Why Won’t the Open Hi-Hat Stop? A Four-Pass Choke-Group Test

Two correct sounds do not prove one will stop the other

Note mapping answers which sound an input selects. A choke or exclusive relationship answers what the receiver does to an already active voice when another group member arrives. O and C can both audition correctly while their tails still overlap because sound identity and voice behavior are separate settings.

Current manufacturer documentation shows several implementations of a related goal. Ableton Live Drum Rack silences other chains in the same choke group when a member is triggered. Apple Drum Machine Designer uses an Exclusive Group in which a new pad stops the other sounds in that group. Native Instruments’ Battery hi-hat recipe assigns both cells to one new Voice Group and sets the available voices to 1. Maschine adds Send and Receive roles, which can make C interrupt O without treating every relationship as symmetrical. The lesson is to translate the desired behavior, not copy a group number or term across products.

A drum note map can help sender and receiver agree on sound identity. It does not automatically configure voice stealing or exclusive groups in every custom kit. Treat the numeric note and the receiver’s group metadata as separate evidence.

Freeze the kit, routing, and effects before testing

1. Save a copy: duplicate the kit, rack, group, or project and record its original preset name. Do not experiment in the only recoverable version.
2. Name the pair: label one long-tailed open hat O and one short closed hat C. Use the receiver preview to confirm they are not the same sample and are not hidden by mute, solo, or an extreme envelope.
3. Hold monitoring still: use one output, listening level, and device. Temporarily bypass O/C send effects for the first pass, or record their starting values so a wet tail cannot masquerade as the source.
4. Use one interval: trigger O, let its tail become clearly audible, then trigger C after the same delay on every attempt. Do not compare one early C with another C that arrives after O has already ended.
5. Record live input: for a physical MIDI pad path, write the numeric note number, channel, controller bank/preset, and destination track for O and C. Do not rely only on octave labels such as C1 or C2.

O needs an observable tail. If it stops immediately when played alone, inspect the loaded sample, start/end points, envelope, and gate-versus-one-shot behavior first; there is no remaining O for C to interrupt. If C itself has a long tail, learn to distinguish C’s sound from O before calling any residual audio a failed choke.

Four O→C passes move from the receiver to the final mix

Pass 1 — receiver preview: prove the group inside the instrument

Do not use a physical controller or an old MIDI clip. Use the drum rack, sampler, or pad instrument’s preview, audition, speaker, or internal pad control. Trigger O, then C. Ableton explicitly documents its chain Preview control as a way to check a mapping away from a controller; here the same isolation keeps the shortest receiver path under test.

A pass means O’s dry source tail stops at C, or follows the short fade you deliberately configured, while C still plays correctly. If the receiver uses mutual exclusion, also test C→O and decide whether that two-way behavior is intended. If it provides Send/Receive roles, verify the documented direction instead of demanding symmetry.

Pass 2 — live pads: prove they reach the same two targets

After preview passes, repeat the identical O→C sequence with the on-screen pad or physical pad path you use in that receiver. Watch which chain, cell, Sound, or destination highlights while you listen. If this pass fails, the group is no longer the first suspect.

• If on-screen pads pass but a physical controller fails, compare note, channel, bank, and route at the receiver input.
• If both live paths select the wrong sounds, inspect the application layout or receiver input map.
• If the correct labels light but audio differs, check whether the paths reach the same receiver instance and whether one input triggers a parallel layer outside the group.

Do not strike O and C together. The test asks whether a later C interrupts an already active O, so order must remain audible.

Pass 3 — cross-bar loop: prove sequenced playback uses the same behavior

Create a new sparse clip rather than copying an old pattern. Put O on the & of beat 4 and C on the next bar’s downbeat, then loop it at least three times. Crossing the bar line exposes a wrong clip note, route, kit instance, loop boundary, or behavior that exists only during live input.

Key: O = open hi-hat; C = closed hi-hat; dot = no trigger; vertical bar = bar line. Leave kick, snare, swing, note repeat, extra automation, and other hat events out until O is consistently interrupted on the following downbeat.

Keep note lengths unchanged during the A/B test. Do not shorten O or insert an undisclosed note-off to hide a playback-path mismatch. If live input passes but this loop fails, verify the clip’s numeric notes, destination track, receiver instance, duplicates, and automation.

Pass 4 — dry versus effects: identify what is still ringing

Run the same cross-bar loop through two reversible monitoring states: the raw instrument or pre-effect path, then the normal processed path. Do not replace the sample or change its group during this pass.

• Dry O stops at C, but a processed tail remains: the group worked. Inspect reverb, delay, parallel returns, or other downstream processing.
• Dry O continues after C: return to group membership, direction, voice rules, and layers.
• One part of O stops while another dry component remains: solo layers, alternate cells/chains, or parallel instruments.
• C is silent or triggers another sound: resolve mapping or playback before grading choke behavior.

An effect often continues processing audio that entered before the source voice stopped. That decay is not automatically a defect. If the musical task needs a tighter close, A/B one recoverable change—such as O’s send level or a shorter effect preset—rather than continuing to randomize the choke settings.

Let the first failed pass choose one repair

Keep a failed copy F0 and make F1 from it. Change exactly one item in F1: group membership, available voices, Send/Receive role, input note, layer membership, or effect send. Then restart at Pass 1. A five-setting repair may sound better, but it cannot tell you what deserves to be saved.

A passing repair must cross the bar line and survive reload

• Identity gate: O and C audition clearly and are not two labels for the same sample or target.
• Relationship gate: receiver preview follows the tool’s documented mutual or Send/Receive behavior.
• Input gate: the live pad path needed for the task reproduces preview with no hidden layer continuing.
• Loop gate: the cross-bar O→C works for three cycles, and C lands on the next downbeat without a miss or duplicate.
• Effects gate: any residual tail is classified as source voice or downstream effect and kept only intentionally.
• Reload gate: save, close, reopen, then repeat Pass 1 and Pass 3 with the same result.

Save at the scope that owns the relationship: kit, rack, group, instrument, or project according to the receiver. Maschine’s manual, for example, says its choke configuration is saved with the Group rather than each individual Sound; do not assume another tool uses the same boundary.

Common open-hi-hat choke questions

Do familiar MIDI notes make open and closed hi-hats mutually exclusive?

Not necessarily. A note map addresses sound identity; the receiver, kit, or sampler implements the voice relationship. Verify O and C first, then test their group independently.

Can shortening the open sample replace a choke group?

It can be an intentional sound-design choice, but it does not prove the interaction is repaired. If O ends at the same point with or without C, there is no evidence that C interrupted it. Preserve the original and compare the short-sample design with a real O→C relationship.

Why is there still a tail after C arrives?

Solo the dry path. If the source voice stops, the remainder may be reverb, delay, or a short deliberate fade. If the dry sample continues at full length, return to group, voice, and layer diagnosis.

Why does O also stop C?

Ableton and Apple document mutual same-group behavior, so two-way interruption may be the configured result. If you need C to stop O without the reverse relationship, check whether the receiver offers a directional mechanism such as Maschine’s Send/Receive roles. Not every tool does.

Which choke-group number should I use?

There is no cross-device standard number. It is usually a local container inside one kit or receiver. Choose a group that excludes unrelated sounds, record its members and save scope, and verify it after reload.

What if the problem appears only during fast playing?

Keep the tempo and capture the input events first. If O or C is missing or duplicated, use the separate 20-hit velocity and trigger audit. If the events are complete but their audio relationship is wrong, continue with this page’s group and layer path.