Understanding the problem

If we believe Taylor when he says we generally only need shallow processing of the text, then we can state the problem of text-to-speech as simply a matter of deciding what sub-word acoustic units to use, and what contextual features (derived from the text) we need to decorate those with.

6 minutes 15 seconds

This video was
Excellent		0
Very helpful		1
Quite helpful		0
Slightly helpful		0
Confusing		0
No rating		0

Difficulty
My brain hurts		0
Really quite difficult		0
Getting harder		0
Just right		0
Pretty simple		1
No rating		0