AT&T Natural Voices™ Speech Engine for Vox Proxy!

We are proud to announce the availability of AT&T Labs' Natural Voices Text-To-Speech engines for Vox Proxy. The AT&T Natural Voices TTS Engine is the most realistic, human-sounding synthetic speech system available on the market today.

Individual voices are available in U.S. English, German, Latin American Spanish, U.K. English, and Parisian French. The following samples of available voices are "wav" files that were created using the actual installed speech engines.

Listen to samples of Natural Voices
	"Mike" (U.S. English)		"Crystal" (U.S. English)
	"Rich" (U.S. English)		"Claire" (U.S. English)
	"Audrey" (U.K. English)		"Charles" (U.K. English)
	"Rosa" (Latin-American Spanish)		"Alberto" (Latin-American Spanish)
	"Klara" (German)		"Reiner" (German)
	"Alain" (Parisian French)		"Anjali" (Indian-British English)
For comparison, listen to the following Vox Proxy's default voices:
	Susan (Tru-Voice)		Paul (Tru-Voice)

Using Natural Voices in Vox Proxy

Natural Voices are invoked by the TTSEngine command in a Vox Proxy script.

1. From the Script Writer, click on the Wizard button to start the script wizard.
   
   
2. Click on the "Other Agent Commands" button and select TTSEngine.
   
   
3. Scroll through the list of available voices. Natural Voices are listed under "AT&T Labs Natural Voices" followed by the name of the voice.
   
   
4. Select the desired voice
   
   
5. Insert the command immediately after the "Show [character]" command in your script.

The resulting script will look similar to the following:

Show Paul at center TTSEngine "{3c6088d8-8575-444e-8B3F-4BFA3F498C9E}"
Say This is my new voice from AT&T Labs.

The names of these voices come from AT&T, but they can be used with any Vox Proxy characters.

From Version 2 (of Vox Proxy) onwards, there is a new feature that allows you to assign any of these voices as a permanent default for any character, and to facilitate generating wav files using Natural Voices.

Ordering Information

You must be an owner or new purchaser of Vox Proxy to order these speech engines!

To get started, you must order the Natural Voices Engine, which comes with the Mike and Crystal voices. You may add to that initial 2-voice pack any combination of available voices in any languages. Each additional voice is shipped on a separate installable CD.

AT&T Natural Voices™ Speech Engine
	Mike & Crystal (U.S. English)	Each additional voice	4-voice U.S. English pack	All 12 voices
Single-user	S$150	S$50*	S$228**	S$488

  ^* When purchased with the initial 2-voice pack       ^** Limited-time offer

Minimum System Requirements

• Windows 98, Me, or XP (2000 coming soon)
  
• 256 MB RAM (512 MB is better)
  
• Pentium-III or better; 500 MHz or greater CPU speed
  
• Approximately 250 MB hard disk space for EACH voice
  (in other words, 500 MB is required for the initial 2-voice pack)
  

Limitations

• You may NOT redistribute these TTS engines. That means you are not permitted to burn disks containing the Natural Voices TTS engine. You may, however, create audio (WAV) files using the TTS engines and embed those on the CD, using Vox Proxy's "SayWav" command.
  
  
• The files for Natural Voices are BIG (see system requirements above). Make sure you have space on your hard drive for them.
  
  
• This speech engine is demanding of performance. Be sure you have at least 256MB of RAM. The engine does not work well on slides with videos, animated GIFs, or other CPU-consuming tasks.
  
  
• The first time you use the TTS engine (in a given session), it takes several seconds to load. This is unavoidable, but for a live presentation, you can preview a slide prior to starting the actual slide show so that the extra load time is completed before going live.
  
  
• Natural voices does NOT support several of the "speech tags" used in Vox Proxy to modify the pronunciation of text. Specifically, the emphasis, pitch, whisper, and monotone tags are not supported. In addition, the engine does NOT generally modify its pronunciation according to closing punctuation.
  
  Speech tags that ARE supported include: speed, address context, email context, pause, reset and volume.
  
  The result is that you do not have as much flexibility in modifying the pronunciation of sentences as you do using Vox Proxy's default (Tru-Voice) speech engine.