Speech Synthesis tools

I am trying to learn Japanese and as a little side project I wanted to write a program to speak sentences in Japanese to test my comprehension. Now there are lots of example sentences available on the internet but very few of them have attached sound files. So I thought why not use Text To Speech to convert them from romaji or kanji into voices.

I have been looking around and have concluded that the festival project was the best choice (it actually compiled and appears to be in active use). However it is focused on English as you might expect, the simple Japanese voice is incomplete and the download link is broken. Other projects seem to generate voices that 'ring' and can be hard to understand.

Aiko however speaks clearly, what tools does Aiko use to generate her voice?

What other tools exist that I might be able to use?

I am a programmer so I'm not afraid to write some code to get things to work but I know next to nothing about speech synthesis.

Any help much appreciated.

if it's just for learning Japanese, you could just type "text to speech japanese" into google
there are some web tools
http://www.oddcast.com/home/demos/tts/tts_example.php?sitepal
http://imtranslator.net/translate-and-speak/#

kanji-romaji translator
http://www.cs.technion.ac.il/~danken/kanjidic.html
http://kanjidict.stc.cx/hiragana.php

Thanks, those look interesting (especially the voice from imtranslator.net) however I really need this to be a library I can plug into an application. I was thinking (or perhaps dreaming) that it could be extended beyond just simple language drill into something like a personal assistant that could live on a computer or smartphone.

It's good to see that there are such high quality japanese voices in existence but I would like my system to work offline (such as on my commute where connectivity is patchy at best).

Thanks for the links though, the kanjidict link will come in particularly useful.

Peter Hickman wrote:Aiko however speaks clearly, what tools does Aiko use to generate her voice?

She uses Vocaloid, which will sound better on V2 (see global announcement).

Ach, I thought it might have. But for me that is too limited an environment to be working in. All the tools that I have work fine on Linux and OS X so I was hoping not to get locked into a particular (costly) environment. Still the Vocaloid manual was useful as it listed the phonemes that Vocaloid uses for Japanese.

This might help with putting together something in festival.

Thanks for the help.