PDA

View Full Version : When trying to set up Talk to Crew Chief I get an error



bob
11-11-2017, 15:37
Attempting to initialise speech recognition for user specified location en
Unable to initialise speech engine with voice recognition pack for location en. Check that SpeechPlatformRuntime.msi and MSSpeech_SR_en_TELE.msi are installed.
Exception message: No recognizer of the required ID found.
Parameter name: culture
Attempted to initialise speech engine - success = False


I have installed SpeechPlatformRuntime.msi and MSSpeech English.
I have tried changing the speech directory to en-US and the Windows language to English US.

What am I missing here?

mr_belowski
11-11-2017, 16:06
have you tried fiddling with the speech recognition locale property in the app's Properties screen? But default it's en but it's worth trying en-US

bob
11-11-2017, 16:09
Yes, I have tried that. It gives me the same error but en-US instead of en

Attempting to initialise speech recognition for user specified location en-US
Unable to initialise speech engine with voice recognition pack for location en-US. Check that SpeechPlatformRuntime.msi and MSSpeech_SR_en-US_TELE.msi are installed.
Exception message: No recognizer of the required ID found.
Parameter name: culture
Attempted to initialise speech engine - success = False


I really hope I can get the app working, seems very useful for me with Oculus Rift.

mr_belowski
11-11-2017, 16:21
That's odd. It's definitely the language pack initialisation that's failing. Do you have the en-US or en-GB pack installed? I might have got the local string wrong - 'en' should just work but in some cases you might need to specific one which I think should be en-US or en-GB but it's worth trying similar stuff like en_GB etc.

bob
11-11-2017, 16:31
I have the en-US pack installed and the Canadian English pack installed (en-CA?).

When I try with an underscore in place of the dash I get this:

Attempting to initialise speech recognition for user specified location en_CA
Unable to initialise speech engine with voice recognition pack for location en_CA. Check that SpeechPlatformRuntime.msi and MSSpeech_SR_en_CA_TELE.msi are installed.
Exception message: Culture is not supported.
Parameter name: name
en_CA is an invalid culture identifier.
Attempted to initialise speech engine - success = False

I also tried en-CA, en_US, en-GB, and en_GB.

It's very confusing to me that it doesn't work, you would think someone else would have had this problem before if it was something wrong with the application.

mr_belowski
11-11-2017, 16:37
Lots of people have had similar issues but the solution has always been some unknown combination of uninstalling and installing language packs, and faffing with the locale string. The 'correct' config is to have only one language pack installed (en-US or en-GB) and to set the locale to 'en' (without the quotes)

bob
11-11-2017, 16:46
Okay, I tried uninstalling all but the US pack and no luck still with either en-US or en. I also tried uninstalling the US pack as well and only installing the GB pack and no luck there either with en or en-GB.

Interestingly, when I go to windows language options it shows that the language pack and speech pack are installed, but when I try to install the GB or CA packs through windows language options it fails to download.

Originally I got the speech packs from MS website, however.

bob
11-11-2017, 16:52
When I try to manually enter en-US in the config file rather than using the Properties window I get an error saying that I need to talk to the owner of the file in order to get permissions to edit it. Maybe that is the issue?

But it is saving the edit in Properties window on application restart, so idk.

mr_belowski
11-11-2017, 16:56
There's a download link in the apps help file to a version of the runtime and language packs that I know work. Might be worth uninstalling these components and installing them from those links

bob
11-11-2017, 17:02
Okay I did what you said and also installed the x86 installer rather than just the x64 and it works now!!

Thank you for the help!

mr_belowski
11-11-2017, 17:06
Yay :D

Blacktiger
29-11-2017, 10:56
Im having the same problem, I installed everything from the links in the apps help, the x64 runtime for my x64 system and the en-US package. But I still get the error.

12:55:58.835 : Attempting to initialise speech recognition for user specified location en-US
12:55:58.851 : Unable to initialise speech engine with voice recognition pack for location en-US. Check that SpeechPlatformRuntime.msi and MSSpeech_SR_en-US_TELE.msi are installed.
12:55:58.851 : Exception message: No recognizer of the required ID found.
12:55:58.851 : Attempted to initialise speech engine - success = False


I already tried uninstalling, reinstalling, with several reboots between that but it doesnt work. In my OS-Settings the runtime and the voice pack is listed in the Apps & Features, so its definetely installed...

EDIT: When I try to uninstall the runtime while crewchief is running it tells me to close CrewChief first, so crew chief is definetly using the runtime...

mr_belowski
29-11-2017, 12:17
and you've installed the runtime and language packs from the link in the CC help file?

Have you also tried resetting the app to default settings?

Blacktiger
29-11-2017, 13:25
yup did and tried everything already mentioned by bob and resetting and reinstalling and stuff like that.
Im also not dumb when it comes to computers, as I am a web developer (or something like that) but I simply cant find what the problem is / what the solution would be. cant do anything more than reinstalling 1000x times :/

Blacktiger
26-01-2018, 19:47
Do you have any news on this problem?

I tried reinstalling it today but it still doesnt work...

mr_belowski
26-01-2018, 19:55
Sorry fella, I honestly have no idea why these components won't work for you

Seth
02-02-2018, 04:35
I think i figured out why its failing. I was having the same issue. I was installing the en-US language pack and was getting that same error

12:55:58.835 : Attempting to initialise speech recognition for user specified location en-US
12:55:58.851 : Unable to initialise speech engine with voice recognition pack for location en-US. Check that SpeechPlatformRuntime.msi and MSSpeech_SR_en-US_TELE.msi are installed.
12:55:58.851 : Exception message: No recognizer of the required ID found.
12:55:58.851 : Attempted to initialise speech engine - success = False

So I decided to install the en-GB version too and it worked right away. So I think the trick is that you need the en-GB version installed no matter what. the US and GB versions can both be installed. It just doesnt work with only the US version installed. You might want to make that clear in the install instructions.

Hope this helps others!

Nocturne
02-02-2018, 15:58
Hi Guys,

question about language. I have a germin win10, installed th en_GB language pack. But ChrewChief is not able to understand me. I tried to set system default language tu english, which helped a little bit with the problem. Is the application using english voice by default? because when i am trying to set voice recc. to EN it will switch back to german because it is the default system language.

Do i have to create a new config file with german text in it or is there any way around that?

Nocturne
03-02-2018, 12:09
No ideas?

mr_belowski
03-02-2018, 12:55
the application will default to 'en' locale for speech recognition only if it fails to set the requested locale. Do you have the console output?

Nocturne
04-02-2018, 13:51
15:50:21.157 : Starting app
15:50:22.223 : Set Assetto Corsa (64 bit) mode from previous launch
15:50:22.287 : Minimum sound pack version = 105 using sound pack version 142, driver names version 129 and personalisations version 128
15:50:22.296 : Preparing driver names
15:50:22.324 : Prepare driver names completed
15:50:22.326 : Preparing sound effects
15:50:22.327 : Prepare sound effects completed
15:50:22.328 : Preparing personalisations for selected name Marcel
15:50:22.335 : Prepare personalisations completed
15:50:22.336 : Preparing voice messages
15:50:22.599 : Prepare voice message completed
15:50:22.600 : Finished preparing sounds cache, found 5792 driver names and 2541 sound sets. Loaded 347 message sounds with 340 active SoundPlayer objects
15:50:22.600 : 219 sounds have personalisations
15:50:22.626 : Reading driver name mappings
15:50:22.628 : Read 2014 driver name mappings
15:50:22.628 : Reading driver name mappings
15:50:22.628 : Read 422 driver name mappings
15:50:22.630 : Using spotter Geoffrey
15:50:22.630 : Spotter Geoffrey has oval-specific sounds - these will be used for tracks marked as 'oval'
15:50:22.648 : Loading controller settings
15:50:22.667 : Load controller settings complete
15:50:22.672 : Attempting to initialise speech recognition for user specified location en-GB
15:50:22.733 : Success
15:50:22.739 : Loading all voice command alternatives from speech_recognition_config.txt
15:50:22.857 : Attempted to initialise speech engine - success = True
15:50:23.059 : Checking for updates
15:50:23.153 : Took 0.55s to lazy load remaining message sounds, there are now 6642 loaded message sounds with 340 active SoundPlayer objects
15:50:23.510 : Got update data from primary URL: http://thecrewchief.org/
15:50:23.515 : Check for updates completed

looks good

mr_belowski
04-02-2018, 14:54
Yes, it's definitely using the en-GB locale and language pack for speech recognition there. If it doesn't recognize anything, are you sure it's not your microphone?

Nocturne
05-02-2018, 10:51
Some commands work, some others won´t... Mic seems okay, using the rift micro

The Iron Wolf
05-02-2018, 14:39
You may want to tweak recognition confidence in preferences. After you use CC voice commands, save the console output and search for word "confidence". You'll see what confidence CC recognizes commands with. If your confidence value in preferences is above that value, command is not recognized, and Chief will ask you to say again.

I've really poor results with rift mike compared to dedicated microphone, but use it anyway. I personally use confidence of 0.34 (down from default 0.5), and I also noticed that AGC hurts recognition (in Rift mike settings). Lastly, mike volume and external noise (speakers, open headphones) affects recognition too. You'll need to experiment to fine tune it for your environment.

This is becoming a FAQ, I'll add an entry.