Velocity vs Aftertouch on MIDI Drum Pads: A Two-Pad Test

Velocity, Polyphonic Key Pressure, and Channel Pressure carry different evidence

The MIDI Association places Note On, Polyphonic Key Pressure, Channel Pressure, and Control Change in the Channel Voice family, but their payloads are not interchangeable. The Association describes Note On as carrying channel, key number, and velocity for the initial activation. Polyphonic Key Pressure carries a key number plus pressure, while Channel Pressure carries one pressure value for the channel and no key number. Apple labels the latter “aftertouch” in Logic Pro and shows its Number field as unused; Logic's separate poly-aftertouch event retains the MIDI note name.

Ableton's Push manual makes the practical separation visible: its Expression settings offer MPE, Poly Aftertouch, and Mono Aftertouch as different modes, while its Sensitivity settings separately shape incoming note velocity. Yamaha likewise documents separate output, fixed-value, curve, minimum, and maximum controls for Note On velocity and aftertouch, plus a choice between Channel and Poly aftertouch. Hardware labels therefore do not replace a message capture.

Build one direct, observable path before listening to a drum sound

1. Use a raw monitor first: connect the controller through the shortest practical MIDI or USB-MIDI route to a monitor that exposes event type, channel, note number, value, timestamp or order, and input port. Turn the receiver's audio down or bypass it for P0–P3.
2. Name two pads: choose Pad A and Pad B with different note numbers. Confirm their messages rather than relying on printed pad labels; the one-hit mapping audit gives a separate procedure if their identities are not stable.
3. Hold the channel constant: for the classic P2 comparison, use two notes on the same ordinary MIDI channel. If a documented MPE configuration deliberately assigns member channels, mark the route MPE-member, record the actual event type separately, and exit the ordinary same-channel P2 comparison.
4. Record the controller state: write the preset, pad bank, velocity curve, fixed-velocity state, aftertouch output setting, pressure type, and threshold if exposed. Disable note repeat, arpeggiation, MIDI effects, input transforms, and duplicated routes. A fixed onset is a separate question covered by the fixed-velocity audit.
5. Prevent echo: monitor one physical input only, turn MIDI Thru or local echo off where it would return events to the input, and clear the event list before every pass. Unexpected duplicate Note Ons make pressure evidence ambiguous.
6. Use comfortable force: test only within the manufacturer's normal playing guidance. The goal is a repeatable message pattern, not the largest possible value.

Save a screenshot or text export of each raw list if the monitor permits it, but do not merge passes. “The sound got brighter” is an audio observation; “Polyphonic Key Pressure, channel 10, note 38, values changed while note 38 remained held” is message evidence. Keep them in separate columns.

P0: use quick taps to establish the onset-velocity baseline

Clear the monitor. Tap Pad A once and release it without deliberately pressing farther after the strike. Repeat with four comfortable taps that span light to firm playing, then do the same on Pad B. Preserve the event order for each isolated tap.

P0 is not a calibration contest. If onset values bunch together, finish this protocol with the setting documented, then use the dedicated 20-hit velocity calibration to investigate response consistency.

P1: strike once, keep one pad held, then change pressure

Clear the monitor. Strike Pad A once, keep the finger on the pad, increase pressure smoothly, ease it back without lifting, and then release. Do not make a second impact. Repeat the gesture four times, then run four equivalent passes on Pad B.

• Onset gate: exactly one intended Note On starts the gesture. If further pressure creates another Note On, mark RTRG and solve that before treating the stream as continuous pressure.
• Order gate: any pressure events must be logged after the onset and before or around release in the order actually received; do not sort the list by value.
• Type gate: write PPA only for Polyphonic Key Pressure and CPA only for Channel Pressure. A CC or other raw expression stream is ALT, not “close enough” aftertouch. Record MPE-member separately as route context, never as the event type.
• Identity gate: a PPA event should expose the note number in the raw record. A CPA event has no note-number field, even if the monitor's interface visually associates it with the last-played pad.
• Absence gate: if no pressure-type event appears, write NONE. Check the manual, output toggle, selected mode, and threshold once; absence is not automatically a hardware fault.

Yamaha's FGDP-30 guide is a useful example of why the settings record matters: aftertouch transmission can be switched off, its message type can be Channel or Poly, and its curve, fixed value, minimum, and maximum are configured separately from Note On velocity. That does not imply another controller has the same controls; it shows why an undocumented preset is insufficient evidence.

P2: two held pads reveal whether the message identifies a note or only a channel

Use Pad A and Pad B on the same ordinary MIDI channel. Strike and hold A, then strike and hold B. Keep B steady while changing pressure on A; then keep A steady while changing pressure on B. Finally release both. Run four passes in total, alternating which pad is struck first.

Channel Pressure can affect all active notes on an ordinary shared channel when a receiver maps it that way, but the MIDI message itself only supplies channel and pressure. In MPE, assigning one note to each member channel can give that same Channel Pressure event per-note meaning. That route is outside this same-channel P2 comparison, but the raw event is still classified CPA; MPE-member is recorded beside it as context.

P3: follow the same events through record, storage, and replay

A live-input monitor proves only what reached that monitor. P3 asks three different questions: did the recorder receive the event, did the clip or file retain its type and fields, and did playback emit it again?

1. Record one known pass: capture the complete P2 gesture into an otherwise empty MIDI region or sequence. Keep input transforms, retrospective merging, quantization, and MIDI effects off.
2. Inspect before playback: open the recorder's event list. Compare event count, order, channel, note number where applicable, and value with the live P2 capture. Apple documents separate Logic event types for aftertouch and poly aftertouch, including the unused note-number field for channel aftertouch and the retained note field for poly aftertouch.
3. Store and reopen: save the project or a duplicate test file, then reopen the stored region if the workflow allows it. Do not overwrite the raw P2 export. Confirm that the same pressure event type and fields remain.
4. Replay to a second monitor: take your hands off the controller, disable the live input route, and play the stored region to an output monitor. A replayed event must come from the stored path, not a doubled live-thru path.
5. Compare four columns: mark live input, recorded event list, reopened storage, and replay output separately. One checkmark cannot stand in for all four.

If pressure exists live but not in the recorded list, or appears in that list but is missing after the saved copy is reopened, use CAP and inspect record filters, event support, or persistence. If the reopened event is present but absent from output, use OUT and inspect playback filters or routing. This is the same reason that choosing between MIDI evidence and audio evidence matters: an audio waveform cannot preserve the original MIDI message fields for later inspection.

A controlled receiver A/B separates message delivery from audible response

Only run this comparison after P3 shows a repeatable PPA or CPA event through replay. Use the stored performance so Note On velocity, timing, and pressure values are identical in both conditions. Monitor at a conservative level.

If the output monitor receives pressure in both conditions but only B changes, the transmitter and replay path were not the cause of A's silence. If neither changes while the messages arrive, investigate receiver support, message-type selection, and mapping before blaming the pad. If the receiver changes without the tested pressure message arriving, inspect velocity response, envelopes, MPE, CC, automation, and duplicated routing. Even when B changes amplitude, a MIDI value of 127 is not 127 dB and does not predict a fixed acoustic level.

Keep one evidence ledger, assign result codes, and stop at the first unresolved layer

Do not summarize the session as “aftertouch works.” State what message was observed, its scope, whether it survived storage and replay, and whether one controlled mapping produced an audible response. Replace the illustrative values below with your own raw evidence.

Codes can be combined when they describe different layers: CPA+OUT, for example, means the raw type was Channel Pressure and the first failed downstream layer was replay output. Never combine codes to hide an unresolved event type.

Velocity and aftertouch questions for drum-pad testing

Is velocity just the first aftertouch value?

No. In this MIDI 1.0 test, Note On velocity is part of the note's onset message. Aftertouch arrives as later pressure data: either Polyphonic Key Pressure with a note number or Channel Pressure without one. Preserve the event types instead of combining their numeric values.

Does poly aftertouch guarantee that only one drum sound will change?

No. The message identifies a note, but the receiver decides whether it supports the message and what it controls. A plug-in could ignore it, map several notes to the same parameter, or use it for filter, pitch, amplitude, an effect, or another destination.

Why can I see aftertouch but hear no change?

The monitor proves message arrival, not receiver mapping. Run the same stored performance through neutral Receiver A and deliberately mapped Receiver B. If B still does not respond, check whether it supports the exact PPA or CPA type and whether any filter removes it.

Does no aftertouch mean my pad is broken?

No. The controller may not implement pressure after the strike, the current preset or port may disable it, the threshold may not be crossed, or the route may filter it. Report NONE for the tested configuration; a fault diagnosis requires separate device-specific evidence.

Is MPE the same as polyphonic aftertouch?

No. MPE is a performance and channel-allocation method, not a raw Channel Voice message. A member channel may carry ordinary Channel Pressure, CC, or Pitch Bend for its assigned note. Record the exact event first and add the MPE member-channel context separately; per-note behavior alone does not prove Polyphonic Key Pressure.

Can a pressure gesture be sent as a CC instead?

A controller can expose expressive data through a Control Change route, but CC carries a controller number and value and remains a different message type. Record it as ALT, preserve the CC number, and test its documented mapping separately.

Does 127 mean maximum volume?

It is the top value in the 7-bit field shown by these MIDI event views, not a sound-pressure measurement. The receiver's instrument, velocity curve, pressure mapping, gain staging, amplifier, speakers or headphones, and listening position determine the audible result. It is never “127 dB” by definition.