Voice data collecting for english native (USA) - Montgomery (Alabama)
$30-80 USD
Closed
Posted over 2 years ago
$30-80 USD
Paid on delivery
Hi everybody!
We collect audio data to develop voice recognition software (like Siri, or Alexa).
We’re currently looking for native american people (from USA) to help developing voice recognition software.
The task need to be done by 2 native people with a smartphone working together in the same room. It takes 2 hours and it consists into record free conversation about different arguments.
The job is paid 80USD per couple (20USD/person/hour).
Parecipants requirements:
1) The recorder must be native American and speaks American accent (USA).
2) Speak naturally, fluently and clearly. Do not read a script or deliberately simulate others tune.
3) Do not talk about violence, pornography, political sensitivity, etc.
4) Each recorder should talk about half time of the recording. if a person only has a large number of simple words ("um", "oh", and "right"), then it is invalid.
5) Avoid too much overlapping in conversations.
6) Each recorder can only work once for the project.
Recording environment:
1) Choose a quiet environment, avoid sudden noise during recording. Don’t select a big open spaces to avoid echo.
2) If you cannot record the first sentence, please contact us, do not skip to the next recordings.
Recording content and duration requirement
1) Select topics from the topic list (we'll send it to you), and talk freely on the selected topics.
2) Each group should talk about the topic for 5-30 minutes, and the total duration should be 2 hours.
Recording requirements:
1) Mobile phone: download DTRecorder. Make sure there are enough phone memory to prevent the phone memory from running out when the recordings are not completed.
2) The phone needs to be placed about 20-30 cm from the speaker, do not put the mic directly and closely to the speaker to avoid spraying wheat. Please don’t move phone often while recording.
3) Once you downloaded the app we'll provide you with a personal code to access the task.
After you finished the recording process your work will go through a quality check (it's gonna take 1-2 working days). Once your recording has been approved we'll release the payment.
Fixed $20 USD/hr x 2 hours = $40 per person / $80 per pair
We are looking forward to collaborate with you!
I have a background in theater and stream on Twitch so I am comfortable behind a mic. I was raised in the south by a mother from Michigan, so I have fairly neutral, non-accented voice.