bigjake620
Can someone help me with getting my speech recognition to work in Windows 10 on my new PC? It works fine on an older windows 8 PC that I had been using.
I added the microphone control to the project to make sure that it hears the microphone. That works.
In the speech recognition configure menu, when I speak the levels tab on the side goes up and down. So the microphone seems to be working. I'm using an external microphone.
There isn't a line readout in the speech recognition control in the upper right corner as there is on my windows 8 version. So I don'tbelieve the module is connecting to the speech recognition part of microsoft. No confidence number or anything. This is what my project looks like:
When I open microsoft word, I'm able to get the dictation working. So I believe the speech recognition part of microsoft works, but it doesn't appear to connect to ez-robot.
Any help would be appreciated. I've been struggling with this for a few days.
click on the question mark ? in the speech recog to get info .also check the gear icon to see if there a englich language there.
I had the same problem when I bought my new Windows 10 computer. Consensus was that MS changed something. Try updating your Windows version. That may help.
Here's my original thread with the question
https://synthiam.com/Community/Questions/help-speech-recognition-problems-17647
Thanks, perry. I was looking for that thread. There was an update that broke a lot of things for windows programs about a year ago. Use windows update and make sure everything is the latest as Perry advised
I've updated my windows 10 this morning to vers 1903. I restarted my computer and updated the ez-robot software. Unfortunately the speech recognition still isn't working. I don't have the line graph as before. I do have the en-us selected and my microphone is working. See below. Do you have any other ideas? Thanks.
did you save your project on your pc and cloud? also the volume must be pritty high in win10.atleast thats what mine says.
Try the beta version of ARC
I downloaded the beta version 2019.06.25.00. The speech recognition still isn't working. This is how the workspace looks. Again there isn't a line graph for speech recognition, the microphone is still registering. I tried both the microphone on my lap top and also an external one connected to a webcam. The webcam one makes the green bargraph increase more than the built in one on my laptop.
I don’t think anyone was able to pinpoint how theirs just started working before. I wish there was more information I could give you on what to do. For others that experienced this, windows had an update a few days late and it just started working again. However, in your case it’s not and I haven’t a clue why
For the record, Mine never did start working again. I still don't have this functionality. Perhaps it is a manufacturer bios update or something that needs to happen. I have an Alienware and it is always kept up to the minute updated.
Really Perry?! Awe man that’s too bad. Geez I wish I could figure it out but I’m really at a loss. It’s far too simple of a .net function and not much else I can try
what about the other speech recognition options like Microsoft azure?
mayby you can try this.1 save project in cloud 2 delete ARC 3 delete browser historie 4 reboot pc 5 download ARC. the red line is good.mine is same .and when click on it becomes green.i use a separt usb mic.
DJ - yes I still use the others but I had certain functionality tied to this plugin. I had it set up to activate Bing SR so the bot wasn't constantly responding to me. Always different ways to do stuff though.
Nomad - no more installing and uninstalling for me. I've kinda moved on from the problem. Thanks though.
PErry
Perry, I’m going to revisit the code again and see what I can find. Might have to get you to try a few apps for me. Wonder if ptp has any thoughts on this?
Hallo!
original thread with the question: https://synthiam.com/Community/Questions/help-speech-recognition-problems-17647/comments#comments-list
I witnessed the same behavior but it's random and sometimes is not immediately. I tried to follow up the issue with Perry but he had no time and it seems he moved on e.g. (gave up).
short story: I have two Intel NUCs (I7 and I5) both have windows 10, the NUC i7 slows down the environment (foreground applications) when the screen is off (idle after a few minutes) although i have all the power savings OFF the exception is the save screen / screen off. I've tried different patches, i changed different registry settings and the only way is to disable the save screen and keeping on the monitors (burning) the other NUC i5 has the same windows 10 version, the same Intel tools (different drivers) and no issues. Windows 10 is the common element... I blame Intel !
Returning to the initial discussion:
If we combined all users we have Perry, Will, Abrown, bigjake620.
I wrote some test code but I'll need someone able to reproduce the issue to run the tests.
@ptp, I can see the audio device and sound wave produced by my external mic with your audio tool box plug in. But, I get nothing in the native speech recognician. Greyed out and no wave form. It does not respond to my commands. Various computers with various versions of Windows 10 updates either work or don't work. I have 3 laptops, and two desktops.
I've read the older thread. This seems like a strange bug. I've tried changing my region an my speech to en_GB, but still haven't gotten the Speech Recognition to work. I still am not getting a line chart in the Speech recognition. The Bing Speech Recognition works well, but it's annoying having to hit the 'Start Recording' every time you want to issue a command.
i use external mic from about 20 dollars brand is mx crypt .works perfect.
Hey does this plugin work? https://synthiam.com/Products/Controls/Audio/Total-Speech-Recognition-17688
Not on this machine. I'm 100% confident that it worked on 1803 but broke on the last update to 1903. And was broke before 1803.
Both of my laptops are 1803 and it works... hmm
No, unfortunately the Total Speech recognition plugin didn't work either. The bar looks a little different then the Speech Recognition, but still no green line. I've tried an external microphone too. Bing Speech works, so I know I'm getting sound.
Bigjake, what windows build are you running? You can see it it WINVER
Here is my Windows version
Can you try this? Seems there’s a lot of issues on speech rec with build 1903
mine is same as perry.
I can reproduce the issue by changing this from ON to OFF. It should be ON
Yes, I have the switch "Allow apps to access your microphone" turned on. My build is the same as Perry_S.
Unfortunately did not fix the issue on this machine. Mine was "ON". I turned it off then, back to on. No change.
below are more setting to put in on position.
Yeah - good find nomad. I also found this... http://maxedtech.com/fix-microphone-not-working-in-windows-10-1803/
mine destop appvieuwer says off.
OK....when i looked where nomad pointed...i had ARC open but was not showing up there, but the Speech runtime exe was showing in use. I turned off the "on" to "off" then back to on. The Runtime no longer showed it was in use and would not come back online. So i shut down ARC and rebooted the computer. On reboot a windows update was applied. I restarted ARC and revisited the microphone privacy settings and I had ARC now listed....and IT WORKS in ARC!
So was it the windows update on reboot, or killing the speech runtime and rebooting that did the fix?! Someone will have to confirm which one did it.
in the link from dj says after check on ,close and reopen app.
here it says only reopen app.but its possible you have to reboot
Hmm, I checked that the ARC can access the microphone. Also rebooted as fxrtst suggested. I'm still not getting either of the speech recognition modules to work. The microphone does see wave patterns.
the lower speech recog is not complete.i would delete bolt and take a new window.
@bigjake, did you turn off the "on" slider and turn back on?
Deleted......................................................
Maybe turning it off and on?
the microphone control doesn’t use the same libraries as speech recognition control. The microphone control talks directly to the device. However, the speech recognition control uses the builtin Microsoft Windows speech recognition library and that’s the trouble. It’s because the speech recognition system is separate from ARC and built into windows.
We’ll figure it out
I've moved the switch back and forth. when I remove the microphone control the ARC and Speech Run Time aren't in-use.
Correct. Now try to reboot and open ARC and see if you have the wave.
Tried everything listed above with no solution. Rebooted and restarted and flipped the switch on and off too many times to count.
Noticing Will gets the red text saying 'Currently in use'. I do not. Also, curiously it says last accessed 10/8/2019 even though I have beenb using it many times today.
It may indeed have to do with the latest windows update then? Not sure if this is latest latest, but just what installed on this machine, displays:
Interesting. I am on version 418 so 388 came and went for me. I don't understand Windows update at times. If version 418 is out why is it serving up 388 for you?
I know i never understand the way they deploy the updates...windows updates says i have the most current update?!....well I certainly was hoping for a universal fix..but apparently nothing I did managed to help others. Dang.
bigjake620 try a new blanck ARC.you have only two windows.
perry did check number 3 see pic.
I rebooted and restarted ARC with a blank project. The "Speech Runtime executable" is in use, but not ARC.
i had to do this reboot thing for six month in a row.its very annoing.but it needs to be done. there is not much els you can do then delete history and ARC.reboot and redownload ARC. i think only one speech recog per project is allowed.but am gessing.
Well I have solved my problem. Since this is happening to a lucky select few I imagined it might be the factory bios or system software. I disabled my Alienware Sound Center at startup and that fixed it. I imagine Alienware needs to change their software to avoid this conflict. Probably the same for some other manufacturers. Disabling it had no adverse effect on my system.
Now I have the coveted green squiggle line
Perry S
great find and solution.love the pictures.
Alan
It will be interesting to see what type of computer bigjake has. This might be his issue as well.
Nice find!!
bigjake - could you press ctrl-shift-esc to load task manager. And show us what’s in the Startup list?
I've been out of the country for a few weeks, and am back now. Thanks for pursuing this. I don't have alienware running at startup. Here's what I have:
Do you see anything that I should turn off?
Try disabling the Realtek HD Audio Universal.... I believe it is similar to the Alienware sound center (in fact, may be the same software that Alienware re-brands).
Alan
I disabled the Realtek HD Audio on startup, rebooted and relaunched ARC. Still no squiggle on the speech recognition module
Did you notice your Microsoft Speech Recognition is disabled? I haven't read back through this whole thread, but was there a reason you disabled that?
Alan
when i used the microsoft speech for the first time.i had to follow a speech test.i needed to read some centens. after the pc recog my voice,i could choose ,use my pc with my voice or not.i did choone not to use. but after that the speech recog in ARC worked.not before the test.
Good news, I updated my computer again and now the Speech Recognition works. I'm not sure what the update fixed, but it's great to talk to my robot again. Thanks for the help.