Hroof Arabic Alphabet Game Study Guide

Student Details: Huda Ayman (Student ID: $3456646$ ).
Supervision: Supervised by Mr. William, March $2026$ .
Academic Submission: Dissertation submitted for the degree of Bachelor of Science with Honours Software Engineering at the University Of Stirling.
Document Statistics: Word count of $10,095$ words.

Background of the Arabic Language: * Arabic is a widely spoken global language. * The alphabet consists of $28$ letters written from right to left. * Complexity arises because letter shapes change depending on their position within a word (initial, medial, or final forms).
Challenges in Traditional Learning: * Traditional methods often rely on repetition, memorization, and textbook exercises, which can be boring for young learners. * The Arabic script and pronunciation rules are inherently difficult for beginners and non-native speakers.
The Role of Technology: * Technology provides interactive systems like the Hroof game to make learning enjoyable and engaging. * Active participation, pictures, and sounds enhance information retention and understanding, particularly for children.
Project Goal: To develop a version of the Hroof game for children aged $7$ to $12$ that uses pictures, sounds, and speech recognition to teach the alphabet.

Problem Statement: * There is a lack of simple, focused digital tools for early-stage Arabic alphabet learning. * Existing applications are often too complex, lack interactivity, or do not provide pronunciation support. * Limited use of speech recognition restricts speaking practice opportunities.
Aims and Objectives: * Design a user-friendly graphical interface for young learners. * Implement a level-based system for structured learning. * Integrate audio playback for correct letter pronunciation. * Implement speech recognition for speaking practice. * Evaluate system effectiveness and usability.

Game-Based Learning: This approach provenly enhances motivation and engagement by integrating content into environments with levels, rewards, and challenges.
Comparative Analysis: Platforms like Duolingo and Rosetta Stone use repetition and feedback but focus on widely spoken languages like English or Spanish, with limited emphasis on specialized Arabic alphabet learning for kids.
Modality of Learning: Combining visual (pattern recognition), auditory (pronunciation), and kinesthetic (active participation) approaches caters to different learning styles.
Speech Recognition in Education: Allows for immediate feedback on pronunciation, which is difficult to assess via traditional methods, though accurate recognition can be hindered by background noise or accents.

Modular Architecture: The system uses a modular design where components communicate through defined interfaces for better maintainability and scalability.
Technological Integration: * Java: Used to manage the Graphical User Interface (GUI), level logic, and user interactions. * Python: Used for complex tasks like audio playback and speech recognition. * MySQL: Used for user account management and progress tracking.
Key Interface Screens: * Home Screen: Entry point with Sign In and Sign Up options. * Map Screen: Displays levels in a grid layout; each level represents an Arabic letter and is unlocked progressively. * Game Screen: Displays the specific letter and features "Listen" and "Speak" buttons.
Data Flow: 1. User selects a level. 2. Java retrieves the Arabic letter from the Levels class. 3. User interacts (Listen/Speak). 4. Request sent to Python via ProcessBuilder. 5. Python processes audio and returns recognized text. 6. Java compares text to expected letter and provides feedback.

Java GUI (Swing): Uses components like JFrame, JPanel, JButton, and JLabel. Layouts are managed via BorderLayout and GridLayout.
Java-Python Communication: Achieved through the ProcessBuilder class. * The system executes Python as an external process. * Standard output is captured via a BufferedReader. * Final communication used a file-based approach (result.txt) to avoid pipe buffering issues.
Speech Recognition Workflow (Listen.py): * Utilizes the SpeechRecognition library and Google Web Speech API. * Uses the sounddevice library for audio capture to avoid driver conflicts. * Includes a normalization process to handle letter variants (e.g., "أ", "إ", "آ", "ا").
Class Structure: * Game: Main controller for navigation and interaction. * Levels: Stores alphabet data and mappings to sound files. * Buttons: Handles reusable UI components. * Home: Manages the initial login/registration interface.

Database Table (ArabicGameUser): * id: unique user ID ( $int$ ). * Full_name: user's full name ( $Varchar$ ). * Age: user's age ( $Int$ ). * Country: user's country ( $Varchar$ ). * Username: unique login ID ( $Varchar$ ). * Password: user password ( $Varchar$ ). * Last_Level: tracker for completed levels ( $int$ ). * score: user's score ( $int$ ).
Input Validation Constraints: * Age must be between $5$ and $50$ . * Country field must contain valid text (no numbers/symbols). * Usernames must be unique (checked against database before registration).

Speech Recognition Timeout: * Issue: Consistently returned "Listening timed out" errors. * Solution: Corrected microphone device selection (configuring specifically for "Intel SST Microphone Array"), lowered energy threshold (default $300$ was too high), and increased recording duration from $3$ to $4$ seconds.
PyAudio WASAPI Subprocess Error: * Issue: PyAudio failed with [Errno -9999] Unanticipated host error when run as a Java child process due to Windows blocking microphone access. * Solution: Replaced sr.Microphone with the sounddevice library. Audio is recorded, saved as a temporary WAV file, and then processed via sr.AudioFile.
Python Compatibility: * Python version $3.10$ was used specifically because newer versions had installation issues with PyAudio.

Usability: User testing indicated the interface is intuitive for the target age group ( $7$ - $12$ ). Large buttons and cloud-based design were effective.
Functionality: Real-time communication between Java and Python was successful. Audio playback was a highly praised feature.
Performance: The system runs smoothly on standard hardware with minimal delay.
Success Metric: Recognition reliability increased significantly after implementing noise calibration and switching to the sounddevice approach.
Limitations: * Dependency on external Python scripts and Google Web Speech API (internet required). * Lack of advanced feedback (hints/explanations) and advanced gameplay (animations/scoring). * Sensitivity to environmental noise levels and individual pronunciation variations.

IDEs & Languages: Eclipse, Python $3.10$ .
Libraries: SpeechRecognition, PyAudio, sounddevice, SciPy, MySQL Connector/J.
APIs: Google Web Speech API.