In my experience the little recordings you can download as samples of a TTS always sound better than what you can actually produce. It is best to search for demo web sites that actually let you type in text and then have it spoken with the voice you choose. I thought that I might want to buy the AT&T natural voices when I listened to something prerecorded, but when I found a website that allowed me to test them I was not so impressed. I have the same problem with not being able to tollerate choppy speech that makes it sound like many words are made up of disassociated syllables. At least AT&T's voices don't sound overly emotional as Neo Speech Paul and Kate do. They also no better than to pronounce things without vowels. I don't really like hearing about the .berf files on my computer--sounds too much like barf. lol They also seem to know how to deal with punctuation very well. So far my opinion is still, "Long live synthesized speech!" And spare the human readers that must work for hours to provide voice samples to make modern TTS voices. I could sware that AT&T Rich is actually made with the voice of one of the British readers for NLS, but I can't think of his name in order to ask all of you if you know what I mean. Sarah Van Oosterwijck http://home.earthlink.net/~netentity/