The snare hits, and then it keeps going — a pitched tail that was charming in the room and is in the way in the mix. You try a gate and the drum stops sounding like a drum. You try a notch and the stick goes soft. Every fix costs you the part you wanted to keep.
The reason is that the ring and the hit overlap in both the dimensions your usual tools work in. They share frequencies, and they share time. A tool that cuts by frequency cannot avoid the hit, and a tool that cuts by time cannot avoid the tail. You need a third axis.
The short version. The ring is harmonic — sustained and pitched. The stick is percussive — transient and broadband. Separate the signal on that axis and you can pull one down without touching the other. ENZYME does it causally, so it works on a live insert. 14-day full trial, no account needed.
A snare is a resonant system. The batter head, the resonant head, the shell and the wires all have preferred frequencies, and a hit dumps broadband energy into all of them at once. What comes back out is a short broadband burst — the crack — followed by a decaying, pitched tail as the resonances give their energy back.
That tail is not a defect. A snare with no ring at all sounds like a sample of a click. The problem is usually one of two specific things:
Both are about the sustained, pitched part. Neither is about the crack. That distinction is the whole article.
A gate decides on level over time. Set it to close before the ring finishes and it also removes the natural decay and whatever room was in the tail, so the drum jumps out of the acoustic the rest of the kit is in. Worse, the threshold that kills the ring on a full hit also kills your ghost notes, because a ghost note is quieter than the tail you are trying to remove. You end up automating the gate, which is a lot of work for a snare.
Gates are the right tool when you want a very dry, very controlled snare. They are the wrong tool when you want the drum to still sound like it was recorded somewhere.
The ring has a fundamental, so you can find it by sweeping and notch it. The trouble is that the ring's fundamental usually sits inside the range that gives the snare its body. A notch deep enough to meaningfully shorten the tail is deep enough to hollow out the hit, because the notch is applied to the hit as well — the hit contains that frequency too.
Better, and worth trying first if you have one. A dynamic band only acts while the ring is above threshold, so it leaves the rest of the performance alone. But the ring overlaps the hit in time, not just after it — the resonances start ringing during the hit. So the band is still attenuating during the transient. You have improved the ratio, not removed the conflict.
Look at a spectrogram of a snare hit rather than a waveform. Two patterns are visible:
Those are different shapes, and a shape is something you can test for. Compare each bin against its neighbours along time and along frequency: whichever comparison finds it more typical decides whether the bin is harmonic or percussive. Bins that fit neither cleanly become a residual component — noise, breath, room.
This is harmonic/percussive source separation, and it is normally an offline technique because the obvious implementation looks both forwards and backwards in time. ENZYME runs a median-filter separation causally, so it works on a live insert with no lookahead beyond the analysis window. It is the same separation that backs Intrect's offline work, and it gives you three components you can attenuate independently: H, P and R.
On a snare, that lines up almost perfectly with the problem: the ring is H, the stick is P.
If the output suddenly sounds like one isolated element, a Mon button is still latched. They stay down on purpose so you can leave and come back to the same comparison, which also makes it the first thing to check when something sounds wrong.
The separation stage and the resonance detector are independent, and they compose. A useful pattern on a difficult snare:
Because the suppressor works on protrusion above a local baseline rather than on an absolute level, it does not need to be told where those modes are, and it will keep finding them if the drummer hits harder in the last chorus. The manual covers both stages control by control.
ENZYME — dynamic resonance suppression per frequency bin, per frame, with causal harmonic/percussive/residual separation. VST3, CLAP and a native AUv2 for Logic and GarageBand. 14-day full trial, no account and no email required.
Sympathetic resonance. The drum head, the shell and the snare wires all have resonant frequencies, and the hit excites them. Some ring is what makes a snare sound like a drum rather than a click, so the goal is almost never to remove all of it — it is to shorten or thin the part that clashes with the track.
A gate is the blunt version. It closes on level, so it takes the ring, the room and the natural decay at the same time, and it fights you on ghost notes because a threshold that kills the ring also kills the quiet hits. It is fine when you want a very dry, very controlled snare, and wrong when you want the drum to still sound like it was in a room.
Yes, and if you are still tracking that is the better fix — damping at the source costs you nothing in the mix. This article is about the case where the take already exists.
Only partly. The ring's fundamental is often close to the frequencies that give the snare its body, so a static notch deep enough to shorten the ring usually thins the hit too. A dynamic band is better because it only acts while the ring is present, but the ring overlaps the hit in time as well as in frequency, so a purely spectral tool is always making a compromise.
A way of splitting a spectrogram into a sustained, pitched component and a transient, broadband one. Sustained energy is steady across time at a given frequency; transient energy is steady across frequency at a given instant. Comparing each bin against both patterns lets you assign it to the harmonic or percussive side. It is normally an offline technique — ENZYME runs it causally so it works on a live insert.
It can, if you cut too hard. The ring is the drum's resonant pitch, so attenuating it heavily makes the snare read as less tuned and more like a noise burst. Use the difference monitor: if the removed signal sounds like a pitched drum note rather than an overlong tail, ease off.