Text-to-Song is the Future of Text-to-Speech?

This article was written on August 19, 2007 by CyberNet.

Text To SongOne of the biggest pitfalls of text-to-speech utilities is that people always say that they never sound like human beings. They just aren’t natural, and often have a robotic tone to their voice that makes it difficult to listen to for a lengthy period of time. What if we could index songs and piece together sentences based upon the lyrics? Would that be any better?

As it turns out, no it wouldn’t. :D Over in the forum Pieter noticed a site that is dedicated to converting text to songs using snippets from music. I decided to try it out myself, and it is extremely amusing! To see what I’m talking about click here to listen to a phrase that I put together:

Computers are the future. They will control the world soon! Save yourself!

I also noticed that it isn’t able to pronounce everything, and if a letter is skipped it will be highlighted in red. For insance, it can’t pronounce "CyberNet News" without skipping the letter C. If this does happen, however, they let you give the title of an existing song that uses that word. It will then (hopefully) get added into their database.

Go ahead and try it out. It’s actually pretty fun. :D

Let Them Sing For You [via Download Squad and the CyberNet Forum]

Copyright © 2010 CyberNet | CyberNet Forum | Learn Firefox

Related Posts:


No Responses to “Text-to-Song is the Future of Text-to-Speech?”

Post a Comment