Audio Frequency Spectrum: The Complete Guide

Sound is vibration. When a speaker cone moves back and forth, it pushes air molecules into alternating regions of compression and rarefaction, creating a pressure wave that travels to your ear. The number of times the speaker cone completes a full back-and-forth cycle in one second is the frequency, measured in Hertz (Hz). A tone at 440 Hz means the air pressure oscillates 440 times per second. This particular frequency is the standard tuning reference for the note A above middle C. The human ear can detect frequencies from approximately 20 Hz (a deep rumble you feel in your chest) to 20,000 Hz (a piercing, mosquito-like whine). In practice the upper limit is lower for most adults, and it falls with age and noise exposure. Music occupies the full audible range, but most of the melodic and harmonic content sits between 80 Hz and 8 kHz. The frequency spectrum is not perceived linearly. Humans hear pitch on a roughly logarithmic scale: the distance between 100 Hz and 200 Hz sounds like the same "interval" as the distance between 1,000 Hz and 2,000 Hz (both are one octave). This is why spectrum visualizers use logarithmic frequency axes and why equalizers space their bands logarithmically.

Audio engineers divide the spectrum into named regions, each with distinct sonic characteristics and practical implications for mixing and mastering. While exact boundaries vary between sources, the following divisions are widely accepted and provide a useful framework for understanding what each part of the spectrum contributes to music.

The lowest frequencies in music provide physical impact and harmonic foundation. Sub-bass (20-60 Hz) is the domain of 808 kicks, bass drops, and the lowest organ pipes. These frequencies are felt as much as heard. In clubs and cinemas, powerful subwoofer systems reproduce these frequencies at volumes that create chest-thumping physical sensations. On small speakers like phone and laptop speakers, sub-bass is almost entirely absent, which is why music mixed for these devices emphasizes the bass range (60-250 Hz) instead. The bass range (60-250 Hz) carries the body and warmth of music. The fundamental frequencies of bass guitar, the lower notes of piano and guitar, and the "punch" of a kick drum all live here. A well-balanced bass range gives music its sense of power and groove. Too much bass energy makes a mix sound boomy and muddy; too little makes it sound thin and lacking in authority. For music visualization, the bass range is typically mapped to large-scale visual motion: camera movement, global brightness pulses, and particle system expansion. The physical, visceral nature of low frequencies translates naturally into large, powerful visual gestures. Many visualizers use bass energy as the primary driver for "beat pulse" effects, where the entire visual field contracts and expands with each kick drum hit.

The midrange is where the majority of musical information lives. Vocals, the melodic content of most instruments, harmonic overtones, and rhythmic articulation all occupy this broad band. The human ear has evolved to be most sensitive in the 1-4 kHz range (the range of human speech), which is why the midrange is so critical for mix clarity and emotional communication. The low midrange (250-500 Hz) provides body and warmth to instruments. This is where the "chestiness" of a vocal, the fullness of an acoustic guitar, and the body of a snare drum reside. Overemphasis in this region creates a boxy, muddy sound that is one of the most common mixing problems. The midrange proper (500 Hz-2 kHz) is the core of melodic content. Most of what you hum along to in a song lives in this frequency range. The piano's most expressive register, the human voice's strongest harmonics, and the melodic lines of guitars, horns, and strings all concentrate here. The upper midrange (2-4 kHz) adds presence and articulation. The "attack" of a guitar pick, the consonants of a vocal, the crack of a snare drum, and the articulation of a violin bow change all live in this region. The ear's heightened sensitivity here means that even small boosts or cuts have a significant perceptual impact. In visualization, midrange energy is often mapped to color intensity, geometric complexity, and the density of mid-level visual elements. Because the midrange carries melody and harmony, it provides the richest source of continuously varying data for driving visual parameters.

The upper portion of the spectrum provides clarity, air, and spatial information. The presence range (4-6 kHz) is where vocals achieve intelligibility and instruments gain definition. The brilliance range (6-20 kHz) adds shimmer, sparkle, and the perception of acoustic space. Cymbals, the sizzle of hi-hats, the "air" around a recording, and the highest harmonics of instruments all occupy this region. High frequencies are directional: unlike bass frequencies, which radiate in all directions, treble frequencies travel in a relatively straight line. This is why speaker placement and room acoustics disproportionately affect the treble quality of a listening environment. It is also why headphones often reproduce high frequencies more accurately than speakers in an untreated room. Age-related hearing loss (presbycusis) mainly affects high frequencies, and the upper limit of hearing keeps falling through adult life, at different rates for different people. This has practical implications for music production: mix engineers must rely on metering and analysis tools rather than ears alone to ensure that high-frequency content is balanced, especially as they age. For visualization, high-frequency energy is typically mapped to fine-grained visual details: particle density, texture intensity, small-scale shimmer effects, and the "crispness" of rendered edges. The ethereal, airy quality of high frequencies translates naturally into delicate, detailed visual elements.

The Fast Fourier Transform (FFT) is the mathematical algorithm at the heart of every spectrum visualizer. It takes a slice of audio (a "window" of samples) and decomposes it into its constituent frequencies, producing a spectrum that shows how much energy is present at each frequency. This is the same principle behind equalizers, spectrum analyzers, and spectrograms. The FFT works by correlating the input signal with sine and cosine waves at each frequency. If the input contains a strong component at a given frequency, the correlation will be high, producing a tall bar in the spectrum display. The output of an FFT is a set of frequency "bins," each representing a narrow frequency range. The number of bins equals half the FFT window size: a 2048-sample FFT produces 1024 frequency bins. The resolution of an FFT involves a tradeoff between frequency precision and time precision. A longer window (more samples) produces finer frequency resolution but smears temporal detail. A shorter window captures rapid changes but has coarser frequency resolution. In practice, music visualizers use windows of 1024 to 4096 samples (23 to 93 milliseconds at 44.1 kHz), which provides a good balance between seeing individual notes and responding to rhythmic events. For visualization, FFT data is typically processed before display. The raw frequency bins are logarithmically grouped (to match human pitch perception), converted from amplitude to decibels (to match human loudness perception), and optionally smoothed over time (to reduce visual noise). The result is the familiar spectrum analyzer display: a bar graph or smooth curve showing the frequency content of the music in real time. The MusicMind report uses the same method for its spectral balance: it averages Hann-windowed FFT frames over the whole file and shows the energy in six bands, in dB relative to the loudest band.

Understanding the frequency spectrum is the foundation of audio mixing. Every mix decision is, at its core, a frequency decision: EQ shapes the spectral balance of individual tracks, compression controls the dynamics of specific frequency ranges, and spatial effects like reverb and delay affect different frequencies differently. A well-mixed song has a balanced frequency spectrum with no single band overwhelming the others. The bass provides foundation without muddiness, the midrange is clear and present without harshness, and the treble adds air and detail without sibilance. Achieving this balance requires both technical knowledge of frequency ranges and trained ears that can identify problems. Mastering engineers use spectrum analyzers to compare their work against reference tracks, ensuring that the overall tonal balance matches industry standards for the target genre. Electronic music and hip-hop often carry more sub-bass and treble energy than acoustic and folk recordings, which is why references should come from the same genre. Understanding these conventions is essential for producing music that sounds right within its genre context.

The frequency spectrum is the richest source of continuously varying data available to a music visualizer. While BPM provides a rhythmic pulse and key provides a harmonic baseline, the real-time frequency spectrum provides hundreds of data points per frame that can drive fine-grained visual animation. A common mapping strategy divides the spectrum into bands and assigns each band to a visual layer. Bass frequencies (20-250 Hz) drive large-scale elements: background color pulses, camera shake, and global particle system behavior. Midrange frequencies (250 Hz-4 kHz) drive mid-level elements: geometric shapes, color gradients, and structural patterns. High frequencies (4-20 kHz) drive fine details: particle sparkle, texture intensity, and post-processing effects like film grain. More sophisticated visualizers go beyond simple band splitting. They track the spectral centroid (the "center of gravity" of the spectrum, which correlates with perceived brightness), spectral spread (how wide the frequency content is, which correlates with timbral richness), and spectral flux (how rapidly the spectrum is changing, which correlates with rhythmic activity). These higher-level features provide more musically meaningful control signals than raw frequency band energy.