Screens are meant to be touched, games consoles are designed for us to wave our limbs in front of them and Google and Apple are desperate for us to talk to our phones. Amongst all of this, speaking to our PCs has fallen somewhat by the wayside, but the company providing the speech recognition technology behind Siri wants us to try again.
Somewhat predictably 'Dragon NaturallySpeaking 12' is the successor to the excellent NaturallySpeaking 11.5 and somewhat predictably Nuance is calling it the "fastest, most accurate and easy-to-use version of Dragon yet".
What causes these proclamations are claims of a 20 per cent improvement in out of the box speech recognition accuracy, faster performance and new technology which looks to learn your preferences as you use the software - for example how you format words, phrases and numbers. In addition is full Gmail and Outlook.com support, correction options respond to more naturalistic language as well as more natural text-to-speech reading of your work by Dragon which allows you to review without returning to the screen.
These features build on existing core features such as social network integration ("Post to Facebook..." / "Post to Twitter..."), web search and opening and closing of programs. Perhaps most useful of all, however, was v11.5's integration with Dragon smartphone apps which allow dictation via your phone's microphone letting you ditch a headset completely.
Voice recognition software has a reputation for infuriating, lengthy setup routines and while Nuance Dragon NaturallySpeaking 12 doesn't drop them completely the process is quick and – most importantly – educational. A nice tweak Nuance has made from v11.5 is that the setup process now takes even more care in teaching you key commands and usage scenarios while it learns your voice so it is well worth doing. In all it took about 15 minutes and even new users would feel confident about using punctuation, switching between programs, composing emails and performing web searches in this time. The training process and additional training exercises can be accessed at any time.
Of course it is worth pointing out that smartphone voice recognition has no setup process whatsoever which may make some impatient. The counterpoint is the setup is both a tutorial and should be far more accurate than just trying to adapt to the generic tones of your nationality. At least that's the theory...