Amharic Scripted Speech Dataset
Budget / Salary$750–1,500
TypeFreelance project
LocationRemote
Posted1 hour ago
I am compiling a clean, high-quality Amharic speech corpus for automatic–speech-recognition research. Your role is simple: record yourself reading a set of Amharic sentences that I will provide and return the audio files together with the exact written sentences. No time-marked transcript or speaker metadata is required—only the audio and its matching text.
Deliverables
• One WAV file per sentence, 16-bit / 44.1 kHz (or higher)
• A plain-text file that lists each sentence exactly as spoken, in order
• File names that match the sentence IDs supplied in the script
Recording guidelines
• Quiet environment, minimal echo, no background noise
• Consistent microphone position and volume
• Native or near-native Amharic pronunciation
Please mention any prior experience with speech data collection tools (Audacity, Praat, etc.) and how many minutes of usable audio you can provide within one week. I will review a short sample first; once the quality is approved, we will proceed with the full batch.
Deliverables
• One WAV file per sentence, 16-bit / 44.1 kHz (or higher)
• A plain-text file that lists each sentence exactly as spoken, in order
• File names that match the sentence IDs supplied in the script
Recording guidelines
• Quiet environment, minimal echo, no background noise
• Consistent microphone position and volume
• Native or near-native Amharic pronunciation
Please mention any prior experience with speech data collection tools (Audacity, Praat, etc.) and how many minutes of usable audio you can provide within one week. I will review a short sample first; once the quality is approved, we will proceed with the full batch.
Apply on Freelancer →
Project sourced from Freelancer.com. Applications happen directly on the original platform — we never collect your data.