SOLVED How do I get "Allison" to read/speak my script?

[MrBostn]

Okay I just fired up a Incredible13-2? CentOS 6.9 server for conferences only. It's running on a physical computer.

So I listened to the IVR demo and it sounds so much better than my shitty voice. All I want is to have Allison say Welcome to our conference service, please enter the conference room number you want attend (or something like this)

For now I've recorded the IVR but Allison sounds so much better. I just need some pointers on where to begin.

Thank you.

 

[dicko]

You can use anything inside of your /var/lib/asterisk/sounds with Alison's dulcet tones, if you need more, then contact her, she will say whatever you want for a buck or two (actually a lot more, she is a professional VO artist but she has bills to pay, just like you!! ) and add to "the mix "

 

[MrBostn]

Boy was I off track. I thought the baked in Allison IVR Demo was text to speech. I was thinking I could type the script, run it through something and I get a voice reading it.

Thanks

 

[mainenotarynet]

Cepstral ('Allison' and other voices on version 4?) was $30 per voice (I think) license. When they upgraded to version 5 the license jumped to $250 and you needed channel licenses too to use Allison on more than one call at a time - BAD

Filte, espeak, Festival? etc are robotic sounding or to tone down your words really really bad.

You gave GoogleTTS, AmazonParrot, IBM BlueMix, PicoTTS, and M$TTS to choose from now (Nerdvittles I think has articles on how to install all of them and make them useable).

The old PWhite Texttospeech module in FreePBX (if you can even find it now) used to work swimmingly but for some reason the Filename checker thingy in it Barfs on perfectly good names (Firefox used to accept them, now it too Barfs) and now all I get is Conversion errors and no recordings

The Texttospeech module that FPBX put out is two pieces the engins (where are they section (and if you don't know the folder in which they are in (like for Cepstral I think it was /opt/swift/bin/swift or something like that) you are SOOL unless you like playing Linux hide and seek. Then the build the TTS files part where you enter your text - but if the engines part isn't configured first then no engines come up in the box. Bad all around

Suggestion: Use AmazonParrot (on the outside (see how to find parrot by searching on Nerdvittles) Write your text - It will spit a mp3 file out and load that into /var/lib/asterisk/sounds/(custom or texttospeech) folder works nicely then go to system recordings to add it to the system

 

[MrBostn]

Cepstral ('Allison' and other voices on version 4?) was $30 per voice (I think) license. When they upgraded to version 5 the license jumped to $250 and you needed channel licenses too to use Allison on more than one call at a time - BAD

Filte, espeak, Festival? etc are robotic sounding or to tone down your words really really bad.

You gave GoogleTTS, AmazonParrot, IBM BlueMix, PicoTTS, and M$TTS to choose from now (Nerdvittles I think has articles on how to install all of them and make them useable).

The old PWhite Texttospeech module in FreePBX (if you can even find it now) used to work swimmingly but for some reason the Filename checker thingy in it Barfs on perfectly good names (Firefox used to accept them, now it too Barfs) and now all I get is Conversion errors and no recordings

The Texttospeech module that FPBX put out is two pieces the engins (where are they section (and if you don't know the folder in which they are in (like for Cepstral I think it was /opt/swift/bin/swift or something like that) you are SOOL unless you like playing Linux hide and seek. Then the build the TTS files part where you enter your text - but if the engines part isn't configured first then no engines come up in the box. Bad all around

Suggestion: Use AmazonParrot (on the outside (see how to find parrot by searching on Nerdvittles) Write your text - It will spit a mp3 file out and load that into /var/lib/asterisk/sounds/(custom or texttospeech) folder works nicely then go to system recordings to add it to the system

Thank you for the detailed reply. Much appreciated. I'm a frequent traveler to Maine. Portland, Booth Bay Harbah, etc. Love it up there.

 

[MrBostn]

I just signed up for AWS fired up Amazon Polly and bam! sweet sounding IVR. It's basic but this is just a conf. computer. Thank you

 

[Red RedMan]

Okay I have looked every where I would like to know how to integrate polly tts into incrediblepbx2020 on centos7
Ive seen instruction from 2017 with wazo but not for 2020 on cent os 7
Please help if it can be used

 

[mainenotarynet]

Old thread but,

Google search for Amazon Polly. They have a free tier (I think), I also use IBM TTS (Watson) that is PAYG, but not much -- great for the weather (mine still barfs my zipcode - 04401 - Maine), Yahoo news and such on the fly. But for the system recordings stuff I just use Polly.

Type your text, save as an mp3 (or wav 8000Hz 16-bit pcm encoded if you wish), download it, the upload to your box in the var/lib/asterisk/sounds/[language code]/custom folder.

Go to the system recordings, name your recording and select the recording from the use system recording drop menu -- it's a long one though. You can start typing the file name to make it quicker.

Submit, Apply Changes. Done

Now you can use your recordings.

 

[Red RedMan]

Old thread but,

Google search for Amazon Polly. They have a free tier (I think), I also use IBM TTS (Watson) that is PAYG, but not much -- great for the weather (mine still barfs my zipcode - 04401 - Maine), Yahoo news and such on the fly. But for the system recordings stuff I just use Polly.

Type your text, save as an mp3 (or wav 8000Hz 16-bit pcm encoded if you wish), download it, the upload to your box in the var/lib/asterisk/sounds/[language code]/custom folder.

Go to the system recordings, name your recording and select the recording from the use system recording drop menu -- it's a long one though. You can start typing the file name to make it quicker.

Submit, Apply Changes. Done

Now you can use your recordings.

The only one i found was for wazo and pbx2020 and can't seem to locate the one for ibm all i find is how to sign up for the service
any help we be appreciated

 

[mainenotarynet]

The only (What) You found? I believe that incredible 2020 has the 'bits' built in for you to use IBM (in /etc/asterisk/extensions_custom.conf look for Yahoo news, or weather by zip. In those codes are several TTS to read stuff to you,you just have to uncomment the one you want and comment out the others.

search nerdvittles.com for IBMTTS. You'll find it

 

[ostridge]

need some pointers on where to begin.

a bit late to the party but there are already a large number of could be relevant files.
I just did a search as follows

## This creates a text file results of the search:
## this searches the /en/ directory for conf* files and lists the result.
##then you can read them in nano or something

   ls /var/lib/asterisk/sounds/en/conf* > /root/sounds-conf.txt
some key words to search:-
welcome    please-enter-the      at      enter-conf-call-number     you-entered      please-try     conference      one-moment-please     confirm-number-is
please-enter-the-access-code     please-enter-your     office-code     please-stay-on-line-to-be-connected   try-again-later    chat-room
digits/0 to  9   10 to 19 20 30 to 90    hundred    thousand     million
oclock     today    tomorrow     yesterday     day-1 -2 -7
dollars     pound
________________________


## Here is a 'welcome' search
ls -x *welcome*
dir-welcome                       dir-welcome.gsm                         dir-welcome.ulaw
dir-welcome                   privacy-stop-calling-not-welcome2      privacy-stop-calling-not-welcome.gsm
privacy-stop-calling-not-welcome     privacy-stop-calling-not-welcome.wav     privacy-stop-calling-not-welcome2
privacy-stop-calling-not-welcome2   privacy-stop-calling-not-welcome2
unwelcomecall         unwelcomecall           unwelcomecall.ulaw
unwelcomecall.wav                      welcome                           welcome
welcome                     


##   ls options from    man ls
##        -w, --width=COLS     set output width to COLS.  0 means no limit

##       -x     list entries by lines instead of by columns

##       -X     sort alphabetically by entry extension

You might find what you need just sitting there in the directory. Or may be you can use the Audacity programme or annother sound editor to Cut and Paste the bits of files you need. Several prompts can be chained to play together such a this line from my Lenny dialplan:

; 53669 => n,Playback(en_GB/this-call-may-be&en_GB/recorded&lenny/Lenny01)  ; Enable this to use existing ( joined-up with '&' ) sound files(3)
## (your language (en) may differ)