25,000 Modern Standard Arabic Speech Clips with Full Diacritization

25,000 fully-diacritized Modern Standard Arabic text and audio pairs synthesized by a single Saudi male neural voice. The dataset was created by HeshamHaroon and was last updated on April 20, 2026. Audio clips are rendered at 48 kHz / 16-bit PCM and are organized across 10 thematic categories.

Use Cases

• Training Arabic speech recognition models based on the 25,000 high-quality audio clips.
• Developing or fine-tuning text-to-speech systems based on the neural voice synthesis and diacritized text pairs.
• Linguistic analysis of Modern Standard Arabic based on the fully-diacritized (tashkeel) text corpus.
• Building pronunciation dictionaries or grapheme-to-phoneme models based on the aligned text and audio.

Strengths

• 25,000 audio-text pairs provide a substantial corpus for model training.
• Full diacritization (tashkeel) of the text supports precise phonetic and linguistic analysis.
• High-quality 48 kHz / 16-bit PCM audio synthesis from a consistent neural voice (ar-SA-HamedNeural).
• Data is categorized into 10 thematic groups, which may aid in balanced sampling or topic-based tasks.

Limitations

• Column-level documentation is absent; field semantics must be inferred after download.
• The dataset consists of synthetic speech from a single voice, which may not capture natural speech variations or accents.
• Freshness should be verified as the last update date is in the future (2026-04-20).

Provenance

Source

HuggingFace user HeshamHaroon, synthesized using Azure Neural TTS.

Collection Method

Synthesized by the ar-SA-HamedNeural voice (Azure Neural TTS).

Time Range

null

Freshness

Last updated 2026-04-20 05:14:04

Geography

Saudi Arabic voice suggests a Saudi linguistic focus, but the text is Modern Standard Arabic.