Phonics Exercise Audio Dataset

The Phonics Exercise Audio Dataset is a custom-built audio corpus developed to support speech-based learning and pronunciation correction. It currently consists of 198 audio samples that represent correct and incorrect pronunciations of Arabic letters from "أ" to "ت". These samples are recorded in a standardized format to support consistent feature extraction and model training. The dataset enables the development of phoneme classification systems aimed at improving auditory discrimination skills in dyslexic learners. Ongoing data collection efforts seek to expand this resource to include more letters and speakers, increasing the system’s effectiveness in varied learning contexts.