I recall one VR based program (can't remember whether it was Hal or not at the moment) that knew I had laughed aloud and it made a remark about the topic being funny.
Therefore, why couldn't the software be able to detect other sounds like a sneeze, coughing, a hic-cup / hic-cough, clapping, whistling, snoring, bells / phone ringing, music, etc.
I might have mentioned before but once I was critiquing a program for another vendor and as it was listening to me, I picked up my acoustic guitar and played some finger style intro. In a couple of seconds, she said, "I've never heard that song before."
Well, of course I was floored and began to wonder how much intel the software actually had.
But I digress, the recognition is within the software because the hardware will "hear" the sounds. What it does with them or how it interprets them is the key.