Selective Leveling in Action
The most powerful mixing tool isn’t a plugin. It’s the fader.
Before you reach for an EQ or compressor, ask: can this be solved with a level change? A vocal that’s buried doesn’t need more high-end boost — it might just need to be louder. A synth pad that’s overwhelming the mix might just need to come down 2 dB. The simplest solution is almost always the best.
Remember the selective leveling framework from Chapter 7: what (frequency), when (dynamics), where (space). The fader is the blunt instrument — it changes everything at once. EQ, compression, and panning are the precision tools. This chapter is about knowing when the blunt instrument is enough and when you need to get specific.
Selective leveling in practice means working through a hierarchy:
-
Balance the 2-mix first. Get the overall level relationships right using just faders. No plugins. No processing. Just volume. This is where you establish the recipe — how much of each ingredient.
-
Identify problems that faders can’t solve. A vocal that’s the right level but sounds muddy. A kick that’s loud enough but doesn’t cut through the bass. These are frequency or dynamics problems — they need tools beyond the fader.
-
Apply the minimum effective processing. An EQ cut to remove the mud. A compressor to tame the kick’s inconsistent dynamics. Each move should solve a specific problem you identified. If you can’t name the problem, don’t add the plugin.
-
Re-check the levels after processing. Every plugin changes the level. A 3 dB EQ boost adds 3 dB of level. A compressor changes the dynamic envelope. After each move, re-balance the faders. Processing and leveling are an iterative loop, not sequential steps.
Mixing hierarchy flowchart: balance faders → identify problems → apply minimum processing → re-check levels → loop.
The Listen-Compare-Act Loop
Good mixing is mostly listening and deciding, with occasional brief actions. The mixer isn’t constantly tweaking. They’re listening, comparing to the reference, identifying one thing to change, making the change, then listening again.
That rhythm — listen, compare, identify, act, listen — is the entire method. If you internalize that loop, you can mix anything. The specific tools (EQ, compression, effects) are just the vocabulary. The loop is the grammar.
Circular workflow diagram: Listen → Compare (to reference) → Identify → Act → Listen.
One related concept: don’t mix with your eyes. It’s tempting to stare at the waveform, watch the spectrum analyzer, check the gain reduction meter. But your meters confirm what your ears decide — they don’t replace them. Close your eyes sometimes. Listen. The mix lives in the air between your speakers, not on the screen.
When two instruments fight for the same frequency space — when one masks the other — the fix is usually a combination of level and frequency targeting. Bring one down, carve a small EQ pocket in the other, or both. The Drop and Cut technique article demonstrates this in detail.
Level Matching for Honest Comparison
Every comparison in mixing must be level-matched. This applies to:
- Reference tracks (Chapter 10): match your mix level to the reference level before A/B-ing
- Before/after processing: bypass a plugin and listen. If the processed version is louder, turn it down until it matches the bypassed signal. Then compare. If it still sounds better, the processing is working. If it sounds the same, you just made it louder — not better.
- Different mix versions: when comparing v1 to v2, match the levels. The louder one will always sound “better” even if it isn’t.
The louder-is-better illusion (Chapter 8) undermines every mixing decision if you don’t control for it. Level-matching is not optional — it’s the foundation of honest evaluation.
Identifying Specific Qualities
When you A/B against a reference, don’t just think “theirs sounds better.” That’s not actionable. Identify the specific quality that’s different:
- “Their low end is tighter — my bass is boomy below 100 Hz”
- “Their vocal has more presence — mine is being masked by the guitars in the 2-5 kHz range”
- “Their mix has more width — my instruments are all clustered in the center”
- “Their dynamics breathe more — my mix feels flat and lifeless”
Each of these observations leads to a specific action. Boomy bass → high-pass filter or EQ cut. Masked vocal → carve a notch in the guitars. Narrow width → panning adjustments. Flat dynamics → less compression or more dynamic arrangement. The specificity is what makes the comparison useful.
When to Stop
A mix is never finished — it’s abandoned. At some point, the returns from additional tweaking become vanishingly small. You’re moving the vocal up half a dB, then back down, then up again. You’re swapping reverb plugins for the third time. You’ve lost perspective.
Signs you should stop:
- You’re making changes you can’t hear without soloing. If the change isn’t audible in the context of the full mix, it doesn’t matter.
- You’re undoing what you did ten minutes ago. You’re going in circles.
- You’ve been working for more than 2-3 hours without a break. Your ears are fatigued. Take a walk. Come back tomorrow.
- The mix meets your reference goals. You set targets (Chapter 10), you hit them. Export.
Do a final check on multiple systems — earbuds, laptop speakers, car stereo if possible. If it works everywhere, it’s done. If it doesn’t, note the specific problems, go back and fix those specific things, then check again. Don’t reopen every decision.
What to Practice
- Mix with faders only. Open a multitrack session and mix it using nothing but faders and pan knobs. No EQ, no compression, no effects. Get the balance as good as you can. This teaches you how far level relationships alone can carry a mix — and where their limits are.
- Level-match a before/after. Add an EQ or compressor to a track. Bypass it. Match the levels. Now compare. Did the processing actually improve the sound, or did it just get louder?
- Set a timer. Give yourself one hour to mix a song. At the end of the hour, export it — no extensions, no “just one more thing.” Listen back the next day. Notice what matters and what doesn’t. The time constraint builds decisiveness.
- Write down three observations from a reference comparison. Not “theirs is better.” Three specific, actionable differences. Address the most important one. A/B again. Repeat.