3

Why is Windows Speech Recognition (the app) in Windows 8 and 8.1 so pathetic?

I've been using it since they first put it on their OS years ago and it has hardly improved in the slightest.

It still has a hard time understanding the most basic things. And I don't want this to turn into a hardware debate. My hardware is fine and I've tried cheap junk and really expensive hardware. I've tried headphones (with a Mic) and I've tried desk microphones and I am now using a mic that's built into my new all in one pc. So far this is the best mic I've had, by far. But still - Windows Speech Recognition is pretty much useless for any task.

I know Speech is difficult. But I also know that they can do MUCH better than this. I know this because I own a (one of the cheapest models) HTC 8S (Windows Phone) and it has the BEST recognition I have ever used in my entire life. It almost always understands what I say - 100% of the time. What I don't understand is why can't they use their brains for a change instead of letting their products die. Why can't they realize they have great Speech recog on their phones and find a way to incorporate this tech into the Windows OS on the PC?

And if by any chance they are using the same stuff on the PC that is on the phone, then they must have made changes because it's completely ******.

So I've been looking for maybe an update or a new version of Windows Speech Recognition app, but have not been able to find one. Is there any way I can drastically improve the accuracy of this thing?

1

2 Answers 2

2

The only solution for now is to "Train Computer to better understand you":

Control Panel\Ease of Access\Speech Recognition

In other case, you can modify some issues by Speech Dictionary: Right Click Speech Recognition bar > Open the Speech Dictionary

And I think Microsoft is bringing a solution to above problem in next Windows code-name Threshold. see http://www.bing.com/dev/en-us/dev-center. They are also bringing Siri like voice assistant over next release of Windows Phone OS.

3
  • Thanks for the links. I've used the Speech API in my Windows Phone apps. I spent about an hour last night "training" the computer to better understand me but it just seems to be a never ending struggle. Jan 26, 2014 at 13:37
  • 1
    And I have also tried Mac OSX's Speech Recognition, (Enabled in Accessibility settings) I was expecting Good outcomes but it's also with similar issues as Windows Speech Recognition have. Just to make a conclusion: Until the Speech recognition is done offline, these issues persist, but when an online handler is available (Bing, Google, Apple)'s voice recognition then the results are more accurate, because of bigger and accurate voice database. Jan 26, 2014 at 13:41
  • 1
    Just hope that someday they all get better and better :) Jan 26, 2014 at 13:45
0

Microphones built into devices like notebooks and screens are NOT suitable for speech recognition use. They are usually made only for voip calls where the listener is human and much more able to understand what is being said.

Built-in microphones suffer from severe interference from moving parts in the notebook or computer (fans, hard drives) and have almost no noise cancelling. So any environmental noise is not filtered. This is different to mobile phones that have microphones dedicated to recording speech.

When using dedicated USB microphones, accuracy improves markedly for Windows Speech Recognition. Some people even use it for everyday dictation.

Also understand that Windows Speech Recognition has been about the same since the days of Windows Vista and won't improve anytime soon. More modern versions of desktop speech recognition programs like Dragon handle bad microphones much better.

You must log in to answer this question.

Not the answer you're looking for? Browse other questions tagged .