[RELEASED] Unity Google Cloud Speech Recognition [VR\AR\Mobile\Desktop]

hello I just update Google Cloud Machine Learning Kit, the GCSpeechRecognition seems not working, yes it detect the audio but did not show the text , not error or any, the old version still working

 

Hello,

this example you showed in screenshot mostly for comamnds detection.
does it work in another example?

Best

 

Hello,

next update of microphone webgl asset will directly work with GC spech recognition with just importing and deleting one file.

Next update will come soon.

alpha package I'll send you in discord as discussed.

Best

 

Hello.

if you have any questions - fill free to ask us directly in our Official Discord Server
There you will see announces, tutorials and our community that could help you.

Best Regards

 

Hi
I purchased "Google Cloud Translation" and "Google Cloud Speech Recognition"
Try to use both of them in the one project but I got the class conflict under the “_Generic”
Can you let me know how to fix these error?

Also I email frostweep@gmail.com
but none write me back

 

Hello,

looks like issues fixed.

Best

 

please help how can i increase the microphone sensitivity?
I think that should be responsible for this the property "Audio Volume Multiplier" in "GCSpeechRecognitonConfig", but it is not using in code

 

@FrostweepGames I am working on an IOS app that relies on the continuous use of voice input. When using this plugin i notice that voice detection is causing a large drag on the system as it runs every frame. I am no audio engineer but i was wondering if there was a better way to check if the user is talking that would be less of a drag on my system?

Code (CSharp):

1.  
2. if (DetectVoice)
3.                 {
4.                     bool isTalking = _voiceDetectionManager.HasDetectedVoice(AudioClip2ByteConverter.FloatToByte(_currentAudioSamples));
5.  

 

Hello,

there is no way tooptimize code but you could try to record less size of audio chunk.

I'm not sure that this function is targeted for mobile but I'm sure it works well on phones with good cpu.

Currently we havent any possibility to optimize it, but we will work on it in future.

Join our discord to get news faster. Or ask other customers about anything!

Best Regards

 

Thanks for the reply. We are targeting IOS. Using iPhone 11 Pro Max's. We're using this in conjunction with a web rtc system. When profiling and the phone is lagging it is showing that a majority of the usage is coming from that one line of code checking if the user is speaking. AudioClip2ByteConverter.

 

Hello,

hmm iphone 11 is quite enough for it..

how long you record? does it start lag directly during firsst record?

not sure that there a way to optimize it now and I would recommend to not use it till we find better solution for detecting voice level.
conenct to our discrod server to get news faster!
Best

 

The recording goes anywhere from 5 to 30s depending upon the speaker. We are using the AutoDetectVoice feature and running this as a sort of transcription for a conference call.

Can you provide me details to the discord channel?

 

Perhaps the solution is changing the frequency that it checks for the user talking.

 

Hello,

https://discord.gg/TZdhnWy

 

Hello!

We are happy to say to You that the first release of Google Cloud Streaming Speech Recognition is now live! check it out right now at http://u3d.as/18RU

Thanks for your support.

Best Regards

 

Directly recognise works perfect for us but our code is in combination to detect voice and then directly recognise is on by default. Results with runtime are not accurate. How can we edit the code to have just directly recognise and give a timeframe to user to record and then detect and return. Also we want to check for the right answer in alternates detected how to do that?

 

Hello,

you could nodify MediaManager to extend time to end talking event which will fire when user stop talking.

To check alternatives you could look at Example script which includes sample of how to do that.

Best Regards

 

Hello everyone.

I'm currently testing this asset, and I'm having trouble making the GCSR_Example scene work on Android. It is working perfectly fine on Desktop. When testing on my Google Pixel 2 XL, however, it freezes to a black screen right after the Unity logo appears. Through testing, I've identified that it has something to do with line 124 in GCSpeechRecognition.cs:

Code (CSharp):

1. ServiceLocator.Instance.Update();

Has anyone here encountered the same problem and found a solution?

 

Hello, I love the asset and have been using it for years!
Is there a plan to release an update for Cloud Speech-to-Text V2 API?

 

Hello,

yes, we plan to do that

Best Regards

 

Hi, in our game we have audio files that we want players to exactly pronounce. Would this plugin work in our case?