The Verge has sources claiming that the next Xbox will be able to perform speech-to-text, natural language recognition, and wake on voice tasks. If Durango ships paired with a next-gen Kinect that offers substantially better voice and motion detection, Microsoft could have something special here. There are roughly 75 million Xbox 360s in the wild, but only a small fraction of those have a Kinect paired with them. If every Durango console has a next-gen Kinect, developers will actually be incentivized to create games that take full advantage of those features instead of haphazardly slapping on Kinect functionality. Financially speaking, reaching an entire user base is much more compelling than reaching the fraction that went out to buy a peripheral device.

Information is still scarce, so we don’t know all of the implementation details quite yet. The consoles will undoubtedly be much more powerful than current phones and tablets, so it seems plausible that it could handle the speech recognition on the hardware itself instead of sending it to the cloud like Siri does. It’s not a forgone conclusion, though, and having everything pass through servers does have benefits. It allows for better analysis of how people use the service, and offers on-the-fly updating to the software. Even with an iPhone sitting next to a wireless router, Siri still sometimes chokes on very basic commands. With the current always-on DRM rumors, don’t be surprised if it needs a constant network connection in either scenario.

In the cloud or handled locally, better speech is a safe bet going forward from Microsoft. Let’s hope that this time around, Redmond knocks it out of the park. It’d be nice to see the Xbox team finally deliver on the full potential of the Kinect concept since they certainly didn’t do that with the current hardware.