Guide · 8 min

Twin-Track Audio and 3-Second Echo Hiding

Why Metis Meet records Them and Me on separate tracks, and how the pending-line echo filter preserves your real replies.

All guides

Why two tracks matter

Cloud coaches attribute speaking time from merged transcript segments. Metis Meet records system audio (Them) and your microphone (Me) as separate 48 kHz tracks so coaching metrics, WPM, and pause detection run only on audio you actually spoke.

Echo hiding without headphones

When room speakers bleed into your mic, pending Me lines wait 3.0 seconds, compare word-set overlap and word-pair similarity (≥ 0.60) against recent Them lines, and drop loudspeaker echo while keeping genuine overlaps.