Kindle Accessibility and Speech Synthesis
Sidebar to Jakob Nielsen's Alertbox, March 9, 2009 on the usability of Kindle 2.
The Kindle's best feature might be its ability to turn any book into a large-print book. Having 6 font sizes is truly a boon to all elderly readers and many low-vision users.
For blind users and those with very low vision, the device also offers a text-to-speech synthesizer. As an experimental feature, this sounds pretty good — and indeed, it works better than many screen readers I've heard. The synthesizer can read books at 3 different speeds, though it's not as fast as experienced blind users would likely want.
I found two main problems with the synthetic voice:
It doesn't recognize document structure, so, for example, it'll run a headline and the start of the body text together as if they were one long sentence. Not having a pause after the headline makes the information hard to follow.
It's hard to follow unfamiliar proper names. One book I partly read myself and partly had read aloud included names like Camaban, Cathallo, Derrewyn, Ratharryn, Saban, and Slaol. It was easy to recognize and differentiate these names visually, but when the synthesizer butchered them, I quickly lost track of who did what where.
Obviously, for Kindle to truly become a book reader for the blind, the user interface itself must become accessible. Currently, if you can't see the screen, you can't select a book to have Kindle read out loud. But the beginnings are here, and even at its current stage, Kindle presents huge accessibility benefits for an immense audience of readers with somewhat reduced vision.
Finally, Kindle 1 was roundly criticized for making it too easy to accidentally turn the page. Although the new release fixes this usability problem, the designers went overboard: Turning the page now requires considerable pressure directed with fair precision. This is something that might be hard for users with arthritis to do.