XHTML+Voice in Style

How to style spoken text using CSS. Part 2 of an introduction to X+V.

( Read the article )

Hi everyone,
I was reading this article about voice styling which I found really interesting. Searching through the net, I found a test site where most of these styles where presented (http://alexandre.alapetite.net/phd-risoe/mxml/style.php) However, although the documentation of Opera 9.10 says it supports the -xv-phonemes style attribute of CSS3, it didn't seem to work on my machine. Or maybe the syntax is wrong... I tried different variations (external style sheet, style attribute in at the tag, internal style)but it didn't seem to work... Phonemes are working when I put them directly into the regular phoneme tag within the vxml part, which would be my work-around, if there is not a better solution... Is the bug in front of the keyboard? Has anyone a working example I could have a look at?
I try to include basic pronounciation into a bilingual dictionary application I'm develloping and liked the idea of the browser actually reading it out without me creating thousands of sound samples. (Dictionary is Tagalog(Pilipino)-English. Tagalog pronounciation can easily be derived from the spelling. (at least roughly, the usual caveats apply: lean words and so on...) ).
Cheers and thank you very much for any constructive suggestions!
Jó