Evil geniuses and world domination are 2 of our goals... we also like Dr Who

Giving a Ubuntu-ebook a voice

no comment

I have been experimenting with the functionality of the Ubuntu ebook template.  The idea is that if you are writing your book using the template, that the same content can be published to an ebook and an audio book.  The new command to accomplish this is :quickly read.

The read command will eventually produce an m4b file complete with chapter files, allowing listeners to skip to a particular chapter.  To accomplish this :

  1. The read command pases the content of each file through to festival to read. Each chapter file is read into it’s own wav file using the festival speak command.
  2. The script constructs a chapter marker file  from the individual wav files.
  3. The script uses sox to join the chapter wav files into a  single audio file.
  4. Eventually the script will create chapter marker points within the audio book, allowing a user to skip directly to chapters.

At the moment there is no method to directly set the chapter points on the new audio book, so it was fortunate that the recent episode of FLOSS Weekly #240 mentioned a new online audio service – Auphoric.com

Auphoric provides a web service and API to make basic sound engineering as accessible as blogging is right now.  I used this service to insert the chapter markers. I was able to take the Auphonic web services and the files produced by the quickly read script and create an audio book.

Auphonic Home page

Auphonic Home page

Empty Auphonic Screen

Empty Auphonic Screen


  • In the Audio / Video Source selected file as my upload method and chose the out.wav file produced by the quickly read script.
  • In the Output format I chose AAC (M4A, MP4,M4b), and left the bitrate at the default 80kbps HE AAC. I selected the m4b extension.
  • In the chapter Marks section I clicked the “import chapter marks from a file” and loaded the chapter file created by the quickly read command.( The Auphoric API describes the format of the file)
Auphonic output settings and chapter marks read from a chapter file.

Auphonic output settings and chapter marks read from a chapter file.

  •  I left all the other options, and ran the job.
Auphonic finished result

Auphonic finished result


  • Following process the download option is available.

So following this process, from a single ebook project I can create a epub, and at the same time generate an audio book file.   Now to my ear, the voice used in festival is just horrible.  So I have contacted the Accessible Computing Foundation.  I first became aware of the ACF through the work of Jonathan Nadeau – having heard him interviewed firstly on Linux Outlaws and also through his own Frostbite media podcast.  If anyone can offer a developer some advice regarding recommended voices for festival, then it should surely be the ACF.

Hi thought I’d drop you guys a line.

– I’m currently working on an ebook project for the ubuntu framework – I’m still working on it, but as part of the solution the framework also generates an audio book (in m4b format).

The idea is to harness the power of the quickly project templating environment, allowing authors to quickly create ebooks using commands such as quickly add-chapter, or quickly publish.

it’ still a little rough at the moment, and I’m relying on the auphonic.com site to set chapter markers – I hope to remove that requirement eventually.

I’m currently using festival to produce the output, and to my untrained ear it sounds terrible. Do you guys have any guidance for using generated speech for accessibility?


Mike Hingley


Tags: ,

Comments are closed.