SpeechCore

SpeechCore is a cross-platform C++ library that abstracts the process of communicating with various screen readers, managing low-level details and providing a clean, simple-to-use interface.

Features

• Simple and intuitive API with straightforward usage
• Cross-platform support for various screen readers, with the option to support additional ones

Installation

1. Download the matching library build for your platform from the releases page.
2. Link against the library (static or shared, depending on your choice).
3. Copy the contents of the include folder to your project directory.
4. Include SpeechCore.h in your project.

Building

The GitHub repo includes a Visual Studio solution and a SConstruct file. If you're using a different build system, keep the following notes in mind:

• Define either __SPEECH_C_EXPORT (shared) or SPEECH_C_STATIC (static) when compiling.
• JNI files are included for Java support. Ensure you have the required Java files, or exclude them if not needed.
• Include only platform-specific files for your target platform.
• Linux builds require the Speech Dispatcher library.
• Windows builds need to link against SAPI.LIB.
• Macos builds need to link against object library. And the AVFoundation and Foundation frameworks.
• Documentation generation requires Doxygen and Sphinx.

Usage

Simple usage example:

#include <iostream>
#include <SpeechCore.h>

int main() {
    Speech_Init();
    if (Speech_Is_Loaded()) {
        std::cout << "Current screen reader: " << Speech_Get_Current_Driver() << std::endl;
        std::cout << "Speaking some text" << std::endl;
        Speech_Output(L"This is a test for the SpeechCore library. If you're hearing this, it indicates the library is functioning properly.");
    }

    Speech_Free(); // Free resources when you're done.
    return 0;
}

See the documentation for more detailed usage examples.

Documentation

Documentation can be found here.

Library Notes

• Windows screen readers (NVDA, JAWS, Zhengdu, PCTalker, and System Access) require specific binaries, included with the source code.
• Enhanced control over SAPI 5, including voice configuration and speech parameters.
• Braille output functionality is included and implemented for the screen readers that support it. Mainly NVDA and Jaws

Supported Screen Readers

• Windows: NVDA, JAWS, System Access, Zhengdu Screen Reader, PCTalker, SAPI 5
• macOS: AVSpeech
• Linux: Speech Dispatcher

Bindings

Bindings are available for Python, .NET/C#, and Java.

For Python Install python bindings via pip:

pip install SpeechCore

Dotnet Install via NuGet Package Manager:

dotnet add package SpeechCore.CrossPlatform

Or via Package Manager Console in Visual Studio:

Install-Package SpeechCore.CrossPlatform

Credits

This library was inspired by Tolk, with adaptations for more flexibility. It was initially developed for personal projects and later expanded to be cross-platform.

Change Log

Version 1.1.1

• The Speech_Detect_Driver now rescans for drivers on Windows even if one is currently running.

Version 1.1.0

• Fixed NVDA speech interrupt functionality.
• New functions available on api:
  • Speech_Set_Pitch and Speech_Get_Pitch. For controling pitch parameter for drivers that support it.
  • Speech_Output_Text. An enhanced speech function which includes an with_ssml parameter to use if the speech driver supports it. The SC_SSML_SUPPORT an be used to check if it does.

Version 1.0.2

• Added PCTalker screen reader support.
• Implemented braille functionality for screen readers that support it (NVDA, jaws)
• Modified Encoding handling for Unix platforms, it should function properly now.
• Python bindings have been rewritten with pybind11 an is now a a part of the main repo.
• Dotnet bindings are now available as a nuget package.

Contributions

Contributions to the library or support for additional screen readers are welcome.

License

MIT License

Todo List

• Implement Android/iOS support
• Add Meson build files
• Add CMake build files