Dessa processes direct audio signals to capture emotional tones, sarcasm, laughter, and pauses in real time