Short Command Speech Dataset

via Freelancer ·

Budget / Salary₹12,500–37,500
TypeFreelance project
LocationRemote
Posted36 minutes ago
I need a clean, well-organized corpus of short spoken commands. Each command should be recorded as an individual audio file, then paired with a matching annotation file that states exactly what was said. Please keep background noise to an absolute minimum and normalise levels so every clip is ready for downstream ASR or intent-detection work.

Deliverables
• One folder of audio clips (WAV or high-quality lossless equivalent) – each file contains a single spoken command.
• A companion CSV or JSON file with, at minimum, the filename and the verbatim transcription of that command. Feel free to include extra metadata (speaker ID, environment notes, etc.) if you are already capturing it.

I will validate the set by running a small ASR benchmark; acceptance requires at least 99 % alignment between the provided transcription and the recognised text.
audio services transcription english (us) translator audio production audio editing audio engineering data annotation speech recognition
Apply on Freelancer →

Project sourced from Freelancer.com. Applications happen directly on the original platform — we never collect your data.