The VC microphone gives me nothing but false

The voice computer microphone has simply not been working for me. I've spent about six hours problem-solving to try and get it to work, but I still haven't had any luck. I've read the documentation and tried quite a few things.
The problem that I'm having is that the microphone does not pick up any of my commands whatsoever. When I lower the confidence levels needed, all that happens is that I get a bunch of false positives. For example, when I lowered it to meeting 93% confidence to recognize "voice computer" and "control centre" it interpreted every sentence I spoke as containing one of these commands. This came close to crashing my computer. I tried quite a few other confidence levels as well.


In order to help you, there are a few questions that you need to answer for us.

First, what microphone/soundcard are you using?

Some microphones work better than others, but we need to specific microphone model and soundcard that you're using. Also, what version of Dragon are you using?

Second, go to the Windows Control Panel | Speech Recognition | Set up microphone. In blue at the very top it identifies your soundcard. Keep in mind that Windows does not recognize microphones, it only recognizes soundcards. Regardless, is what is displayed at the top of that screen the same as the soundcard that you're using for Dragon? If not, then go to the Sound dialog and look at the microphones that are listed under "Recording devices". Do you have more than one Recording device listed as active? Is the default soundcard the same soundcard that you're using for Dragon and in the Speech Recognition dialog for WSR. These have to be identical.

Third, if you have more than one Recording device listed in the Windows Sound dialog, be sure that you disable all but the one that you're using, but don't hide any that you disable. One thing that happens with some users of laptops is that you have both the soundcard that you're using for Dragon and the built-in microphone/soundcard that comes with the laptop (i.e., WebCam microphone). Both of these cannot be active in the Windows Sound dialog.

Since the VC microphone is using the Windows Speech Recognition (WSR) speech engine, setting the confidence level for the VC microphone to open the "VoiceComputer Control Center" can be tricky. Lowering the confidence level for the "Control Center" too low will cause what I think is your problem. Don't assume that the confidence level for the "VoiceComputer Control Center" where it doesn't seem to respond always means that you have to lower that confidence level. Sometimes, particularly on boot up of your system and under some other conditions, you may have to say "VoiceComputer Control Center" several times before it will open. You need to try that several times before you conclude that you need to change the confidence level. The reason for this is that you don't want spontaneously spawning the VoiceComputer Control Center, which is what I suspect is happening with you.

Nevertheless, without the initial information requested above, we can't help you take the next step.

Addendum: Max, I found out from the KnowBrainer forum what microphone you're using (i.e., Blue Yeti) and the fact that is connected via a USB SoundPod.

Your system configuration is fine, but your microphone is highly suspect. As you noted on the KnowBrainer forum, you're having a problem with extra characters in the Spelling Window. This is not uncharacteristic of the Blue Yeti microphone.

As another example, we know that the Buddy USB 7G desktop microphone does not work with VC microphone at all. The probability is not that Blue Yeti microphone falls into this category, but that is not ideal, particularly for use with the Windows Speech Recognition (WSR) speech engine.

What you can try is going to the Windows Control Panel | Speech Recognition, and train your microphone and see if that helps. However, you may be caught between a rock and a hard place in this regard. It's a known fact that the Blue Yeti microphone does not work well with Windows Speech Recognition (WSR).

Chucker2017-05-01 09:58:47

It is a blue yeti blackout USB microphone. I believe that there is no separate soundcard associated with USB microphones. Both Windows speech recognition and Dragon are using this microphone.

Yes, I had already disabled all the other recording devices.
What I had said in the post is that it never successfully recognized “voice computer” or “control Centre” when I said them. It only recognizes them by mistake, and when it was at a confidence level that may this happen it would almost always happen, regardless of what I say.


Just out of curiosity say "show VoiceComputer version" and let me know what version of VoiceComputer you're using.

The most current version is 2.19.2.

The reason that I ask is that the commands that you're using for opening the VoiceComputer Control Center are not correct based on the current version above.

Yes, thanks for your reply about the microphone! I will be getting a new one soon.
I am using version 2.19.2 In the window for configuring the accuracy settings for the VC microphone it said that voice computer and control centre were the commands.

I have a better microphone now (a Speechtech GN USB-2). The VC microphone is still definitely unusable, and I have tried toggling the confidence levels some more. At confidence levels around 95% it can still misshear around half the sentences I speak as containing one of the key words, but it still can’t really tell the difference between me saying one command versus another.

I tried to make the configuration as clear as I could but I clearly need to improve it.

Let's take this step by step.

First, have you gone through all the steps in Using the VC Microphone.pdf located in: C:ProgramDataVoiceTeachVoiceComputerDocumentation.

1. You've disabled all other microphone devices.

I’ve gone through the whole document a couple of times. And I’m really not sure if it’s worth my time to troubleshoot something that is working this poorly.

Right now I have the accuracy settings 94.4% for the microphone toggle commands and the commands to bring up the control centre. Right now this is letting the VC microphone correctly recognize these two commands, but every third sentence I say I get a false positive. Usually it works even more poorly than this.
Sorry that you're having such problems with this.
1) The Windows Speech Engine is imperfect; 2) None of us pronounce the same words exactly the same way each time we say them; and 3) The microphone is always on, which will inevitably lead to some false positives (but hopefully only one or two a day).

Setting the confidence levels where you need them is a balancing act. You against the Windows Speech Engine.

Look at the Microphone Log file. What accuracy is it reporting for the microphone and the commands to bring up the control center?

Have you gone through the Windows Speech training?

You can't set the confidence level so low that you can get 100% accuracy every time you say the commands because that will lead to false positives.

I find that I get extremely few false positives when I set the confidence level high enough so that I can only achieve the levels maybe 3 out of 4 times.

That means that the system will often understand me the first time I issue a command but sometimes I have to repeat the command 2-4 times before it understands me.