Saudi Arabic Spontaneous Speech Sample with 849 Hours

A collection of 849 hours of Saudi Arabic spontaneous speech audio covering multiple topics. All audio is manually transcribed into text, with annotations for speaker identity and gender. It was created by Nexdata and last updated in April 2025.

Use Cases

• Train an automatic speech recognition model on 849 hours of manually transcribed Saudi Arabic speech audio.
• Build a voiceprint recognition model using annotated speaker identity and gender attributes.
• Construct a corpus for machine translation research from the transcribed text content of spontaneous speech.

Strengths

• 849 hours of manually transcribed speech audio provides a substantial volume for model training.
• Annotations include speaker identity and gender, enabling speaker-specific or demographic analyses.
• Speech data is spontaneous and covers multiple topics, increasing linguistic diversity.

Limitations

• The dataset is described as a sample of a larger paid dataset, so the full data is not available.
• The sample size in terms of distinct audio files or speakers is unknown.
• No information is provided on audio quality, recording conditions, or potential background noise.

Provenance

Source

Nexdata

Collection Method

Data collection and manual transcription method is unspecified.

Freshness

Last updated on 2025-04-24.

Geography

Saudi Arabia (inferred from language and dataset title).