My experience converting a book to audio using TTS

I wanted to read a new book that was translated into Arabic recently. It was a new publication in the latest Cairo international book fair. I was excited to read it, but there is a problem, I don't know anyone who can buy it for me and send it to me, and the publisher doesn't have an online store. But then I thought, why not get the English version of the book and read the original version?.

I managed to find the English version of the book online and started reading it. However, I spent a lot of time reading small sections and then couldn't continue because I always had to stop and do something else. I wanted to read the book in a more convenient way. Then I thought again, what if I can listen to the book instead of reading it? I can listen to it while doing other things, and I can also listen to it while commuting. So I decided to convert the book to audio using Text-to-Speech (TTS).

Now there is a problem, I didn't follow the TTS advancements for a while, and I didn't know which TTS tool to use. I did some research and found out that there are many TTS tools available online, but I wanted to find the best one that can produce high-quality audio. Should be processed offline and should take books in PDF format. It should also free to use, so ElevenLabs was off the table. I knew about korboro-82M model, it is amazing in how its quality for a model of its size. It supports English mainly. I have tried it before, and I was impressed with the results. But this was using tools that do it for paragraphs, and I wanted to convert the whole book at once.

I found an app called sandbook on App Store, it is a free app that runs korboro-82M model to read the books, but it does need to do it in real-time, and because it is using iOS GPU, it cannot run in background. I tried it, and it worked well, but I had to keep the app open and the screen on while it was reading the book, which was not very convenient. I wanted to find a better solution that can run in the background and…