Room correction explained: what software and DSP monitors can fix, and what they cannot

You buy decent monitors, set them up on your desk, and your mixes still fall apart in the car. The bass is huge on one song and missing on the next. Somewhere along the way someone tells you to buy room correction software, and it sounds like the answer: put a microphone where your head goes, press a button, and the room disappears.

Room correction is genuinely useful. It is also the most oversold tool in the home studio. This guide explains what it does under the hood, how to run a measurement that you can trust, and the hard limits that no algorithm gets around. If you have not yet thought about the walls themselves, start with our guide to acoustic treatment on a budget. This one picks up where that ends.

What room correction actually does

Every room correction system follows the same three steps.

1. Measure. A test signal, usually a sine sweep, plays through each speaker. A measurement microphone at your listening position records what arrives. The difference between what went out and what came back is the frequency response of the speaker and the room together.
2. Compute a filter. The software compares that response with a target curve (flat, or a gentle downward tilt) and calculates an inverse filter: cut where the room adds, boost where it subtracts, within limits.
3. Apply it. The filter runs either as software on your computer, usually a plug-in on the master bus or a system-wide app, or inside the speakers' own digital signal processing (DSP).

The tools sold to producers differ mainly in steps two and three. Sonarworks SoundID Reference runs on your computer and works with whatever monitors you own. Genelec GLM calibrates Genelec's own Smart Active Monitors and stores the result in the speakers, so the correction stays on even when no computer is involved. Room EQ Wizard, known as REW, is free, measures and analyses rather than correcting automatically, and is what many acousticians use to check everyone else's claims.

Why a room needs correcting at all

At low frequencies the wavelengths are as long as your room. Sound bounces between parallel walls and builds up standing waves, called room modes. At some frequencies the reflections add to the direct sound; at others they cancel it. Hugh Robjohns of Sound On Sound put numbers on it: depending on where you sit, reflections can lift the bass by up to 6 dB or cut it by 20 to 30 dB.

That asymmetry is the whole story of room correction. Peaks are easy. Cutting a 6 dB bump at one frequency is a normal EQ move, and the speaker has to work less, not more. Dips are a different animal. A deep null is not missing energy; it is energy cancelling itself. Boost it by 10 dB and you also boost the reflection that cancels it. You burn amplifier headroom and push the woofer harder, and the hole stays.

The better correction tools know this and cap how much they will boost. If your software lets you change that limit, leave it low.

A measurement you can trust, step by step

Most bad correction comes from bad measurement, not bad software. Here is a routine that works with any of the tools above.

1. Fix placement first. Move the speakers and desk before you measure. Keep the speakers symmetrical to the side walls, tweeters roughly at ear height, and the listening seat out of the dead centre of the room, where bass cancellation tends to be worst. Our guide to choosing studio monitors covers placement in more detail.
2. Use the right microphone. You need an omnidirectional measurement microphone with a calibration file for that specific unit. A vocal mic will measure the mic, not the room. Sonarworks sells its own measurement mic; REW works with any calibrated measurement mic.
3. Set a sensible level. Measure at roughly the volume you mix at. Too quiet and room noise pollutes the low end; too loud and you measure the speaker straining.
4. Point the mic at the ceiling unless the manual says otherwise. Each tool specifies orientation. Follow it exactly, because the calibration file assumes it.
5. Take several positions. Most tools measure a spread of points around your head, not just one. This averages out the very local peaks and nulls so the filter does not chase a dip that only exists one centimetre to the left.
6. Keep the room as it will be. Chair in place, door shut, the same clutter on the desk. A chair back moved after measurement changes the result.
7. Check the result with REW. After correction, measure again with the correction on. If the corrected curve is smoother through the low mids and the low end is no worse, keep it. If a deep dip is still there, that is your signal to move something, not boost something.

What correction cannot do

It corrects one seat. The filter is built for the listening position. Lean back, stand up or move to the sofa and the room's response changes, but the filter does not. That is why it pairs badly with a client couch behind you.

It cannot remove time. Correction mostly alters frequency response. A room that rings, where a kick decays for half a second after the speaker stops, still rings. You will hear smoother tone and the same smear. Only absorption, mostly thick bass traps, shortens decay.

It adds latency. Linear-phase filters, which change level without shifting phase, delay the signal. That is fine for mixing and annoying when you record or play a software instrument. SoundID Reference offers a zero-latency filter mode; switch to it, or bypass correction while you track.

It is only on when it is on. If the correction is a plug-in on your master bus, it must come off before you bounce. Forgetting is a real way to ship a master with a room curve baked in. System-wide apps and speaker DSP avoid that trap.

Where it fits in the order of operations

The order that gets the most from your money:

1. Place the speakers and the seat properly. Free, and often the biggest change.
2. Treat the room: corner bass traps, then panels at first reflection points.
3. Measure, then apply correction to trim what is left.
4. Learn the result on reference tracks you know well.

Run that order backwards, correction first, and you are using a filter to argue with physics.

What we think

Room correction is sold as a shortcut, and the marketing leans on the fear that you cannot trust your own ears. We think that framing is backwards. The point of correcting a room is not to make a measurement graph flat. It is to make your decisions travel, so that the bass you choose in your bedroom is the bass people hear in their earbuds and cars.

That changes how you should judge it. The test is not the after-graph. It is whether your mixes need fewer revision passes after you play them elsewhere. Give it a month of references before you decide.

It also explains why we are wary of correcting everything. A heavily corrected small room can feel sterile, and some producers find they make worse choices because the low end becomes something they look at on a curve rather than feel in the chair. Trimming the obvious peaks and stopping there is often the more musical result. Your room is part of how you hear, and the aim is to know its accent, not erase it.

If money is tight, spend it in this order: REW and a calibrated microphone first, so you can see the problem; then a pair of corner traps; then correction software if the measurement still says you need it. Plenty of rooms, once measured honestly, need a move and a trap more than a licence. And for a second opinion that ignores the room entirely, a good pair of headphones remains the cheapest reference there is; see our guide to mixing on headphones.