Compression Translation & Pitfalls

Going deeper

The loudness war — what happened, and why it ended

From the late 1990s through the 2000s, the music industry waged what's now called the "loudness war." Mastering engineers competed to make CDs as loud as physically possible — heavy compression, brick-wall limiting, and minimum dynamic range. The reasoning: louder records seemed to sound "better" on first listen, and labels believed perceived loudness drove sales.

The result was 15 years of mixes with crushed dynamics, fatiguing listening experiences, and ironically — songs that all sounded the same volume because everyone hit the ceiling. Albums like Death Magnetic by Metallica and Vapor Trails by Rush became case studies in how heavy compression destroyed music.

The war ended around 2014 when streaming platforms started loudness normalization. Spotify, Apple Music, YouTube, Tidal — all of them analyze every track and play it back at a target loudness (around −14 LUFS). A track mastered at −6 LUFS gets turned down by 8 dB to match. The old "louder is better" arms race no longer wins; quiet, dynamic mixes now play back at the same perceived level as loud ones, and listeners can hear the difference clearly: dynamic mixes have more punch and movement at the same playback volume.

The implication for your mixing decisions: aggressive bus compression and heavy limiting no longer add value on streaming. They just remove dynamic range without making the mix louder when it actually plays. Aim for −14 LUFS integrated, leave 6+ dB of true peak headroom, and let dynamics carry the music.

Genre-specific compression norms

Different genres have different norms for compression intensity. Hitting the genre's typical range matters because listeners are calibrated to it — a singer-songwriter mix at EDM compression levels sounds wrong; an EDM track at jazz compression levels sounds underproduced. Use the table below as a starting point.

"The loudness war ended in 2014 when streaming services said 'we don't care how loud you mastered it.' The trick now is making the mix sound great at the platform's target — not at maximum possible loudness." — FTM, on the post-loudness-war era

Mastering vs. mix-bus compression — different goals, different settings

"Compression on the master bus" is two different things, depending on what stage of the workflow you're at.

Mix-bus compression (during the mixing stage, applied while you're still working on the song) is glue. Settings: 2:1 ratio, 1–3 dB GR, gentle. The job is to make the tracks feel like they're moving together. Use the SSL G-style compressor or an emulation; the goal is invisible glue.

Mastering compression (during the mastering stage, applied to the final stereo bounce) is final polish — making the song sit consistently across an album or playlist, controlling occasional peaks, adding subtle character. Settings depend on the genre and the source mix. Goal: get to target LUFS without over-compressing dynamics. Often combined with a final limiter for true-peak control (−1 dBTP max).

Don't try to do both jobs with one compressor. The mix-bus comp during mixing is for cohesion; the mastering comp later is for delivery polish. They're separate tools doing separate work.

Multiband compression — when single-band isn't enough

A standard (single-band) compressor reduces gain on the whole signal when ANY frequency exceeds threshold. Most of the time that's fine — but sometimes you want to compress only one frequency range while leaving the rest alone. Multiband compression is the solution.

A 3-band multiband compressor splits the signal into low, mid, and high bands (with crossover points typically at 200 Hz and 2 kHz). Each band has its own compressor with its own threshold, ratio, attack, release. You can compress the lows aggressively (controlling boomy bass) while leaving the mids and highs untouched. Or compress only the highs (taming sibilance and harshness) without affecting the body of the source.

Common multiband uses:

• De-essing — compress only 5–8 kHz when sibilance peaks. Better than full-spectrum compression for vocals.
• Boomy bass control — compress only 60–150 Hz when the bass is loud, leaving the upper octaves of the bass unaffected.
• Harsh top-end taming — compress only 8 kHz+ when the source gets harsh, leaving the body of the sound untouched.
• Master bus gentle multiband — split the master into 3–4 bands, apply 1–2 dB GR per band as needed for tonal balance correction.

The catch: multiband compression is easy to over-use and harder to dial in than single-band. Most issues can be solved with a well-placed single-band compressor or a smart EQ move; reach for multiband when those aren't enough. Stock multiband compressors: Logic's Multipressor, Ableton's Multiband Dynamics, Pro Tools' Pro Multiband Dynamics, Reaper's ReaXComp, Studio One's Multiband Dynamics. Free third-party: TDR Nova (excellent dynamic EQ alternative).

Dynamic EQ — the modern alternative

Dynamic EQ is a hybrid of EQ and compression. An EQ band only engages (boosts or cuts) when the signal in that band exceeds a threshold; otherwise it's transparent. The result is similar to multiband compression but more surgical and less pumping-prone.

Use cases:

• Tame a vocal's sibilance — dynamic EQ cut at 6 kHz that engages only when sibilance hits, releases instantly. More natural than a static EQ cut.
• Control resonant frequencies — a kick drum that rings at 220 Hz can be controlled with a dynamic cut at 220 Hz that only engages on the loudest hits.
• Mastering tonal correction — a 2 dB cut at 4 kHz that only engages when the mix gets harsh, leaves quiet sections unaffected.

Most modern EQ plugins now offer dynamic mode. FabFilter Pro-Q 3, iZotope Neutron, Logic's Channel EQ (in newer versions), Ableton EQ Eight (basic), TDR Nova (free) all support dynamic EQ. Dynamic EQ has largely replaced multiband compression for surgical tonal work — same result, easier to dial in, fewer pumping artifacts.

When NOT to compress — the restraint catalog

Not every track wants compression. The following sources often sound BETTER without any compression at all:

• Synth pads and ambient textures. Already compressed by their envelope; adding more flattens them.
• Programmed drum machines. Already perfectly consistent; compression only colors them, doesn't control anything.
• Heavily distorted electric guitars. Distortion IS compression — adding more squashes the dynamic interest.
• Loops & samples. Already compressed at their source. Use sparingly if at all.
• Solo classical / acoustic recordings. Dynamic range IS the music. Compression flattens the emotional arc.
• Tracks that already sound good. If a track sits well in the mix without compression, leave it. Compression for the sake of "I should compress" is a common beginner mistake.

The bypass test: turn off the compressor. Does the track sound clearly worse? If yes, the compressor is helping. If no — bypass and move on. Working engineers do this check on every track, every time.

The compression pitfalls catalog

• Compression for its own sake. Adding a compressor because "this should have one" instead of because the track needs it. Symptom: dozens of compressors at 1–2 dB GR each. Cure: bypass test every plugin; remove the ones that don't earn their place.
• Compression to fix recording problems. A poorly recorded vocal with inconsistent levels can't be fully fixed with compression. Re-record if possible; compress only what's salvageable.
• Stacking too much compression. Track comp + bus comp + mix bus comp + mastering comp = 4 stages of compression on the same signal. Each one might be 2–3 dB GR alone, but cumulatively that's 8–12 dB and the dynamics are gone. Pull back on individual stages.
• Heavy mastering compression on a streaming target. Above −12 LUFS doesn't make a streaming-distributed track louder. It just removes dynamics that platform normalization will turn down anyway.
• Same compression on every song. A "default chain" doesn't account for genre, song dynamics, or production style. Reach for compression because the source needs it, not because the template has it.
• Bypassing only at the END of mixing. Compression that drifts becomes a problem 4 hours into a mix session. Bypass-check throughout, not just before bouncing.
• Compression chasing reference tracks too literally. A reference track was mastered for a finished mix; you can't reverse-engineer their compression chain. Use references for tonal reference, not as a compression recipe to copy.
• Pumping at unwanted spots. Aggressive sidechain or bus comp can pump unintentionally on dense passages. Listen for the pump on full-mix sections; if you hear breathing where you don't want it, pull back.

In your DAW

Translation tools and metering for compression decisions: