If you follow the tech blogs, you will know that the buzz this week has been about the Google Nexus phone that runs Android 2.1. One of the new features of Android 2.1 is speech-to-text for any text input field in any application on the device. Earlier versions of Android support speech recognition in search and maps. During my recent trip I extensively used Google maps and navigation and attempted to use the speech recognition, and what I found is that because the speech recognition is done on Google’s servers rather than on the phone, you must have a good network connection for the recognition to work. Unfortunately, I had several instances where the network connection wasn’t good enough and I ended up having to type in the text.
I have long been a champion of speech recognition on mobile devices, so I am very happy to see Google incorporate it in Android. However, speech recognition has got to be on the device so that it is as reliable has accessing the on-screen keyboard. Most people will not be patient enough to keep trying to use it if it is constantly failing.
Notes From The Cave 