Critical Bands: The Bark Scale
The human ear does not perceive frequencies linearly. The cochlea divides incoming sound into "critical bands" — frequency ranges that are processed as perceptual units. The Bark scale, named after acoustician Heinrich Barkhausen, models these bands. There are approximately 25 critical bands spanning the audible range from 20 Hz to 20 kHz. Each band is roughly equal in perceptual width despite varying hugely in Hz width — the lowest band covers just 80 Hz while upper bands span 3,500+ Hz. The visualization engine uses Bark bands instead of linear FFT bins to create perceptually meaningful visual responses.
Click a band to hear its center frequency. Height shows bandwidth.
20 Hz (Sub bass)1 kHz (Presence)20 kHz (Ultra high)
Equal Loudness & Frequency Weighting
Human hearing is not equally sensitive across frequencies. We are most sensitive around 2–4 kHz (the range of speech consonants and baby cries) and much less sensitive to sub-bass and extreme highs. Equal loudness contours (Fletcher-Munson curves) map these sensitivity differences. The chart below shows perceptual weight multipliers at key frequencies — a 31 Hz bass tone needs to be 50x more powerful than a 2 kHz tone to be perceived as equally loud.
Click a column to hear that frequency. All tones play at equal digital volume — but they won't sound equally loud.
Your ears are most sensitive around 2–4 kHz (speech range). Bass and extreme highs need far more power to sound equally loud.
Temporal Integration
The brain integrates audio information over different time windows depending on the attribute. Loudness perception integrates over ~200ms — very brief sounds seem quieter than sustained ones at the same amplitude. Pitch perception needs only ~50ms of signal. Rhythm perception operates over ~100ms windows, and our sense of tempo maxes out at around 600ms between beats (100 BPM) — faster than that, we start to hear individual beats as a continuous stream rather than a rhythmic pattern.
Auditory Masking
Louder sounds can mask quieter sounds at nearby frequencies — a phenomenon called simultaneous masking. The upward spread of masking (25 dB/Bark) is greater than downward spread (10 dB/Bark), meaning low frequencies mask high frequencies more effectively than vice versa. This is why a heavy bass line can make hi-hats harder to hear. The visualization engine accounts for masking by reducing visual reactivity for frequency bands that are likely masked by louder neighbors, preventing misleading visual feedback.
A 31 Hz bass tone must be 50x more powerful than a 2 kHz tone to sound equally loud. Your ears aren't flat — and neither should your visualizer be.