Jeff,
Thanks for the feedback, I've set ipm_getxmitslot to sync mode and the timeslot info is now correctly set which now makes the record functionality work as expected.
That's excellent - than you very much.
Mike
Jeff,
Thanks for the feedback, I've set ipm_getxmitslot to sync mode and the timeslot info is now correctly set which now makes the record functionality work as expected.
That's excellent - than you very much.
Mike
Hi,
I’m currently using HMP 4.1_213 on a 64-bit RHEL 7.1 system. I have located the dialogic gc_basic_call_model_voice_video_3g_cnf_nbup demo application and have been trying to get this up and running to review the basic call control and voice functions.
I’ve managed to successfully compile the demo and can play out a G.711 A-law WAV file (after modifying static void dxPlayAudioPrompt(int index)) and the wav plays out as expected.
I’ve added a function to review recording audio to a wav file (static void dxRecordAudio(int index)) – this function uses the dx_reciottdata function to record a g.711A wav.
When I run the demo, dx_reciottdata returns success but the functionality appears to “hang” and the only way I can generate the TDX_RECORD termination event is if I hang-up the call. The recording wav file is actually created but it only contains 46 byes of header information.
I’ve reviewed the rtf logs (see attached) and I can’t see any errors being reported.
I’ve also attached the modified demo application which I’m using – does anyone have any pointers on where I may be going wrong or if I’ve missed anything obvious?
Many Thanks
Mike Sung
[View:~/cfs-file.ashx/__key/communityserver-discussions-components-files/7/1588.0310.gc_5F00_basic_5F00_call_5F00_model_5F00_voice_5F00_video_5F00_3g_5F00_cnf_5F00_nbup.c.txt:550:0]
[View:~/cfs-file.ashx/__key/communityserver-discussions-components-files/7/8154.5807.rtflog_2D00_LOCAL_2D00_20160629_2D00_09h13m26.270s.txt:550:0]
Does the card support host based stack without configuring the onboard NIC with an IP address.
We are using the current PCD file of @ipvs_evr_isdn_4ess_ml11_311.pcd
and FCD of @ipvs_evr_isdn_4ess_ml11_311.fcd
@ipvs_evr_isdn_ni2_ml11_311.fcd gave us issues starting. this card is connected with a CTBUS cable to other cards in an expansion box
Hi Jeff,
Thanks, you helped me very much.
One question remains, is how to properly convert raw b/w file to tiff? I tried to use different approaches but failed. The most obvious way was to use fax2tiff utility from libtiff library that "...creates a TIFF file containing CCITT Group 3 or Group 4 encoded data from one or more files containing ‘‘raw’’ Group 3 or Group 4 encoded data (typically obtained directly from a fax modem)" but it couldn't do anything either.
You need to configure the IP address for the board in the DCM no matter what as that controls the RTP stream. In this case host based stack, means the SIP stack is running on the physical computer and the RTP stream is running on the DM3 board. Otherwise known as split call control mode.
Jeff
Hi friends, i have the same question, im using the D120JCT-LS and i dont know how it tests, or what software i need to use, do you can tell me how do it, pls?.
Thanks!.
Hi at all!.
I have a D120JCT board, an i cant make a call from this ¿ do you have any software to test the board ? .
I use 2 analogic lines, Line 1 is uses the pin 3(-) and pin 4(+), Line 2 is uses the pin 2 (+) and pin 5 (-).
When i call to the line 1, it response with very high noise.
When i call to the line 2, it not response, its keeping ringing.
I m using the software “Polled Mode Phone Answering” to make tests, it comes with the Dialogic Release Software package.
Thanks for your help.
Regards.
Yes, you described everything quite correctly, just don't forget to call ipm_Listen from the back-end IPM resource to the receiving timeslot, when ASR comes to the picture. Therefore, you will need to use 2 RTP licenses per call, one for a caller, and another for Nuance. This is all true when using 1PCC mode.
Do you use CSP resource for bi-directional communications during ASR session?
I asked because with CSP and MRCP you don't need to use additional IPM resource. CSP streams audio data to the host, and your app can use MRCP to forward it to ASR server.
Is it about T38 faxing? If so, oes your license contain T38 resources, and do you specify T38 codec in your answer SDP on the first INVITE? It is necessary to prepare HMP to handle a switchover to T38 from audio call
No it's all about G.711 fax, I know the 2nd INVITE is superfluous, but it's impossible make the provider omit it.
Do I need to use T.38 fax instead? Will it help?
Hi,
I’m looking into the integration of HMP with Nuance currently – in particular, ASR. We actually have a MRCP client that we have developed so we can communicate with the Nuance server and retrieve the RTP Port that the Nuance session is listening on. From HMP we need to establish an IP connection to the Nuance session so that speech can be directed to Nuance.
We have a HMP application set up which supports G.711 SIP calls. I understand that when the SIP call is set up, the ipm_StartMedia function is used to start RTP streaming and in order to playback announcements - the various receive and transmit timeslots will need to be set for the voice and IP channel (dx_getxmitslot, ipm_Listen, ipm_GetXmitSlot, dx_listen).
The above is all working and I can place a call into the HMP application and have it play back an announcement. What I want to do now is to be able to create another RTP connection which will link with the Nuance session and have the audio from the Sip call diverted to it (if that makes sense).
Do I need to use a new IP device (i.e. call ipm_StartMedia on the new device with the RTP info in the IPM_MEDIA_INFO structure set to reference the IP address and port of Nuance)? If so, is there anything else that I need to do?
Thanks,
Mike Sung
Hi Leonid,
Thank you for your quick response to this. We have not currently looked into using CSP as yet but it is something which we will explore. For the time being we are looking to migrate from an existing solution (Natural Access that uses 2 IP streams for SIP and ASR) to HMP with a similar architecture so we’re currently reviewing the ipm functionality.
Regarding the call to ipm_Listen – I understand that this function connects the receive timeslot of the IP channel to the transmit channel of another device – I know that if I were to connect the voice and IPM devices - I would use dx_getxmitslot to get the transmit timeslot for the voice channel. Can you please advise on how I would set this up for ASR?
My thoughts are that for the 2nd (ASR) IPM device connection – I need to reference the transmit timeslot of the 1st (SIP) IPM connection (via ipm_GetXmitSlot) so that I can use this in ipm_Listen for the 2nd (ASR) connection? Is this correct?
Many Thanks,
Michael
There are two things I don't see often: first, unlike initial invite with G711 and RFC 2833 digits, the re-invite comes with G711 only. If the app does not change the DTMF xfer mode to inband before responding the second invite, HMP would reject such request due to incapable capacities.
Second, the switch adds silenceSupp off attribute, which is meaningless for HMP G711, since it is always off for this codec, but it should be OK, I guess, HMP would just ignore it.
RTF logger is set for errors only, so I can't tell what was the reason for rejection. Please enable Global Call, Global Call for IP, gc_h3r and SIP stack and retest. It should show you the exact reason for 488 Not Acceptable Here response.
Thanks Leonid, I will do as you proposed and let you know the results.
Hi all,
I'm looking for details about port density and performance for HPM 4.1 Linux (SU213) in relation to the hardware specs of the server, physical and virtual.
I'd like to have a table for example in which find info like "on a server with x CPU cores and y GB of RAM you can run up to n HMP ports", with particular attention to virtual environment, and how to optimize performance.
Thanks,
Roberto
Leonid, this is the rtflog file with all required options turned on, the SIP 488 response from Dialogic is here.
Thanks for the log. I think, I see where the problem is.
Your app is setting IPPARM_OPERATING_MODE to 0x2:
08/03/2016 19:29:05.151 2476 2668 gc_h3r SIP_CA..GR DEBG sip_callmanager:189 ! 2 ! >> callModifyStateReInviteReceived(), OperatingMode=2
It is defined in gcip_defs.h as
#define IP_T38_MANUAL_MODE 0x02 /* switch thru GCEV_EXTENSION event */
In this mode, HMP will send GCEV_EXTENSION on reinvite, and only when the re-invite is used for audio-t38 switchover; in audio-to-audio re-invites it is not is not handled, and the app will not receive GCEV_REQ_MODIFY_CALL event.
Your app must use IP_T38_MANUAL_MODIFY_MODE, even it does not use T38:
#define IP_T38_MANUAL_MODIFY_MODE 0x04 /* switch thru GCEV_REQ_MODIFY_CALL event */
it must be done on a board level device, something like this:
gc_util_insert_parm_val(&target_datap, IPSET_CONFIG, IPPARM_OPERATING_MODE, sizeof(int),
IP_T38_MANUAL_MODIFY_MODE);
gc_SetConfigData(GCTGT_CCLIB_NETIF, m_hBoard, target_datap,
0, GCUPDATE_IMMEDIATE, &req_id, EV_ASYNC);
Please give it a try and let me know if this resolve the problem
Also, please note
08/03/2016 19:29:05.151 2476 2668 gc_h3r ERR1 sip_callmanager:1276 ! 2 ! >> onReInviteToVoice: already in voice media, ignore
08/03/2016 19:29:05.151 2476 2668 gc_h3r WARN sip_callmanager:293 ! 2 ! << callModifyStateReInviteReceived(): failed: m_pChannManager->isMatchPerformed()=true, chanState=4 rc=-998
So, most likely, there is no need to modify your codec settings on such re-invite, the stack will do it.
Hello,
Encountered a problem with weird behavior of Dialogic HMP while handling incoming fax call over SIP. The provider's gateway after the first INVITE sends the 2nd one with more specific call settings. But HMP thinks that the 2nd INVITE is wrong and rejects it with the SIP 488 error. Then the gateway sends BYE cause he decides that the call was abandoned and breaks the session.
Is there a workaround to make HMP handle the 2nd INVITE politely?
I attached here the rtflog file and wireshark log file