Nothing has changed - what control are you using? The Speech Recognition or Bing Speech Recognition? or Advanced Bing Speech Recognition?

What version of ARC are you using?

Can you share a project that i can take a look at?

Is your microphone set as the Default Default in windows?

Do you see the VU meter displaying the graph window when you speak? That shows it can hear audio

Is there any errors or warnings in the DEBUG dialog when you load ARC and the control? Add the debug dialog to view. Project -> add Controls -> Genera -> Debug

Using the latest EZB Mic is set as default. VU meter works fine and I adjusted my mike input to 87, like the tutorial. I went through the Windows 10 training lessons as well. I even tried re configuring the mic via the option in the plugin.

ex.EZB

1. I see you're using the generic speech recognition control. What variable were you expecting to see?

2. The Auto Position is not an Auto Position Movement Panel, so there's no way any other control knows how to instruct it to move - this might be what you were expecting?

I made a little video that might help? Hope so anyway

1++ DJ. I love it when you jump in and help. I always learn something.

There is a limitation to the procedure DJ has suggested. Your project can only have one [url=https://synthiam.com/Software/Manual/Auto-Positioner-Gait--16057][color=#ce3991][font=OpenSans, -apple-system, BlinkMacSystemFont, "Segoe UI", Roboto, "Helvetica Neue", Arial, sans-serif, "Apple Color Emoji", "Segoe UI Emoji", "Segoe UI Symbol", "Noto Color Emoji"]Auto Position[/font][/color][color=#ce3991][font=OpenSans, -apple-system, BlinkMacSystemFont, "Segoe UI", Roboto, "Helvetica Neue", Arial, sans-serif, "Apple Color Emoji", "Segoe UI Emoji", "Segoe UI Symbol", "Noto Color Emoji"][color=#ce3991][font=OpenSans, -apple-system, BlinkMacSystemFont, "Segoe UI", Roboto, "Helvetica Neue", Arial, sans-serif, "Apple Color Emoji", "Segoe UI Emoji", "Segoe UI Symbol", "Noto Color Emoji"][color=#ce3991][font=OpenSans, -apple-system, BlinkMacSystemFont, "Segoe UI", Roboto, "Helvetica Neue", Arial, sans-serif, "Apple Color Emoji", "Segoe UI Emoji", "Segoe UI Symbol", "Noto Color Emoji"][color=#ce3991][font=OpenSans, -apple-system, BlinkMacSystemFont, "Segoe UI", Roboto, "Helvetica Neue", Arial, sans-serif, "Apple Color Emoji", "Segoe UI Emoji", "Segoe UI Symbol", "Noto Color Emoji"][font=OpenSans, -apple-system, BlinkMacSystemFont, "Segoe UI", Roboto, "Helvetica Neue", Arial, sans-serif, "Apple Color Emoji", "Segoe UI Emoji", "Segoe UI Symbol", "Noto Color Emoji"] [/font][/font][/color][font=OpenSans, -apple-system, BlinkMacSystemFont, "Segoe UI", Roboto, "Helvetica Neue", Arial, sans-serif, "Apple Color Emoji", "Segoe UI Emoji", "Segoe UI Symbol", "Noto Color Emoji"][url=https://synthiam.com/Software/Manual/Movement-Panels-17372][color=#ce3991][font=OpenSans, -apple-system, BlinkMacSystemFont, "Segoe UI", Roboto, "Helvetica Neue", Arial, sans-serif, "Apple Color Emoji", "Segoe UI Emoji", "Segoe UI Symbol", "Noto Color Emoji"]Movement Panel[/font][/color][font=OpenSans, -apple-system, BlinkMacSystemFont, "Segoe UI", Roboto, "Helvetica Neue", Arial, sans-serif, "Apple Color Emoji", "Segoe UI Emoji", "Segoe UI Symbol", "Noto Color Emoji"], control. However you can have as many Auto Position controls as you want. I personally have several [size=3][font=OpenSans, -apple-system, BlinkMacSystemFont, "Segoe UI", Roboto, "Helvetica Neue", Arial, sans-serif, "Apple Color Emoji", "Segoe UI Emoji", "Segoe UI Symbol", "Noto Color Emoji"]Auto Position controls. Each control operates a different part of my robot. We can add a Auto Position Movement Panel but we can't import more then one of the Auto Position Controls if they use different servos. [/font][/size][/font][/font][/font][/color][/font][/color][/font][/color]

Also, you can use a Auto Position Control with your Speech Rec control. You need to write a script behind each phrase you want to say to call the action in the Auto Position control you want to run. So inside the Speech Rec Control and next to the phrase, click onthe pencil icon and write something like this:

ControlCommand("Auto Both Arms", AutoPositionFrameJump, "Center_Both_Arms") #This will initiate servos to keep them from jumpung.

Sleep(1000)

ControlCommand("Auto Both Arms", AutoPositionAction, "Both Ars Out")

Dave - you can still have as many Auto Position controls for different parts of your robot following the standard ARC procedure that I outlined above. That procedure being "Using a Movement Panel to handle moving the robot". And by moving, i mean physical location movement of the robot (i.e. how the robot moves Forward, Left Right Stop, etc). If you don't use a movement panel, then you miss out on a huge benefits of the software. Such as the camera tracking, joysticks, wiimote, default speech recognition commands, and more...

The Auto Position control I used in the video is a Movement Panel. Again, it registers itself as a Movement Panel throughout the software so all controls know how to make the robot move with simple commands like Forward() or Reverse() or Stop()

A movement control is how your robot moves. Your robot can only move with one method of locomotion and THAT is the movement panel.

Now even though you have an Auto Position movement panel, there’s nothing stopping you from adding a hundred regular Auto Position panels.

And, you can chain them together with ControlCommand(), if you needed to.

But David, in your application you would NOT use an Auto Position Movement Panel because your robot moves through a different means of locomotion. In your case, David, you would use an HBridge or Sabertooth movement panel. (I'm guessing that you use one of those two physical locomotion methods)

All I’m doing in these responses is duplicating exactly what’s in the manual pages or Movement Panel description.

Thanks DJ, I appreciate the help. Unfortunately I think we leap frogged my original problem. I understand how to use the AutoPosition function. I usually use the speech recognition window to trigger a script that executes various AutoPosition calls. I did not have them in this example program because I had not gotten there yet. I could not get any reaction out of the speech recognition box. This is not an AutoPosition question. I was just trying to get the speech recognition control to respond to the items that are already baked in there, for example "Robot show phrase list". My control remains dark with no response. Yours shows confidence values and an indication it was doing something. I just have a non responsive dark box.

I figured I could track the control variable $Speechphrase or $confidencevalue in the variable monitor but they do not show. I can not remember if they were ever there in the past.

I know my mic is working because I can set up the Bing speech recognition with no problem.

For now my control just looks like this no matter what I say to it.

Your control does not have a VU meter displaying the audio input. That leads me to believe the mic is not set as default? You did mention the VU meter works, but i don't see it in the screenshot. Also, if that's the case, check the debug log as mentioned in my previous post. Can you view the debug and see what it says?

Thanks DJ,  When I referred to the VU meter I was referencing the VU meter in the mic setup. I do not see that in the control itself, you are correct. I also believe this must be with my computer as this started when I switched to a new laptop.

Here is the results of the debug window.

EZB 0: 2/17/2019 5:22 PM - Windows version: Microsoft Windows NT 10.0.17134.0

2/17/2019 5:22 PM - WELCOME TUTORIAL disabled by user

2/17/2019 5:22 PM - Screen Resolution: 1536x864 (96x96 dpi)

2/17/2019 5:22 PM - Version: 2019.01.26.00

2/17/2019 5:23 PM - Speech Recognition Culture: en-US

2/17/2019 5:23 PM - Loaded speech recognition successfully

EZB 1: 

EZB 2: 

EZB 3: 

EZB 4: 

Below is a pic showing the mic VU and that it is functioning and set to default.

Hmmmmm - when you right click in the sound settings on the recording mic device, is there an option to make it the default communication device? I'm at a loss to what windows needs to select that microphone as the input device. The white VU in the speech recognition window is because it can't find the default device. The code in the speech recognition control does this..

          _sre.SetInputToDefaultAudioDevice();

Which simply sets the input to the default audio device, which i see you have selected by the green checkbox. But for some reason it's not finding it... You're in an unknown territory to me, i've never come across the specific setting that is required. You're going to have to share the solution with me when you find it. But it is definitely a windows setting

Perry,  I has a situation just as you described.  To fix it I just deleted the Speech Recognition Control , closed ARC, restarted ARC and reloaded the Speech Recognition Control.  Worked for me.

Thanks 3DGuy. Tried that a few different times.

DJ, thanks for the troubleshooting help. I'll let you know what I find. I'm pretty religious about updating drivers so maybe there is a setting somewhere. All other programs recognize my mic as the default as do other controls. Kinda crazy.

Perry - try this latest beta release. I added a debug message if there's no default audio device. If that's the case for your situation, you should see a debug message in the debug window. Otherwise i'm still out of ideas: https://synthiam.com/Software/Release/Beta-Release-2019-02-17-00-17660

I can confirm this. I tried adding speech recognician and i get a white box with no waveform and does'nt hear anything. If I add bing i get the wave form and it hears me fine. (not the latest beta). Tried removing and rebooting multiple times same thing happens.

There must have been a Windows Update that changed some sound card configurations. I've seen news posts about that happening from a windows update. Check if there's a newer update for your pc? This is using the default windows drivers so there's unfortunately nothing i can do from this end that i can think of hmmmmm...

Yeah I'm up to date on windows. Maybe it sounds as tho they messed up something. Just have to wait for a newer update from them. At least bing works just fine.

That’s the rare frustration I experience often from depending on other infrastructures. It’s quick for ARC users to rush to file a bug and it’s anti climate for you to hear me say ugh sorry there’s no nothing I can do and don’t know why it’s doing it lol

It’s fine that it works for me and our computers though haha

MS just released a pretty big update to Windows today. Unfortunately it did not fix the problem. My computer is an Alienware 15r3 so maybe there is a driver issue.

No additional debug information with the latest beta update?

I do not see a difference in the debug window results with the Beta

EZB 0: 2/17/2019 11:52 PM - Windows version: Microsoft Windows NT 10.0.17763.0 2/17/2019 11:52 PM - WELCOME TUTORIAL disabled by user 2/17/2019 11:52 PM - Screen Resolution: 1536x864 (96x96 dpi) 2/17/2019 11:52 PM - ARC BETA Version: 2019.02.17.00 2/17/2019 11:53 PM - Speech Recognition Culture: en-US 2/17/2019 11:53 PM - Loaded speech recognition successfully

EZB 1: EZB 2: EZB 3: EZB 4:

No messages in debug means no errors. So windows is returning "yup, everything is fine" even though something else is not. What about in the advanced configuration of the speech recognition setup? Are the input devices set there? (see two images below)

Yeah mine is brand new Asus. I'm pretty sure DJ is right with the windows update. I bought this for CES and i remember doing some testing and Speech recognician worked fine then. Windows broke something..of course it will only be for a few million people...

All looks good there. Level meter moves as I speak.

I just checked and my Cortona works just fine with the laptop mic and still works with bing....so who knows?

Perry, does your mic work with bing plug in if you add it..ie do you see the waveform?

Not cortana. We're talking about the built in windows speech recognition. You'll have to test with the actual thing that we're discussing not working.

Point was, my microphone is functioning on my laptop, in both Cortana and works with Bing speech Recognician. I was not implying that Cortana used the speech recognician system built into windows.

Okay good - want to make sure we're testing the same thing

FYI... All my PCs are up to date with all Windows 10 updates and the speech recognition control continues to work fine on all 3 of them...

Check the Microphone Setup button located inside the SR control when you click on the Gear Icon. It at the top when the new window opens. It will take you right the sound control section of windows without exiting ARC. If it's working in there it should be attached and working in the SR control also.

In the past I've had to delete ARC controls (including the SR control) and reinstall them to get them to work. Try doing that if all else fails.

EDIT: I didn't see where Will installed a new SR control and couldn't get it to work. I tried this yesterday and it worked for me in ARC. However I don't think my Windows 10 version is totally up to date. It's downloading an update now.

I just installed the latest EZB on my work computer, a Dell Latitude E6440 which is an older machine. The recognition control works on it. So I am pretty much at the %100 point in believing that the latest Windows update screwed up something for newer computers or at least computers that may have unique driver sets. My Alienware is supposed to be fully Windows compliant but there were other updates that temporarily broke it. Like MS make a patch that works for 95% of the machines then hits the last 5% corner cases at some future point.

That's crazy - hope we figure it out soon for ya! Any progress?

No real progress here for me. I will just have to wait for the next Windows patch to see if it fixes it. As a reminder for others, this functionality can somewhat be duplicated by using the google speech recognition and the AIML bot with a little work. Those are both still working fine.

Perry, FYI, I have an HP Notebook - 17-bs153cl that's only a couple years old. I have all the latest system drivers versions installed directly from HP's website. The Laptop is running Win 10 and was just fully updated yesterday. My SR control still works fine in ARC. I started a new project and installed a SR control. Still worked just fine.

You may be right, Probably a issue with yourunique driver set.

Perry,

Quoted "I will just have to wait for the next Windows patch to see if it fixes it."

One of your posts you mentioned windows version "10.0.17134.0" released last April and July/LT (my work computer has the same version).

Your last post you mentioned "10.0.17763.0" this is the current version is not LT version (not stable).

Every time i update to a new version even being conservative (only LT) i need to spend a few days dealing with minor but annoying collateral damages, I still keep a windows 7 for those dark days.

If i understood well the problem exists with two windows 10 versions, if it's a windows problem and is not solved yet, i would look for another solution maybe different vendor drivers ?

Can you post the specs for the motherboard and the microphone ?

Ptp - nice to see you!

The strange thing with perry’s Experience is the sound card microphone is working fine with naudio audio capture in bing speech recognition. It doesn’t work with Microsoft.speech. That’s the part I find strange. Has me wondering if it’s not a sound card issue but is a speech configuration issue instead.

Oh I had the same thing happen - but it turned out to be a USB video capture device.

The default mic input was the video capture device - so that was an easy fix

On my teacher's computer, the Speech Recognition Control does work.  The field does display sound waves when capturing sound, and JD does recognize the commands.

Does anyone have an idea what might be causing my problem?

Hi abrown,

I believe it may be something that was broken during the last Windows update. There is a thread discussing the problem here. It is not fixed yet.

https://synthiam.com/Question/17647

Thank you for your response on this, Perry_S.  I guess I'll have to wait until the next Windows Update to see if it clears up the problem.

Microsoft broke their speech recognition - it’s very unfortunate because many are suffering from the issue. We don’t even know what they did that caused the issue and have been trying to figure out a work around. At this time the only advice we have is to wait for Microsoft’s fix

Thanks for expanding on the seriousness of the problem, DJ.

DJ,  Are you using the commands for Microsoft Speech Platform 11 ( https://docs.microsoft.com/en-us/windows/uwp/design/input/speech-recognition ) or are you still using the SAPI 5.4 format commands ( https://stackoverflow.com/questions/11935533/c-sharp-sapi-5-4-languages )?  I am wondering if Microsoft finally ended support for SAPI without announcing it.  It hasn't been updated since Windows 7.

ARC is using the System.Speech namespace. The namespace you linked to is for UWP, which is universal windows platform - a different flavor of the .Net eco-system and incompatible. There's no way they could discontinue System.Speech, because everything using Speech is built on it that isn't UWP.

Thanks for clarifying.  Surprisingly difficult to research speech on Windows 10.

Ya it’s super difficult - and Microsoft has figured out how to trick google to have old articles appear as new. I searched for system.speech recognition not working, and a bunch of articles showed up saying 5 days ago. But the article was dated 2007...

Because Microsoft tricks google to be on top with incorrect update time stamps, it’s impossible to find current messages on the subject.

I just got the same issue, windows 10 1803

No new hardware.

and no significant updates:

If the update list is correct ... I don't think is a major Windows Update fault...

I'll try to understand what is going on.

Wait - that's actually picking up your voice? Just with no graph... is that right? So speech recognition is working, but only the VU graph is not. That is what i get from your results. Can you confirm?

No, it stop working.

It stopped working while you were using the control? Because i see in the screenshot the words "Robot Turn Left" were detected with low confidence. How did those words get there?

I open the project with a few controls SpeechRecognition, TextToSpeech and one of my plugins: AudioToolBox the control was working and i believe detected the a sentence. I left the PC to grab something i comeback and the level was gone and the SR was not working too.

I think it is the VU graph. Down your Mic input volume way down and restart ARC. See if the graph goes black... there might be a rounding error on the scaling calculation which is throwing an exception that isn't being displayed because it's in a different thread - it's a real old piece of code that i'll have to update.

Do you have this running in the taskmanager when the speech recognition control is loaded?

I have it running.

Try this release: https://synthiam.com/Software/Release/Beta-Release-2019-02-24-00-17700

I have a dark box now but no activity. Here it is next to the Bing plugin

Here's the log EZB 0: EZB 1: EZB 2: EZB 3: EZB 4: Windows version: Microsoft Windows NT 10.0.17763.0 WELCOME TUTORIAL disabled by user Screen Resolution: 1536x864 (96x96 dpi) ARC BETA Version: 2019.02.24.00 Speech Recognition Culture: en-US Loaded speech recognition successfully

Geez - i have no idea. I was hoping it had something to do with the graph, but nope. It would be nice to know if it was an issue on our side so it could be fixed. But it's such a simple piece of code, as ptp knows. Merely turn on speech recognition and wait for an event. No rocket science...  hmm

Is it the same mechanism that the Bing plugin uses? Just so odd that that continues to work.

It's not the same mechanism as the bing plugin. The bing plugin uses a different library.

The Speech Recognition control uses its own library in the System.Speech.SpeechRecognition namespace. The audio in the graph comes from that library. If the audio graph isn't updating, the library isn't sending audio, which means the library isn't receiving audio either.

The strangeness is PTP witnessed Windows change and stop working when he got up and left the computer for a minute