Visual classroom noise meters are a staple in elementary and middle school classrooms (grades 1–6), helping students self-regulate volume during independent study and group work. However, most browser-based noise meters fail when deployed on a real classroom setup: a smart TV or projector connected to a teacher’s PC with a webcam mounted 2 to 4 meters away from the students’ desks.
To solve this hardware gap, we built Quiet Class (Тихий клас) — an autonomous, offline-first PWA noise meter, task timer, and star-jar reward system with a custom Web Audio API digital signal processing (DSP) pipeline.
Why Traditional Browser Noise Meters Fail on Classroom Webcams
During classroom testing with distant TV-mounted webcams, four acoustic bottlenecks consistently broke naive RMS volume meters:
- 90th-Percentile Calibration Trap: When calibrating room silence over 3 seconds, taking the 90th percentile meant a single cough, chair scrape, or whisper pushed the “silence baseline” to
-35 dB. Subsequent thresholds jumped to-17 dB— requiring students to shout directly into the microphone to trigger a warning. - Unrealistic Far-Field Thresholds: Requiring
+10 dBfor the yellow zone and+18 dB(a 63× jump in acoustic energy) for the red zone ignores inverse-square law attenuation. Normal speech at 3 meters adds only+4.5 to +9 dBabove the room floor. - Excessive Debounce Windows (
700 ms): Requiring nearly a full second of continuous shouting caused short exclamations (200–350 ms) to be discarded completely. - Low-Frequency PC & Projector Fan Rumble: Cooling fans, AC units, and desk vibrations (
50–110 Hz) artificially inflate broadband time-domain RMS readings, masking human speech frequencies.
The Web Audio DSP Pipeline
Instead of reading raw microphone samples directly, Quiet Class routes the audio stream (echoCancellation: false, noiseSuppression: false, autoGainControl: false) through a four-stage Web Audio graph:
1. 120 Hz Highpass Filter (Fan & Rumble Cut)
A Biquad highpass filter at 120 Hz (Q = 0.7) strips out low-frequency hum from desktop fans, ceiling projectors, and desk bumps before level evaluation.
2. 2.2 kHz Speech Presence Peaking Filter (+3.5 dB)
A peaking filter centered at 2200 Hz (Q = 1.0, +3.5 dB gain in the Webcam profile) accentuates human voice consonants and whispers, creating a sharp signal-to-noise ratio even 3–4 meters from the screen.
3. Software Preamp Gain (0.5× to 4.5×)
Teachers can switch between three hardware presets (Webcam 2–4m at 2.2×, Laptop / PC at 1.5×, and Headset / Close at 1.0×) or fine-tune the preamp gain while watching a live monochromatic VU meter with real-time dB telemetry.
4. 25th-Percentile Calibration & Leaky Silence Tracking
- Spike-Immune Calibration: Room calibration samples 3 seconds of audio and selects the 25th percentile clamped to
[-75 dB, -44 dB], ignoring transient noises and guaranteeing the meter never becomes “deaf”. - Leaky Silence Tracking: During quiet stretches (
> 2.5 sbelow baseline), the ambient floor gently adapts downward toward the true acoustic minimum of the room.
Roly-Poly Verlet Physics & Classroom Motivation
- 60 Hz Verlet Physics with Roly-Poly Restoration: Characters react with proportional tremors in the Yellow Zone (
+4.5 dB,120 msdebounce) and a startledscareBouncein the Red Zone (+9.0 dB,250 msdebounce). World-space lighting stays fixed at the top-left while character faces tilt within±25°like physical roly-poly toys. - 7 Visual Themes & Projector Light Mode: Includes Sleepy Critters, Solar System (two-pass Saturn rings and day/night terminators), Soap Bubbles, Emoji, Billiards, Numbers, and Eyes, plus a high-contrast Light Room mode for sunlit classrooms.
- Glass Star Jar & Multi-Class Profiles: Every 30 seconds of continuous quiet launches a golden star along a quadratic Bézier curve into the class jar, saving weekly goals and rewards across multiple class rosters (
4-B,3-A,5-C) inlocalStorage. - 100% Local Privacy & Offline PWA: Audio is processed strictly in RAM via Web Audio API — nothing is recorded or transmitted over the network.
Launch the tool directly in your browser:
👉 Open Quiet Class (Тихий клас) Online
Explore our other open educational tools for schools and STEM clubs:
Interested in 1-on-1 online lessons in Unity game development, C#, Python, Scratch, or 3D modeling?
👉 Explore 1-on-1 Tutoring & Mentorship with Sivochka Ihor
