Voxeo's Text-to-Speech Guide

That's right, you guessed it. This chapter delves into the wonderful world of text-to-speech (TTS) -- specifically, the types of TTS that are offered on Voxeo's Staging and Production platforms. You'll find everything you need here to get you on your way to fine-tuning your text-to-speech applications, but don't take my word for it, let's take a look at what we have to play with here. Below you'll find a comprehensive listing of the TTS voices available in our Staging environments. In addition, you will notice that the voices vary depending on whether you are using our Prophecy platform or Prophecy with Nuance platform.

Note: Developers should be aware of the fact that the only TTS voices that are available without charge for the US production platforms are the Prophecy Female, and Prophecy Male TTS voices. While all voices listed below for the US are available to test on our staging platforms for no charge, invocation of non-standard TTS voices in a production capacity are subject to additional fees: Contact the Voxeo Sales Team for specific pricing details for any non-standard TTS voices that you may want to use for commercial grade applications well in advance of a production deployment.

As the UK staging and production platforms require a wider range of TTS and ASR offerings, the voices listed in the tables below are not subject to additional fees when used in a production capacity. However, we request that you notify your dedicated Account Manager when invoking these TTS voices so that we can ensure that we stay well ahead of the curve in terms of available TTS ports.

Before we jump into the available TTS languages, we'll first want to know how we should reference them in our code. There are two methods that we can utilize in order to play the desired voice:

xml:lang
voicename

xml:lang

The "xml:lang" attribute of the prompt element is the language identifier for the specified prompt. If there is no xml:lang given, then the platform default (en-us) will be used. To modify the TTS voice via xml:lang, we would do something similar to the following example:



<prompt xml:lang="en-us-fm3">

  Hello there!

</prompt>

We can also set this at the vxml level, should we want to use the TTS voice throughout the entire document:



<vxml version="2.1" xml:lang="en-us-fm3">

Either of these methods will change the TTS voice to English-Female3. Another way to do this would be via the "voicename" property.

Voicename

The "voicename" property allows us to set the TTS voice at different scopes in our application as well:



<property name="voicename" value="English-Female3"/>

Now that we know how to specify different TTS voices in your code, let's take a look at the actual voices. But first...

Things you should know...

TTS Engine: The TTS Engine pertaining to the TTS voice listed.
Language: Self-explanatory.
Voice Name: This is the value to be used when specifying the "voicename" property.
Mapping: This is the xml:lang mapping which invokes the corresponding voice.