Home > Kinect News > Kinect finally understands the Australian accent

Kinect finally understands the Australian accent

December 6, 2011

The Xbox 360 is getting a massive update today, arguably the biggest since its inception. But this dashboard update is not the only update that is being pushed through because the Xbox 360 finally learned how to understand the Aussie accent!

Australian voice control available today

We already know that the dashboard update that is bound to happen today is pretty exciting, but Australian users are really in luck. A big part of the new dashboard is that it can be almost completely controlled by voice, that is…if Kinect supports your language. Strangely enough Australia did not have voice control support yet (last time I checked they speak English just like in the UK and US) and thus would get a seriously gimped new dashboard. Fortunately Microsoft decided it would be time for Kinect to understand the Aussie accent and thus the whole of Australia is getting voice control today! Enjoy!

australian voice control,Dashboard update,Kinect,update,voice control

http://twitter.com/XboxVictorious Alan Walsh

Ahh its not just Australia theres loads more regions getting it too , Australia, Canada (French), France, Germany, the Republic of Ireland, Italy, Japan, Mexico, New Zealand and Spain.
Corellianrogue

“Xbox, put another shrimp on the barbie.” :p
Guest

“Strangely enough Australia did not have voice control support yet (last time I checked they speak English just like in the UK and US)…” is quite an oversimplification.

Actually Australians DON’T speak English just like the UK OR the US. And, come to think of it, the UK and US don’t speak like each other either. Even if the language is the same, a different underlying acoustic model is needed for each major dialect to ensure good recognition. And even then, you will have regional differences within that geography that mean people will get better or worse results. It’s all a matter of balancing how many (quite costly) acoustic models you want to construct to get reasonable results for most of your audience. Likewise, Spain and Mexico are different models, and Canadian and French French.

Anonymous

Thanks for clarifying that. It’s quite interesting that so much work goes into make the recognition work for different dialects. Does all that have to be developed from scratch or are those acoustic models reusable and can be obtained from for example research companies.

Guest

Theoretically, there is at least some opportunity for reuse, but not practically. Acoustic models need to consider the environment, and ideally, the microphone/audio system capabilities. Up to know, almost all acoustic models have been built for PCs, smartphones, or cars. Kinect in a living room setting is different from all of those.