Mid-Side Mastering: 10 Steps to a Complete Mix

Mid-side processing lets you treat the centre of your stereo mix and the edges as two separate signals. You get surgical control that left/right processing can’t touch. The mid channel carries your kick, bass, lead vocal, and snare. The sides carry width, room, and stereo instruments. Process them together and you’re blunt. Process them apart and you’re precise. Here’s exactly how to do it.

Step 1, Understand the M/S Matrix Before You Touch a Fader

Mid-side isn’t a plugin mode. It’s a conversion. Your stereo signal (L and R) gets encoded into sum and difference components.

Mid = L + R. Side = L minus R.

Alan Blumlein figured this out in 1933. The maths hasn’t changed. When you decode back, you get stereo again. Any imbalance you introduce in the M/S domain shows up as phase and width changes in the final stereo output.

This is why reckless M/S work creates mono compatibility problems. The side signal only exists in stereo. Cut the sides aggressively and your track collapses on a mono speaker. Boost the sides and it falls apart on earbuds.

Before touching anything, drop Voxengo MSED at the end of your chain. Toggle between mid-only and side-only monitoring. Listen to what’s actually in each channel. That reconnaissance will save you from processing blindly.

Step 2, Set Up Your Monitoring Correctly

Frustrating fact: most engineers skip this step and then wonder why their masters sound different on different systems.

Check your mix in mono first. Then in stereo. Then in M/S solo mode using MSED or your EQ’s solo function. You need a baseline.

Set your metering plugin to show stereo correlation. A reading above +0.5 is healthy. Below 0 means phase problems already exist in the mix you’ve been handed. If the mix arrives broken, M/S processing won’t fix it. Flag that to the mix engineer before you start.

We’ve taken in mixes where the correlation was reading -0.3 before we touched a single control. M/S EQ on top of that made everything worse. Catching it early saved the session.

Step 3, Apply a Low-Cut to the Side Channel First

This is the one move every master benefits from. Low-frequency energy in the sides creates pumping, phase weirdness, and translation problems on club systems that sum to mono.

In your M/S EQ, switch to side channel only. Apply a high-pass filter at 100Hz with a 12dB/octave slope. Some engineers push this to 120Hz on dense electronic mixes.

Don’t go steeper than 24dB/octave here unless you’re deliberately shaping. 12dB sounds more transparent on program material.

This single cut makes your low end hit harder on mono speakers. The kick and bass stay locked in the centre. The width in your mid-range and highs is unaffected. It’s the most satisfying M/S move you’ll make.

Step 4, EQ the Mid Channel for Clarity and Focus

The mid channel is where most of the energy lives. Be conservative. Maximum 3dB of boost or cut. Wide Q settings only: 0.6 to 1.2 for anything corrective.

Common moves:

A gentle shelf cut at 300-400Hz in the mid by 1-2dB clears the boxy buildup that crowds vocal intelligibility. A +1dB shelf boost starting at 8kHz in the mid adds air to the vocal without blasting the room sound on the sides.

We had a mastering session where the mix had a honky 350Hz build in the lead vocal. Pulling 1.5dB at 350Hz with a Q of 0.8, mid channel only, cleaned it without touching the drum room bleed sitting in the sides. That’s a move you can’t make in L/R mode.

Step 5, EQ the Side Channel for Width and Dimension

The sides carry your reverb tails, stereo synths, room mics, and panned guitars. This is where the perceived space of your master lives.

A gentle boost in the 2-7kHz region on the sides, 0.5 to 1dB maximum with a Q of 1.0, opens up the stereo image without sounding hyped. Keep it under 1dB. Seriously. We’ve heard engineers push this to 3dB and the mix starts to feel unnatural on headphones.

A slight high-frequency roll-off on the sides above 16kHz can tame harshness if the mix was mixed bright for streaming. Use a shelf at -0.5dB. Subtle.

Check mono compatibility after every side EQ move. Toggle mono in MSED. If the mix changes character significantly, you’ve gone too far.

Step 6, Apply Mid-Side Compression Strategically

M/S compression is underrated and mostly misused. The two applications that actually work:

First, light mid compression to control dynamic vocal energy without touching the stereo width. Try a ratio of 2:1, attack of 30-50ms, release of 150-250ms. Aim for 1-2dB of gain reduction maximum.

Second, side compression to control width pumping on dynamic mixes. Ratio of 1.5:1. Attack slower: 60-80ms so transients pass through. Release of 200-300ms. You want the sides to breathe, not clamp.

Faster attack times on the side compressor create an ugly mono-pumping effect on loud transients. We’ve tested it. It sounds like the mix is collapsing inward on every kick hit. Annoying in a way that’s hard to diagnose if you’re not monitoring it properly.

Step 7, Use M/S Saturation Sparingly

A touch of saturation on the mid channel only adds harmonic density to the centre image. We’re talking 0.1 to 0.3% THD equivalent. Nothing audible as distortion. Just presence.

Brainworx bx_digital V3 has mid-only saturation built in. Dial it in with the solo button active, then compare with the full stereo signal. If you can clearly hear the saturation with the mix playing at full volume, back it off.

Never saturate the side channel. The harmonic content it adds smears into the stereo image and creates correlation artifacts. Hard pass.

Step 8, Check the Width Balance with a Metering Plugin

ADPTR Metric AB and iZotope Insight 2 both show stereo width visually. Check your processed master against the reference track.

A healthy pop or electronic master typically sits between 30-50% width on a vector display. Wider isn’t better. Wider is just wider. The question is whether the width is intentional and controlled.

If your sides are louder than your mid in RMS terms, the mix will feel thin on mono playback. If the correlation drops below +0.3 after your processing, you’ve introduced phase problems. Neither is good.

Step 9, Compare in Mono, Stereo, and on Headphones

This is where you audit the full picture. Three passes minimum.

First: mono reference, A/B your processed master against the dry mix. The tonality should be close. The mid should feel cleaner. The low end tighter.

Second: stereo, at the same loudness. Check that the width feels natural, not exaggerated.

Third: headphones. M/S processing affects headphone listening differently than speakers. Boosted sides on headphones can sound like the mix is coming from outside your skull in an ugly, disorienting way. If it sounds brilliant on speakers and ridiculous on headphones, back the side boosts off by half.

Step 10, Render and A/B the Full Master

Bounce both versions: one with your M/S processing and one bypassed. Match the loudness with your metering plugin. Loudness bias is real and it’ll fool you every time.

Listen back tomorrow. Fresh ears catch things session ears miss. We’ve approved masters that sounded great at 2am and felt overcorrected the next morning.

If the M/S processing made a difference you can clearly hear, make sure that difference actually serves the track. Remember: on average, maybe 1 in 10 masters genuinely needs M/S EQ. Most mixes just need clean gain staging and good limiting. M/S is a precision tool, not a default step.

Pro Tips

Use linear phase EQ for M/S processing on the master bus, not minimum phase. Minimum phase EQ introduces frequency-dependent phase shifts that compound at the sum and difference stage. Linear phase mode costs more CPU and adds latency, but the phase integrity on a final master is worth it. FabFilter Pro-Q 3 and Ozone 11 both offer this switch per band.

Keep a dedicated M/S decoder after your M/S processing chain. Voxengo MSED in decode mode at unity gain is a zero-cost insurance policy. It shows you the true stereo output and catches any unintentional encoding artifacts.

Automate nothing in M/S processing on a master. Static settings only. Dynamic M/S changes across a track create width modulation that sounds like an effect, not mastering.

Use the Brainworx bx_digital V3 for complex M/S work if you want character alongside control. For clean, transparent M/S work, FabFilter Pro-Q 3 in M/S mode is the cleaner tool.

Common Mistakes to Avoid

Boosting the sides by more than 3dB in any region. We’ve heard this destroy mono compatibility on tracks that sounded fine in the studio. The mix implodes on Bluetooth speakers and phone speakers. Keep side boosts below 1.5dB, prefer 0.5-1dB.

Using minimum phase EQ for M/S mastering work. The phase rotation compounds across the matrix decode. This is a workflow detail almost no tutorial addresses directly. It creates subtle smearing in the stereo image that you’ll hear on high-quality playback systems even if it passes casual listening.

Skipping the mono check. Every single time. It’s not optional.

Applying M/S compression with attack times below 15ms on the side channel. The result is a pumping, narrowing effect on transients that sounds like a defect. Slow attacks on side compression. Always.