Voice Transcription

So I know I am late to the party, but I have just discovered voice transcription in on my phone. (I still have the Nexus S 4G, just so you know.) It is great! As long as you have a data connection you can say what you want to type and you phone will transcribe it! I love this feature. And Google's technology works very well. Of course it isn't perfect, missing a word here and there. It is available on the stock Android keyboard on most devices as a mic button next to the space bar. The keyboard is replaced with a giant mic and you can say what you want. If you keyboard doesn't have it you can download a number of other keyboards that does have the button. My favorite is Thumb Keyboard 4. It is highly customizable with a number of other features.

The technology is interesting because instead of doing the transcription on the device like Windows Vista and 7 do the audio is recorded and sent to Google's server farm. There powerful machines with lots of memory and plenty of examples of human speech to compare to figure out what you are trying to say. Then the text is sent back to your phone.

I know this is a feature of the new iPad and the current iPhone 4S, but it has been in Android for a while now. Apple uses a speech transcription engine from a company called Nuance (makers of the famous Dragon Naturally Speaking) while Google made their own. The thing about Apple's technology is that it is only available on those two devices mentioned earlier while the Google technology is available on anything running Android 2.3 or above. The best part is that the Nuance engine is available on Android through a program called T9 Flex. It is a keyboard app that has a standard QWERTY keyboard with Swype functionality and a handwriting recognition engine as well. It works pretty good as well.

One of the things I have heard people talk about is that Google's engine doesn't work as well as Nuance's technology. I have yet to see that be the case. In my personal experience they are about equal. But I have both and so I am still experimenting.

It is also possible in Android to replace the engine entirely, installing an app that handles speech recognition itself. Vlingo installs itself like that. I don't know who's engine they are using now since Nuance owns them, but they used to use one made by AT&T. Still need to play with that one more.

Unfortunately my iPod touch will have to do with the Dragon Dictate app that doesn't integrate into the phone like the other options on Android. Just one more example of the openess of Android providing more functionality.

UPDATE 13 April 2012 So I discovered that Vlingo installs itself as a voice recognition handler on Android. This means that in the settings you can change the speech recognition engine from Google's to Vlingo and since Vlingo is now owned by Nuance they are probably using the Nuance engine. This means that you can use Nuance's engine anywhere that speech recognition is allowed. Android openess FTW.