Nuance TTS with multiple voices

[chad_meyer]

Hey everyone,

We recently installed Nuance with a single voice file (male) and I have a CVP application that plays audio with no problem.  I am wondering if it is possible for the CVP app to choose which audio voice to use?  I see the option to set the language (ENglish/Spanish..etc) but nothing that would allow to set which voice to use.  Any help would be appreciated.  

[bill.king1]

Hi, in your example, you're saying you have (for instance) 2 en-us voices, one male and one female?

I've not seen it where it has multiple of the same voice per language, as opposed to say en-us is female and es-mx for Spanish for instance is female.

If you look on the Nuance server, how do they show (since they typically list the voice and then the language)?

[chad_meyer]

Bill,

That is correct.  Assuming we are only using English but one CVP application is specifying to use voice "Ethan" while another CVP application for a different business unit uses voice "Heather", for example.

Can you advise on the following question?

"If you look on the Nuance server, how do they show (since they typically list the voice and then the language)?"

I am fairly new to Nuance and their support said they won't assist as this is a CVP/UCCE issue, not Nuance....

[bill.king1]

Hi, if you go to the Nuance server under Windows control panel, for programs, you will typically see the voices (names) as well as the languages that are supported that are installed on the server.

[ptindall]

Try including an SSML element with gender attribute (e.g. <voice gender="female">) at the start of your TTS string.   Plus append a  closing </voice> at the end.

[chad_meyer]

@ptindall Thanks for the direction of using SSML.  Looking through some of those docs it seems as though I would want to use the following to specify a different voice name:

<voice name="Ethan">Text to be played</voice>

or

<voice name="Heather">Text to be played</voice>

Would you be able to confirm this or am I interpreting that incorrectly?  I referenced the following doc, located here.

[ptindall]

Absolutely, yes, you can use a number of different options including specifying the voice name explicitly, as in your example.  Gender is something you're more likely to use combined with language to automatically select a voice by required attributes rather than explicitly.

[chad_meyer]

Thanks.  I will get another voice file installed and report back.