-
-
Notifications
You must be signed in to change notification settings - Fork 80
Azure Speech Service
This video just runs through the instructions on this page.
- For Speech Recognition and TTS to work you must have an Azure Subscription Key.
- Option 1: Free Azure Account
- Completely free for the first month. After first month you will be asked to upgrade your account to "pay as you go" but still have access to free monthly limits

- Option 2: Free Azure Account for students (no credit card required)
- Completely free, have to renew every year

-
After making your account you will need to create a speech service to get your Key and Region. You will enter this information in in the "Microsoft Azure Cognitive Service" tab located in "Settings"
-
Follow this video to get your key and region information:

-
The pricing tier for your speech service in azure should be set to F0 Free if you wish to take advantage of azures free monthly limits and not be charged

-
I am not responsible for any charges you receive if you upgrade from a Free Azure Account and use S0 Standard pricing! It is up to you to monitor your own usage if you are using a pay-as-you-go azure account
-
Spending Limits
-
Avoid Charges
- Azure Speech Service Pricing and Free Monthly Limits (This program uses Speech to Text Standard, Text to Speech Neural, and Speech Translation Standard)
- Monitor Usage (the location to see your speech service usage can be hard to find, this post should help!)
- Budgets and Alerts
-
Spending Limits
-
-
Your key and region go in the "Microsoft Azure Cognitive Service" tab located in the "Speech Provider" tab
- Make sure to click the change button for both key and region

- Make sure to click the change button for both key and region
- Many azure voices have options for selecting Speaking Styles
- These can drastically change the sounds of voices (try them out)
- Spoken Language is the language that you speak natively
-
Translation Language is the language that you wish to translate to.
- It should be set to No Translation (Default) when not in use
- Speak to text hours and translation hours are separate you get 5 hours each
- Pro Tip: You can technically use all your 5 for your free monthly speak to text hours and then use Translation Language set to your Spoken Language for an extra 5 extra hours
- Profanity filter is on by default turn it off in azure settings.
-
Dictionary takes advantage of Azure's Phrase List feature to allow users to add new words to be recognized.
- For instance, it can be used for words like "Pogchamp" or user names that Azure wouldn't know otherwise
- Separate different words or phrases with commas.
- Phrase List Example:
VRChat, Sippbox, Poiyomi, Pogchamp, Suss E Baka
- Continuous Recognition (Azure) allows a user to continuously speak and have their words transcribed without constantly pressing the speech to text button
- WARNING: You will quickly use up your free azure limit with this feature enabled.
