There is no data of urdu speech on the internet where individual words are spoken by a large number of different people. This makes it tough for Pakistanis to build good open source examples of speech recognition.
To build an open source speech dataset in Urdu I Hamza Iqbal created this webpage where everyone can contribute there voice to make a good urdu speech recognizer. I am hoping to gather single spoken words from several hundred people and then release the resulting data set under an open license, together with examples of how to use it to create speech recognizer.
To make sure that I will be share your data as part of a public dataset with no privacy or other concerns, I need to get your agreement formally. If you would like to participate in providing some samples, please review and confirm your agreement to the following:
1. By participating, this website will ask for permission to use your microphone and prompt you to speak single words which we will record (the “Duniya”, “Hiffazat”). There will be total 3 levels, in every level different words will appear on-screen, one after another with short pauses and you will be only have 2 seconds to speak that words. When the microphone is active, you’ll see an oscilloscope display showing a waveform reflecting the audio received. The Record button will turn red when a word is being recorded (each a “clip”). Clips will be displayed in the bottom half of the web page where you can review and delete any of them locally before submitting them to our server. Clips will only be sent to our server when you press the ‘Upload’ button, or the ‘OK’ button on the dialog that appears once you’ve completed all the requested words. At any point you can press the ‘Stop’ button to pause the recording process. Navigating away from this website will remove all local copies of the clips.
2. Your participation is voluntary and you may choose to cease participation at any time. You will not be compensated for your participation. If you choose to participate, you must be in a quiet room with no other voices and only speak the requested words. A closed room is a good place for this.
3. By participating, you agree that I will use these clips to develop different type of services, and share the clips with others, including the general public, for example, as part of a public dataset to facilitate research. You understand that your voice alone or in combination with other information could identify you. You specifically waive and release any potential intellectual property right, right of publicity, or right of privacy claim against me or others for use or sharing of your clips.
4. This agreement does not create any agency or partnership relationship. This agreement is not assignable or transferable by you. This agreement is the parties’ entire agreement on this topic, superseding any prior or contemporaneous agreements. Any amendments must be in writing. Failure to enforce any of provisions of this agreement will not constitute a waiver.
By clicking ‘I Agree’ you understand the points above, and consent to the collection, use, and sharing of the clips.