Thanks, but, no, those don't help. Let me try to repeat the question. According to the documentation for the speech engine, which hasn't been updated since about 2006, it would seem, one is supposed to be able to embed commands in text to affect the generation of speech. An embedded command looks like [[command params]]. There are (were?) commands to introduce silence, change rate, volume, or other parameters, to indicate the context of a word or phrase -- presumably to tell part of speech or tense -- supply a phonetic pronunciation, which includes stress, or even give something like a melody. What little discussion I've found on the topic indicates that most of those functions don't work with the newer voices, and possibly not even with the older ones any more -- so far, the only command I've been able to get to work with newer voices is the one to introduce a pause [[slnc <duration>]]. (Phonetic spelling does work with some voices.) So, my question is, which of the other commands in fact still work, and with which voices? A broader question would then be -- will the ones that don't work ever be brought back? I have a large document I convert to audio. I have been able to get almost all of it to be pronounced correctly by replacing the original spelling of a handful of words with "creative" respelling. The biggest problem is the speech engine can't tell the part of speech or tense of words, so a word like "read" will always come out as "reed", even when it should be "red." One word, for instance, that I can't get to come out right is "content" -- it is always treated as a noun, with the emphasis on the first syllable, and I can't force it to be treated as an adjective, with the emphasis on the second syllable, nor can I find any "respelling" that gives what I want. Does anyone know anything useful about all that?
