250-Hour Saudi Arabic Recordings
Budget: $250 – $750 USD
I’m assembling a 250-hour corpus of Saudi Arabic narration to train a Generative AI model, and I want every minute captured by native speakers of the three key dialects in the Kingdom—Najdi, Khaleeji, and Hijazi.
The recordings must sound clear, natural, and studio-quality. Please deliver clean 48 kHz / 16-bit WAV files with consistent volume, no background noise, and no processing other than gentle normalization. Simple, conversational narration is all that’s required; no dialogues, interviews, or advertising reads.
To keep the dataset balanced, I’d like roughly equal coverage of each dialect. You may record alone or coordinate a small team, as long as every speaker is a genuine native of the dialect they read in.
Deliverables
• 250 total hours of narrated audio, evenly split across Najdi, Khaleeji, and Hijazi
• A spreadsheet listing file names, length, speaker dialect, and a one-line description of the content for quick reference
• All raw takes as separate files plus a “final” trimmed version for each segment
I will review a five-minute sample from each dialect before we move on to full production to be sure the audio chain and accent are spot-on. Once approved, we can break the work into manageable milestones and keep progressing until the full 250 hours are complete.
The recordings must sound clear, natural, and studio-quality. Please deliver clean 48 kHz / 16-bit WAV files with consistent volume, no background noise, and no processing other than gentle normalization. Simple, conversational narration is all that’s required; no dialogues, interviews, or advertising reads.
To keep the dataset balanced, I’d like roughly equal coverage of each dialect. You may record alone or coordinate a small team, as long as every speaker is a genuine native of the dialect they read in.
Deliverables
• 250 total hours of narrated audio, evenly split across Najdi, Khaleeji, and Hijazi
• A spreadsheet listing file names, length, speaker dialect, and a one-line description of the content for quick reference
• All raw takes as separate files plus a “final” trimmed version for each segment
I will review a five-minute sample from each dialect before we move on to full production to be sure the audio chain and accent are spot-on. Once approved, we can break the work into manageable milestones and keep progressing until the full 250 hours are complete.