Natural text-to-speech voice

Reading text on a screen, especially my academic text books is a grind, lately I've come across Edge's amazing 'read aloud' functionality that's seriously improved my experience, and ability to absorb information, even reading long articles.

However, I find that the text-to-speech voices that edge leverages, have the same clunky robotic accents that thy did back in Windows XP, which - after a while - can become quite tedious and become an effort to understand and absorb.

I have come across an abundance of open source projects that provide a much more natural sounding text reading voice, and I'm wondering if there has ever been an interest in giving the text-to-speech functionality an overhaul. We are in the year 2019 after all, and still relying on accessibility features from pre-2010 is becoming obstructive .

Is this a change that we will see in future updates, or will it be left in the hands of third parties?