Text to Speech Detect Voicemail

[ctc_olsen]

Hi Everyone. What we are looking to do is that, when doing an outbound dialling, if Vicidial detects that it's a voice mail, it would call the "Text To Speech" function and will say something like "Hi Mr. Smith, we are XXX company calling to offer you a service".

Where Mr. Smith is the last name on the lead. We already have a port registered.

Code: Select all

[root@go ~]# swift -V

Cepstral Swift v6.0.1, March 2012

Default Voice:  Callie-8kHz          v6.0.0
Language:       US English           v5.1.0
Lexicon:        unknown              v0.0.0

Concurrency:    1 Port(s) Registered
                0 Port(s) In Use

Distribution:   No audio distribution license was found.
                Saving audio to a file is disabled.

Copyright (C) 2000-20012, Cepstral LLC.


Can someone point us to the right direction please?

[ctc_olsen]

Additionally, we tried to test this.

Under Admin - Text to speech, we already have a tts ID active.



Under SURVEY SETTINGS FOR THIS CAMPAIGN:, we already set it to the TTS ID


We get this on the logs

Code: Select all

 file.c: Unexpected control subclass '-1'
(2:12 AM) ctc_rain: file.c: File testtts does not exist in any format


It works if we use a plain audio file. Am I missing something?

[mflorell]

Have you read the Vicidial Manager manual?

[ctc_olsen]

Thanks for the reply Matt. Yes I read the managers manual(release 2.2.1) but it does not show how to put the TTS sound file to the campaign survey. Right now we set it as under Survey First Audio File to TTS_testtts(TTS ID=testtts) but we get this at the error logs.

[Apr 2 12:10:03] VERBOSE[11718] logger.c: [Apr 2 12:10:03] -- Playing 'sip-silence' (escape_digits=) (sample_offset 0)
[Apr 2 12:10:03] WARNING[11718] file.c: File TTS_testtts does not exist in any format
[Apr 2 12:10:03] VERBOSE[11753] logger.c: [Apr 2 12:10:03] == Manager 'sendcron' logged off from 127.0.0.1

NOTE: The "Answering Machine Message" field is able to be set up as a Cepstral Text-tospeech(
TTS) prompt, meaning that you can take information from the vicidial list and read it back
within the message that is being played. For more information on TTS functionality, see the chapter on
Cepstral TTS further down in this manual.

I see that in the manual, I kept checking everything but nothing is said about setting this up. Help please.

[mflorell]

Are you still using a 2-year-old version of Vicidial?

If you are, then you need to upgrade.

[ctc_olsen]

In the upgraded Vicidial, can we use TTS on AMD to leave a voicemail? Additonally, what specific version number should we upgrade to please?

[mflorell]

Yes, you would get that functionality. You should upgrade to the most recent svn/trunk version.

[williamconley]

if you listed your installation method and version it would be easier to suggest your upgrade path. for instance, if you installed with Vicibox 3.X and chose SVN during installation, you can go to /usr/src/astguiclient/trunk and issue "svn up" to get the new code (it won't change anything, just download the new code) and then read the UPGRADE document in the code for what to do next.