Multithreaded Outbound Calls

[Guest BlackBird]

Hello Support Team,

 

I plan to use VG for Dialogic to make outbound calls through a system that has a dialogic card with 12 lines. I looked at older discussions but could not find one to answer the following question/scenario:

 

Can VoiceGuide multithread outbound calls? For example, if I have 12 telephone lines and I have 6 outbound call jobs registered in the Outbound db and I defined the lines to use per outbound job (row) via the “LineSelection” field as follows:

 

Outbound call #1 use phone lines 1, 2

Outbound call #2 use phone lines 3, 4

Outbound call #3 use phone lines 5, 6

Outbound call #4 use phone lines 7, 8

Outbound call #5 use phone lines 9, 10

Outbound call #6 use phone lines 11, 12

 

Assuming that the telephone lines are all available and that the outbound calls are all set to occur on the same date and time, would the calls occur simultaneously since they are assigned to specific telephone lines?

 

Regards,

Ralph

 

Btw, kudos on the enhanced tools (editor) to post discussion, nice upgrade...

[SupportTeam]

would the calls occur simultaneously

Yes. they will.

 

When you load the calls as you describe this will happen:

 

Outbound call #1 will use phone line 1

Outbound call #2 will use phone line 3

Outbound call #3 will use phone line 5

etc.

 

all calls will be made at the same time.

 

Unless you have some requirement to direct certain calls to be made on certain lines then the easiest option is not to specify anything for the LineSelection field. Then calls will be made as soon as any line is available.

 

You can try all this for yourself with the demo version of the software (although the demo version is limited to about 10 outbound calls before needing restarting)

[Guest BlackBird]

would the calls occur simultaneously

Yes. they will.

 

When you load the calls as you describe this will happen:

 

Outbound call #1 will use phone line 1

Outbound call #2 will use phone line 3

Outbound call #3 will use phone line 5

etc.

 

all calls will be made at the same time.

 

Unless you have some requirement to direct certain calls to be made on certain lines then the easiest option is not to specify anything for the LineSelection field. Then calls will be made as soon as any line is available.

 

You can try all this for yourself with the demo version of the software (although the demo version is limited to about 10 outbound calls before needing restarting)

 

Currently the trial version works on my system, however, I am currently trying VG with a single voice modem thus I can not try the above scenario, however the final server will use a dialogic card and a licensed VG install, so we'll test the above scenario.

 

With the respect to your response (above) please clarify - Would a TTS engine slow/interfere with the simultaneous executions of the calls? I am thinking that the TTS engine will need to convert the text for all calls, one phone message at the time, correct?

[SupportTeam]

Would a TTS engine slow/interfere with the simultaneous executions of the calls?

TTS generation is quite fast, and the person on the phone line would not hear any appreciable delay when they are listening to it.

TTS generation is done on a separate thread to the threads which control the telephone lines themselves and - TTS generation does not affect the responsiveness of the system, how fast the calls are made etc. How long are the sound files which you want to generate using TTS?

[Guest BlackBird]

Would a TTS engine slow/interfere with the simultaneous executions of the calls?

TTS generation is quite fast, and the person on the phone line would not hear any appreciable delay when they are listening to it.

TTS generation is done on a separate thread to the threads which control the telephone lines themselves and - TTS generation does not affect the responsiveness of the system, how fast the calls are made etc. How long are the sound files which you want to generate using TTS?

 

The sound files to be generated should be about 15 seconds (maximum) in length or less.

 

In the interest of ensuring system performance, an option I am considering is that, when there are multiple outbound calls that use the same message (example 2000 calls with the same promotional marketing offer) is to capture the sound file generated by the TTS engine for the first call as a static .wav file and reuse the .wav file for the remaining 1999 calls (as long as the message does not change) at the end of the call job, then delete the sound file. This leaves me with the following questions:

• From the VG performance perspective, would re-usage of the sound file be more efficient in terms of execution of the job rather than having the TTS regenerate the sound file on the fly for every call (within a call job of say 2000 phone numbers)? or would there not be significant gains still with this approach?

• Would VG be able to handle the above scenario, that is, having the TTS generate a sound file and VG recycle the sound file for a number of phone calls? If yes, could you suggest a method on how would I go about ensuring that when I have multiple call jobs VG will use the appropriate sound file that matches the respective call job?

[SupportTeam]

From the VG performance perspective, would re-usage of the sound file be more efficient in terms of execution of the job rather than having the TTS regenerate the sound file on the fly for every call (within a call job of say 2000 phone numbers)? or would there not be significant gains still with this approach?

The re-generation of file would not affect the speed with which calls are made. The generation of a 15 second file would take less then half a second (depends on how fast the CPU is), so listener would hear a gap of about half a second between end of one .WAV file and the start of the generated TTS file.

 

Would VG be able to handle the above scenario, that is, having the TTS generate a sound file and VG recycle the sound file for a number of phone calls? If yes, could you suggest a method on how would I go about ensuring that when I have multiple call jobs VG will use the appropriate sound file that matches the respective call job?

The best way to do it would be to just pre-generate the file using the one of the applications installed by the SAPI SDK (download link is on our WWW) and in scripts just point to that pre-generated .WAV file.

[Guest BlackBird]

The re-generation of file would not affect the speed with which calls are made. The generation of a 15 second file would take less then half a second (depends on how fast the CPU is), so listener would hear a gap of about half a second between end of one .WAV file and the start of the generated TTS file.

 

 

The best way to do it would be to just pre-generate the file using the one of the applications installed by the SAPI SDK (download link is on our WWW) and in scripts just point to that pre-generated .WAV file.

 

Thanks for the pointer, so the idea would be to pre-generate the sound using the SAPI coupled with the TTS engine (for a better voice than those supplied by MS, albeit its free) - I will look further into how to automate this option.

 

cheers!