Originally Posted by AlPe
Also, SMIL does not seem suitable for what I want.
Correction: SMIL is the right way to go, as, per spec, it provides a way of "limiting" the reproduction of an audio clip to a certain portion of text.
1) iBooks does not support it (not even iBooks 3) and ADE neither
2) AZARDI offers some minimal support to SMIL
3) Readium offers support for SMIL, but --- to date --- the "stable" version (0.5.3 9/29/2012) has problems with internal links and SMIL. They have already been solved in the sources, so you might want to get the source code via git, and load it into Chrome as "unpacked extension".
So, for the moment, for my needs (iBooks, unfortunately), I will simply embed the audio file, add an <audio> tag, but I will not add SMIL yet.
I hope SMIL will gain traction soon.