Deepspeech recognition rate

Hi,
Isn’t it possible to get the same result using sox only, without using arecord?