Updating Text To Speech files in bulk?

While this feature SOUNDS good (pun intended), the new voices do not match the timing of the old voices; you still have to resync your elements on each slide individually. Further, having done several dozen slide updates at this point, it isn't clear what drives the timing change; one screen the TTS in the new voice will be longer, the next shorter. I've also found that even if the overall timing is about the same, the pauses for periods, commas, etc., shift around within the overall time for the same TTS wording.