The most common source of data for audio collection is recording several audio samples from human speech and voice.