I’ve created audio versions of the various sections of my PhD Confirmation document (PDF ~1Mb). They are read by a AT&T natural voice called “Mel”:
- Summary.mp3 (3Mb, 3mins) – gives an overview of the document; helps you decide if this is going to be your cup of tea.
- Section1.mp3 (46Mb, 48mins) – discusses the issues with widgets and how standardization could help fix them.
- Section2.mp3 (14Mb, 15mins)- discusses the methodologies that can be used to fix the problems with widgets, and other relevant stuff.
- Seciton 3 is the Widgets 1.0 Requirements, which you can read at the w3c.
Files are all encoded in 32Khz, 16Bit, mono…. perfect for listening on any Mp3 Player. Document was converted to Mp3 using NaturalReader by Natural Soft.
UPDATE: added the latest draft of the widgets 1.0 requirements document.
Getting the document to be read smoothly required me to make a few minor editorial changes to the document:
- Replaced words like “I’ve” to “I have”, as the text-to-speech (TTS) engine was reading them as “I ve ee”.
- Put a full stop “.” on the end of every heading to force the TTS engine to pause before reading the next paragraph.
- Put either comas or full stops on all bulleted or numbered lists. The TTS engine reads lists as if they are a sentence.