The Physics of Reaction Audio
Human reactions have unique acoustic signatures:
- Laughter: Characterized by rhythmic staccato bursts in the 250 Hz to 2 kHz frequency range with sharp attack times.
- Gasps: Rapid intake of breath causing a brief high-frequency sibilant whoosh followed by total silence.
- Cheers & Shouts: Sustained harmonic vocal energy between 500 Hz and 3.5 kHz exceeding +15 dB above the ambient baseline.
In contrast, background television noise or video game explosions are either constant or diffused throughout the room.
Environment Calibration Guide
| Environment | Ambient Noise Level | Recommended Mode | Behavior |
|---|---|---|---|
| Quiet Bedroom / Late Night | 25 dB - 35 dB | High Sensitivity | Triggers on subtle giggles, whispers, and soft gasps. |
| Living Room Movie Night | 45 dB - 60 dB | Standard Dynamic (Default) | Adapts baseline to TV volume; triggers on genuine vocal outbursts. |
| Party / Live Sports Gathering | 65 dB - 80 dB | High Threshold | Ignores background conversation; triggers only on shouting, clapping, and leaping. |
Why Vision + Audio Fusion is Essential
Relying on audio alone is always flawed: an explosion in an action film will trigger a recording even if you are falling asleep. Relying on computer vision alone is also flawed: looking away or rubbing your eye can look like a reaction.
🧠 The Power of Sensor Fusion
Fesi runs a dual-modality validation layer. An audio spike only triggers a save if accompanied by facial landmark movement (mouth opening, smile curvature, eyebrow velocity). This fusion eliminates 99.4% of false positives.
Experience Smart Noise-Filtering Reaction Capture
Download Fesi and enjoy intelligent, calibrated reaction recording on iOS and Android.