What Is Frequency Response in Audio?

Human hearing spans roughly 20 Hz at the deep-bass bottom to 20,000 Hz at the high-treble top, though the upper limit falls with age — many adults hear little above 15-16 kHz. Frequency response measures how uniformly a device reproduces that range. A theoretically perfect system would output every frequency at exactly the level it received, so a flat input yields a flat output; no real transducer achieves this, because headphones, speakers, and microphones all impose their own character, and rooms add more. Response is shown as a graph with frequency on the horizontal axis (almost always a logarithmic scale, since we hear pitch logarithmically — the octave from 100 to 200 Hz occupies the same visual width as 1,000 to 2,000 Hz) and level deviation in decibels on the vertical axis. A flat line at 0 dB means perfect accuracy; a peak means that region is emphasized (louder), a dip means it is recessed (quieter). Reading that shape tells you, at a glance, how a device will color everything you play through it.

A response graph looks intimidating but reads simply once you know the conventions. The x-axis runs left (low frequencies) to right (high), logarithmically, typically 20 Hz to 20 kHz. The y-axis shows deviation in dB, with 0 dB as the reference; the useful information is the shape and how far it strays from flat, not the absolute numbers. Bumps and dips translate directly to tone: a rise centered around 60-100 Hz means extra bass punch, a scoop around 200-500 Hz thins the body of instruments, a peak near 2-5 kHz adds presence and can become harsh, a boost above 8-10 kHz adds air or sibilance. The tolerance figure is everything: '20 Hz-20 kHz ±1 dB' describes a nearly ruler-flat, accurate device, while '20 Hz-20 kHz ±10 dB' technically covers the same range but could be wildly colored within it. Watch also for how gently the extremes roll off, and treat manufacturer graphs skeptically — smoothing, scaling, and measurement method dramatically change how flat a product appears.

Engineers divide the audible range into named regions, and knowing them makes frequency response concrete. Sub-bass (20-60 Hz) is felt more than heard — the rumble of a kick drum's body or a synth's foundation. Bass (60-250 Hz) carries the weight and punch of bass guitars, kick drums, and the low end of most instruments. Low-mids (250-500 Hz) hold the warmth and body of vocals and guitars, and are where 'muddiness' lives when overdone. Mids (500 Hz-2 kHz) contain the fundamental character and intelligibility of most sounds, including the core of the human voice. High-mids or presence (2-5 kHz) govern clarity, attack, and consonant definition — the ear's most sensitive region, easily made harsh. Highs (5-10 kHz) add brightness and detail, and brilliance or air (10-20 kHz) contributes sparkle and openness. When a review says headphones are 'bass-heavy' or 'bright,' it is describing which of these bands the frequency response emphasizes, and the same vocabulary describes EQ moves and format limitations.

The audio format is almost never the weak link in a frequency-response chain. Uncompressed formats (WAV, AIFF) and lossless formats (FLAC, ALAC) have perfectly flat response — they store every frequency exactly as recorded, so the file imposes no coloration at all. Lossy formats at reasonable bitrates are effectively flat across the audible range too: a 192 kbps or higher MP3, or a 128 kbps AAC, preserves the full spectrum for practical purposes. The one real format effect is a low-pass cutoff at low bitrates, where the encoder discards the top of the spectrum to save bits — a 64 kbps MP3 typically rolls off everything above roughly 11-16 kHz, which is why very low-bitrate audio sounds dull and closed-in. But at any bitrate you would actually use for music, the format preserves the whole audible band, and the coloration you hear comes entirely from the transducers and room downstream. In a modern chain, worrying about a codec's frequency response is optimizing the wrong link.

'Flat' and 'colored' name the two philosophies of audio reproduction, and neither is universally correct. A flat response is accurate: it reproduces the recording as it actually is, adding nothing, which is exactly what a mixing or mastering engineer needs — they must hear the truth to make good decisions, and a colored monitor would lead them to compensate for problems that only exist in the speaker. A colored response is flavored: bass boosted for impact, treble lifted for perceived clarity, mids scooped for a 'smile curve' that flatters casual listening. Consumer headphones and speakers are almost always deliberately colored, because most listeners prefer a little extra bass and sparkle to clinical neutrality. The key insight is that these serve different goals: accuracy for production, pleasure for consumption. Trouble arises only when the two are confused — mixing on colored consumer headphones bakes their coloration inversely into your work, so a bass-boosted headphone leads you to under-mix the bass, and the track sounds thin everywhere else.

Transducers are where frequency response matters most, because their deviations dwarf anything the format or electronics contribute. Even excellent headphones stray from flat by several dB across the spectrum, and the differences between models are largely differences in their response curves: one boosts sub-bass for a fun, thumpy sound, another lifts the treble for perceived detail, another aims for neutrality. Studio monitors and reference headphones chase flatness so engineers hear the recording accurately; consumer earbuds and lifestyle headphones deliberately shape the sound to please the average buyer. This is why a spec line like '20 Hz-20 kHz' is nearly meaningless on its own — it states the range but not the accuracy within it, and every product on the shelf claims roughly that range. The number that matters is the tolerance (±1 dB versus ±10 dB) and, better still, the actual measured curve. When two headphones 'sound different,' you are almost always hearing two different frequency-response shapes.

A subtle but important point: a headphone measured as perfectly flat does not actually sound neutral to a listener, because of how our ears and head interact with sound. Loudspeakers in a room deliver bass and treble to the ear differently than headphones clamped to it, and research — most famously Harman International's listener-preference studies — established a target curve that headphones should follow to sound subjectively neutral and pleasing to most people: a gentle bass shelf boost of several dB, a flat midrange, and a mild, controlled treble contour. In other words, the 'correct' headphone response is not a flat line but this deliberately shaped curve, because it mimics the experience of good speakers in a good room. This is why headphone reviews increasingly compare products against the Harman target rather than against flat, and why 'neutral' headphones still show a bass bump on a graph. It is a reminder that frequency response must be interpreted through human perception, not just measured against a ruler.

Frequency response is not only a playback concern — it shapes sound at the capture end too. Every microphone has its own response curve, and engineers choose mics partly by how that curve flatters a source. A 'flat' or neutral measurement microphone captures the source truthfully, ideal when you want to color it later with EQ or when measuring a system. But many prized studio mics are deliberately non-flat: a large-diaphragm condenser with a gentle presence lift around 5-10 kHz adds sheen to vocals; a classic dynamic mic with a tailored midrange and rolled-off extremes suits close-miked instruments and stage use. There is also the proximity effect, where directional mics boost bass as the source moves closer — a response change that depends on distance, used intentionally by broadcasters and singers for warmth. So the frequency response you ultimately hear is a chain: the mic's curve colors the recording, the format preserves it, and the headphones or speakers color it again. Understanding response at both ends explains why the same voice sounds different through different microphones.

For anyone listening on speakers, the room alters frequency response more than the speakers themselves. Sound reflects off hard walls, floors, and ceilings, and those reflections combine with the direct sound to reinforce some frequencies and cancel others, producing large peaks and dips that vary as you move around the room. Corners amplify bass through boundary loading, and every room has resonant modes — specific low frequencies that ring and pile up because they fit the room's dimensions — which is why untreated rooms often sound boomy or uneven in the bass. Professional studios spend heavily on acoustic treatment (broadband absorption, bass traps in corners, diffusion on reflective surfaces) precisely to flatten the room's contribution so the monitors' true response reaches the ears. The practical consequences: if your audio sounds wrong, suspect the room before the gear; small changes in speaker and listening position can dramatically alter the bass; and switching to headphones removes room effects entirely, which is one reason so much modern mixing happens on well-understood headphones.

Equalization is the deliberate tool for reshaping frequency response, and understanding response makes EQ intuitive rather than mysterious. An equalizer applies boosts and cuts to chosen frequency regions, letting you correct problems or create character: cut around 200-400 Hz to clear 'mud' from a dense mix, dip a harsh peak near 3 kHz to tame sibilance, add a gentle high shelf above 10 kHz for air, or roll off sub-bass rumble below 40 Hz. The bands map exactly to the spectrum vocabulary — you are literally redrawing the response curve of the signal. EQ also compensates for the response failures described above: room-correction systems measure a room's frequency response and apply an inverse EQ to flatten it, and headphone-correction software nudges a known headphone curve toward a target like Harman. The guiding principle is subtractive first (cutting problem frequencies usually sounds more natural than boosting), and always in service of the final response you want the listener to hear. EQ is, in essence, frequency response made adjustable.

Translate all of this into decisions. For format choices: any modern lossy format at 128 kbps or above preserves the full audible frequency range, so the format is not your bottleneck — spend your attention on transducers and room, not codecs. For recording: pick microphones whose response suits the source (neutral when you want to shape later, flattering when the character helps), and record with adequate headroom so nothing distorts. For listening and mixing: know your headphones' and speakers' response — if they boost bass, your mixes will come out bass-light; if they lift treble, you will under-brighten — and reference finished work on multiple systems (studio monitors, consumer earbuds, a phone, a car) to average out any single device's coloration. For problem-solving: when audio sounds wrong, investigate in order — room first, then transducers, then EQ, and only last the format, which is almost never the culprit. Frequency response is the thread connecting microphone, format, EQ, speaker, and room into the sound you finally hear.