AlgoRythmetic/hufazalmatoun

Huffaz Al Matoun Dataset

A high-quality Arabic speech dataset containing 98,256 audio clips of classical Islamic texts recited by various narrators.

Dataset Description

This dataset was extracted from the "Huffaz Al Matoun" (حفاظ المتون) application, containing audio recordings of classical Arabic Islamic texts with their corresponding transcriptions.

Features

Feature	Type	Description
audio	Audio	MP3 audio file of the recitation
text	string	Arabic transcription of the audio
speaker_id	string	Name of the narrator/reciter
matn_name	string	Name of the Islamic text being recited
audio_id	string	Unique identifier for the audio clip
duration_sec	float	Duration of the audio in seconds
channel	string	Source channel ("huffaz al matoun")
gender	string	Speaker gender ("male")
dialect	string	Arabic dialect ("classic")

Statistics

• Total Samples: 98,256
• Total Speakers: 296 different narrators
• Language: Classical Arabic
• Audio Format: MP3

Usage

from datasets import load_dataset

dataset = load_dataset("AlgoRythmetic/hufazalmatoun")

# Access a sample
sample = dataset["train"][0]
print(sample["text"])
print(sample["speaker_id"])
print(sample["matn_name"])

License

This dataset is released under the CC-BY-4.0 license.