Book to Audiobook with Mozilla TTS

Python
DaVinci Resolve

I was looking for the audiobook "Alone" by Richard Byrd. After getting audible because I thought I found it and getting dissapointed when it was something else I decided to try creating it myself. I felt like text-to-speech technology was at a point to make an enjoyable experience. Obviously no comparison to a real audiobook but good enough. It also just sounded like a fun project. I went with Mozilla TTS. To my understanding it's pretty state of the art. I didn't train my own model due to scope, time and skill issues. At the time I recently got an old HP Z420. I had set it up with PopOs because I already used and liked it quite a bit. I used that machine (or an SSH connection to it) because I felt like the combination made sense and I wanted to get more comfortable using Linux distros. Obviously I'm still a linux noob.

I split up the text file into chapters. After I wrote a script to clean up all letters not supported by the Mozilla TTS. This was a lot of trial and error. Who would have thought that ΒΌ was such a common character. Also seperating solely by periods was problematic. Shoutout to the guy only refered to by his initials. After solving these problems I send packages of about 10 sentences to Mozilla TTS and got short .mp3 files back.

I uploaded the first two chapters to youtube so you can listen in on the results. I wanted to upload the whole thing but it's way over ten hours.