Riwayat Identification Audio Dataset: Four Canonical Quranic Recitation

This dataset contains audio samples prepared for Riwayat identification in the Holy Qur’an. It includes four canonical Riwayat recited by Sheikh Mahmoud Khalil Al-Husary, who has completed full recordings of the Quran in several authentic recitation styles. The audio was collected from ten different Surahs and segmented into fixed 20‑second clips. The dataset contains 1400 audio segments in total, with 350 samples for each Qira’at. All recordings originate from the same reciter to ensure consistent voice characteristics and to focus on differences in recitation rules. To facilitate machine learning and classification tasks, all audio segments were converted into MFCC images, which represent the spectral features of each clip. These MFCC images form the final dataset used for Qira’at recognition experiments. This dataset supports research in Quranic audio processing, machine learning, deep learning and automatic identification of Riwayat.