It has been a bit quiet the last two months. We’ve been busy internally working on training an acoustic model which would allow us to highlight each word as it is being spoken. This function is already working in the book section, and I think it’s amazing, but unfortunately it is quite a bit harder to do for most of our videos, which don’t have the clear audio the books have. Instead, there is background noise, music, people talking over each other - all things that throw the system off. I’ve made some progress with the technical implementation, but it’s not there yet.
In the meantime, we are going to spend some time improving and perfecting what we have. As a first step, as of today the website became a bit faster (for some views, more than a bit). Other things on our list are:
- Making it easier to discover the right content.
- Improving the video player (lots of small annoyances there).
- Fixing the iPhone App and getting the Android app published.
- Improving our dictionary, giving more information when looking up words.