How to Sidechain the Beat to the Vocal (Ducking That Nobody Hears)
Vocal mixing · 6 min read
Sidechaining the beat to the vocal means the instrumental bus dips a little, only while the singer is singing, and comes back the instant they stop. Done right it is 1–2 dB of gain reduction that nobody can hear as an effect, but everybody hears as "the vocal sits on top". Done wrong it is the kick-style pumping you hear on bad trap mixes. The difference is the amount, the release, and whether you duck the whole beat or only the 1–5 kHz the vocal lives in.
Gain reduction target on the beat bus
1–2 dB, not 6
Keyed from the vocal · release ~150–300 ms · mids only if you can
The 60-second answer
Put a compressor with an external sidechain on the instrumental bus (or a stereo stem of the beat). Set its key input to the lead vocal track, post-fader. Ratio 2:1, attack 10–30 ms, release 150–300 ms, and lower the threshold until the meter shows 1–2 dB of reduction on the loudest vocal phrases and 0 dB in the gaps. If your DAW has a dynamic EQ with a sidechain input, do the same thing but on a wide bell around 2–4 kHz instead of the whole signal. Bypass it and listen: the vocal should feel 1 dB louder, and the beat should not feel like it moved at all.
Why a static EQ cut is not enough
The usual advice for a vocal that will not sit is to carve a notch in the beat around 2–4 kHz. That works, and it is the first thing to try (the full method is in EQ: carve space for vocals). The problem is that a static cut is there all the time, including the 40% of the song where there is no vocal. The hook with no singer sounds thin. The intro sounds thin. You end up cutting less than you need, so the vocal still fights during the verse.
A sidechain makes the cut conditional. The beat only loses those 1–2 dB when the vocal is actually there; in the gaps the instrumental is at full size. It is the same carve, switched on and off by the singer.
What is really going on: masking, not loudness
The vocal does not get buried because it is quiet. It gets buried because the beat is producing energy in the same 1–5 kHz band at the same moments, and the ear hears the louder of the two. The fix is to reduce the competing energy by a small amount at the exact moments it competes. That is why 1–2 dB is enough: you are not making the beat quieter, you are tipping the masking contest. Check the result with the Vocal–Beat Balance Check, which gives you the vocal-to-beat level number so you are not guessing.
Sidechain the beat to the vocal, step by step
| # | Step |
|---|---|
| 01 |
Route the instrumental to one bus. Every beat element (or your stereo beat stem) goes to a "BEAT" bus. The vocal stays on its own path. Do not include the vocal reverb returns in the beat bus. BEAT bus = all instruments, no vocals, no vocal FX |
| 02 |
Insert a sidechain-capable compressor on the beat bus. Stock works: Logic Compressor, Ableton Compressor/Glue, FL Fruity Limiter (comp mode) or Maximus, Pro Tools Dyn3 all have a key input. Set the key/sidechain source to the lead vocal, post-fader if offered. key input = lead vocal · listen to key once to confirm |
| 03 |
Set the shape. Low ratio and medium times. Fast attack would chop the beat's transients; long release would leave the beat ducked after the phrase ends. ratio 2:1 · attack 10–30 ms · release 150–300 ms · knee soft |
| 04 |
Find the threshold by the meter, not by ear. Loop the densest vocal line. Lower threshold until gain reduction reads 1–2 dB on the words and returns to 0 between lines. If it never returns to 0, the release is too long or the threshold is too low. GR 1–2 dB peak · 0 dB in gaps |
| 05 |
Narrow it if you can. If the compressor has a sidechain filter, high-pass the key at 150–200 Hz so plosives and low notes do not trigger it. Better: use a dynamic EQ on the beat bus with the same key, one bell at 2.5–3.5 kHz, Q ~1, max 2–3 dB cut. The lows of the beat never move. dyn EQ bell 2.5–3.5 kHz · Q 1 · −2 to −3 dB max |
| 06 |
Level-match and A/B. Bypass the compressor and compare. The beat should not sound different in the gaps. The vocal should feel a touch clearer without you touching its fader. If you can hear pumping, halve the reduction. bypass test · if audible as an effect, too much |
The trick: key it from a de-essed, high-passed copy of the vocal
The key signal does not have to be the vocal you hear. Send the lead vocal to a dedicated "KEY" aux with a high-pass at 200 Hz and a de-esser pulling hard at 6–8 kHz, and feed the sidechain from that aux instead of the track. Now the ducking is triggered by the body of the words, not by breaths, plosives or esses, so the beat stops flinching on every "s". Nothing on the KEY aux reaches the mix; it is a control signal only. This is the step that turns a nervous, flickering duck into a smooth one.
Chains with the reasons left in.
The Vocal Chain Bible is 88 real vocal chains from hit records, as the engineers described them — mic, preamp, every plugin and setting, in order, with a free-plugin version of every chain. Every chain comes with the reason behind each setting, so you know what to move for your voice.
Get the Vocal Chain Bible — $48 →Mistakes that kill it
- Ducking 4–6 dB. That is a kick sidechain amount. On a vocal it reads as the beat breathing in and out with the lyrics. Stay at 1–2 dB; if that is not enough, the problem is the vocal chain or the arrangement, not the duck.
- Keying from the vocal reverb or the vocal bus with FX. The tail keeps the compressor open after the phrase. Key from the dry lead only.
- Full-band ducking on a bass-heavy beat. The 808 drops with every word and the low end wobbles. Use a sidechain high-pass on the key, or a dynamic EQ so only the mids move.
- Doing this before the vocal is mixed. Set the vocal chain first, then the beat EQ carve, then this. Sidechain is the last 1 dB, not the first 6.
Keep going
FAQ
Should I sidechain the whole beat or just some elements?
Start with the whole instrumental bus at 1–2 dB. If the drums start to feel soft, remove the drums from the ducked bus and sidechain only the melodic elements (keys, pads, guitars, synths), which are what actually mask the vocal.
Sidechain compressor or dynamic EQ?
Dynamic EQ is cleaner because only the 2–4 kHz band moves and the low end stays put. A full-band compressor is faster to set up and works in any DAW with a key input. If you only have the compressor, high-pass its sidechain at 150–200 Hz and keep the reduction at 1–2 dB.
Can I do this in GarageBand or a DAW without sidechain routing?
If your DAW has no key input, you can approximate it with volume automation on the beat bus: draw −1.5 dB during vocal phrases and 0 dB in the gaps. It is slower, but the result is the same and it is fully under your control.
Who's behind this
Ivan Meshcheriakov · IvanFromRnD
5+ years ago I moved to LA with zero connections and started building from nothing — 14-hour days, meeting people and interning at studios. It took me 5+ years to meet and interview 100+ Grammy-winning engineers — then I turned it all into the Vocal Chain Bible.
Read my full story →