A bit off topic but what’s the status on any research toward leveraging audiobooks to produce text-to-speech? What are the hurdles in making it seamless and imp

A bit off topic but what’s the status on any research toward leveraging audiobooks to produce text-to-speech? What are the hurdles in making it seamless and imperceptible? And what’s the “theoretical” number of recorded words/hour required to produce something reasonable?