Mastering Your MusicLesson 4 of 6

The Mastering Chain

Corrective EQ, glue compression, saturation, stereo work and a limiter at −1 dBTP — what each stage does and how hard to push it.

5 min read

A mastering chain is a short line of processors, each doing one small job. There's no single correct chain, but most masters follow the same logic: fix problems first, shape second, raise level last. The limiter goes at the end because it should only ever see the finished, shaped signal.

1. Corrective EQ

Start by removing what's wrong before adding anything.

  • Low-end cleanup. A gentle high-pass filter somewhere below the lowest musical content (often 20–30 Hz) removes sub-sonic rumble that eats headroom without being heard. Don't set it so high that it thins the kick and bass.
  • Resonances and build-ups. Narrow cuts of 1–2 dB on a ringing frequency, or a broad dip in a muddy 200–400 Hz region, often do more than any boost.
  • Harshness. If the mix is fatiguing, look around 2–5 kHz. A dynamic EQ that only cuts when that band gets loud is often better than a static cut.

Use a linear-phase EQ if you hear the phase shift of steep filters smearing the low end; otherwise a regular (minimum-phase) EQ is fine and has no latency or pre-ringing.

2. Tonal shaping EQ

Now the broad, musical moves — usually wide shelves:

  • A low shelf of +0.5 to +1.5 dB for weight.
  • A high shelf around 8–12 kHz for "air".

If you're boosting more than about 2 dB anywhere, revisit the mix.

3. Glue compression

A gentle compressor on the whole mix makes the parts feel like one performance. Typical starting points:

  • Ratio 1.5:1 to 2:1.
  • Threshold set for about 1–2 dB of gain reduction on the loudest parts.
  • Attack slow enough to let transients through — 10–30 ms is a common range.
  • Release fast enough to recover between phrases; many compressors' auto-release works well here.

Watch the gain-reduction meter. If it never returns to zero, the compressor is just turning the whole song down and flattening it. Match output gain to input gain when you A/B it, otherwise you'll prefer whichever is louder.

4. Saturation (optional)

Subtle harmonic saturation — tape or tube emulation — adds density and helps a mix sound louder at the same peak level because harmonics fill in the spectrum. It's very easy to overdo. If you can clearly hear distortion, back off; most mastering saturation should be felt more than heard. Skip this stage if the mix already has plenty of character.

5. Stereo adjustments (optional)

  • Mono the lows. Below roughly 100–150 Hz, stereo information rarely helps and can cause problems on vinyl and club systems. A mid/side EQ cutting the side channel in the lows, or a "mono below" tool, tightens the bottom.
  • Width. A small lift of the side channel in the high mids and highs can open a cramped mix. Check the result in mono; excessive widening makes elements vanish when summed.

6. The limiter

The limiter raises the level and catches the peaks so nothing exceeds the ceiling.

  • Ceiling: −1 dBTP, with the limiter's true-peak or oversampling mode on. Without it, inter-sample peaks can still exceed the ceiling.
  • Input gain / threshold: push into the limiter until you reach your target loudness. Watch the gain reduction: 1–3 dB on the loudest peaks is transparent on most material; beyond about 4–6 dB, most mixes start to lose punch and you'll hear pumping or distortion.
  • Release: shorter is louder but can distort low frequencies; longer is cleaner but can pump. Auto-release modes are a reasonable default.

If you can't reach your target loudness without heavy limiting, the answer is usually upstream — in the mix's balance or a stage before the limiter — not more limiting. Some engineers use two limiters each doing a little, or a clipper before the limiter to shave the fastest peaks.

Gain staging through the chain

Each processor should output roughly the level it receives, until the limiter. That way bypassing any one stage is a fair comparison, and nothing in the middle of the chain clips.

A sample starting chain

  1. EQ: high-pass at 25 Hz, −1.5 dB at 300 Hz (wide), +1 dB shelf at 10 kHz.
  2. Compressor: 2:1, 20 ms attack, auto release, 1–2 dB reduction.
  3. Mid/side EQ: side channel high-passed at 120 Hz.
  4. Limiter: −1 dBTP ceiling, gain until the integrated target, 1–3 dB peak reduction.

Treat these numbers as a place to start listening, not a preset. Bypass each stage in turn at matched level. Any stage you can't hear an improvement from should come out of the chain.

Sources

  1. Bob Katz (2015). Mastering Audio: The Art and the Science (3rd edition). Focal Press.
  2. Recommendation ITU-R BS.1770-5: Algorithms to measure audio programme loudness and true-peak audio level. ITU, 2023.
  3. Loudness normalization on Spotify. Spotify for Artists.