/daɪˈnæmɪk ˌiːˈkjuː/
Dynamic EQ is an equalizer whose band gain responds automatically to the input signal level, cutting or boosting only when a frequency region exceeds a set threshold. It combines EQ precision with compressor-like reactivity.
Every time you reach for a static notch to tame a nagging 3 kHz harshness, you are also cutting that frequency during every quiet, silky passage where it never bothered anyone. Dynamic EQ fixes that bargain.
A dynamic EQ is an equalizer in which one or more filter bands can be configured to apply gain only when the signal energy at a target frequency crosses a defined threshold. Below that threshold the band remains at zero gain — or at a user-defined resting gain — and the mix hears an unaltered spectrum. Above the threshold the band clamps down (or opens up, in the case of upward expansion) with a speed and depth governed by attack, release, ratio, and range controls. The result is frequency correction that tracks the music's own dynamics rather than imposing a fixed tonal character across the entire duration of a clip.
The distinction from a static parametric EQ is reactive behavior: a static EQ applies its gain curve identically at every moment, regardless of what the signal is doing. The distinction from a multiband compressor is precision and transparency: a multiband compressor operates on the full output level of a crossover-defined band, which means it affects both tonal balance and dynamics simultaneously and can introduce crossover phase artifacts. A dynamic EQ operates on a narrow, precisely tunable filter — often as tight as 0.1 Q — and its gain change is calculated from a detection circuit (the sidechain) that monitors only that target region, leaving the rest of the spectrum structurally intact. Crossover coloration is either absent or minimized depending on implementation.
The practical consequence is that dynamic EQ excels at problems that are intermittent. A vocalist whose chest resonance balloons at 280 Hz on loud belted notes but sounds full and appropriate on softer phrases is a textbook candidate. A snare whose high-mid clang peaks at 4.2 kHz only on rim shots, a bass guitar whose low-mid bark at 350 Hz surfaces only on open-string attacks, a mix bus whose high-frequency content gets aggressive only during a dense pre-chorus — all of these respond better to a frequency-reactive tool than to a permanent surgical cut that robs the signal during benign moments.
Dynamic EQ bands can be configured to cut (downward compression of a frequency band), boost (upward expansion, where a quiet frequency is lifted into prominence above a threshold), or operate bidirectionally. Most modern implementations also allow the user to choose whether the detection signal is the same audio passing through the band (internal sidechain) or a separate routed signal — enabling de-essing-style workflows where, for example, the dynamic EQ on a kick drum is triggered by a room mic rather than the kick itself.
In contemporary production across genres from hip-hop to orchestral scoring, dynamic EQ has become one of the most relied-upon transparent correction tools precisely because it leaves source character intact during the moments when problems are absent. Paired with good gain staging and surgical static EQ, it represents the fine surgical layer of a mature frequency management workflow.
At its core, a dynamic EQ band contains two subsystems working in parallel: a filter path and a detection path. The filter path is a standard parametric or shelf filter — identical in mathematical structure to the filters found in any static EQ — positioned in the main audio signal flow. The detection path is a sidechain circuit that continuously analyzes the signal energy at or around the filter's center frequency. The detection path feeds a gain-control element (functionally analogous to a VCA or digital gain multiplier) that varies the filter's applied gain in real time according to the relationship between detected energy and the user-set threshold.
When detected energy is below the threshold, the gain-control element holds the filter at its rest position — zero gain for a pure dynamic band, or a static offset if the user has dialed in a baseline boost or cut. When detected energy rises above the threshold, the control element applies additional gain in the direction configured: negative gain (attenuation) for dynamic cutting, positive gain for dynamic boosting. The rate of gain change is governed by attack time (how quickly the filter responds after the threshold is crossed, typically 1–100 ms) and release time (how quickly it returns to rest after energy drops below threshold, typically 20–500 ms). The maximum gain change is bounded by a Range parameter measured in dB, preventing the band from over-correcting even if the signal peaks far above threshold.
Ratio determines how aggressively the band responds per dB of overage, mirroring compressor ratio logic: a ratio of 2:1 means that for every 2 dB the detected signal exceeds the threshold, the filter applies 1 dB of additional cut. Infinite ratio (hard limiting behavior) causes the band to clamp immediately to its maximum range value as soon as threshold is crossed — useful for de-essing applications but harsh-sounding on broadband musical material. Detection filter shape (sometimes called the SC filter or key filter) allows the user to narrow or widen the frequency window the detection circuit monitors, separating the listening zone from the processing zone; a narrow detection window on a broad processing band is a common technique for surgical dynamic control without tight filtering of the output.
Phase behavior varies by design philosophy. Linear-phase dynamic EQ implementations (such as those in certain modes of FabFilter Pro-Q 3 or iZotope's Ozone Dynamic EQ) introduce processing latency but avoid the phase rotation inherent in minimum-phase filters, making them preferred on bus and mastering contexts. Minimum-phase implementations introduce frequency-dependent phase shift but have zero added latency, making them appropriate for tracking and mixing in real time. Some tools offer a switchable mode so the producer can choose per band.
The practical signal flow summary: audio enters → detection sidechain measures band energy → comparator evaluates energy vs. threshold → gain control modulates filter depth per attack/release/ratio curves → filtered audio exits. The elegance of the system is that when the threshold is never crossed, the entire dynamic mechanism is inert and the plugin introduces no more coloration than a bypassed filter — which is to say, essentially none.
Diagram — Dynamic EQ: Dynamic EQ signal flow: input splits to filter path and detection sidechain; sidechain compares energy to threshold and modulates filter gain via gain control element; output shows frequency curve at rest vs. active state.
Every dynamic eq — hardware or plugin — operates on the same core parameters. Know these and you can work with any implementation.
Set in dBFS relative to the detected band level. When signal energy in the target frequency region exceeds this value, the band begins to apply gain. A threshold of −18 dBFS on a vocal high-mid band, for example, will leave softer phrases untouched while catching only the loud, brash consonants. Threshold is the primary tuning variable — set it too low and the band is functionally always active, degenerating into static EQ behavior.
Expressed in dB, Range caps the correction depth regardless of how far the signal exceeds the threshold. A Range of −4 dB means the dynamic cut will never exceed 4 dB even during extreme peaks, preventing overprocessing. On de-essing tasks 6–10 dB of range is common; on mix-bus resonance control, 2–3 dB is typically sufficient to stay transparent.
Measured in milliseconds, attack governs the speed at which the gain-control element moves from rest to the computed gain reduction. Fast attacks (1–5 ms) catch transient peaks such as sibilance or consonant bursts cleanly. Slow attacks (30–80 ms) let initial transients pass, preserving punch and naturalness — important when dynamic EQ is used on drums or to control sustain-phase resonances rather than transients.
Release controls the decay of gain correction once detected energy falls below threshold. Excessively short releases (below 20 ms) cause 'pumping' or 'breathing' artifacts as the band flutters between corrected and uncorrected states. Longer releases (100–300 ms) create smoother, more transparent behavior but may hold correction into quiet passages where it is unwanted. Program-dependent or auto-release modes, available in tools like FabFilter Pro-Q 3, adapt release time to signal complexity.
Mirrors compressor ratio: at 2:1, every 2 dB above threshold results in 1 dB of additional filter gain. A ratio of 1:1 produces no effect; infinity:1 clamps instantly to the Range ceiling, functioning like a brickwall filter gate. Most musical applications use ratios between 2:1 and 6:1. De-essing typically benefits from higher ratios (6:1–∞) to eliminate harsh peaks decisively.
Standard parametric frequency selection, identical to static EQ. Precision here determines whether the correction hits the problem region cleanly or bleeds into musically important adjacent content. Use a spectrum analyzer or real-time frequency plot to park the center frequency exactly on the offending peak — even 200 Hz misplacement on a 1 kHz resonance can leave the problem audible while dulling surrounding harmonics.
Controls how wide or narrow the frequency window is affected when the band applies gain. High Q values (4.0–10.0) create surgical notches that isolate a resonance tightly; low Q values (0.5–1.5) apply broader tonal shaping. Note that the Q of the detection filter and the Q of the output filter can often be set independently in advanced tools — allowing a narrow detection window to trigger a broader corrective curve.
Internal detection routes the band's own input signal to the sidechain — the standard mode. External (sidechain) detection routes a separate audio source to trigger the band's gain control, enabling cross-instrument dynamic EQ: a room mic triggering a kick drum's low-mid cut, or a full mix bus triggering a corrective band only when a loud element enters. This mode significantly expands the creative utility of dynamic EQ beyond simple self-detection correction.
Session-ready starting points. These are starting-point values for typical studio material at −18 dBFS RMS; adjust threshold first, then range, then time constants by ear.
| Parameter | General | Drums | Vocals | Bass / Keys | Bus / Master |
|---|---|---|---|---|---|
| Center Freq (typical) | 200 Hz – 8 kHz | 100–400 Hz, 2–5 kHz | 2–4 kHz (bite), 7–10 kHz (sibilance) | 80–300 Hz (low-mid) | 2–5 kHz (air / harshness) |
| Threshold | −18 to −12 dBFS | −16 to −10 dBFS | −20 to −14 dBFS | −18 to −12 dBFS | −22 to −18 dBFS |
| Range (dB cut) | 2–6 dB | 3–8 dB | 4–10 dB (de-ess), 2–4 dB (bite) | 3–6 dB | 1–3 dB |
| Attack | 5–20 ms | 1–5 ms (transient), 20–40 ms (sustain) | 1–3 ms (de-ess), 10–25 ms (tone) | 5–15 ms | 10–30 ms |
| Release | 60–150 ms | 40–100 ms | 50–120 ms | 80–180 ms | 100–250 ms |
| Q | 1.0–3.0 | 1.5–5.0 | 2.0–6.0 (de-ess), 1.0–2.0 (tone) | 1.5–4.0 | 0.8–2.0 |
| Ratio | 2:1–4:1 | 3:1–6:1 | 4:1–∞ (de-ess), 2:1–3:1 (tone) | 3:1–5:1 | 2:1–3:1 |
These are starting-point values for typical studio material at −18 dBFS RMS; adjust threshold first, then range, then time constants by ear.
The conceptual lineage of dynamic EQ extends to the earliest attempts to make tone controls responsive rather than static. The 1950s saw broadcast engineers at institutions like the BBC experimenting with frequency-selective limiters — circuits that would attenuate a specific band when it became problematic on air. These were discrete, hardware-intensive constructions, more akin to band-specific limiters than true parametric dynamic equalizers, but they established the core logic: measure energy at a frequency, compare it to a threshold, apply gain correction. The Neve 2254 compressor/limiter from 1969, though not a dynamic EQ by name, incorporated frequency-selective detection options that engineers used to create reactive tonal control on specific instruments.
The first purpose-designed dynamic EQ in the modern sense is widely attributed to the API 560-era era thinking crystallized in the Gain Brain by dbx (1972) and, more directly, in the Dynamic Equalizer products that appeared in the mid-1970s from companies including Neve and Eventide. Eventide's Instant Phaser and Omnipressor units from 1971–1974 explored envelope-controlled frequency modification. By 1976 the Dolby Cat. 430 spectral processor used frequency-dependent dynamic gain for noise reduction — a commercial application of dynamic frequency control at scale. The SSL 4000 series console, introduced in 1976 and refined through the early 1980s, incorporated a dynamic element into its channel EQ architecture that recording engineers on sessions with artists like Peter Gabriel, Dire Straits, and Fleetwood Mac discovered could be used for reactive high-mid management on vocals.
The plugin era democratized the concept substantially. TC Electronic's System 6000 (1999) brought broadcast-grade dynamic EQ into professional post-production suites. Algorithmix released a dynamic EQ plugin in the early 2000s used by mastering engineers including Bob Katz and Ian Shepherd. The watershed commercial moment, however, came with Waves' introduction of the Linear Phase Multiband Compressor and, more critically, iZotope's Ozone 5 (2012), which packaged a well-designed dynamic EQ section inside a widely adopted mastering suite, exposing tens of thousands of independent producers to the workflow. FabFilter's Pro-Q 2 (2013) included a dynamic band mode that instantly became the industry reference implementation — praised for its zero-latency minimum-phase option, per-band sidechain routing, and the visual clarity of its spectrum analyzer overlay, which made threshold-setting intuitive for engineers at every experience level.
By the mid-2010s, dynamic EQ had moved from a specialist mastering tool to a standard element in mixing templates. Producers working in genres from trap (where controlling low-mid mud in dense 808-heavy arrangements is a constant challenge) to country (vocal sibilance on bright room microphones) adopted it as a first-call transparent correction tool. Plugin developers including Sonnox (Oxford Dynamic EQ, 2016), TDR (Nova, released free in 2015 and widely credited with bringing dynamic EQ to producers who could not afford premium tools), and Brainworx (bx_dynEQ V2) refined the paradigm. By 2020, virtually every major DAW had incorporated dynamic EQ functionality natively — Logic Pro X's Smart EQ learning-assisted curve, Ableton Live's implementation via Max for Live, and iZotope's Neutron suite extended the workflow into real-time machine-learning-assisted frequency management, where AI detection can suggest and apply dynamic bands based on source material analysis.
Vocals are the single most common dynamic EQ application in contemporary mixing. The two primary targets are the 2–4 kHz presence region, which can become harsh and fatiguing during louder, more emotionally intense phrases, and the 7–10 kHz sibilance band. For the presence issue, a dynamic cut centered at the specific offending frequency (found by sweeping with the band soloed) with a threshold set just below where the harshness becomes audible, a Q of 1.5–2.0, and a range of 3–5 dB will smooth high-energy passages while leaving quieter phrases completely open. For de-essing via dynamic EQ, center the band at the vocalist's specific sibilance frequency (commonly 7–9 kHz for female voices, 5.5–8 kHz for male), set Q to 3.0–6.0, attack to 1–2 ms, and ratio to 6:1 or higher. This approach often sounds more natural than a dedicated de-esser because it respects the broader high-frequency character of the voice rather than triggering on a fixed wideband detector.
Drums and percussion benefit from dynamic EQ primarily on the kick and snare. On kick drum, a dynamic cut in the 200–400 Hz range tames the cardboard 'thump' that appears only when the beater hits hard — preserving the warmth of softer hits in a hip-hop groove while preventing the low-mid from muddying the mix on accented beats. On snare, a dynamic cut at the ringing resonance frequency (often 900 Hz–1.8 kHz, found with a narrow static sweep) with a fast attack (2–5 ms) and a range of 4–7 dB removes the boxy ring on hard hits without dulling the crack during ghost notes. Room mics fed through a dynamic EQ can suppress low-frequency buildup that only appears during loud kit passages, maintaining spaciousness during fills without requiring a static shelf that kills ambience during quieter playing.
Bass guitar and synthesizer bass use dynamic EQ to manage the low-mid bark (250–450 Hz) that appears on aggressive picking or plucking attacks. A dynamic cut in this region with a medium attack (8–15 ms) allows the initial click of the attack to pass, then corrects the sustain-phase bloom that can obscure kick drum definition. On synthesizer bass and 808 patterns, dynamic EQ at 60–100 Hz helps manage pitch-to-pitch level variation caused by different fundamental frequencies hitting different room resonances — though this application sits at the boundary between dynamic EQ and multiband compression and requires careful threshold setting to avoid pumping on fast sub patterns.
Mix bus and mastering applications are where dynamic EQ's linear-phase options become especially important. On a mix bus, a dynamic high-shelf cut above 10 kHz — set to activate only when the high-frequency content becomes overly bright during dense sections — maintains air and sparkle during verses while preventing ear fatigue on a loud chorus. In mastering, dynamic EQ is used to address frequency imbalances that occur only at certain energy levels in a recording: a mix that is tonally correct at lower levels but accumulates harshness at full commercial loudness can be treated with a 2–3 dB dynamic cut at 3–5 kHz triggered only at the loudest moments. Mastering engineers Ian Shepherd and Bob Katz have publicly described this approach as essential for achieving loudness without listener fatigue.
One email a week. The techniques behind the terms — curated by working producers, not algorithms.
Abstract knowledge becomes practical when you can hear it in music you know. These tracks demonstrate dynamic eq used intentionally, at specific moments, for specific purposes.
The kick drum at the top of 'HUMBLE.' sits with unusual clarity in the 60–100 Hz region despite the sparse, dark arrangement. Engineers have noted that frequency-reactive processing was used on the kick bus to control the 200–300 Hz low-mid buildup that can accumulate on hard 808-adjacent hits, allowing the fundamental to punch without clouding the vocal pocket. Listen on headphones to how the kick's body stays consistent across accented and unaccented hits — a clear indicator that dynamic rather than static equalization was used. The vocal chain shows similar treatment in the 3–5 kHz region: Kendrick's aggressive delivery never turns harsh, suggesting a dynamic cut in the presence region activated only on the loudest syllables.
Finneas O'Connell has discussed in multiple interviews (including Mix magazine, 2019) that the vocal production on 'bad guy' relies heavily on transparent dynamic processing rather than heavy static EQ. Billie's close-mic whisper technique generates large variations in sibilance energy and low-frequency proximity effect across different syllables. The sibilance sits cleanly in the mix without the over-de-essed lisping quality that a static high-frequency cut would cause — consistent with dynamic de-essing via a narrow-band dynamic cut at approximately 8 kHz. The bass line's low-mid presence shifts noticeably between the verse and the drop at 1:00, suggesting dynamic control of the 200–400 Hz region that clears space for the sub when the arrangement thins.
The interaction between Nile Rodgers's rhythm guitar and Pharrell Williams's vocal on 'Get Lucky' demonstrates a textbook case of what dynamic EQ solves in dense arrangements. The guitar's 2–4 kHz chop and Pharrell's vocal presence region share similar frequency real estate. In the final mix (mastered by Bob Ludwig), the two elements coexist without the vocal being buried or the guitar needing a permanent presence cut. This kind of transparent separation — where both sources maintain full character yet avoid masking — is consistent with dynamic EQ on one or both channels, activating a cut in the guitar's 2.5–3.5 kHz region only when the vocal arrives, or vice versa. The consistency across the three-and-a-half-minute runtime without listener fatigue supports reactive rather than static processing.
Nigel Godrich's production on Kid A, particularly the vocal processing throughout 'Everything in Its Right Place,' was an early high-profile instance of deliberate frequency-reactive manipulation of Thom Yorke's voice. The heavily processed vocal exhibits a quality where the formant-related resonances shift in a way that tracks Yorke's dynamic level — softer notes retain full spectral content while louder phrases are dynamically shaped in the 1–3 kHz region. The production was completed using Pro Tools with third-party dynamic processing tools available in 2000, and Godrich has cited Eventide and TC Electronic hardware used alongside software for this album's unique tonal character.
One dynamic band applied to a precisely targeted frequency region. The dominant use case in mixing: a single reactive notch on a vocal resonance, a single dynamic cut on a snare ring, or a single de-essing band. Computationally lightweight and sonically transparent, single-band dynamic EQ is the most common configuration and the default starting point for most producers.
Multiple simultaneous dynamic bands, each with independent threshold, range, attack, release, and ratio controls. Used on mix buses and in mastering to address several frequency-region issues simultaneously — for example, taming low-mid muddiness, controlling upper-mid harshness, and managing high-frequency brightness with three reactive bands in parallel. Tools like FabFilter Pro-Q 3 and iZotope Ozone Dynamic EQ support up to eight or more simultaneous dynamic bands.
Implements the filter using linear-phase FIR processing, which eliminates frequency-dependent phase rotation at the cost of added latency (typically 20–120 ms depending on settings). Preferred on mix bus and mastering chains where phase coherence between corrected and uncorrected moments is audible. Not suitable for live monitoring or tracking due to latency, but considered the gold standard for final-stage dynamic frequency correction on full program material.
Uses standard IIR filter topologies, introducing frequency-dependent phase shift but operating with zero or near-zero latency. Appropriate for tracking, mixing in real time, and any context where monitoring latency must remain below the perception threshold. The phase characteristics are consistent with how analog hardware EQs behave and are generally considered musically acceptable; some engineers prefer the subtle coloration that minimum-phase behavior adds.
A modern category in which machine-learning models analyze source material and automatically suggest or apply dynamic band positions, threshold levels, and gain values. Tools like iZotope Neutron's Track Assistant and Sonible's smart:EQ 3 use source classification (vocal, drums, guitar, bass) to generate starting-point dynamic EQ curves that the producer then refines. Particularly useful for beginners building an intuition for frequency-reactive correction and for rapid workflow in high-volume production environments.
Frequency conflicts — two instruments in the same range at similar levels — are the root cause of muddy mixes.
These MPW articles put dynamic eq into practice — specific techniques, real tools, and applied workflows.