Place a brick wall at the end of a corridor and throw a ball at it. The ball stops dead. Now replace that wall with a thick curtain and throw again. The ball still slows down, but it pushes through, deformed by the resistance rather than halted by it. That distinction—between stopping and shaping—captures the fundamental difference between limiting and compression, even though both processes use nearly identical control parameters.

Producers and engineers routinely encounter compressors and limiters sharing the same interface paradigm: threshold, ratio, attack, release, makeup gain. This surface similarity breeds a persistent misconception that limiting is simply more compression—that you turn the ratio up and cross some invisible threshold into limiting territory. The reality is more nuanced and more consequential for your signal chain. The behavioral differences between these tools aren't just matters of degree. They reflect fundamentally different design philosophies about what should happen to audio dynamics.

Understanding those differences isn't academic. Misapplying a limiter where compression belongs destroys micro-dynamics that give a mix its sense of life. Reaching for a compressor where limiting is required leaves transient peaks unchecked, risking distortion in downstream conversion or broadcast processing. Each tool solves a different problem, and conflating them creates new problems that are often difficult to diagnose after the fact. What follows is a closer examination of why these tools diverge in practice despite converging in their control surfaces.

The most obvious parameter distinguishing compression from limiting is ratio, but the implications of high ratios are more profound than the numbers suggest. A compressor operating at 4:1 allows signal above the threshold to continue rising, albeit at a reduced rate—for every 4 dB of input above threshold, 1 dB passes through. The dynamic contour of the original signal survives, compressed but recognizable. The louder parts are still louder than the quieter parts. Musical intention is preserved in miniature.

A limiter operating at ratios of 20:1, 50:1, or true infinity:1 does something categorically different. It doesn't reshape the dynamic contour above threshold—it eliminates it. Signal that crosses the threshold is held at or imperceptibly above that ceiling regardless of how much energy is pushing from below. This isn't a gentler version of what compression does. It's a different operation entirely, closer to clipping than to dynamic range reduction in its mathematical behavior.

This distinction has cascading consequences for how program material behaves. A compressed drum bus retains the relative dynamic relationships between hits—a hard snare strike still reads as louder than a ghost note, just with less distance between them. A limited drum bus flattens everything above threshold to the same output level, fundamentally altering the performer's dynamic intention. The information isn't reduced; it's removed.

The knee parameter becomes critically important here. Soft-knee compression creates a gradual transition into gain reduction, which helps preserve the perception of natural dynamics even at moderate ratios. Most dedicated limiters employ hard-knee behavior by design, because the entire purpose is to establish an unambiguous output ceiling. When producers use a compressor with extreme ratios but soft-knee settings, they often get a hybrid behavior that neither truly limits nor transparently compresses—a source of confusion in many mixing scenarios.

Worth noting is how modern look-ahead limiters further separate themselves from compressors. By reading the signal slightly ahead of real-time processing, they can anticipate peaks and apply gain reduction before transients arrive, achieving true brickwall behavior without the artifacts that instantaneous gain reduction would cause. This architectural difference has no parallel in traditional compression design, underscoring that these are tools built around different engineering objectives.

Takeaway

Compression reshapes dynamics proportionally—the musical contour survives in reduced form. Limiting erases dynamics above a threshold absolutely. The difference isn't one of degree but of kind, and treating them as points on the same continuum leads to processing decisions that damage rather than serve the material.

Attack and release times on a compressor and a limiter might share the same label and even the same unit of measurement, but they serve different functions within each tool's operational logic. A compressor's attack time is fundamentally a creative parameter. Setting a slower attack on a drum compressor lets the initial transient punch through before gain reduction engages, preserving percussive impact while controlling sustain. Faster attacks catch more of the transient, producing a rounder, more controlled envelope. The engineer is sculpting the sound's temporal shape.

A limiter's attack time serves a protective function. It must be fast enough to catch the fastest transients in the program material before they exceed the ceiling. In mastering limiters, attack times routinely operate in the sub-millisecond range—often between 0.01 and 0.5 milliseconds. At these speeds, the limiter is responding to individual sample values rather than to the perceptual envelope of the sound. This is a fundamentally different scale of operation from a bus compressor working at 10-30 milliseconds.

The consequences of these different time scales are audible and significant. When a limiter's attack is too slow for the material, transients overshoot the ceiling—exactly the failure condition the tool exists to prevent. When a compressor's attack is set as fast as a limiter's, it clamps down on transients so aggressively that percussive material loses its spatial cues and impact, often introducing an unpleasant pumping artifact as the gain reduction constantly engages and releases at audio-rate speeds.

Release times diverge similarly. A compressor's release is another creative tool—fast releases create energy and excitement, slow releases produce smooth, transparent gain reduction. A limiter's release must be calibrated to avoid intermodulation distortion, a phenomenon where the gain reduction envelope itself becomes audible as a low-frequency artifact modulating the program material. Sophisticated limiters use program-dependent release algorithms that adapt to the spectral and dynamic content in real time, a complexity unnecessary in most compression scenarios.

The interaction between attack and release in limiters also introduces considerations around successive peak handling. When peaks arrive in rapid succession—as they do in dense modern productions—the limiter must recover from one gain reduction event fast enough to accurately catch the next peak without either overshooting or sustaining excessive gain reduction between events. This temporal precision requirement is why dedicated limiter algorithms are architecturally distinct from compressor circuits, even when the front-panel controls appear interchangeable.

Takeaway

A compressor's time constants are expressive tools for shaping the feel of dynamics. A limiter's time constants are engineering constraints that must be fast and precise enough to guarantee ceiling integrity. Using compressor-speed timing on a limiter invites overshoot; using limiter-speed timing on a compressor destroys musicality.

The clearest way to understand when each tool is appropriate is to identify what problem you're actually solving. Compression addresses dynamic range—the distance between the quietest and loudest moments in a signal. If a vocalist's delivery varies too much for a consistent mix position, compression narrows that range so quieter phrases sit closer to louder ones. The goal is balance and consistency within a musically acceptable dynamic window.

Limiting addresses peak level—the absolute maximum amplitude a signal is permitted to reach. If a mastered track must not exceed -1 dBTP for streaming delivery, a limiter enforces that ceiling with mathematical certainty. The goal isn't musical shaping but technical compliance. These are different problems requiring different solutions, and this is where misapplication causes the most damage.

Using a limiter for dynamic range management—setting a low threshold to pull down an entire vocal performance—produces a flat, lifeless sound because the tool eliminates rather than reshapes dynamic variation. The vocal becomes uniformly loud, stripped of the expressive micro-dynamics that convey emotion and intentionality. Conversely, using a compressor to control peak levels for broadcast compliance is unreliable—moderate ratios allow transients to exceed the target ceiling, potentially causing downstream distortion or normalization issues.

In practice, many professional signal chains use both tools in series, each solving its designated problem. A mix bus compressor at 2:1 or 3:1 gently shapes the overall dynamic feel of the mix, adding cohesion and controlling broad dynamic swings. A subsequent limiter with a brickwall ceiling ensures no peaks exceed the delivery specification. The compressor handles the musical problem; the limiter handles the technical one. Collapsing both tasks into a single processor—as happens when producers push a compressor to extreme ratios or set a limiter's threshold inappropriately low—compromises both objectives.

The emergence of multi-stage limiting algorithms in tools like those from iZotope, FabFilter, and Sonnox reflects this understanding. These processors internally separate the tasks of dynamic shaping and peak control into distinct processing stages, even within a single plugin interface. Recognizing that compression and limiting solve different problems isn't just theoretical clarity—it's the foundational logic behind the most sophisticated dynamics processing available today.

Takeaway

Before reaching for any dynamics processor, identify whether you're solving a dynamic range problem or a peak level problem. Compression shapes how dynamics feel. Limiting defines where dynamics stop. Conflating the two means neither problem gets solved properly.

The shared interface of compressors and limiters is a historical artifact of analog hardware economics, not a reflection of shared purpose. These tools diverge in ratio behavior, time constant requirements, and appropriate application scenarios in ways that matter every time you instantiate one on a channel strip or master bus.

Treating limiting as extreme compression isn't just imprecise—it leads to processing decisions that damage the very qualities you're trying to control. The transient information that compression preserves in miniature, limiting removes entirely. The creative latitude that compression's time constants offer, limiting's speed requirements foreclose.

The discipline here is diagnostic before it's technical. Identify the problem—dynamic range or peak level—and reach for the tool designed to solve it. That clarity of purpose is what separates processing that serves the music from processing that merely acts upon it.