I use various Linux's past RH6.2, and use ATT Natural Voices for TTS. It reads me stories from the Guteburg project while I write machine tool code.
There were no special services needed. I had to run sndconfig to get the sound to work at all, but the app itself was just a matter of unyarring and using. I needed to write a script to simplify the long command line but that's all.
AT&T Natural Voices is a commercial product, the base price for two good quality voices is near 50$ US.
I chose AT&T after listening to many many synth voices. There's a lot of personal choice here but I think they (AT&T( sound way better than L&H, IBM, or M$.
But, dont be surprosed by weird readings as I often hear
"Chapter Virgin Islands" for 'Chapter VI' and more frequently, the Homer Simpson syndrome: 'He didnt know what to DOH".
Even if you use EMAC SPEAK ( I didnt yet ), you'd best look into the gutenberg project with toms of free ascii books. I've listened to all of the tom Swift series, and lots of PGWodhouse, and now into Burmese Days by ERBurroughs (? was it Burroughs?).
Anyways, from Confuscius to Alice in wonderland, entire books for free!
What TTS system are you thinking of using? and do you want input as well as output?