ChangeLog

8.9: Clean code for release of Debian package
8.5: Support MacOS (pulseaudio needed)
8.4: Support speech-dispatcher 0.9. Fix issue of English too fast when Chinese is set a base rate. Treat space not a Chinese symbol if it's within English Text. Improve punctuation reading.
7.7: Speak single alphabet with Huang voice.
7.6: Fix English sample rate issue when Chinese voice sample rate is not
16000.
7.4: Toisanese is supported.
7.0: New overlap algorithm applies. eSpeak is set as default English engine.
6.3: support [[pin1 yin1]] style
6.2: use kal male voice as default English voice of Mandarin because it's more clear. Other female voices are in huge size.
6.0: Voice data is re-format into one file.
5.7: Much noise is removed. word context is added as an extended feature.
5.6: Festival and estools lib will be detected automatically. Festival static libraries are no longer a must.
5.3: speeched API is practically usable. Run `install.pl`
5.2: Naturalness of voice is greatly improved.
5.1: Female (cmu_us_slt_arctic_hts) English voice is added
5.0: Tibetan is supported
4.11: speech-dispatcher API implemented
4.10: Add feature to handle numbers
4.8: Hakka language is added
4.7: Festival is disable by default
4.2: sonic is used as the default library for changing speed instead of SoundTouch. SoundTouch can be enable with `configure --enable-soundtouch`
4.0: Dictionary module is rewritten. Much faster. Android is supported
3.0: Server Mode is supported
2.6: Ubuntu(deb), Fedora(RPM), FreeBSD are supported
2.0: Rewritten in C++. SAPI5 supported
0.91: A stable Windows Gtk version with Cantonese, Mandarin and Korean
0.75: Korean is supported