Anatomy of an Audio Book


How do you create an audio book using Text To Speech (TTS) technologu? Well here's how I did The saadhu testament.

The book was only available as a microsoft word document or as html - the HTML is exactly what is available on the Saadhu website for free online reading and this is what I decided to use.

Next was a decision between using FreeTTS or festival. I had already used Free TTS once before in some serious work but not had much of an opportunity to work with Festival so this decision went in favour of festival.

The next step was a create simple PHP script that stripped out all the HTML tags from the e-book and replace them with Sable Markup. Since PHP is well suited for this sort of work the script was hardly 100 lines in length.

After the .sable files were generated they were handed over to the text2wave program which is part of the festival distribution. The output generated by this program was in the wave format and as a result some files were in excess of 60 Mega Bytes - far from ideal.

Now it was the time to turn to audacity. Though there are other tools available for conversion from wave to mp3 or wave to ogg on linux, I couldn't find any good ones that could do both while giving me a chance to playback (and edit if needed) the wave or compressed files that I am working with. Audacity can do all that and many more besides.

After the MP3 and OGG conversions the files were much smaller than they used to be and it's these files that you can download at

comments powered by Disqus