/ˈreɪ.ʃi.oʊ/
Ratio is the number that determines how many dB of input signal above the threshold results in 1 dB of output gain increase. A 4:1 ratio means every 4 dB over the threshold produces only 1 dB of output level increase.
Ratio is the single number that separates a mix that breathes from a mix that suffocates — get it wrong by even a few notches and a snare that should crack like a whip instead thuds like a wet newspaper.
In dynamic signal processing, ratio defines the relationship between how much a signal exceeds a compressor's threshold and how much gain reduction is applied in response. It is expressed as a proportion — input change versus output change — and written in the form X:1, where X represents the number of decibels the input must rise above the threshold to produce a single decibel of increase at the output. A ratio of 4:1 therefore means that for every 4 dB the signal climbs above the threshold, the compressor allows only 1 dB of that increase to pass through, reducing the remaining 3 dB as gain reduction. This one parameter, more than attack or release, defines the character of a compressor's response and the ultimate loudness ceiling it imposes on a signal.
Ratio works in concert with the threshold, attack, release, and makeup gain parameters but it operates on a fundamentally different axis: while threshold determines when compression begins and attack/release determine how fast it responds, ratio determines how hard the compressor squeezes once it is engaged. A low ratio — 1.5:1 or 2:1 — is so gentle it barely alters perceived dynamics; it is most audible on instruments with wide natural dynamic range, such as orchestral strings or acoustic guitar. A high ratio — 8:1, 10:1, 20:1 — dramatically flattens transients and sustains, useful for taming erratic bass performances or aggressively shaping drum room ambience. Above approximately 10:1, the device crosses the informal industry threshold into limiting territory, where the output is effectively capped at or just above the threshold level regardless of input level.
The perceptual effect of ratio is inseparable from its mathematical action. A signal compressed at 2:1 with a high threshold may sound completely unprocessed because only the loudest peaks are touched and they are barely attenuated. The same ratio applied with a low threshold — set so that the compressor is active almost continuously — produces a noticeably dense, pumped, or hyper-present quality because the entire dynamic envelope is being continuously shaped. This interplay explains why experienced producers rarely set ratio in isolation; they adjust it while watching gain reduction meters and listening to how much of the original transient energy is surviving, a technique sometimes called riding the knee.
From a mathematical perspective, ratio is a slope applied to the gain transfer function of the compressor, expressed in dB-in versus dB-out space. Below the threshold, the slope is 1:1 — every decibel of input produces exactly one decibel of output, meaning no compression is occurring. Above the threshold, the slope flattens to 1/X, where X is the ratio setting. A ratio of ∞:1 (infinity-to-one) makes the slope exactly horizontal above the threshold: no matter how loud the input gets, the output does not rise at all — this is the mathematical definition of a true brick-wall limiter. Understanding this transfer function geometry helps producers predict how a compressor will behave with unusual program material, such as a sine wave sweep or a full-band mix buss, and explains why certain compressors with soft-knee designs apply a gradual transition between 1:1 and the set ratio rather than a hard corner, resulting in a more natural, less artifact-prone sound.
Every compressor contains a gain-reduction element — a VCA, an opto cell, a FET transistor, or a tube-based vari-mu stage — that reduces the signal's amplitude when the sidechain detector detects a level above the threshold. Ratio is the instruction that tells this element exactly how steeply to attenuate. Internally, the ratio is computed from the difference between the instantaneous input level (measured after the detection filter and before the gain element) and the threshold value. If the signal is 8 dB over the threshold and the ratio is set to 4:1, the compressor calculates that the output should rise by only 2 dB above the threshold level (8 ÷ 4 = 2), and the gain-reduction element therefore attenuates the signal by 6 dB at that moment. This calculation is continuous and updating at audio rates on modern digital compressors, though analog circuits approximate it with time-domain charge/discharge behaviors that introduce pleasing nonlinearities.
The knee setting modifies how ratio is applied around the threshold boundary. With a hard knee, the full ratio engages the instant the signal crosses the threshold — there is a sharp corner in the transfer function curve. With a soft knee, the compressor gradually transitions from 1:1 to the full ratio across a range of, typically, 10 to 20 dB centered on the threshold. The practical result of a soft knee is that the onset of compression is gentler and less perceptible: material that hovers near the threshold — like a vocalist who drifts between loud and soft phrases — sounds more natural because there is no abrupt shift in dynamic character. Hard-knee settings are preferred on transient-heavy material like drums, where the clear demarcation between uncompressed and compressed signal helps preserve the attack phase of hits while controlling their sustain.
Peak versus RMS detection also interacts critically with ratio. A peak-detecting compressor measures the instantaneous maximum amplitude of the signal and applies ratio-based gain reduction based on that peak value. This makes it fast and suited to transient limiting. An RMS-detecting compressor averages the signal's power over a short window — typically 50 to 300 ms — and bases gain reduction on this average, which more closely matches how human hearing perceives loudness. An RMS compressor with a ratio of 4:1 will behave less aggressively on a snare hit and more aggressively on sustained vocal phrases than a peak compressor with the same ratio, because the snare's peak is brief enough to sit partially inside the integration window while the vocal's sustained energy accumulates. Producers should always consider detection mode when interpreting ratio values, since the same numeric ratio can produce radically different amounts of audible compression depending on how the detector is configured.
In parallel compression — a technique popularized on the New York drum bus — the ratio setting on the wet compressor is typically very high (8:1 to ∞:1) precisely because the heavily compressed signal will be blended with the uncompressed dry signal. The blend attenuates the harsh artifacts that an extreme ratio would otherwise introduce when heard in isolation: the attack transients are preserved by the dry path while the sustain and density contributed by the compressed path fill out the low-level content. Understanding ratio as one axis of a multidimensional dynamic design space — where parallel blend, attack, release, knee, and makeup gain all interact — is what separates producers who use compression intuitively from those who simply copy presets.
A final note on units and scale: ratio is dimensionless and linear in its X-to-1 expression, but its perceptual effect is not linear. The audible difference between 2:1 and 4:1 is considerably larger than the difference between 8:1 and 10:1, because gain reduction is already severe by 8:1 and further increases in ratio push toward a ceiling that is already nearly flat. This is why most compressor hardware spaces the high-ratio numbers more closely together on the control knob — the increments of perceptual change grow smaller. Producers mixing by ear should remember that doubling the ratio does not double the perceived compression effect; the relationship is logarithmic in practice.
Diagram — Ratio: Compressor transfer function showing four ratio curves (1:1, 2:1, 4:1, ∞:1) and the threshold point on a dBin vs dBout graph.
Every ratio — hardware or plugin — operates on the same core parameters. Know these and you can work with any implementation.
Expressed as X:1, ratio defines how many dB of input level increase above the threshold produce 1 dB of output level increase. Practical values range from 1.1:1 (barely audible compression) to ∞:1 (hard limiting). For general-purpose tracking compression, 3:1 to 4:1 covers most use cases without inducing pumping artifacts.
Threshold sets the input level — in dBFS or dBu — at which the compressor's ratio curve departs from the 1:1 unity line. A lower threshold means the ratio acts on more of the signal for more of the time. Setting threshold so the gain reduction meter reads 3–6 dB on peaks is a practical starting point for transparent dynamic control.
Knee controls whether the ratio engages abruptly (hard knee, 0 dB width) or gradually over a range of levels approaching the threshold (soft knee, typically 6–20 dB width). Soft knees sound more natural on vocals and acoustic instruments; hard knees are preferred when precise transient control is needed, such as on kick drums or limiting stages.
Because ratio-based gain reduction lowers the signal's average and peak level, makeup gain (also called output gain) is applied post-compression to restore perceived loudness to a level comparable with the unprocessed signal. Without makeup gain, higher ratios will always sound quieter and therefore subjectively worse — a classic source of A/B bias. Some compressors offer auto makeup gain, which is convenient but should be understood as an approximation.
Detection mode determines whether the ratio is applied based on the instantaneous peak amplitude (peak detection) or the averaged energy over a short window (RMS detection). Peak detection is more responsive and suited to transient limiting; RMS detection tracks loudness more perceptually and tends to produce denser, more sustained gain reduction at the same ratio setting. Many hardware compressors have a fixed detection mode that is part of their sonic character.
Attack does not change the ratio itself but controls how quickly the full ratio-based gain reduction is reached after the signal crosses the threshold. A fast attack (0.1–1 ms) allows the ratio to engage almost immediately, clamping transients. A slow attack (30–100 ms) lets the transient pass uncompressed before the ratio kicks in, preserving punch. The interplay of attack and ratio determines how much of a drum's impact survives compression.
Session-ready starting points. These are starting points calibrated for modern mixing contexts; always adjust threshold and makeup gain in tandem with ratio to achieve target gain reduction.
| Parameter | General | Drums | Vocals | Bass / Keys | Bus / Master |
|---|---|---|---|---|---|
| Light/Transparent | 1.5:1 – 2:1 | 2:1 (room mic) | 1.5:1 – 2:1 | 1.5:1 – 2:1 | 1.5:1 – 2:1 |
| General Purpose | 3:1 – 4:1 | 4:1 (overheads) | 3:1 – 4:1 | 3:1 – 4:1 | 2:1 – 3:1 |
| Aggressive/Character | 6:1 – 8:1 | 6:1 – 8:1 (snare) | 6:1 – 8:1 | 6:1 – 8:1 (bass) | 4:1 (parallel) |
| Hard Limiting | 10:1 – ∞:1 | 10:1 (kick peak) | 10:1 – ∞:1 (DI) | 10:1 (synth bass) | ∞:1 (clip limiter) |
| NY Parallel Compression | 8:1 – 20:1 | 10:1 – 20:1 | 8:1 – ∞:1 | 8:1 – 10:1 | 4:1 – 8:1 |
| Typical GR at Setting | 2–4 dB | 3–8 dB | 3–6 dB | 4–8 dB | 1–3 dB |
These are starting points calibrated for modern mixing contexts; always adjust threshold and makeup gain in tandem with ratio to achieve target gain reduction.
The concept of automatic gain control — the engineering ancestor of the ratio parameter — dates to the mid-1920s, when radio broadcast engineers at AT&T and the BBC needed a way to prevent transmitter overload as performers moved toward and away from microphones. These early AGC circuits had no adjustable ratio; they simply slammed gain down hard whenever levels exceeded a threshold. The Fairchild 660 and 670, designed by Rein Narma and introduced in 1959, were among the first devices to offer operator-configurable compression behavior through a six-position time-constant switch, which varied the effective attack and recovery response. True continuously variable ratio controls would emerge in the following decade as VCA (voltage-controlled amplifier) technology matured.
The dbx 160, introduced in 1971 and designed by David Blackmer, was instrumental in popularizing the ratio knob as a discrete, front-panel control alongside separate threshold and output gain knobs. Blackmer's design used his proprietary RMS detection and a true VCA gain element to produce a mathematically clean compression curve — the first widely available compressor where the transfer function closely matched its labeled ratio setting across the full dynamic range. Engineers including Bob Clearmountain used the dbx 160 extensively in the late 1970s and early 1980s, notably on drum buses for artists including Bruce Springsteen and the Rolling Stones, where ratios of 4:1 to 6:1 were used to achieve the dense, punchy drum sound that defined the era.
The UREI 1176LN, introduced in 1967 and designed by Bill Putnam Sr., offered four fixed ratio positions — 4:1, 8:1, 12:1, and 20:1 — selected by four buttons on the front panel. The notorious "all-buttons-in" mode, discovered accidentally by engineers who pressed all four ratio buttons simultaneously (which the circuit was not designed to do), produced an unusual compression behavior that has since been estimated to result in a ratio somewhere between 12:1 and 20:1 with an altered release characteristic, creating an extremely aggressive, almost explosive transient behavior. This setting was used on John Lennon's vocals on several Imagine sessions and appears on countless rock recordings from the 1970s onward. The 1176 established the idea that discrete ratio selections with distinct sonic characters were preferable to continuous knobs for certain creative applications.
The SSL G-Bus compressor, introduced on the SSL 4000 G console in the early 1980s, shifted ratio usage toward bus and master processing. With ratio options of 2:1, 4:1, and 10:1, the G-Bus became the defining tool for mix-bus compression at 2:1 to 4:1, gluing together multi-track mixes with gentle, consistent gain reduction. Mix engineers including Andy Wallace, Chris Lord-Alge, and Michael Brauer built their signature mix compression approaches around the G-Bus, and its characteristic 2:1 ratio with a medium attack and auto release became the default starting point for bus compression through the 1990s and 2000s. The proliferation of digital audio workstations from the late 1990s brought the ratio parameter to millions of producers through bundled software compressors, though the conceptual vocabulary — and the starting-point ratios — remained anchored to hardware conventions established by these earlier designs.
Drums: On individual drum hits — kick, snare, toms — producers typically use higher ratios (6:1 to 10:1) with fast attacks and medium releases to enforce a consistent dynamic ceiling without losing the transient energy that defines the hit's character. A common approach is to set the ratio to 8:1, then use a slow enough attack (15–30 ms on a snare) that the initial crack passes uncompressed before the compressor engages. On overhead and room microphones, much lower ratios (2:1 to 3:1) are preferred because the cymbal content and room wash have complex, sustained dynamics that heavy compression turns unpleasantly static and harsh.
Vocals: Vocal compression is almost always applied in at least two stages, and ratio plays a different role in each. The first stage — often a peak limiter or hard-knee compressor at 8:1 to 10:1 — catches extreme dynamic spikes from syllables, breath sounds, and performance inconsistencies. The second stage — a gentler 3:1 to 4:1 VCA or opto compressor — provides consistent, musical gain shaping across the phrase. Using two moderate stages rather than one extreme stage produces less total distortion and avoids the pumping artifacts that a single high-ratio compressor generates when working hard. Classic opto compressors such as the LA-2A, which has a fixed program-dependent ratio that averages approximately 3:1 to 4:1 depending on input level, remain preferred vocal tools precisely because this behavior is built into their circuit architecture.
Bass and Low-Frequency Instruments: Bass guitar and synthesizer bass lines combine sustained low frequencies with percussive transient attack; ratio choices must balance both. A ratio of 4:1 to 8:1 with a moderate attack (5–10 ms) is a common starting point, allowing the pick or pluck transient to pass before compression engages to level out the note-to-note dynamic variation. High ratios (10:1 and above) on bass create a very dense, almost sample-like consistency — useful in electronic music and hip-hop where the bass needs to lock tightly to the kick grid — but at the cost of losing the natural performance feel of acoustic or DI-recorded instruments.
Bus and Master Processing: The most consequential ratio decision in a modern mix is on the stereo bus, where ratios of 1.5:1 to 4:1 are standard. The goal of bus compression is not gain reduction per se but cohesion — the phenomenon where compressing all elements simultaneously through a shared gain-reduction element causes them to breathe together, reinforcing the sense that the mix occupies a single acoustic space. Ratios above 4:1 on the bus risk audible level pumping in time with the kick drum or other dominant transient elements, which is perceived as a rhythmic artifact rather than a stylistic choice unless specifically intended. The 2:1 ratio on the SSL G-Bus running 2–4 dB of gain reduction is a widely documented reference point for this purpose across rock, pop, and R&B production.
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 ratio used intentionally, at specific moments, for specific purposes.
The bass guitar performed by Nathan East and the rhythm guitar are compressed with what engineers at Electric Lady Studios described as moderate ratio settings (3:1 to 4:1) through a UA 1176, chosen to preserve the organic, lightly grooved feel of the performances rather than creating a modern hyper-compressed bass texture. Listen to the way individual notes retain their pick attack while the sustain remains consistent across the low register. The drum bus compression at approximately 2:1 creates the sense that all elements of the mix occupy the same physical space without audible pumping.
The drum samples are crushed with high-ratio compression (8:1 to 10:1 range, likely through a dbx 160 or similar VCA) to produce the hyper-dense, almost distorted snare and kick textures that give the track its aggressive urgency. Compare the kick in the verses to the brief moments of silence — the makeup gain compensating for the ratio's gain reduction creates an almost violent contrast between compressed hits and silence. The vocal also sits in a heavily compressed (6:1 to 8:1) space that keeps Jay-Z's dynamic delivery consistently forward in the mix.
Finneas has discussed applying intentionally visible compression to Billie's vocal chain, using a ratio that reaches into the 6:1 to 8:1 range on peaks to push the vocal into a confined, intimate dynamic space consistent with the bedroom-pop aesthetic. The result is a vocal that barely varies in perceived loudness across the verse despite her naturally wide dynamic range as a performer. The sub bass line is similarly high-ratio compressed, producing a locked, immovable low end that sits precisely on the grid with no note-to-note variation in level — a signature of modern pop production.
The 808 bass on HUMBLE. exhibits textbook high-ratio compression behavior: the sub hits with a consistent, flat dynamic envelope suggesting a ratio of 10:1 or higher on the bass bus, likely combined with a limiter. The hi-hat pattern, by contrast, sounds relatively uncompressed with audible velocity variation intact — producer Mike WiLL Made-It has described leaving the high-frequency percussion elements with minimal processing to create textural contrast against the crushed low end. Listen on headphones for the moment the 808 enters at 0:13; the suddenly flat, dense sub is the clearest ratio artifact in the track.
Voltage-controlled amplifier compressors apply ratio with mathematical precision and very fast response times, making them the most 'accurate' implementation. The ratio setting on a VCA compressor closely matches the actual gain reduction slope measured on a transfer function analyzer. VCA compressors at high ratios (8:1 and above) can introduce audible distortion on bass-heavy program material due to the speed and precision of their gain reduction element.
Opto compressors use a light-dependent resistor (LDR) whose resistance changes with the light output of an electroluminescent panel driven by the sidechain signal. The nonlinear response of the LDR means that ratio is not a fixed value but varies with signal level and history — lower-level signals experience a gentler effective ratio while louder signals engage a steeper slope. This program-dependent behavior produces a self-adapting compression characteristic widely regarded as particularly natural and musical on vocals and acoustic instruments.
FET (field-effect transistor) compressors use the nonlinear behavior of a FET in its gain element, producing a fast, slightly aggressive character that adds harmonic excitement at higher ratios. The 1176's four ratio positions (4:1, 8:1, 12:1, 20:1) each have a subtly different sonic color beyond the ratio value itself, due to how the FET's operating point changes. The 20:1 setting approaches limiting behavior and is commonly used on drum room microphones to create a heavily squashed, ambient effect.
Vari-mu compressors use a variable-bias vacuum tube as the gain element; as the tube is driven harder by the sidechain control voltage, its gain decreases in a logarithmic manner. The result is a soft, program-dependent ratio that increases as the signal exceeds the threshold by greater amounts — gentle at moderate overdrive, progressively steeper as levels climb. This behavior is considered ideal for mix bus and mastering applications where transparent, musical compression is paramount.
Digital compressors implement ratio as a precise mathematical slope in the gain computer, with no circuit-derived nonlinearities unless specifically modeled. This allows them to exactly realize any ratio from 1.01:1 to ∞:1 and to combine ratio with features like lookahead detection, mid-side processing, and frequency-selective sidechain filters in ways impossible in analog hardware. The tradeoff is that the absence of circuit color means the ratio behaves exactly as labeled — which is useful for mastering but can feel sterile in creative tracking contexts.
These MPW articles put ratio into practice — specific techniques, real tools, and applied workflows.