Documentation for replace field in "Edit speech"

I can see from the forum entry about it spelling out all caps, that the "Replace" field in the "Edit speech" feature supports more than just plain English.  Is there documentation somewhere about everything it supports?  
 
In particular I'd love a way to tell it a pronunciation at a phonetic level, rather than having to try to figure out an English spelling that will yield the correct pronunciation.  For instance, is there some way to directly represent a schwa?  
 
Thanks

so, there does appear to be a standardized TTS markup language, described here
My initial experiments of having the edit speech feature insert <phoneme> tags have not been successful, probably because the document itself is not wrapped in the correct tags.  I'll let you know if I find a way to get it to work.  

so, according to this

the android tts engine supports phoneme tags (which are the ones that would be relevant in replacement text).  
how hard would it be to make it leave these tags alone?  or alternately to have it leave alone any tags that were placed by the replacement feature?
btw here's the spec for the tts markup.  I apparently left it out of my original post

thanks
John

sent.  Subject is "TTS markup missing links"

Correction. It started to work once I added a <speak>… </speak> wrapper. Specifically, for Google and Acapela voices, the tag:
 
                <break time=”3s”/>
 
Did work fine, but phoneme, at least with “xsampa” alphabet, was ignored. However the Pico voice did execute it correctly, I believe, for this sample:
 
                <speak>Hello! <break time=”3s”/> Testing <phoneme alphabet=”xsampa” ph=”&#34;{k.t@`”/></speak>
 
it said something that sounded like:
 
                Hello! … Testing actor
 
Now we would need to discover, which SSML tags are really supported by each voice, and when phoneme is supported, which alphabet to use…
 
Greg