VoiceGuide IVR Software Main Page
Jump to content

Questions On Lumenvox Speech Recognition

Recommended Posts

Hi,

 

I purchased the Lumenvox speech recognition engine, and I have a few questions:

 

1.

In the VG Config.xml, I put this at the end of the file before </VoiceGuideConfig>:

 

<Devices_ASR>

<Server_MRCPv2>

<ResourceType>speechrecog</ResourceType>

<ClientIP>10.1.1.9</ClientIP>

<ClientPort>8062</ClientPort>

<ServerIP>10.1.1.8</ServerIP>

<ServerPort>8060</ServerPort>

<RtpMin>49200</RtpMin>

<RtpMax>65300</RtpMax>

</Server_MRCPv2>

</Devices_ASR>

 

I'm not sure what to put for the ServerIP and ClientIP and the respective ports. I've installed everything on the same machine that the VG service & Dialogic card is. What should I put? Thanks.

 

2.

Does the speech recognition work with both Play module and Get Numbers module? If it only works with the Play module, how do I do multi-digit input (like 8465 for an extension)? Can the recognition recognize multiple words and multiple numbers?

 

3.

If my system is in English, how am I supposed to know what languages/accents to use in my grammar files? (e.g. Australian or American English)?

 

4.

Do you know how I can do voice to text, e.g. leave a voicemail message and have the system convert the speech to text which can be emailed?

Share this post


Link to post
I'm not sure what to put for the ServerIP and ClientIP and the respective ports. I've installed everything on the same machine that the VG service & Dialogic card is.

Your machine's IP address.

 

Can the recognition recognize multiple words and multiple numbers?

Yes. It all depends on how you define the grammar.

 

If my system is in English, how am I supposed to know what languages/accents to use in my grammar files? (e.g. Australian or American English)?

... you'd use Australian if the system is in Australia and/or callers have Australian accents, and you'd use the American grammar if the system is in US and/or callers have American accents...

 

Do you know how I can do voice to text, e.g. leave a voicemail message and have the system convert the speech to text which can be emailed?

You would not be able to do full voice to text with a Lite license, and you would need to do your own research to see if the full LumenVox license would meet your requirements.

 

Sounds like you should do some research into grammars as well.

Share this post


Link to post
QUOTE
QUOTE
I'm not sure what to put for the ServerIP and ClientIP and the respective ports. I've installed everything on the same machine that the VG service & Dialogic card is.

Your machine's IP address.

QUOTE
Can the recognition recognize multiple words and multiple numbers?

Yes. It all depends on how you define the grammar.

QUOTE
If my system is in English, how am I supposed to know what languages/accents to use in my grammar files? (e.g. Australian or American English)?

... you'd use Australian if the system is in Australia and/or callers have Australian accents, and you'd use the American grammar if the system is in US and/or callers have American accents...

QUOTE
Do you know how I can do voice to text, e.g. leave a voicemail message and have the system convert the speech to text which can be emailed?

You would not be able to do full voice to text with a Lite license, and you would need to do your own research to see if the full LumenVox license would meet your requirements.

Sounds like you should do some research into grammars as well.


1. For my machine's IP address, should I put in something like localhost or 127.0.0.1, or should I use the one from www.whatismyip.com? And what should I use for the port?
2. For the language/accents, what I meant is if some of my callers will be using Australian accents, and some will be using American accents... can I set it up to recognize both?
3. Again: Does the speech recognition work with both Play module and Get Numbers module? If it only works with the Play module, how do I do multi-digit input (like 8465 for an extension)?

Share this post


Link to post
1. For my machine's IP address, should I put in something like localhost or 127.0.0.1, or should I use the one from www.whatismyip.com?

Use ipconfig.exe from the DOS Command line.

 

And what should I use for the port?

Leave them as they are.

 

2. For the language/accents, what I meant is if some of my callers will be using Australian accents, and some will be using American accents... can I set it up to recognize both?

No.

 

3. Again: Does the speech recognition work with both Play module and Get Numbers module?

Only works with Play module.

 

If it only works with the Play module, how do I do multi-digit input (like 8465 for an extension)?

If your grammar allow multi digit/word input then this will allow multi-digit/word input in the Play module.

 

Share this post


Link to post

Okay. If the speech recognition only works with the Play module, and I want the caller to be able to say AND type in 8465, then how would I do that?

 

The play module won't accept multi-digit keypad inputs.

Share this post


Link to post

I tried putting in the correct IP addresses, and the sample yes/no script I downloaded from the VG website is still not working. I've attached log files - could you please help? Thanks.

log.zip

Share this post


Link to post
If the speech recognition only works with the Play module, and I want the caller to be able to say AND type in 8465, then how would I do that?

The ASR engine should support both DTMF detection and speech. See: http://www.w3.org/TR/speech-grammar/#AppE

 

I tried putting in the correct IP addresses, and the sample yes/no script I downloaded from the VG website is still not working. I've attached log files - could you please help?

Configuration looks fine, but the card that you are using is not Speech Capable (it does not support Echo Cancelled Streaming) so looks like it cannot be used for Speech Recognition:

 

from ktTel log:

 

180515.093 0896 001 ERROR ec_stream Function not supported, 8

Share this post


Link to post

I'm using a Dialogic® D/120JCT-LS Combined Media Board. It supports speech recognition.

According to 120JCT Data Sheet:

 

"This globally approved product offers functionality such as... full duplex echo cancellation..."
"Supports DSP-based onboard fax and host-based speech recognition to maximize the number of boards in the system"

Could it be that I have to enable Echo Cancelled Streaming somehow? Please help me get this to work. Thanks.

Share this post


Link to post

Go to Dialogic Configuration Manager and view the properties for your card. On the Misc tab see if the "CSP_Enabled" parameter is set to "Yes".

 

If it is and you are still seeing the "ec_stream Function not supported" in the ktTel trace then you should contact the card supplier.

 

Are you using the current model D/120JCT? It is possible that older models of the D/120JCT were not Speech Enabled.

Share this post


Link to post

Yes, the CSP_Enabled parameter is set to yes. Do you know of any other parameters I need to check?

 

I have a D/120JCT Rev. B. I believe it is the latest... do you know of any other versions?

 

Thanks.

Share this post


Link to post

You may need to directly specify the Firmware file on that card.

 

Stop the Dialogic service and in the Dialogic Configuration Manager bring up the prioperties for that card. Go to the Misc tab and select this to be the Firmware file:

 

d120csp.fwl

Share this post


Link to post
You may need to directly specify the Firmware file on that card.

 

Stop the Dialogic service and in the Dialogic Configuration Manager bring up the prioperties for that card. Go to the Misc tab and select this to be the Firmware file:

 

d120csp.fwl

 

I tried that and speech recognition still doesn't work.

 

Here is a segment from the logs: now it's saying that FT_CSP flag is not present, but it also says ec_stream ok.

 

Please help.

 

 

 

204044.546 4708 001 play(1, 0x6b208d8, tpt=0x0, xpb=0x1e26d4) => 0, hli=0x1def40

204044.546 4708 001 asr ec streaming enabled as MRCPv2 connector specified

204044.546 4708 001 EcStream_Start begin

204044.546 4708 FT_CSP flag not present. ec_stream may not be supported on this voice device. ft_e2p_brd_cfg=0x0, FT_CSP=0x100

204044.546 4708 001 ec_stream call hli->voicedev=1 recformat=7 pool=06B2C890

204044.546 4708 001 voice stream format: ULaw (7)

204044.546 4708 001 ERROR ec_setparm(1, DXCH_EC_TAP_LENGTH). Err Msg = Device busy, Lasterror = 9.

204044.546 4708 001 check if the current board and channel are CSP enabled.

204044.546 4708 001 ERROR ec_setparm(ECCH_NLP). Err Msg = Device busy, Lasterror = 9

204044.546 4708 001 check if the current board and channel are CSP enabled

204044.546 4708 001 InitParams_EC end

204044.562 4708 001 ERROR ec_setparm(DXCH_BARGEIN). Err Msg = Device busy, Lasterror = 9

204044.562 4708 001 ERROR ec_setparm(DXCH_BARGEINONLY). Err Msg = Device busy, Lasterror = 9

204044.562 4708 xpb wDataFormat=0x7(7) (1=adpcm, 3=alaw, 7=ulaw, 8=pcm), wFileFormat=1 (1=vox, 2=wav)

204044.562 4708 001 ec_stream ok

204044.562 4708 001 ec streaming started. buffpool=0x6b2c890

204044.562 4708 001 CtEventProcess (from store) idx=16, iDev=1, lEvtType=134, pEvtData=0x6ab7878, lEvtDataLen=4, (store: evinque=0, maxever=1)

204044.562 4708 001 ev TDX_CST (CST Event Received)

204044.562 4708 001 TDX_CST DE_LCON data=7583

204044.562 4708 001 raise Dialogic TDX_CST 134 (7583 0 0 DE_LCON )

204045.515 4708 001 CtEventProcess (from store) idx=17, iDev=1, lEvtType=129, pEvtData=0x6ab78a8, lEvtDataLen=0, (store: evinque=0, maxever=1)

204045.515 4708 001 ev TDX_PLAY (Play Completed)

204045.515 4708 001 EvHandler_TDX_PLAY hPlayRec_IottChainStart=0x6b208d8, hPlayRec_IottChainStart_Copy=0x6b2a970

204045.515 4708 001 play free(0x6b7a0e8, 0x6b208d8) eot

204045.515 4708 001 cleared dwPlayId: hli=0x1def40, hli->dwPlayId=0(0x0)

204045.515 4708 001 CTelProxy::Event_PlayEnd begin

204045.515 4708 001 raise PlayEnd 712250

204045.515 4708 001 CTelProxy::Event_PlayEnd end

204049.546 4312 001 ERROR mrcp recog thrd recognizer start failed. Timout awaitng for channel create completion. Recognizer not active.

204049.546 4312 mrcp CMrcpClientSpeechRecog::SessionTerminateAndDestroy begin

204049.546 4312 mrcp SessionTerminateAndDestroy calling mrcp_client_context_session_terminate

204050.984 4680 001 fn PlayStart(iLineId=1, sFileList=,C:\Program Files\VoiceGuide\system\voice.wav, sXMLOptions=, lActionKeysEnabled=0)

204050.984 4680 001 PlayStart(hLine=1, strSoundFile=,C:\Program Files\VoiceGuide\system\voice.wav, iPlayId=718687(0xaf75f), iParam1=0, iParam2=0, zsParam1=, zsParam2=, keys[NotUsed])

204050.984 4680 001 PlaySetControlKeys(1,0,,,,,,,,,,)

204050.984 4680 001 PlaySetControlKeys end

204050.984 4680 001 play start (hli=0x1def40, strSoundFile=,C:\Program Files\VoiceGuide\system\voice.wav, iPlayId=718687(0xaf75f), iParam1=0, iParam2=0, zsParam1=, zsParam2=)

204050.984 4680 data tag found 15 chars after bits per sample field.

204050.984 4680 wav: format=7, channels=1, hz=8000, bytes/sec=8000, bytes/sample=1, bits/sample=8, DataBlockSize=7154 C:\Program Files\VoiceGuide\system\voice.wav

204050.984 4680 leading: 0x76 0x76, remaining data_size=7152 (from start=2)

204050.984 4680 001 iXpbSampleRateDRT_FirstFile = DRT_8KHZ

204050.984 4680 001 iXpbBitsPerSample_FirstFile = 8 (as per returned data)

204050.984 4680 001 XpbDataFormat: DATA_FORMAT_MULAW, rate_const=0x40, bits=8 (iWavFilesEncodingFormat_LastFile=7)

204050.984 4680 001 dx_setevtmsk(1, [MASK]) => 0

204050.984 4680 001 dx_setdigtyp(1, DM_DTMF) => 0

204050.984 4680 001 dx_clrdigbuf(1) => 0

204050.984 4680 001 dx_clrsvcond(1) => 0

204050.984 4680 001 xpb: wDataFormat=7, wFileFormat=1, nSamplesPerSec=0x40, wBitsPerSample=8 ()

204050.984 4680 001 play(1, iott=0x6b27cc0, tpt=0x0, xpb=0x1e26d4) call

204050.984 4680 001 play(1, 0x6b27cc0, tpt=0x0, xpb=0x1e26d4) => 0, hli=0x1def40

204051.546 4312 mrcp SessionTerminateAndDestroy calling mrcp_client_context_session_destroy

204051.546 4312 001 raise CTelProxy::Event_Generic m_pTelProxyClient=00691460

204051.546 4312 001 raise mrcp|4313 SESSION_THREAD_COMPLETED|0 0 (0 0 0 source:RecogThread )

204051.953 4708 001 CtEventProcess (from store) idx=18, iDev=1, lEvtType=129, pEvtData=0x6ab78d8, lEvtDataLen=0, (store: evinque=0, maxever=1)

204051.953 4708 001 ev TDX_PLAY (Play Completed)

204051.953 4708 001 EvHandler_TDX_PLAY hPlayRec_IottChainStart=0x6b27cc0, hPlayRec_IottChainStart_Copy=0x6b27c98

204051.953 4708 001 play free(0x6b7a0e8, 0x6b27cc0) eot

204051.953 4708 001 cleared dwPlayId: hli=0x1def40, hli->dwPlayId=0(0x0)

204051.953 4708 001 CTelProxy::Event_PlayEnd begin

204051.953 4708 001 raise PlayEnd 718687

204051.953 4708 001 CTelProxy::Event_PlayEnd end

204054.468 4708 001 CtEventProcess (from store) idx=19, iDev=1, lEvtType=134, pEvtData=0x6ab7908, lEvtDataLen=16, (store: evinque=0, maxever=1)

204054.468 4708 001 ev TDX_CST (CST Event Received)

204054.468 4708 001 TDX_CST DE_TONEON cst_data=192

204054.468 4708 001 tone description retrieve hli=0x1def40, iToneUserID[1]=192 sToneUserName[1]=[DISCONNECT_TAPI1]

204054.468 4708 001 raise Dialogic TDX_CST 134 (192 0 0 DE_TONEON DISCONNECT_TAPI1 )

204054.468 4708 001 CTelProxy::Event_CallState m_pTelProxyClient=00691460

204054.468 4708 001 raise CallState LINECALLSTATE_DISCONNECTED-DISCONNECT_TAPI1

204055.468 4888 001 fn DropCall(sLineId=1, sXMLOptions=[], iParam1=0)

204055.468 4888 001 TelDriver_DropCall(sXMLOptions=[])

204055.468 4888 001 dx_sethook(1) call

204055.468 4888 001 dx_sethook 1 DX_ONHOOK ok

204055.500 4708 001 CtEventProcess (from store) idx=20, iDev=1, lEvtType=135, pEvtData=0x6ab7938, lEvtDataLen=4, (store: evinque=0, maxever=1)

204055.500 4708 001 ev TDX_SETHOOK (SetHook Completed)

204055.500 4708 001 raise Dialogic TDX_SETHOOK 135 (0 0 0 DX_ONHOOK )

204055.500 4708 001 CTelProxy::Event_CallState m_pTelProxyClient=00691460

204055.500 4708 001 raise CallState LINECALLSTATE_IDLE

204055.500 4708 001 fn ReleaseCall(sLineId=1, crn=0x0, Param1=0)

204055.500 4708 001 TelDriver_ReleaseCall channelType=analog. just signal back that ReleaseCall completed

204055.500 4708 001 CTelProxy::Event_CallState m_pTelProxyClient=00691460

204055.500 4708 001 raise CallState GCEV_RELEASECALL

204055.500 4708 001 TelDriver_ReleaseCall Event_CallState returned.

204055.500 4708 001 ec_stopch call

204055.500 4708 001 ec_stopch ok

204055.515 4704 001 ec_cb 384

204055.515 4704 001 ec_cb 127

204055.515 4708 001 CtEventProcess (from store) idx=21, iDev=1, lEvtType=224, pEvtData=0x6ab7968, lEvtDataLen=0, (store: evinque=0, maxever=1)

204055.515 4708 001 ev TEC_STREAM (termination event for ec_stream/ec_reciottdata)

204055.515 4708 001 ev TEC_STREAM ec_stream/ec_reciottdata completed

204055.515 4708 001 raise Dialogic TEC_STREAM 224 (0 0 0 )

204055.531 4752 001 iTimer_RingIgnore_InGuardTimeAfterHangup=20

204056.531 4752 001 iTimer_RingIgnore_InGuardTimeAfterHangup=10

 

 

 

Share this post


Link to post
You may need to directly specify the Firmware file on that card.

 

Stop the Dialogic service and in the Dialogic Configuration Manager bring up the prioperties for that card. Go to the Misc tab and select this to be the Firmware file:

 

d120csp.fwl

 

I tried that and speech recognition still doesn't work.

 

Here is a segment from the logs: now it's saying that FT_CSP flag is not present, but it also says ec_stream ok.

 

Please help.

 

 

 

Hi, I am waiting for your reply in order to continue working on this project.

 

Please respond as soon as you can. I'd most appreciate it.

Share this post


Link to post
You may need to directly specify the Firmware file on that card.

 

Stop the Dialogic service and in the Dialogic Configuration Manager bring up the prioperties for that card. Go to the Misc tab and select this to be the Firmware file:

 

d120csp.fwl

 

I tried that and speech recognition still doesn't work.

 

Here is a segment from the logs: now it's saying that FT_CSP flag is not present, but it also says ec_stream ok.

 

Please help.

 

 

 

Hi, I am waiting for your reply in order to continue working on this project.

 

Please respond as soon as you can. I'd most appreciate it.

 

 

I posted this question below on the Voiceguide support forums. It has been

over 3-4 days and I still haven't received a response. What is the problem?

It's not like support hasn't been responding to other people's questions.

 

This happened just after I went ahead and bought Voiceguide Enterprise. Why

does this sudden lack of support coincide with my buying the license?

 

Does that mean you don't provide support any more after clients make purchase?

 

Please let me know why my post hasn't been responded to and help me resolve

this problem. If you don't provide support after the purchase is made to your products, then you should say so so that customers know before they buy!

 

 

 

 

Share this post


Link to post

We are still awaiting response from the relevant engineer on this. We are in the middle of the Easter holiday break now, so we would expect some feedback early next week.

Share this post


Link to post

We first need to configure this card so that it will correctly respond to the applications attempts to setup the "CSP" - which is required for Speech Recognition.

 

Your system is showing this:

 

204044.546 4708 FT_CSP flag not present. ec_stream may not be supported on this voice device. ft_e2p_brd_cfg=0x0, FT_CSP=0x100

 

And a correctly setup system would shows something like this in the log trace:

 

212257.506 3108 001 ec FT_CSP flag is present. ec_stream is supported on this voice device. ft_e2p_brd_cfg=0x126, FT_CSP=0x100

212257.506 0916 ec buff forwarding thread starting ID=ktTel parent=0x06D05DA8 1024x388

212257.506 3108 001 ec ec_stream call hli->voicedev=1 recformat=7 pool=06D05DA8

212257.506 3108 001 ec voice stream format: ULaw (7)

212257.521 3108 001 ec_setparm(1, DXCH_SPEECHPLAYTHRESH, -4) ok

212257.521 3108 001 ec ec_setparm end

212257.521 3108 xpb wDataFormat=0x7(7) (1=adpcm, 3=alaw, 7=ulaw, 8=pcm), wFileFormat=1 (1=vox, 2=wav)

212257.521 3108 001 ec ec_stream ok

212257.521 3108 001 ec streaming started. buffpool=0x6d05da8

 

(above trace taken form our test system running on a D/41JCT card)

 

So we'd first need to figure out why the Dialogic drivers are saying that this card does not support CSP.

 

What Dialogic drivers are you using on this system? Can you give us access to this system using www.logmein.com or VNC or similar?

 

 

 

Share this post


Link to post

CAn you tried running the CSPLive.exe test application from Dialogic to confirm that the CSP is enabled on this card?

 

The CSPLive.exe can be found in Dialogic's \demos\SpeechProcessing\CSPLive directory.

Share this post


Link to post

Here is the CSPLive.exe which you can use to test if your card has the CSP enabled. It has been slightly modified to include more logging and to fix a bug in the original CSPLive.exe from Dialogic which occasionally reports an error when setting the DXCH_EC_TAP_LENGTH parameter.

CSPLive.zip

Share this post


Link to post

1. I'm using Dialogic System Release 6.0 PCI in a Windows XP machine.

2. I ran CSPLive.exe with your modified file and I got the errors in the attached file. Please take a look. Thanks.

post-2066-1207114288_thumb.jpg

Share this post


Link to post

Screenshot shows that all the calls to the CSP related functions failed.

 

If in the Dialogic Configuration Manager you have done all of the following:

 

- selected to use the CSP firmware,

- set the 'CSP Enabled' set to Yes,

- set the Country code set to United States

 

then it looks like there is a problem with this card and you should contact the card supplier.

 

Are you using the latest SR6.0? (currently SU184).

Share this post


Link to post

I just set the Dialogic config manager to those settings you mentioned, and now it doesn't give me those errors anymore. However, it still doesn't seem like a successful result - please look at the attached file.

 

After it says "GlobalCall channel is waiting to receive a call" nothing happens for about 45 seconds and then it says "no calls or events were received in the last 45 seconds..."

post-2066-1207117465_thumb.jpg

Share this post


Link to post
I just set the Dialogic config manager to those settings you mentioned, and now it doesn't give me those errors anymore. However, it still doesn't seem like a successful result - please look at the attached file.

 

After it says "GlobalCall channel is waiting to receive a call" nothing happens for about 45 seconds and then it says "no calls or events were received in the last 45 seconds..."

 

Oh, and I'm using SR 6.0 build 181

Share this post


Link to post
After it says "GlobalCall channel is waiting to receive a call" nothing happens for about 45 seconds and then it says "no calls or events were received in the last 45 seconds..."

Have you tried placing a call into the system. The call must arrive on the 'first' port. Only the first port gets opened (dxxxB1C1).

Share this post


Link to post

Ah, I wasn't calling in. I just tested it and called in at the appropriate time. It seemed to work - it recorded my voice and played it back. The screenshots are attached.

 

Please advise of the next step - does this mean CSP is working? After it plays back my voice, it again seems to be waiting for an incoming call.

post-2066-1207197442_thumb.jpg

post-2066-1207197447_thumb.jpg

Share this post


Link to post

Yes, this confirms the card is now correctly setup for CSP operation and CSP is working on that card.

Please download and install this version of VoiceGuide v7:

[old link removed]

This version has some extra CSP related trace debugging which should let us better see CSP operation in the VoiceGuide traces.

Please install this new version and then modify the Config.xml appropriately to connect to LumenVox MRCPv2 server and use the "SpeechRecog_Digits_Multiple" script, and then place a call into the system and please post the traces generated by VG for that call (ktTel, ktMrcp and vgEngine traces).

Share this post


Link to post
QUOTE(SupportTeam @ Apr 2 2008, 10:28 PM) <{POST_SNAPBACK}>
Yes, this confirms the card is now correctly setup for CSP operation and CSP is working on that card.

Please download and install this version of VoiceGuide v7:

[old link removed]

This version has some extra CSP related trace debugging which should let us better see CSP operation in the VoiceGuide traces.

Please install this new version and then modify the Config.xml appropriately to connect to LumenVox MRCPv2 server and use the "SpeechRecog_Digits_Multiple" script, and then place a call into the system and please post the traces generated by VG for that call (ktTel, ktMrcp and vgEngine traces).


Ok, I did this. When I call into the SpeechRecog_Digits_Multiple script, it tells me to enter the 4 digit VMB number followed by the pound key. I said 1234 and it didn't go anywhere.

The logs are attached.

log.zip

Share this post


Link to post

KtTel trace shows us:

 

220615.707 2832 001 ERROR mrcp recog thrd recognizer start failed. Timout awaitng for channel create completion. Recognizer not active.

 

Which means that the LumneVox MRCPv2 server did not respond.

 

Is the MRCPv2 LumenVox service installed on this machine? Are all the LumenVox services started? Please try restarting the LumenVox services (or just reboot the system if you do not want to restart them manually) and then try another call.

 

For debugging MRCPv2 conectivity it is usually easier if the LumenVox engine is located on a different machine as you can then use WireShark to monitor the SIP calls between VoiceGuide's MRCPv2 drivers and the LumenVox MRCPv2 server.

 

Please confirm that this machine's IP address is 192.168.1.9

Share this post


Link to post
KtTel trace shows us:

 

220615.707 2832 001 ERROR mrcp recog thrd recognizer start failed. Timout awaitng for channel create completion. Recognizer not active.

 

Which means that the LumneVox MRCPv2 server did not respond.

 

Is the MRCPv2 LumenVox service installed on this machine? Are all the LumenVox services started? Please try restarting the LumenVox services (or just reboot the system if you do not want to restart them manually) and then try another call.

 

For debugging MRCPv2 conectivity it is usually easier if the LumenVox engine is located on a different machine as you can then use WireShark to monitor the SIP calls between VoiceGuide's MRCPv2 drivers and the LumenVox MRCPv2 server.

 

Please confirm that this machine's IP address is 192.168.1.9

 

My machine is connected to a LAN, and when I do ipconfig in cmd.exe, I get 192.168.1.9.

 

And yes, LumenVox and VG are installed on the same machine.

 

I also tried the IP from whatismyip.com, but it didn't work - same error. I tried restarting the Lumenvox services already, and it didn't work.

 

I attached my config.xml file.

Config.zip

Share this post


Link to post
And yes, LumenVox and VG are installed on the same machine.

If LumenVox is installed on the same machine it would not be possible to use WireShark to debug the communication between the MRCPv2 server and the MRCPv2 client.

 

Looks like we'll need to create a test app which you will be able to use to test the response of the LumenVox MRCPv2 server, as looks like that is the only way to confirm correct operation of the server when it is installed on the same machine.

Share this post


Link to post

Okay - when will this test app be ready?

 

Also, I can try installing LumenVox on a separate computer on my LAN. For the IP addresses, would the 192.168.x.x format be correct? Should I try this?

Share this post


Link to post
Also, I can try installing LumenVox on a separate computer on my LAN. For the IP addresses, would the 192.168.x.x format be correct? Should I try this?

Yes.

Not sure how the LumenVox licensing works to allow this. It may be easier to move the Dialogic card and VG to the other machine.

 

Either way, separating the LumenVox and VoiceGuide will let you the use WireShark to trace the MRCPv2 comms between them and you will then be able to see what is the problem.

 

Share this post


Link to post
Also, I can try installing LumenVox on a separate computer on my LAN. For the IP addresses, would the 192.168.x.x format be correct? Should I try this?

Yes.

Not sure how the LumenVox licensing works to allow this. It may be easier to move the Dialogic card and VG to the other machine.

 

Either way, separating the LumenVox and VoiceGuide will let you the use WireShark to trace the MRCPv2 comms between them and you will then be able to see what is the problem.

 

I'll try doing that. Meanwhile, please answer: when will this test app be ready?

 

 

 

Share this post


Link to post

We have made a requires to our developers for creation of this app, but do not have an ETA at this stage.

Share this post


Link to post
We have made a requires to our developers for creation of this app, but do not have an ETA at this stage.

 

When I go to Lumenvox License Administrator and click connect, then View Current Licenses, I see that the VoxLite license says 1 port (0 used). So there is 1 unused port. Doesn't that mean the MRCP server is not using the port?

 

Please let me know right away.

Share this post


Link to post

Please answer this question right away so I can make progress - I need this working ASAP.

Share this post


Link to post

This a LumenVox related question, rather then a VoiceGuide related one.

 

Best way to make progess on this is to capture the SIP call traces between VoiceGuide and LumenVox. We will then be able to see what is happening between the two.

Share this post


Link to post
This a LumenVox related question, rather then a VoiceGuide related one.

 

Best way to make progess on this is to capture the SIP call traces between VoiceGuide and LumenVox. We will then be able to see what is happening between the two.

 

I bought LumenVox through voiceguide - are you saying you don't provide support for LumenVox? If so, where would I get LumenVox support?

 

And how would I capture the SIP call traces? Please give me specific directions - thanks.

Share this post


Link to post

The MRCPv2 SIP calls between the LumenVox and VoiceGuide can be captured using WireShark ( http://www.wireshark.org/ )

 

Just start WireShark capture and select the Network/Ethernet interface to capture packets over.

 

LumenVox needs to be installs on another machine.

 

After setting MRCPv2 server IP address in Config.xml and starting VG just place a call into the system and the MRCPv2 communications between VoiceGuide and LumenVox will then be captured by it. Save the whole captured trace and then .ZIP it up and post it here.

 

When I go to Lumenvox License Administrator and click connect, then View Current Licenses, I see that the VoxLite license says 1 port (0 used). So there is 1 unused port. Doesn't that mean the MRCP server is not using the port?

That just means that you have 1 port available to handle incoming Speech Recognition requests. This is what you would expect to see when a 1 line license is at that time not processing any Speech Recognition requests. ie. It all looks fine.

Share this post


Link to post

Could you tell me how to do the licensing so that I can install LumenVox on another machine? I currently have VG and LumenVox on the same machine and cannot figure out how to deactivate the LumenVox license to install it on a new machine.

Share this post


Link to post

You would have been supplied with a registration email for LumenVox which would have provided you with the login name and password for the LumenVox registration centre.

You would have needed to login there to obtain the LumenVox download and license in the first place.

You should be able to unregister the exiting license and get a license issued for the new system.

Share this post


Link to post
You would have been supplied with a registration email for LumenVox which would have provided you with the login name and password for the LumenVox registration centre.

You would have needed to login there to obtain the LumenVox download and license in the first place.

You should be able to unregister the exiting license and get a license issued for the new system.

 

I was able to install LumenVox on another computer and I generated these log files from VG. I will try to get Wireshark logs as well.

 

Any updates on the test app development?

log.zip

Share this post


Link to post
You would have been supplied with a registration email for LumenVox which would have provided you with the login name and password for the LumenVox registration centre.

You would have needed to login there to obtain the LumenVox download and license in the first place.

You should be able to unregister the exiting license and get a license issued for the new system.

 

I was able to install LumenVox on another computer and I generated these log files from VG. I will try to get Wireshark logs as well.

 

Any updates on the test app development?

 

Updates?

Share this post


Link to post

Ultimately the only way to debug the actual MRCPv2 communications is using WireShark. Do you have the WireShark traces capturing the comms between VoiceGuide and the Speech Recognition engine?

Share this post


Link to post

The ktMrcp trace shows that the LumenVox server did not respond to the MRCPv2 call setup (SIP INVITE). The WireShark traces should confirm this.

 

You may want to run the WireShark traces on both the VoiceGuide and the LumenVox machines to confirm if the MRCPv2 messages do reach the machine on which the LumenVox is installed.

 

 

223553.617 2540 SIP Event [nua_i_state] 0:INVITE sent (client)

223553.617 4864 mrcp_sofia_agent_signal_handler (client)

223553.617 4864 mrcp_sofia_agent_signal_handler 2 (client)

223553.617 4864 Process SIP Event [nua_i_state] Status 0 INVITE sent

223553.617 4864 SIP Call State [calling] (mrcp_sofia_on_state_changed client)

223553.617 4864 mrcp_sofia_on_state_changed end

223558.617 4864 mrcp_client_context_msg_process MRCP_CLIENT_SESSION_TERMINATE mrcp_client=06AAA798, session=06B260A8

223558.617 4864 Terminate MRCP Session fnptr call function=046A30A0 <new>

223558.617 4864 mrcp_sofia_session_terminate begin

223558.617 4864 Terminate MRCP Session fnptr returned function=046A30A0 returned

223600.617 4864 mrcp_client_context_msg_process MRCP_CLIENT_SESSION_DESTROY session=06B260A8

Share this post


Link to post

Create an account or sign in to comment

You need to be a member in order to leave a comment

Create an account

Sign up for a new account in our community. It's easy!

Register a new account

Sign in

Already have an account? Sign in here.

Sign In Now
×