MIDI and sequencing

A producer records a keyboard part as MIDI and later decides it should be a different instrument, transposed up a tone. Explain why both changes are trivial, and what would be needed if it had been recorded as audio.

1. 01MIDI is data

   The MIDI track stores only which notes were played, when, and how hard - not the sound itself.

2. 02Change the instrument

   Because the notes drive whatever instrument is assigned, swapping the virtual instrument re-voices the whole part instantly, with no re-recording.

3. 03Transpose

   Transposing up a tone simply adds 2 to every note number in the data, so the DAW replays the part two semitones higher with no artefacts.

4. 04If it were audio

   An audio recording is a fixed capture of one instrument, so re-voicing would need re-recording, and transposing would need pitch-shifting with its artefacts.

Result: As MIDI, re-voicing and transposing are instant data edits; as audio, they would require re-recording or artefact-prone pitch-shifting - because MIDI is instructions, not sound.

You are programming a string part and want (a) a vibrato that swells on held notes, (b) a smooth crescendo within a sustained note, and (c) an expressive upward pitch slide. Which MIDI message serves each?

1. 01Vibrato on held notes

   Use the modulation wheel (CC1), which most instruments map to vibrato depth, drawn in to swell on the held notes; aftertouch could also drive it from key pressure.

2. 02Crescendo within a note

   Use expression (CC11) to raise the level smoothly through a sustained note, giving a crescendo without re-triggering it.

3. 03Pitch slide

   Use pitch bend, the dedicated high-resolution message, to slide the pitch up smoothly and expressively.

4. 04Note the distinctions

   Velocity would only change the initial strike, so it cannot swell a held note - the continuous CCs and pitch bend are the right tools for within-note gestures.

Result: Modulation (CC1) or aftertouch for the vibrato swell, expression (CC11) for the crescendo, and pitch bend for the slide - each gesture has its own message.

You must program a loose, felt electric-piano chord groove and, separately, a machine-tight sixteenth-note synth arpeggio that is too fast to play. Which input method suits each?

1. 01The electric-piano groove

   The feel and dynamics are the point, so record it in real time on a keyboard against a click, capturing the human timing and velocities.

2. 02The fast arpeggio

   It is too fast to play accurately and needs machine-tight timing, so use step input at a sixteenth-note value, entering each pitch in turn on the grid.

3. 03Refine each

   Lightly quantise the real-time groove if it drifts too far; add a real-time controller pass (filter, expression) over the step-entered arpeggio for movement.

4. 04Justify

   Real-time preserves feel where feel matters; step input guarantees precision and complexity where accuracy matters.

Result: Real-time input for the felt chord groove and step input for the too-fast arpeggio - matching method to whether feel or precision is wanted.

A sixteenth-note hi-hat was step-entered dead straight and sounds mechanical. Give it a swung, human feel.

1. 01Apply swing

   Apply a swing quantise (say 55-62%) at the sixteenth value, which delays the off-beat sixteenths to create the long-short shuffle feel.

2. 02Vary the velocities

   Edit the velocities so the on-beats are a little stronger and the off-beats lighter, adding a natural accent pattern instead of identical hits.

3. 03Humanise slightly

   Add a small amount of timing and velocity humanisation so no two hits are identical, removing the machine-gun uniformity.

4. 04Audition

   The hi-hat now lilts with a swung, accented, slightly varied feel rather than sounding rigidly programmed.

Result: A swing quantise plus velocity shaping and slight humanisation turn a dead-straight step-entered hi-hat into a swung, human-feeling groove.

You must (a) shape the rising dynamics of a synth melody, (b) program a tight drum beat, and (c) hand a notated part to a session player. Which editor suits each?

1. 01The melody dynamics

   Use the piano-roll: draw the notes and shape their velocities in the velocity lane, seeing the melodic contour and dynamics together.

2. 02The drum beat

   Use the drum/step editor: a grid of drum sounds against time makes programming and adjusting a beat quick and visual.

3. 03The notated part

   Use the score editor to display the MIDI as staff notation and print a readable part for the player.

4. 04Consistency

   All three views share the same MIDI data, so edits made in one appear in the others.

Result: Piano-roll for the melody's dynamics, drum editor for the beat, score editor for the notated part - each the fastest view for its task, all on the same MIDI data.

A MIDI string section plays the right notes but sounds obviously programmed. Describe the edits that would make it convincing.

1. 01Shape velocities

   Vary the note velocities to give phrasing and accents, rather than every note at the same strength, so the section has dynamic shape.

2. 02Fix articulation

   Adjust note lengths and overlaps for legato phrasing where the strings should be smooth, and shorten notes where they should be detached.

3. 03Add expression (CC11) and vibrato (CC1)

   Draw expression swells within sustained notes for crescendos and add modulation for vibrato, so the sound breathes like real players.

4. 04Automate the arrangement

   Automate the section's level and perhaps a filter or reverb send to swell into the chorus, giving the part shape over the arrangement.

Result: Varying velocity, fixing articulation, adding expression and vibrato controllers, and automating swells turn a flat MIDI string part into an expressive, believable performance.