PDA

View Full Version : Can't get Voice Recognition working, please help.



Griffin-5
03-04-2017, 04:19
Great application this. I'm loving it with rFactor 2.
However I can't seem to get Voice Activation to work despite going through the necessary steps. I've installed
SpeechPlatformRuntime.msi and then MSSpeech_SR_en-GB_TELE.msi (I used the GB variety as I am in Australia) as per the instructions. I've also tested that my default microphone is working properly so that isn't an issue.

When I try and activate Voice Recognition in CC v4.6.4.2 I get this error message:
14:14:08.696 : Attempting to initialise speech recognition for user specified location en
14:14:08.703 : Success
14:14:08.708 : Unable to configure speech engine grammar
14:14:08.709 : Exception message: The language for the grammar does not match the language of the speech recognizer.
14:14:08.711 : Attempted to initialise speech engine - success = False

Loving this app, just wish I could get this feature working. Any help is greatly appreciated.

mr_belowski
03-04-2017, 06:58
There's a setting in the properties screen for 'speech recognition locale' or something. It defaults to 'en'. Try setting it to 'en-GB' (but en really should work - it just means 'use whatever en language is available). Perhaps your windows install expects en-AU?

Griffin-5
03-04-2017, 09:27
Thanks so much. Changing that to en-GB worked perfectly. I had already changed my setting to GB in Windows but I suspect that if I still had it at my default, then en-AU would have worked too.
Really appreciate your help. I'll donate to your cause to keep this app's development going. Cheers!

oOBATOo
09-04-2017, 16:00
Hi
Its been a while but Ive come back to Raceroom and trying CChief again wow the apps been updated, but the propertise windows still a mess.
How to find these files again? It reads,
16:54:04.136 : Attempted to initialise speech engine - success = False
16:54:04.162 : Attempting to initialise speech recognition for user specified location en
16:54:04.165 : Unable to initialise speech engine with voice recognition pack for location en. Check that SpeechPlatformRuntime.msi and MSSpeech_SR_en_TELE.msi are installed.
16:54:04.165 : Exception message: Retrieving the COM class factory for component with CLSID {49428A60-C997-4D0E-9808-9E326C178D58} failed due to the following error: 80040154 Class not registered (Exception from HRESULT: 0x80040154 (REGDB_E_CLASSNOTREG)).

Now this is in windows 10 I need where to dig those Speach plaform files out of, an how to turn on shared memory Raceroom. pls

mr_belowski
09-04-2017, 16:11
There are some instructions on the 'help' screen (click the button). You need to install the Microsoft speech recognition runtime and English language pack - the help screen should have URLs for these

oOBATOo
09-04-2017, 17:55
There are some instructions on the 'help' screen (click the button). You need to install the Microsoft speech recognition runtime and English language pack - the help screen should have URLs for these

Did as you suggested, but talk to me mr....
Crew cheif is not responing, the only thing I get is(raceroom win 10) two beeps crosing the start/finish line, Its now not saying it cant setup the engine.

18:48:16.860 : Unable to set default audio device
18:48:16.860 : Exception message: Cannot find the requested data item, such as a data key or value.
18:48:16.860 : Attempted to initialise speech engine - success = False
18:48:16.891 : Attempting to initialise speech recognition for user specified location en
18:48:16.907 : Success
18:48:16.907 : Unable to set default audio device
18:48:16.907 : Exception message: Cannot find the requested data item, such as a data key or value.
18:48:16.907 : Attempted to initialise speech engine - success = False
18:48:16.923 : Checking for updates
18:48:17.845 : Check for updates completed

not sure what the fault is mate, but as a clue 18:48:16.907 : Unable to set default audio device might be the thing.

mr_belowski
09-04-2017, 18:32
The app will use the default audio output device, which is set in the windows sound control panel. Normally windows will automatically select one to be default but perhaps you need to do this manually

oOBATOo
09-04-2017, 19:53
The app will use the default audio output device, which is set in the windows sound control panel. Normally windows will automatically select one to be default but perhaps you need to do this manually
Its set as default far as I can see......
I get default replies(spotter no spotter messages)but nothing else.
:confused:
it seams confused

20:54:06.218 : Starting app
20:54:06.265 : Set Race Room mode from previous launch
20:54:06.265 : Minimum sound pack version = 105 using sound pack version 112, driver names version 114 and personalisations version 104
20:54:06.280 : Preparing driver names
20:54:06.280 : Prepare driver names completed
20:54:06.280 : Preparing sound effects
20:54:06.280 : Prepare sound effects completed
20:54:06.280 : Preparing personalisations for selected name Al
20:54:06.280 : Prepare personalisations completed
20:54:06.280 : Preparing voice messages
20:54:06.405 : Prepare voice message completed
20:54:06.405 : Finished preparing sounds cache, loaded 1006 single sounds and 578 sound sets, with 267 active SoundPlayer objects
20:54:06.405 : 90 sounds have personalisations
20:54:06.405 : Reading driver name mappings
20:54:06.405 : Read 1163 driver name mappings
20:54:06.405 : Reading driver name mappings
20:54:06.421 : Loading controller settings
20:54:06.452 : Load controller settings complete
20:54:06.452 : Attempting to initialise speech recognition for user specified location en
20:54:06.499 : Success
20:54:06.499 : Unable to set default audio device
20:54:06.499 : Exception message: Cannot find the requested data item, such as a data key or value.
20:54:06.499 : Attempted to initialise speech engine - success = False
20:54:06.546 : Attempting to initialise speech recognition for user specified location en
20:54:06.562 : Success
20:54:06.577 : Unable to set default audio device
20:54:06.577 : Exception message: Cannot find the requested data item, such as a data key or value.
20:54:06.577 : Attempted to initialise speech engine - success = False
20:54:06.687 : Checking for updates
20:54:07.632 : Check for updates completed

mr_belowski
09-04-2017, 20:36
so do you get regular messages, like 'green green green', and the 'radio check' at the start? You also need to set a default sound input (recording) device for speech recognition to work

oOBATOo
09-04-2017, 21:40
Just done a breif session I did not finish the race Igot,
allsorted message when I first left the pit BOX
I got Green green green
And that was all dispite being told to spot......... .

HIH Ive tried reinstalling CC but cant redownload the three green packages at the top says there all ready downloaded.
thought you had a lot of driver names but only see about 10 in the list or less..

no radio check

oOBATOo
09-04-2017, 21:42
was trying to [post the log but your forum does not seam to like it I keep getting 403 messages
16

oOBATOo
10-04-2017, 16:51
Right got it sorted, its all running perfectly.
I had to uninstal the app,
Clear out appdata, whiloe I was their I cleared out the crewcheif four directory thus totaly resetting your app.
I then installed and downloaded each sound pack that was present all three of them.
Then everything works as expected.

Its worth noting for future that this can happen although these language packs where shown as installed their must have been an error some place.
You live and learn.

mr_belowski
11-04-2017, 16:56
Glad to hear man :). I started a new job last week and am still getting up to speed so might be a bit quiet for a while

belaki
12-04-2017, 22:46
Gentlemen, do be advised I could not (despite using all of my significant and well exercised technical capabilities) get voice to initialize. The failed process: install CF4, install MS SPR, install MSSpeech_SR_en-US_TELE.msi, update CF4 properties to en-US.

The fix was to download and install MSSpeech_SR_en-GB_TELE.msi and update CF4 properties to en-GB. After the restart CF4 become severely happy with the whole notion.

This debugging episode complete, on a lark I went back to properties and updated to en-US just for shucks and grins.

Voila! My British pit boss can now understand me and I don't feel compelled to convert a southern US drawl into a cockney just to find out if the last LMP2 driver I ran over tore up my aero.

On a more serious note, my deepest thanks to Jim Britton and the rest of the contributors CF4 for a truly astounding piece of programming. A donation is headed your way.

mr_belowski
13-04-2017, 07:12
I think this issue may be caused by Microsoft being, in technical terms, 'a bit shit'. Crew Chief Mega Corp apologizes for any inconvenience this may have caused. Thanks for the support :)

inthebagbud
16-07-2017, 07:42
Hi I have started to look at the talk to crew chief option .

I have installed the language packs and speech engine as per the opening post on pcars website, I fire up crew chief and hear the events radio check message but then the following unhandled exception is thrown

20

I am using a turtle beach handsfree headset which is Bluetooth connection to pc and I have checked it is set as default recording device and it is

EDIT

I rebooted windows a couple of times and did get CC to recognise a command but it then hung, so thought I would disable the listening aspect for time being but now when I hit the stop button CC stops responding and as I have CC starting automatically I am stuck in a loop

21

22

EDIT 2

OK a few more reboots/hanging apps and speech on and off and we seem to be working. Windows 10 sometimes seems very slow when installing updating programmes, it's as if it is still working in background and you don't realise it ?

EDIT 3

Back again and its stopped working again - no sound at all and when I try to stop CC it goes into not responding, CAD stops the app butthen it wont fir up again and I also cant open the windows "sound mixer" window

The Iron Wolf
16-07-2017, 19:59
Hi I have started to look at the talk to crew chief option .

I have installed the language packs and speech engine as per the opening post on pcars website, I fire up crew chief and hear the events radio check message but then the following unhandled exception is thrown

20

I am using a turtle beach handsfree headset which is Bluetooth connection to pc and I have checked it is set as default recording device and it is

EDIT

I rebooted windows a couple of times and did get CC to recognise a command but it then hung, so thought I would disable the listening aspect for time being but now when I hit the stop button CC stops responding and as I have CC starting automatically I am stuck in a loop

21

22

EDIT 2

OK a few more reboots/hanging apps and speech on and off and we seem to be working. Windows 10 sometimes seems very slow when installing updating programmes, it's as if it is still working in background and you don't realise it ?

EDIT 3

Back again and its stopped working again - no sound at all and when I try to stop CC it goes into not responding, CAD stops the app butthen it wont fir up again and I also cant open the windows "sound mixer" window

Can you try using Shared Memory mode for pCars instead? You need to enable shared memory in pcars, and use Project cars (64bit/32bit) game choice.

mr_belowski
16-07-2017, 20:29
Not sure that'll make any difference tbh. There may be something weird in the sound drivers for the headset. Does it behave properly with a regular mic and speakers?

The Iron Wolf
16-07-2017, 22:54
Not sure that'll make any difference tbh. There may be something weird in the sound drivers for the headset. Does it behave properly with a regular mic and speakers?
my bad, my original response was rushed (had to go watch F1 with my neighbor). It indeed seems like a problem with sound device rather than connection to the game itself.

I am using bluetooth on my computers to connect wireless speakers, and it's finicky. One thought I have is to make sure that both bluetooth output and input devices are connected and are marked as default in Sound control panel, before launching CC.

inthebagbud
18-07-2017, 18:05
Ok I have had a few days of good speech recognition but it failed tonight and it appears that the exception that is thrown is if the windows speech recognition has not been initialised properly . I generally leave my laptop on and for the usual unknown reason the speech recognition had stalled and therefore CC threw the exception tonight . When speech recognition was back up and running everything is ok so I think its just the way CC is telling me that speech recognition is not correctly initialised .

I will keep my eye on it and if I see any other issues will let you know

Thnaks for the replies

Oh and I am using on xbox if that makes a difference

inthebagbud
23-07-2017, 16:32
Thought I would provide further update which is on a negative as I have had to stop voice recognition as it failed more times than it worked.

I was not able to isolate why as the failure was not constant so I was unable to go "Ah that's it" my only suggestion is that maybe one of the programmes CC/MS speech is going to sleep and losing the connection . MS does seem a little flakey as currently I stopped MS voice recognition and it wont start again unless I reboot

Sorry I couldn't be more help

I was trying to use it in always on mode, Bluetooth headset, xbox with pc app

mr_belowski
23-07-2017, 16:34
That's really really odd. It's been rock solid for me since I started using it back when all this was fields. All I can think is that it's something in your sound drivers :(

inthebagbud
23-07-2017, 18:37
That's really really odd. It's been rock solid for me since I started using it back when all this was fields. All I can think is that it's something in your sound drivers :(

yep definitely a little weird but I will keep my eye on it and maybe revisit in the future. I might look for the headphone lead and see if connecting direct to laptop makes any difference

sure looks like something MS/headset issue

inthebagbud
30-07-2017, 05:45
Crew chief uses the default audio playback device.

Noticed this quote in another thead which may explain my issues as win 10 recently is having issues keeping the Bluetooth connection established to my headset . The headset stops being the default device at random times and CC goes quiet presumably because it has lost its default device .

mr_belowski
30-07-2017, 09:27
Nasty. That sounds like quite a bad driver bug - like the device is constantly disappearing and reappearing

YoMopho
02-02-2018, 19:58
I've tried everything on this thread to get it to initialize, and it wont' :(.

I've reinstalled/wiped out app data. Switched between en-US and en-GB. Nothing can get it to initialize. Any other suggestions on other fixes?

YoMopho
02-02-2018, 23:35
I've tried everything on this thread to get it to initialize, and it wont' :(.

I've reinstalled/wiped out app data. Switched between en-US and en-GB. Nothing can get it to initialize. Any other suggestions on other fixes?

Nevermind. So I did this to fix my own problem. I tried to install the es-MX package and change the properties to es-MX to see if it would stick, and it did. As much as I wish I was Fernando Alonso, I don't speak spanish, so I changed it back to en-US, but it didn't work.

So this was the fix:

1. Uninstall CrewChief v4
2. Delete Apps Folder for CrewChief
3. Uninstall Language Packs
4. Reinstall Language Pack (en-US).
5. ReInstall CrewChief
6. Change Properties to en-US.
7. Then enabled voice control and it initialized.

:)

Artistichalo
16-02-2018, 00:02
How does this work if I'm using a Blue Yeti mic and Voice Meter software .

aritanbm
18-02-2018, 00:04
Hi all!

I'm coming from Voice Attack/DRE and everything works great there, but I heard so many great things about CC that I had to try. Checked out some videos and loved everything. Some say it runs smoother than DRE. Anyway...

Unfortunately, I'm another one who can't the voice recognition to initialize. I've already checked and tried everything on this topic. Installed, reinstalled, wiped, downloaded different language packs, deleted folders and all that. Nothing seems to work!! I always get the message "Unable to initialize speech engine with voice recognition pack for location en. Check that SpeechPlatformRuntime.msi and MSSpeech_SR_en_TELE.msi are installed."

Downloaded and installed both SpeechPlatformRuntime.msi and MSSpeech_SR_en_TELE.msi (before and after installing CC)
Checked the default recording device
Reinstalled everything, wiping the data and folders before
Tried the speech recognition locale change
Tried different language packs

Really really don't know what else to do. Anything anyone suggests?

Btw, I use windows 10.

Thank you in advance.

taosdt12
18-02-2018, 02:18
no radio check

mr_belowski
18-02-2018, 12:52
If you're struggling to get speech recognition to work, go through the steps in this post, one by one and exactly as written

http://thecrewchief.org/showthread.php?547-Speech-recognition-trouble-shooting&p=3354#post3354

aritanbm
18-02-2018, 13:56
mr_belowski, it worked! Thanks a lot!

Is there a way to make the speech recognition engine better understand me? I mean, I'm not a native speaker, but I'm not that bad either. There are some commands that CC just can't understand when I say them, lol. Like "Whats P#'s last lap?". It simply can't understand! Frustrating" Lol.

asner
22-02-2018, 14:21
Just a quick note: The current MS speech framework download did not work for me. Only after I downloaded the version from the link above everything was set up in no time. Obviously the API has changed or something.

kevindang
26-02-2018, 23:00
i have this same problem, thank for share

MaXyM
24-03-2018, 16:54
Hello Mr Belowsky
My story is a bit different, because finally I figure out how to make voice recognition working. Actually I narrowed it to one condition which has to be fulfilled to make it working.
Voice recognition doesn't work until I open Recording devices window (the one from systray). This window has more tabs (playback devices and so). But it works only if recording devices tab is active. Window itself doesn't have to have a focus. It might be on the bottom, but must be opened with recording devices tab active. It will remain working even if mic properties modal window will open.
Once I close recording devices window, or change the tab, CC stops recognize the voice.

I did a test: I activated playback devices tab. Pressed and hold PTT button and then switched to recording devices - CC wasn't able to hear me.
But If I start with recording devices, press PTT and then change tabs - the recognition works.

Strange. I know.
Any idea why? What it can be connected to?
Of course I can use CC with Recording devices window opened. But still.. it's not how it should be.

johnsher
18-04-2019, 22:27
Voice recognition doesn't work until I open Recording devices window (the one from systray).

new user here. Voice recognition was driving me crazy as I'd mostly get "sorry mate, can't hear you". A bit of googling revealed the message above so I figured, nothing to lose here and gave it a go. It now works perfectly but as stated isn't really an optimal solution. Has someone come by a better one in the past year?

johnsher
25-04-2019, 08:17
new user here. Voice recognition was driving me crazy as I'd mostly get "sorry mate, can't hear you". A bit of googling revealed the message above so I figured, nothing to lose here and gave it a go. It now works perfectly but as stated isn't really an optimal solution. Has someone come by a better one in the past year?

in case it helps anyone it appears the problem is with using "hold button" for voice recognition. As suggested in another thread I switched to "press and release" and that has sorted things.

Revan
18-06-2019, 14:32
Voice recognition doesn`t work for me in most of the cases.

Most times it doesn`t ecognize the trigger word "chief" when i say it but oddly enough it kicks in when i breathe sharply.

and if it actually works then the app does not understand my commands.

When i set my mic so high that the app understand my commands the "Beep" for understanding the trigger word constantly beeps for the smallest noise in the room like the ticking of my wall clock

This is so annoying.

mr_belowski
18-06-2019, 14:50
have you tried recording what your mic is picking up when you make a command, and playing it back? I might help you work out the right input level / gain for your mic. A application like Audacity (free open source sound recorder) can really help here - you can see the waveform so it's easy to spot clipping or very low input level

Revan
18-06-2019, 16:30
I double checked with teamspeak recording test. Sound quality is good for my amateur point of view.
I have a asus xonar sound card with an external device that includes a second mic for noise cancellation. This works actually quite good except for crew chief.
For now i assigned a button on my wheel but as there are not much buttons left on my new wheel that arrives in a couple of weeks i would prefer the trigger word instead.

mr_belowski
19-06-2019, 12:46
so the recognition works OK when you hold a button, but not for trigger-word mode?

JoakimBL
21-06-2019, 09:45
I have had to change from "Push and release" to "hold button" for CC to recognize my voice commands. Push and release used to work. Not a big issue for me, but maybe releated, or something you would like to know.

Dark_PT
18-12-2019, 14:44
[I am using CC for Raceroom and i cant use the voice comands this is the log that show when i try to use it:
15:43:33.755 : Invoking speech recognition...
15:43:33.756 : Cancelling wait for speech
15:43:36.583 : No speech input was detected
15:43:36.585 : PlaybackModerator: Opening radio channel as Chief
15:43:36.985 : Playing sounds, events: acknowledge/didnt_understand
15:43:36.985 : (acknowledge/didnt_understand)
15:43:37.991 : PlaybackModerator: Closing radio channel as Chief

I would like some help in this situation. And keep if the good work.