Hello, this week I've added quite a bit to the speech synth. The architecture has been modified quite a bit and now consists of a processing pipeline where you can switch the elements if you want to experiment. I've integrated number and character name reading. CMUDICT, the freely available pronounciation dictionary, is supported. You will need to get CMUDICT.0.6 from Carnegie-Mellon and scan it using the class CMUDict. The VoiceSynthesizer plugin is not affected, so you can use it as before. Before you file in the new stuff, you should remove the category "Speech Synthesis" because the structure of some classes has changed significantly.
For download simplification, I've added my SqSpeech shared lib for the Mac (as a MacBinary file) and Andreas Raab's DLL for windows to the zip file.
As before, it can be found at http://www.heeg.de/~hmm/squeak/SpeechSynthesis.zip
Have fun, Hans-Martin
If you've fetched the file since my last announcement and now, you will notice that the change set does not file in cleanly. Squeak seems to be confused by the various renamings and has added rename expressions into the change set which don't work. In addition, if you load cmudict in a non-Squeak-lineend mode, it won't work. I've changed the reading code to use CrLfFileStream, which works whatever line endings your file has.
Hans-Martin
squeak-dev@lists.squeakfoundation.org