AndroidSpeechToText is a project demonstrating the integration and usage of Google's native speech recognition feature in Android applications. This project was developed using Delphi 12 and tested on Android 13.
-
Android Dependency: This project relies on native Android APIs for speech recognition and is specific to the Android platform. It will not work on other platforms such as iOS or Windows.
-
Online and Offline Speech Recognition: The project utilizes Google's speech recognition service, which can operate both online and offline. However, for offline speech recognition, it's necessary to pre-download the appropriate language packages on your device.
-
Language Compatibility: The language for speech recognition is set to Brazilian Portuguese (pt-BR). Ensure that the selected language is supported by Google's speech recognition API.
-
Event Subscription: The class utilizes TMessageManager to subscribe and unsubscribe from messages. Ensure compatibility with other parts of your application using TMessageManager.
-
Testing and Adjustments: Before integrating this class into a larger project, it's recommended to test it on various Android devices and usage scenarios to ensure proper functionality.
To use this project, simply include the FMX.SpeechToText unit in your Delphi project and call the StartRecognition method to initiate speech recognition.
// Example usage
TSpeech_Text.StartRecognition('Speak Now', ResultSpeech);For additional details about Android speech recognition APIs, please consult the official Android documentation.
Demonstration.mp4
This project is designed for Android devices and uses native Android features for speech recognition. Requires audio recording permission in your app "android.permission.RECORD_AUDIO".