Audio Annotator- Recording Malay in Malaysia

About WizWiz AI

At Wiz AI we believe the future of artificial intelligence is built on more than powerful algorithms. It is built on the quality context and expertise of the people behind the data.

We connect human intelligence with AI technology to help businesses build smarter more reliable and more capable AI systems. From data annotation and evaluation to localization content expertise and AI model improvement we provide the human expertise needed to transform raw data into meaningful intelligence.

Task description

We are currently recruiting local Malaysian participants for a Malay speech recording project.

Participant Requirements:

• Participants must be local Malaysians.

Recording Requirements:

• Recordings must be made in a quiet enclosed indoor environment.
• The background noise level must be below -40.8 dBFS with no echo reverberation or other abnormal sources of interference.
• Participants should maintain a distance of approximately 515 cm from the microphone during recording and the speech amplitude should remain within the smpl range.

• The overall SNR must be above 20 dB.
  
  

Recording Process:

• After claiming the task participants should follow the instructions provided in the task to complete the consent form and fill in their personal information.
• Participants should read the assigned sentences naturally clearly and completely while trying to ensure that the wording and expression are consistent with local Malaysian language usage.
• To make the recordings sound more natural participants may adjust the wording slightly according to their usual speaking habits when appropriate.

Rate: $20 per valid audio hour
Note: Payment is not calculated based on the actual operating time of the recording equipment but on the actual valid audio duration of the final submitted recordings. Only audio that meets all recording specifications and passes the quality check will be counted toward the payable duration. No payment will be made for rejected or invalid audio segments. Any audio that does not meet the requirements must be re-recorded.