The number of audio samples per second. Many speech-to-text services perform best with 16000 Hz audio.