MIDI 1.0 represents note velocity on a 0-to-127 scale, with zero commonly serving as a note-off message. Your keyboard does not send softness or aggression as musical ideas. It sends numbers, and the piano plugin decides how those numbers change volume, attack, and tone.
A poor mapping can make an instrument feel cheap as quiet notes refuse to speak, ordinary chords bark, or firm accents hit the same ceiling. Raising the output volume cannot repair that behavior because loudness occurs after the performance has already been misread.
The receiving instrument applies another mapping, and a sampled piano may select or blend recordings made at different playing strengths. A modeled engine can instead calculate continuous changes in hammer behavior, brightness, and level from the same input. That means the fine-grained piano response remains inaccessible when your keyboard sends almost everything between 55 and 95.
Start with the controller and plugin curves set to their neutral positions. Open a MIDI monitor or record a short passage, then play repeated quiet, moderate, and forceful notes across the bass, middle, and treble. You are looking for the values you can produce deliberately and repeatedly, not the occasional extreme created by hitting a key unnaturally.
Keep the monitoring level comfortable and unchanged because loud speakers can soften your touch, while a quiet setting encourages extra force and shifts measurements. Record first and inspect afterward so watching the velocity meter does not influence each strike.
Save this untouched take because comparing each adjustment against the same MIDI performance removes changes in touch and exposes whether soft notes gained control. It also reveals whether hard accents simply reached the ceiling sooner.
Map those practical boundaries toward the instrument's useful minimum and maximum, then shape the middle. If normal playing crowds the upper half, lower or flatten the middle of the output curve. If moderate playing remains dull and timid, raise that region without dragging the quiet endpoint upward and destroying soft control.
Labels such as soft, normal, and hard are easy to misunderstand. A hard touch setting usually asks for more physical effort before producing a high output, while a soft setting reaches louder values sooner. Judge the result by the MIDI readings and sound together because manufacturers do not promise identical curve shapes behind those names.
Black and white keys deserve a separate pass because their positions change leverage, while obvious jumps in scales or balanced chords can expose uneven sensing. Apply per-key or black-key compensation only after repeated recordings show a stable bias, since correcting one careless performance can create a permanent imbalance.
Next, repeat the test on a second piano preset with a different tonal character. Similar physical gestures should still produce a usable progression, even though the presets need not sound equally bright or loud. If only one preset jumps abruptly, its internal mapping is the likely problem, so changing the controller globally would damage instruments that already respond well.
Leave audio compressors and limiters bypassed during calibration. They can flatten the level differences you need to hear, while the piano's changing timbre may still make stronger samples seem deceptively dramatic.
Do not use velocity compression merely to make a take look tidy. Compression narrows the differences between notes, and expansion can push values into the upper and lower limits where distinct gestures collapse together. The goal is not an impressive histogram but enough room for a whisper, ordinary playing, and a forceful accent to remain recognizably different.
Test the finished curve with a soft chordal passage, gradual crescendo, and repeated accents while keeping the piano preset and monitor level fixed. Save the curve with the controller and instrument names because changing either side changes the relationship you just calibrated.
A poor mapping can make an instrument feel cheap as quiet notes refuse to speak, ordinary chords bark, or firm accents hit the same ceiling. Raising the output volume cannot repair that behavior because loudness occurs after the performance has already been misread.
Your controller and plugin map dynamics differently
The controller creates its own response before the plugin receives anything, with key geometry, sensor timing, firmware, and touch settings influencing its values. A weighted action can still bunch most notes into a narrow band, while a light controller can sometimes provide a broad, predictable range.The receiving instrument applies another mapping, and a sampled piano may select or blend recordings made at different playing strengths. A modeled engine can instead calculate continuous changes in hammer behavior, brightness, and level from the same input. That means the fine-grained piano response remains inaccessible when your keyboard sends almost everything between 55 and 95.
Start with the controller and plugin curves set to their neutral positions. Open a MIDI monitor or record a short passage, then play repeated quiet, moderate, and forceful notes across the bass, middle, and treble. You are looking for the values you can produce deliberately and repeatedly, not the occasional extreme created by hitting a key unnaturally.
Keep the monitoring level comfortable and unchanged because loud speakers can soften your touch, while a quiet setting encourages extra force and shifts measurements. Record first and inspect afterward so watching the velocity meter does not influence each strike.
Save this untouched take because comparing each adjustment against the same MIDI performance removes changes in touch and exposes whether soft notes gained control. It also reveals whether hard accents simply reached the ceiling sooner.
Calibration reveals the missing range
Find the softest notes you can play reliably without silence, then the strongest comfortable accents you can repeat without pounding the keybed. Those boundaries define your usable dynamic range, which matters more than touching 1 and 127 once during a test.Map those practical boundaries toward the instrument's useful minimum and maximum, then shape the middle. If normal playing crowds the upper half, lower or flatten the middle of the output curve. If moderate playing remains dull and timid, raise that region without dragging the quiet endpoint upward and destroying soft control.
Labels such as soft, normal, and hard are easy to misunderstand. A hard touch setting usually asks for more physical effort before producing a high output, while a soft setting reaches louder values sooner. Judge the result by the MIDI readings and sound together because manufacturers do not promise identical curve shapes behind those names.
Black and white keys deserve a separate pass because their positions change leverage, while obvious jumps in scales or balanced chords can expose uneven sensing. Apply per-key or black-key compensation only after repeated recordings show a stable bias, since correcting one careless performance can create a permanent imbalance.
Next, repeat the test on a second piano preset with a different tonal character. Similar physical gestures should still produce a usable progression, even though the presets need not sound equally bright or loud. If only one preset jumps abruptly, its internal mapping is the likely problem, so changing the controller globally would damage instruments that already respond well.
Leave audio compressors and limiters bypassed during calibration. They can flatten the level differences you need to hear, while the piano's changing timbre may still make stronger samples seem deceptively dramatic.
The right curve keeps your performance intact
Place the correction where it causes the least collateral damage, using a controller-level curve only when its basic output is consistently cramped. That mapping follows the keyboard into every instrument. When one piano alone responds badly, use its plugin curve or a real-time DAW processor that preserves the original MIDI for later revision.Do not use velocity compression merely to make a take look tidy. Compression narrows the differences between notes, and expansion can push values into the upper and lower limits where distinct gestures collapse together. The goal is not an impressive histogram but enough room for a whisper, ordinary playing, and a forceful accent to remain recognizably different.
Test the finished curve with a soft chordal passage, gradual crescendo, and repeated accents while keeping the piano preset and monitor level fixed. Save the curve with the controller and instrument names because changing either side changes the relationship you just calibrated.
