I've used text-to-speech to make myself sound files for interpretation
practice, and occasionally to help me check translations in the manner
you describe. I believe you need both a TTS engine and a voice. For
the engine, I've used TextAloud (
http://www.nextup.com/), which is
very serviceable and pretty inexpensive. There are probably free ones
too, and I think newer versions of Windows (at least XP and above)
include a basic one.
Getting a good Japanese voice might be trickier. There are good free
English ones, but after some pretty exhaustive searching around this
time last year, the only really good Japanese ones I found were Misaki
and Show by NeoSpeech, the company that did Stephen Hawking's current
voice (
http://www.neospeech.com/). When I contacted their sales
department they weren't even interested in selling to an individual,
though they said that would change before long, and that was a year
ago. It still wouldn't be cheap, though. If you don't mind if it
sounds more robotic, it looks like the company that makes TextAloud
also sells a mediocre Japanese voice called Kyoko for $45 (http://
www.nextup.com/nuance.html).
Maybe someone else has more suggestions or a clarification about
whether Windows' built-in TTS engine is good enough for what you want.
Jonathan Michaels
Monterey, CA