Activate microphone not working

I have tried everything I can think of, to no avail.


The activate microphone command simply does not work for me anymore. It used to work fine.

I can configure voicecomputer, set the primary and secondary figures just fine. I can configure the microphone, and am getting 94% figures quite routinely.

I backed down the secondary figure manually to 80% for the activate microphone command, still nothing happens.

This has been going on for a couple weeks. Today I got tired of it not working, have tried everything. This includes running the reconfigure voicecomputer command. Has not affected this problem.

I am running voicecomputer 2019 lite, version 1.3.1

Please advise, thank you.
By the way, the command “activate voicecomputer” with the Dragon microphone off works just fine.

Thomas Reid2019-04-30 16:38:01
Please email me the
VC_MicrophoneLog.txt file located in
C:ProgramDataVoiceTeachMicrophone
and the
commands.json file located in
C:ProgramDataVoiceTeachMicrophoneVoiceComputer



The fact that 'Activate VoiceComputer' works and 'Activate Microphone' doesn't work even if you set the secondary accuracy to 80% indicates that this might be a Dragon problem (which I've never seen before).

The 'Activate Microphone' command sends a command to Dragon to toggle the mic on/off. My guess is that the Dragon mic is not responding to the command.

I'll look at the log files but you could try creating a new Dragon user profile and see if the commands work. (With the secondary setting set so low (80%), the command should be recognized every single time you repeat the command. And, the microphone will probably toggle on/off as you dictate a sentence.

A few weeks ago I had the following problem:

When I am scrolling down a window, I say , “page down”, if I say it within the 3 second apart protocol (that is, those 2 words together, repeated), then it turns my microphone off. Absolutely consistently. In fact, as I was typing this post, if I said “page down”, those 2 words, too closely together, then it would not dictate anything into this post. I had to say “page” wait a couple seconds, then say, “down” in order for it to dictate here. When I look at the dictation recognition history in Dragon, I see <???>.


I resolved this by removing the hotkey to turn the microphone on and off. I just restored that back to the default value. This clears up the current issue. I have also retested the above problem, I’m also not having any problems with that, as far as I can see.

Thank you again for pointing me to the point of the problem. Of course, I’m still curious as to what was causing the above original “page down” problem. At this point I recognize that there is nothing that we have to go on to figure that one out. I’ll keep my eyes open. Even writing this response, my microphone has turned off a few times. Possibly due to the very low threshold. So, I will adjust that and keep you posted if I can’t recognize what’s going on.

You can certainly consider this matter resolved at this point for microphone issue. I have used the command several times at this point to turn it back on.

Yes, you posted about the page down problem.

"Interesting, the microphone off hotkey was set as the default {numkey+}, I deleted that, leaving the field blank. Now it is working just fine."

I don’t know why page down would press the {numkey+} and turn off the Dragon mic, but try using another hotkey.


I no longer have the problem with the default setting in Dragon since restoring it this afternoon. The activate command has been working just fine all afternoon. I’m also having no problem with the page up and down command now as well. I will put it down to a glitch in the Dragon profile somewhere along the line.


Again, I appreciate your help with pointing me in the right direction. I am sorry to bother you when it’s somebody else’s problem. So again, thank you!

I need some more advice. Please take a look at my log snippet from this morning. I have a conference call every morning where I was reading some text. I kept getting voicecomputer popping up wanting to restart. Here is a snippet of the log:


Using Default Audio Input Device. - (09:53:17,460 INFO VoiceComputer_Microphone.Speech.SapiEngineListener)
ToggleDragonMicrophone Command Now Active - (09:53:17,493 DEBUG VoiceComputer_Microphone.Speech.SapiEngineListener)
ActivateControlCenterCommand Command Now Active - (09:53:17,494 DEBUG VoiceComputer_Microphone.Speech.SapiEngineListener)
Loaded. - (09:53:17,495 INFO VoiceComputer_Microphone.App)
Heard: ‘activate microphone’ (85.1458% Confidence) - (09:53:48,769 DEBUG VoiceComputer_Microphone.Speech.SapiEngineListener)
Heard: ‘activate microphone’ (92.7426% Confidence) - (09:53:57,079 DEBUG VoiceComputer_Microphone.Speech.SapiEngineListener)
Heard: ‘activate voice computer’ (92.6324% Confidence) - (09:54:00,078 DEBUG VoiceComputer_Microphone.Speech.SapiEngineListener)
Heard: ‘activate microphone’ (88.5133% Confidence) - (09:54:03,014 DEBUG VoiceComputer_Microphone.Speech.SapiEngineListener)
Heard: ‘activate microphone’ (93.6117% Confidence) - (09:54:05,640 DEBUG VoiceComputer_Microphone.Speech.SapiEngineListener)
Heard: ‘activate microphone’ (90.8715% Confidence) - (09:54:07,953 DEBUG VoiceComputer_Microphone.Speech.SapiEngineListener)
Heard: ‘activate microphone’ (93.1073% Confidence) - (09:54:10,640 DEBUG VoiceComputer_Microphone.Speech.SapiEngineListener)
Heard: ‘activate microphone’ (91.7739% Confidence) - (09:54:13,640 DEBUG VoiceComputer_Microphone.Speech.SapiEngineListener)
Heard: ‘activate microphone’ (88.7296% Confidence) - (09:54:16,017 DEBUG VoiceComputer_Microphone.Speech.SapiEngineListener)
Heard: ‘activate microphone’ (91.6342% Confidence) - (09:54:19,141 DEBUG VoiceComputer_Microphone.Speech.SapiEngineListener)
Heard: ‘activate microphone’ (90.8445% Confidence) - (09:54:22,015 DEBUG VoiceComputer_Microphone.Speech.SapiEngineListener)
Heard: ‘activate microphone’ (92.0672% Confidence) - (09:54:24,583 DEBUG VoiceComputer_Microphone.Speech.SapiEngineListener)
Heard: ‘activate microphone’ (88.1057% Confidence) - (09:54:27,267 DEBUG VoiceComputer_Microphone.Speech.SapiEngineListener)
Heard: ‘activate microphone’ (92.5454% Confidence) - (09:54:30,077 DEBUG VoiceComputer_Microphone.Speech.SapiEngineListener)
Heard: ‘activate voice computer’ (93.1700% Confidence) - (09:54:33,267 DEBUG VoiceComputer_Microphone.Speech.SapiEngineListener)
Heard: ‘activate microphone’ (89.7218% Confidence) - (09:54:37,080 DEBUG VoiceComputer_Microphone.Speech.SapiEngineListener)
Heard: ‘activate microphone’ (93.2348% Confidence) - (09:54:39,391 DEBUG VoiceComputer_Microphone.Speech.SapiEngineListener)
Heard: ‘activate microphone’ (92.8021% Confidence) - (09:54:42,332 DEBUG VoiceComputer_Microphone.Speech.SapiEngineListener)
Heard: ‘activate voice computer’ (93.2260% Confidence) - (09:54:47,203 DEBUG VoiceComputer_Microphone.Speech.SapiEngineListener)
Heard: ‘activate microphone’ (92.0029% Confidence) - (09:54:49,516 DEBUG VoiceComputer_Microphone.Speech.SapiEngineListener)
Heard: ‘activate microphone’ (92.0528% Confidence) - (09:54:52,521 DEBUG VoiceComputer_Microphone.Speech.SapiEngineListener)
Heard: ‘activate microphone’ (92.2989% Confidence) - (09:54:55,265 DEBUG VoiceComputer_Microphone.Speech.SapiEngineListener)
Heard: ‘activate microphone’ (90.7463% Confidence) - (09:54:58,087 DEBUG VoiceComputer_Microphone.Speech.SapiEngineListener)
Heard: ‘activate voice computer’ (91.4531% Confidence) - (09:55:00,891 DEBUG VoiceComputer_Microphone.Speech.SapiEngineListener)
Heard: ‘activate voice computer’ (93.1931% Confidence) - (09:55:04,265 DEBUG VoiceComputer_Microphone.Speech.SapiEngineListener)
Heard: ‘activate microphone’ (93.2695% Confidence) - (09:55:07,141 DEBUG VoiceComputer_Microphone.Speech.SapiEngineListener)
Heard: ‘activate microphone’ (90.4100% Confidence) - (09:55:10,514 DEBUG VoiceComputer_Microphone.Speech.SapiEngineListener)
Heard: ‘activate microphone’ (93.3083% Confidence) - (09:55:13,577 DEBUG VoiceComputer_Microphone.Speech.SapiEngineListener)
Heard: ‘activate microphone’ (89.7328% Confidence) - (09:55:16,391 DEBUG VoiceComputer_Microphone.Speech.SapiEngineListener)
Heard: ‘activate microphone’ (92.1252% Confidence) - (09:55:19,641 DEBUG VoiceComputer_Microphone.Speech.SapiEngineListener)
Heard: ‘activate microphone’ (90.0156% Confidence) - (09:55:22,952 DEBUG VoiceComputer_Microphone.Speech.SapiEngineListener)
Heard: ‘activate voice computer’ (90.9380% Confidence) - (09:55:26,081 DEBUG VoiceComputer_Microphone.Speech.SapiEngineListener)
Heard: ‘activate microphone’ (62.7938% Confidence) - (09:55:29,203 DEBUG VoiceComputer_Microphone.Speech.SapiEngineListener)
Heard: ‘activate voice computer’ (87.6637% Confidence) - (09:55:32,329 DEBUG VoiceComputer_Microphone.Speech.SapiEngineListener)


How in the world is this happening, when I never said either of those phrases? Obviously, it is interpreting every single word I say now as these phrases.

Help :slight_smile:

Please note, I should have said in my original post that this snippet was taken after I bumped both numbers up for both commands in order to avoid the secondary command from being recognized.

Thomas Reid2019-05-02 08:40:18

Since the two commands are Activate Microphone and Activate VoiceComputer, the speech engine is trying to match everything you say to these two phrases. In other words, when you say NaturallySpeaking, how confident is the speech engine that you said, “activate voice computer”. Logically, it should be only 5% confident. But, the engine will probably report that it’s 80% confident that you said "activate voice computer."


So, you need to take its confidence level with a ton of salt as you do with everything that Trump says.

Note that the highest confidence level the system reports (when you didn’t say the command) is 93%,
Thus, 94%, 94.6% or 95% is probably a great primary setting so that you don’t get any false positives.
And 93% might be a good secondary setting, so the system will consistently recognize you when you issue the command twice within 3 seconds.

We’ve developed a second way of doing this, with more words in the vocabulary. When we run this, the system will correctly understand many more words so you wouldn’t get all these false confidence levels, which makes for a much cleaner (and more logical) log file.

However, the bottom line is that everything works the same in the second program as it does in the first. Typical user still needs to set the primary setting to 94 or 95% and the secondary setting to 92 or 93%. We may introduce the second version soon but it seems to work the same.

The big issue is that I need help clearly explaining this. Or have a very simple configuration that is self-explanatory. (I need help. I’ve done the best I can…)







Thanks again Ron,


Thank you for the above explanation, that helps tremendously.

I was sort of expecting that your algorithm was doing the process of translating what I speak to English, then making an accuracy determination. I am assuming that after reading your explanation is that your process analyzes more of a digital pattern, not the English translation. I assume for the sake of speed, you don’t actually do the translation to text. Would that be accurate?

Thomas Reid2019-05-02 10:41:17

I have been thinking about the confusion that I’ve been having with this issue. I have used the configure microphone command several times and thought I was using it properly. However, I believe the confusion is that the number that it is producing should be the secondary number.


I was reading the instruction to mean that the number suggested by the configuration command should be the primary, and that a slightly lower number would be safe for the secondary. In fact, it is the secondary number that actually trips the lookout for a second call for the instruction. So, if the secondary number is too low, then almost any utterance trips a call to the command (false positive).

I was reading the instruction to mean that the number suggested by the configuration command

> should be the primary, and that a slightly lower number would be safe for the secondary.
This is correct.

> In fact, it is the secondary number that actually trips the lookout for a second call for the instruction.
> So, if the secondary number is too low, then almost any utterance trips a call to the command
> (false positive).

I believe that all three statements are correct. But, we’re not on the same page because you see a contradiction and I don’t.

I typically set my primary setting to 95% and my secondary to 93%.

When I issue a command, it’s typically understood with a confidence level of between 93 and 96%.
So my primary command is only understood 60-70% of the time.

If the primary command is not recognized.
(Let’s say the command is only recognized with a confidence level of 93.2%.)
The 93.2% confidence level trips the lookout for a second call for the instruction.

I then repeat the command within 3 seconds. (which gives me plenty of time to see that the command is not recognized and to say the command a second time.)
If the program understands my second command with a confidence level of 93% or higher (the program typically understands my secondary command 98% of the time) then the program carries out the command.

I get very few false positives with these settings.

But, if you get a lot of false positives, look at the log file.
1) If the program is giving you false positives because it’s misunderstanding random words as a command word, increase the confidence level of the primary command.
2) If the program is giving you a false positives because it’s misunderstanding random words as command words twice within 3 seconds of one another (with a confidence level higher than your secondary settings, increase the confidence level of the secondary setting.

In the meanwhile, we hope to introduce an update to the microphone feature in the coming months. The update will include a large vocabulary and maybe a short training. Both these features will increase performance of the VC Mic. (The training may come in a later update.)

In our own testing, these features don’t seem to make a lot of difference in the accuracy of the Mic.
But, we get such good accuracy with the feature, that it’s hard to tell.
However, the improvements should help those who aren’t getting the great results that we are.

















Once I set my secondary to a much higher level than I had before (as mentioned, to the number suggested during the optimize process), my false positives have gone away. My success using the double command process works extremely well now. It now works exactly as expected.






Great!

My documentation needs to be improved because some get it and some don’t
If you have time, please write a short explanation of how to configure the mic feature or how it works.
Thanks!




To optimize the Voicecomputer microphone (manually)

Note: there are two confidence levels, primary and secondary. Either number will activate the program
to start looking for the command ("activate microphone" or "activate voice computer").
If you set either number too low, you will encounter several "false positives" (microphone
turning off or on, or the screen to request a voice computer restart" at undesired times).

1. Start the "VC Microphone" application, or make sure it is running.
2. Issue the voice command "activate microphone" a few times

Optional steps
a. Open the file C:/programData/voiceteach/microphone/VC_microphonelog.txt
b. Don't worry that there are too many "heard: 'activate microphone" entries
c. Note the confidence levels in general. Note particularly where you see "not a secondary match"
d. This will give you a good idea of where to set the secondary match confidence level


4. With the Dragon Mic on, issue the voice command "optimize microphone"
5. This will open up a second log file that has confidence levels when the activate command was
successfully heard.
6. I recommend setting the primary confidence level quite high, say around 98%
7. Set the secondary confidence level at about the numbers that you observed in either the
successful heard log (step 5), or the higher numbers that you notice in the
VC_microphoneLog file in the optional steps above. Start a little high, you would repeat
this process from step 4 through step 7 by slightly lowering the secondary number until
you consistently get the results you're looking for.

Great explanation!!! I’m glad that I asked.

This should be a great help to our users!

Thanks!

here we go again. I was not having any problems after getting the right numbers in the configuration file. The current logs are telling me that the “activate microphone” command is being heard quite consistently at around 94.5%. I have set my values near that number, and well below that number. However, after installing the 2020 version, I’m having problems with just this one command as far as action. Like I said, the program is recognizing the command just fine, but nothing is happening-if the microphone is off in Dragon, it will not restart when I give this command. I was having no problem just before the installation.


Can you give me any advice?

Please email me the VC_MicrophoneLog.txt in the C:ProgramDataVoiceTeachMicrophone directory


The “activate microphone” will not activate the Dragon mic if the “VoiceComputer Control Center” is active" (Will soon release an update that will fix this.)

Check that you have disabled all but one of your microphones.
Right click on the speaker icon on the system tray > Open Sound Settings > Sound Control Panel > Recording


Check the system volume (which the VC Mic uses).
Right click on the speaker icon on the system tray > Open Sound Settings > Sound Control Panel > Recording > Microphone > Levels

Set it to 80%. (Various utilities will change the volume. The update I’m working on will set the volume to 80% and keep it there.)

RonK2019-09-07 10:12:10

looks like the culprit was that my Dragon hotkey for turning the microphone on and off was empty.


it is working fine now that I have reset the Dragon options back to default.

What threw me off initially was that when I went into the configuration for VoiceComputer, option seven, that process was telling me that my confidence levels were 94% plus. However, when looking at the log file that you suggested here, the confidence values are much lower. I have now adjusted my levels to reflect what is in the above log file.

I just rediscovered why I changed my default hotkey values in Dragon. When I give the command “press enter”, dictated twice, it turns off my microphone. I don’t know if that’s a dragon thing, or a VoiceComputer thing, or something else on my computer.

another issue that is quite confusing, microphone command 1 is “activate microphone”, so we give settings for that. The impression given on the screen instructions suggest that the settings for commands 3 can be a little lower. However, one of the recognized strings for commands three is “activate microphone”. These two commands are fighting each other it would seem.




I just rediscovered why I changed my default hotkey values in Dragon. When I give the

> command “press enter”, dictated twice, it turns off my microphone.

Yes, you and/or someone else reported this some time back.

Does the enter key and numkey+ share the same key on your computer?

There’s a ‘Enter two’ (enter <1-20>) command in VoiceComputer, not VoiceComputer lite.

Could you use this instead of “press enter”, dictated twice?

Or, you can try another hotkey for the Dragon mic. (You have to restart the VC Mic)

> However, one of the recognized strings for commands three is "activate microphone"

That wouldn’t be the case in the next update…