Microsoft denies voice control missing at Kinect launch

[mike_intellivision]

daroamer said:

mike_intellivision said:

daroamer said:

mike_intellivision said:

Decisions are still probably being made about what will and will not be included at launch.

My guess is that there will be some limited voice control -- because there would be too much complaining (as well as investigating) if there were not.

Mike from Morgantown

This makes no sense.  AFAIK the voice recognition is being done using software that Microsoft has had for years and is using in the current versions of Windows.  It's not like they developed this just for Kinect.  Why do you think it all of a sudden doesn't work?  Do you think all the videos and demonstrations have been fake?

Microsoft fake a presentation ... never. /sarcasm.  (Microsoft invented "smoke and mirrors" demos).

Any recognition software with which I have ever worked has required "training" the software to recognize the speaker's voice (or person's hand writing).   The issue becomes whether or not the average Kinect user will be able to have their comments recognized. 

By limited, I meant a small vocabulary or a small set of languages recognized. Remember, this is not a product for one country or specialized places. In Europe, it would have to recognize English, French, Spanish, Portuguese, Spanish, German, Dutch, etc.  So that will not be a small task -- and could explain the confusion over voice recognition on what will be implemented and what will not be.

If you expect Kinect to be fully fluent in every language at launch, ... 

 

Mike from Morgantown

You've never called a company and had to use voice commands to navigate the automated messages?  Did it require training to understand you? 

There is a big difference between having to have the computer type exact words that you are speaking and the computer only needing to understand the gist of what you're saying.   Many of the demos that were shown that allowed full sentence communication, such as Milo, were not only using your words but also the tonality of your voice to gauge what you were trying to say.  As far as I know there were no demos shown using complete sentences where ever word needed to be recognized.

I don't expect it being fluent in every language at all, I thought you were talking about it being limited even in English.

I have called in a company and had to use a voice-operated menu. Sometimes it works and sometimes I have been told "I did not understand that last comment."  This proves that such general systems often have limited vocabularies. If you use the words they are looking for, it generally works. But if you do not, there are issues.  I do expect much more of the former than the latter for Kinect because I think Microsoft will tell people what their command should be.

Mike from Morgantown