Not surprisingly (since it is first-and-foremost a phone), integration of voice command and dictation was one of the major areas of innovation for the new Nexus One from Google. While I wasn’t one of the privileged few invited to the press/analyst event that launched the phone (full disclosure – I was not provided with a free demo unit), I was able to attend through the magic of modern streaming. A replay of the entire event is available here, but I’m embedding a snippet from CNET News to make my point. Advance to minute 2:38 and you’ll see how Nexus One product manager Erick Tseng highlights the advantages that arise from Google taking active control over development of hardware in parallel with its software development efforts:

As we’ve pointed out for years now, the combination of voice command and dictation on smartphones is powerful. On the Nexus One, Google has done the best job yet of seamless search and navigation. Pressing the microphone button and saying “Navigate to Ikea” is all you need to do to (a) invoke a search of Google to find the nearest Ikea store and then (b) initiate the routine on Google Maps which ultimately leads to rendering the route from your current position to the nearest store. That’s why I’m calling it seamless.

Also note that invoking speech recognition is not yet completely “hands-free and eyes-forward” as will be required for the ultimate mobile speech experience. But that challenge can be overcome in a car with a simple button on the steering column (a la Ford SYNC). The other major challenge is, and will always be, accuracy. When demonstrating the entry of a dictated email message, there was still an element of suspense as Tseng, under ideal conditions, waited those few seconds to see if the dictation is rendered accurately. In this case, it was. And the crowd applauded accordingly.

Personally, I would prefer to have a demo where a proper name is misspelled or the message requires a change in punctuation (from a period to an exclamation point, for example). To me, that first rendering is, more likely than not, going to contain a misspelling or two and require some tweaking of the punctuation. Knowing the quality of the speech scientists working at Google these days, I’m pretty sure that they have many tricks to promote accurate rendering, punctuation and even the insertion of emoticons, but I suspect they will be introducing those after they build the ground level of acceptance that this modest offering is bound to engender.

Today I’ll sit back and watch as mobile speech benefits from the ripple effect from Google’s mass marketing efforts. Google is doing a lot of things right with Nexus One, including the opening up of marketing and distribution efforts to include direct sale of phones through its newly created channel (Google login required). From a personal economic point of view, the difference between the $100+ and a $500+ price tag speaks for itself most price conscious shoppers will opt for the subsidized phone and a term plan with T-Mobile. Regardless of customer choice, the major step forward from my point of view is the bold “voice enabled” stamp that Nexus One puts on its mobile experience.

I’ll be writing a full discussion of the competitive implications across the mobile speech ecosystem in a forthcoming advisory. Suffice it to say that the tight integration of software, hardware, operating system and voice processing that is exemplified by both the iPhone and Nexus One leaves creates challenges for dozens of independent providers of speech-enabled mobile apps. Just as the automobile mount for the Droid with speech enabled navigation has struck fear in the hearts of all personal navigational device (PND) makers, the showcasing of voice control, search, navigation and dictation by Google will drive alternatives to define specific areas of opportunity (niches) and new partnerships that extend both their reach and life expectancies.

Categories: Articles

Tags: Google, mobile speech, speech processing