IRs vs Real Cabinets — What Actually Matters?

The definitive guide to impulse responses and speaker cabinets. Cut through the forum noise, understand the real science, and finally make the right decision for your rig. 

Deep Dive · Tone Science · 14 min read

In this article


Few debates in the guitar world generate as much heat and confusion as the war between impulse responses and real speaker cabinets. On one side: engineers and bedroom players who swear a great IR is indistinguishable from the real thing. On the other: purists who claim you can always tell, and that nothing replaces moving air.

Here's the honest truth — both camps are partially right, and understanding why will completely change how you approach your tone. This isn't a "try both and see what you like" non-answer. We're going to dissect the actual physics, so you can make smart decisions for your specific situation.

"An IR is a perfect photograph of a speaker cabinet at a single moment in time. A real cabinet is a living, breathing, pressure-sensitive organism."


What exactly is an IR?

An impulse response is a mathematical fingerprint of how a system responds to sound. To capture one, an engineer sends a test signal — typically a sine sweep covering the full audio spectrum — through a microphone placed in front of a real cabinet. The difference between what went in and what came out is the cabinet's "transfer function." That difference is saved as a short audio file, usually a WAV, a few hundred milliseconds long.

When you load an IR into a convolution plugin or your modeler (Quad Cortex, HX Stomp, Fractal, etc.), your dry amplifier signal is mathematically convolved with that fingerprint in real time. The processor asks: "if this audio had been played through that cabinet and microphone, what would it have sounded like?" The answer happens in milliseconds, thousands of times per second.

The IR block replaces everything that would normally come after the amp's power section: the speaker, the air between cone and microphone, and the microphone's own frequency response. That last part is crucial and often overlooked — a Royer 121 ribbon positioned three inches off-axis at the cap edge sounds radically different from a Shure SM57 dead-on the dust cap, and both of those are already baked into most commercial IRs.

Key concept: When you load an IR, you're not just loading a speaker simulation. You're loading a speaker + room + microphone + microphone position simulation. All of those variables are frozen in time from the moment of the original capture.


What a real cabinet actually does to your sound

Before you can intelligently compare IRs and real cabs, you need to understand what a physical speaker cabinet actually contributes. It's much more complex than most players realize.

Frequency response shaping

A guitar speaker is, by audiophile standards, a spectacularly bad transducer. A Celestion Vintage 30 rolls off dramatically above 5kHz. A Greenback peaks aggressively around 2–3kHz and then falls away. This is by design — guitar amplifiers produce copious upper harmonics that would sound shrill and fatiguing without the cabinet's built-in high-frequency rolloff. The speaker is the amp's final tone-shaping stage, and it's a radical one.

Non-linear compression and breakup

Here is where things get genuinely complicated. A real speaker cone is a physical object with mass, compliance, and inertia. When driven hard, its behavior becomes non-linear — the cone starts to compress, distort, and produce its own harmonic content. This "speaker breakup" or "cone cry" is a beloved element of classic rock and blues tones. It's partly why a 1960s Marshall on ten sounds different from the same amp at four through an attenuator.

Critically, this behavior is dynamic and input-dependent. The louder you play, the more the speaker departs from its linear, well-behaved self. An IR captures the speaker in a specific, typically mid-level linear state. It cannot model what happens when you dig in hard on a bend.

Cabinet resonance and air movement

The wooden box itself resonates. A closed-back 4×12 behaves differently from an open-back 2×12, not just because of the speaker complement, but because the cabinet's internal air volume, bracing, and material create their own resonant frequencies — typically felt more than heard, adding warmth and weight to the low-mids. You feel a real cabinet through the floor. That physical coupling is real and meaningful, particularly for how you perceive your own playing in real time.


The science: where IRs succeed and fall short

Linear time-invariant (LTI) systems theory is the backbone of convolution-based IR processing. The math works perfectly — for linear, time-invariant systems. The problem is that a guitar speaker driving a room is neither fully linear nor fully time-invariant.

Where IRs genuinely excel

For everything that is linear — the frequency shaping of the speaker, the frequency response of the microphone, the acoustic reflections of the room — an IR is a mathematically exact representation. There is no guesswork involved. If you record a speaker cabinet in a proper studio with a 251 condenser at 12 inches, and you load that IR into a good convolution engine, the frequency coloration of the recorded signal will be identical. This is why blind A/B tests between great IRs and the real cabinet (at matched SPLs, in a proper studio) are so difficult to distinguish. The EQ coloring is the same.

Where IRs fall short

The non-linear behaviors — speaker compression, thermal changes as the voice coil heats up, cone breakup, and the mechanical behavior of a speaker at high SPL — are not captured by an IR. These are dynamic behaviors that change with input level. They are also, to varying degrees, audible and musically significant. The question is: how audible, in your context?

Practical note: In a dense band mix at recording levels, speaker non-linearities contribute far less than most players expect. The frequency fingerprint of the cabinet dominates. In an isolated bedroom practice scenario where you're listening closely, the dynamic differences may be more apparent. Context is everything.

Some newer approaches try to address non-linearity. Dynamic IRs — used in plugins like the Neural DSP suite and certain Fractal algorithms — vary the convolution kernel based on input level, approximating some of the compression behavior. These are genuinely better than static IRs for capturing speaker dynamics, though they still model rather than physically reproduce the behavior.


Busting the biggest myths

"IRs sound digital and sterile" — This is almost always a symptom of a bad IR, a poor amp sim, or improper gain staging — not a fundamental limitation of IR technology. A well-captured IR through a properly driven preamp model is extremely difficult to distinguish from tape recordings of the real thing. The "digital" artifacts people hear are usually aliasing from poor capture or compression artifacts in low-quality IR packs.

"Real cabs always sound better" — Better for what? Real cabs sound better in the room at high volume, for feel and interaction with your playing. They also present significant challenges for recording — room acoustics, microphone technique, bleed, and time-of-day restrictions all become variables you need to manage. For a direct recording in an untreated room, a stellar IR frequently yields more consistently usable results than an average microphone placement on a real cab.

"Expensive IR packs are always better" — Price is correlated with capture quality, but diminishing returns set in quickly. A free IR captured by a skilled engineer in a well-treated room will outperform an expensive pack captured carelessly. The technique, the room, and the source material matter more than the price tag. Some of the best IRs available are free.

"All modelers handle IRs the same way" — They don't. The convolution engine quality, latency compensation, sample rate handling, and how gain staging is managed before the IR block all vary significantly between platforms. An IR that sounds fantastic in Cab Lab may sound underwhelming on a certain hardware unit that compresses or limits the signal differently before the cab block.

"You need to mic your cab a certain way to match an IR" — An IR already contains a mic position. When you use an IR, you're not adding mic behavior on top of a real speaker — you're replacing the entire speaker/mic/room chain. Many players confuse themselves by trying to use IRs while also running through a real speaker.


Head-to-head: recording vs live vs practice