Use the online Google Speech Recognition Service (test beta)
How to add the Google Speech Recognition robot skill
- Load the most recent release of ARC (Get ARC).
- Press the Project tab from the top menu bar in ARC.
- Press Add Robot Skill from the button ribbon bar in ARC.
- Choose the Audio category tab.
- Press the Google Speech Recognition icon to add the robot skill to your project.
Don't have a robot yet?
Follow the Getting Started Guide to build a robot and use the Google Speech Recognition robot skill.
How to use the Google Speech Recognition robot skill
This skill is a test for the Google Speech Service with ARC. There is some configuration necessary for testing this skill, as the Google Speech Service is paid and not free. There is a 60 day trial, which you can sign up for. This skill requires internet access, so if an I/O controller is also being used a second WiFi connection, wired network or direct UART connection to the PC is needed.
Main Window
1. Load Credentials Button This opens the location of your Google speech recognition key (.json) file. Browse to find it's location.
2. Start Recording Button This activates Google speech recognition which will then listen for words and respond accordingly.
3. Visual Feedback Waveform This visually displays the peaks and valleys of the detected frequency. It also shows you if your default recording device is working properly.
4. Response Field This is where Google speech recognition will respond with its text-based response.
Configuration
1. Response Script A response script can be activated when certain words are detected.
2. Variable Name Here you can change the variable name for the detected words.
How to Use the Google Speech Recognition Skill
If you wish to use this skill with PandoraBot, you can send the detected speech to the PandoraBot control using the ControlCommand(). For example, you will add both a PandoraBot control and this Google Speech skill to your project.Don't forget to PAUSE the PandoraBot skill so it doesn't attempt to pickup any audio! We will only be using the Google Speech skill for audio, so the PandoraBot skill can be PAUSED. It will receive speech phrases by the ControlCommand() from the Google Speech response script.
Your project would resemble this...
Now you can edit the Google Speech skill and add this code to the response script which will send the detected speech to the PandoraBot:
ControlCommand("PandoraBot", SetPhrase, $GoogleSpeech)
The configuration would look like the following image. This will send all received data from the Google Speech skill to the PandoraBot skill.
Configuring Google Services For Credentials These instructions only provide some assistance. Google has their own support if you need assistance Generating a JSON Service Account Key. We have provided some information below.
*Disclaimer: Google's web config is an absolute frustrating experience with outdated documentation and hidden menu options. It's easy to get lost. Venture forward at your own risk
To start, visit https://cloud.google.com/speech/ and select Free Trial
From there, you will need to create credentials
Configure the Google Cloud Speech API when adding credentials
Now create a new credential to get a Service Account Key
Select the Service Account that you had already created earlier and choose JSON
A JSON file will download automatically. This is the file that you will load in this skill for authentication before it will work. So save it somewhere useful that you can find in the future. Your Downloads Folder is probably not the best place...
Enable Billing You will need to enable billing, if you get an error that billing is not enabled. Here is assistance link from google: https://support.google.com/cloud/answer/6293499?hl=en
A credit card is necessary for Google's service. However, they won't charge you after the 60 day trial.
Hi,
I am using "Google Speech Recognition" It works fine. But I couldn't find a way to Start and Stop "Recording" from the code. How this can be done programmatically.
It looks like I need to press "Start Recording " / "Stop Recording " manually.
Is there any way to manage this from the script for "Google Speech Recognition)
What does the Cheat Sheet say? I think there's a PAUSE command. Take a look, otherwise I can add something.
I get an error in EZB when loading my .json credentials which begins 'please check the value of the environmental variable GOOGLE_APPLICATION_CREDENTIALS' could Google have changed the format of their API key files?
Sure could have - I’ll take a look
hi. so when i load my credentials into the project it says this: "This file must be in the root folder of this plugin. Also, if it is already in the same folder, the DLL filename and the assembly name may be different. To correct this error, either ensure the file is in the root folder of the plugin, or rename the DLL to the assembly name as presented here in this error."
can you help?
I am also getting this error message "This file must be in the root folder of this plugin. Also, if it is already in the same folder, the DLL filename and the assembly name may be different. To correct this error, either ensure the file is in the root folder of the plugin, or rename the DLL to the assembly name as presented here in this error." when attempting to set up this function.
It appears the Google library is looking for unity. I’ll look into why that is and resolve it. Stay tuned
I did not find the cost for using the service from Google. Tell me how much does it cost?
they also have a well-tuned voice, many languages with good quality. The price is also unknown. Here is a link to listen to: https://cloud.google.com/text-to-speech
Was the unity error ever resolved? I'm running into this issue now. When i load the json key it throws this error: System.TypeInitializationException: The type initializer for 'Grpc.Core.Internal.PlatformApis' threw an exception. ---> System.IO.FileLoadException: Could not load file or assembly 'UnityEngine' or one of its dependencies. General Exception (Exception from HRESULT: 0x80131500) ---> System.Exception: Unable to find a DLL: C:\ProgramData\ARC\Plugins\ac914bee-8fc2-4fbd-b2b5-4661940e3b23\UnityEngine.dll.
This file must be in the root folder of this plugin. Also, if it is already in the same folder, the DLL filename and the assembly name may be different. To correct this error, either ensure the file is in the root folder of the plugin, or rename the DLL to the assembly name as presented here in this error.
I think this robot skill doesn't work anymore. Doesn't it?
Yes, the Google speech recognition robot skill still works. If you experience issues, some robot skill developers use advanced features of the ARC Pro version, but that is rarely the case. You can try uninstalling ARC Free, downloading a new copy of ARC Free, and reinstalling it to ensure you have a valid installation of ARC Free. Otherwise, there are several other robot skills for speech recognition.
For a list of other robot skills that provide speech recognition capability, check here: https://synthiam.com/Support/Skills/Skills-Overview#Audio