Now Search log events and configuration files using AJAX and a browser. > Download your FREE copy of Splunk now >> http://get.splunk.com/ > _______________________________________________ > opensipstack-devel mailing list > ope...@li... > https://lists.sourceforge.net/lists/listinfo/opensipstack-devel > > > |
From: tomach <to...@dg...> - 2007-07-30 07:48:52
|
Because what I noticed is that: changing microphone level make quality of voice worse with g.729 rather then with g.711. So I thought that when codec g.729 is in use i can disable control responsible for changing voice volume... |
From: Ilian J. C. P. <ip...@so...> - 2007-07-31 11:10:03
|
Just tested this a while ago. This does not happen in my build. Have you eliminated the possibility that this is caused by your mic? I've used a mic with no noise-canceling features before and when I throttled the mic's volume, it became sensitive to noise. Thus a lower voice quality. Btw, how are you setting the volume? - Ilian tomach wrote: > Because what I noticed is that: changing microphone level make quality of voice worse with g.729 rather then with g.711. > So I thought that when codec g.729 is in use i can disable control responsible for changing voice volume... > > ------------------------------------------------------------------------- > This SF.net email is sponsored by: Splunk Inc. > Still grepping through log files to find problems? Stop. > Now Search log events and configuration files using AJAX and a browser. > Download your FREE copy of Splunk now >> http://get.splunk.com/ > _______________________________________________ > opensipstack-devel mailing list > ope...@li... > https://lists.sourceforge.net/lists/listinfo/opensipstack-devel > > > |