Most applications for text-to-speech rely on databases or other text formats which do not correspond directly to the way in which humans speak the information.  For example, databases and other applications often contain text like this:

Morristown City of Health Dept of

On the NYSE, IBM +2 1/2 to 105 3/8...

Date: Fri, 15 May 1998 18:55:48 -0400
From: ksmith@xyz.org (Karen Smith)

Such raw text cannot be fed directly into a speech synthesizer! E-Speech text preprocessors translate such text into:

City of Morristown, Department of Health

On the New York Stock Exchange, I B M was up two and a half to one hundred five and three eighths...

Friday, May fifteenth, nineteen ninety eight; six fifty-five P M.
From: K Smith at X Y Z dot org; Karen Smith

Email
Telephone directories
Driving directions
Traffic reports

We are also happy to develop a custom text-preprocessor for your application.