Speech Translator

This app is speech translator and recorder using Azure AI Speech.

Prerequisites

• .NET 10.0 SDK

How to use

Console app

1. Create Azure AI Speech resource. (Bicep)
2. Copy Subscription Key and Region from Azure Portal.
3. Clone this repository.
4. Create src/SpeechTranslatorConsole/appsettings.Development.json.
5. Setup appsettings.Development.json like below.

   {
       "Settings": {
           "Region": "<Region>",
           "SubscriptionKey": "<Subscription Key>"
       }
   }

6. Set the microphone device for translation as the default input device.
7. Run SpeechTranslatorConsole project. (dotnet run --project src/SpeechTranslatorConsole)
8. Type file name in console.

Desktop app (WPF)

1. Create Azure AI Speech resource. (Bicep)
2. Set the microphone device for translation as the default input device.
3. Run the desktop project.

   dotnet run --project src/SpeechTranslatorDesktop

4. In the Azure AI Service 設定 section, enter the Azure AI Speech Region and API Key, then click 保存.
   • Settings are stored in SQLite at %LocalAppData%\SpeechTranslatorDesktop\speech-translator-desktop.db.
   • The API key is protected with Windows DPAPI before it is written to SQLite.
   • Saved settings are automatically loaded the next time the app starts.
5. If you do not save settings in the app, the desktop app falls back to the SPEECH_REGION and SPEECH_KEY environment variables.

   $env:SPEECH_REGION="japaneast"
   $env:SPEECH_KEY="your-speech-key"

6. Select the speaker language and target language.
7. Optionally enter a simple recording file name stem (letters/numbers/-/_, no path or extension). If empty, no file is saved.
8. Click 開始 to start and 停止 to stop.

Recording files are saved as UTF-8 text files under recordings/ relative to the desktop app executable directory.

References

• https://learn.microsoft.com/en-us/azure/ai-services/speech-service/language-identification
• https://learn.microsoft.com/en-us/azure/ai-services/speech-service/how-to-translate-speech

License

see LICENSE