Chapter 9 Quiz

Closed book first pass — that's the honest score. Multiple choice, then true/false with justification (the justification is most of the point), then short answer, then one applied scenario. Answers hide under each Verify fold; the scoring guide is at the end.

Section 1 — Multiple Choice (2 points each)

1. A MIDI clip fundamentally contains:

A) Compressed audio of the performance B) Instructions — notes, velocities, controller moves — but no sound C) A waveform plus pitch metadata D) Sound, but only at low sample rates

Verify **B.** MIDI is a script, not a recording: note on/off, velocity, CC, pitch bend, program change. Sound only exists after a virtual instrument (or hardware synth) performs the instructions. That's why swapping the instrument leaves the performance intact.

2. MIDI 1.0 was first publicly demonstrated in 1983 by connecting:

A) Two Yamaha DX7s B) A Moog and an ARP C) A Sequential Prophet-600 and a Roland Jupiter-6 D) An Atari computer and a drum machine

Verify **C.** At the January 1983 NAMM show — the fruit of the open-standard collaboration led by Dave Smith (Sequential) and Ikutaro Kakehashi (Roland), who shared a Technical Grammy for it in 2013.

3. Velocity values in MIDI 1.0 range from:

A) 0–100 B) 0–127 C) 1–256 D) -64 to +63

Verify **B.** MIDI 1.0 is a 7-bit language: 128 steps (0–127) for velocity, CC values, and note numbers alike. (MIDI 2.0 raises velocity to 16-bit — 65,536 steps.)

4. On a well-built sampled instrument, raising velocity changes:

A) Loudness only B) Pitch and loudness C) Loudness and timbre — often by switching to recordings of harder strikes D) Nothing unless CC1 is also raised

Verify **C.** Real instruments get brighter as they're struck harder (more upper harmonics, sharper transients — [Chapter 1](../../part-01-sound-fundamentals/chapter-01-what-is-sound/index.md)'s recipe changing with effort), so deep libraries record multiple velocity layers and let incoming velocity select between them. Velocity is a timbre selector wearing a volume costume.

5. In the piano roll, a note rectangle's horizontal length represents:

A) Velocity B) Pitch C) Duration — how long the note holds D) Pan position

Verify **C.** Pitch is vertical, time runs left to right, length is duration, and velocity lives in its own lane. Note *ends* are musical decisions too — the chapter's bassline lesson.

6. Quantizing at 50% strength means:

A) Half the notes snap to the grid, half don't B) Every note moves half the distance toward its nearest gridline C) The grid resolution is halved D) Velocities are reduced by half

Verify **B.** Strength (or "amount"/"iterative quantize") moves all notes partway to the grid, tightening a performance while keeping its push-pull intent — the single most valuable quantize setting in the chapter.

7. Swing, as introduced in this chapter, works by:

A) Randomizing all note timing within a range B) Delaying every other subdivision by a percentage, making offbeats lean late C) Increasing velocity on the backbeat D) Shortening all note lengths

Verify **B.** Swing pushes the offbeat eighths or sixteenths late, turning straight time into shuffle/strut. The full groove vocabulary — amounts, genres, drum-machine lore — is [Chapter 13](../../part-03-recording/chapter-13-programming-beats/index.md)'s territory.

8. The chapter's core criticism of the "humanize" (randomize) button is that:

A) It moves notes too far to be useful B) Human deviation is structured and correlated with the music; random jitter has no intention and reads as malfunction C) It only affects velocity, never timing D) It can't be undone

Verify **B.** Players lean consistently, roll chords directionally, accent phrase peaks — deviations with pattern and meaning. Randomness sounds like a tremor, not a musician: drunk, not human.

9. Which is the best summary of samplers vs synths?

A) Samplers are for professionals; synths are for beginners B) Samplers play back recordings ("recorded truth"); synths generate sound from scratch ("generated possibility") C) Synths sound more realistic than samplers D) Samplers use less disk space than synths

Verify **B.** Recorded truth: realism ceiling high, bounded by what was captured. Generated possibility: no realism obligation, unlimited invention (full synthesis course in [Chapter 14](../../part-03-recording/chapter-14-sound-design-synthesis/index.md)). The reverse of D is true — samplers carry gigabytes of recordings; synths are math.

10. The "machine-gun effect" on repeated notes is cured by libraries that include:

A) More velocity layers B) Round-robins — alternate recordings of the same note, rotated per hit C) Release samples D) Higher sample rates

Verify **B.** Identical samples retriggered back-to-back produce an artificial stutter; round-robins restore reality's micro-variation between consecutive hits. (Velocity layers fix a different flatness — dynamics/timbre.)

11. By common convention in orchestral and many synth libraries, CC1 (mod wheel) controls:

A) Sustain pedal B) Pitch bend depth C) Dynamics — crossfading between dynamic layers (timbre and intensity) D) Stereo width

Verify **C.** CC1 morphs between soft/dark and loud/bright recordings during the note; CC11 (expression) handles finer volume phrasing inside that timbre; CC64 is the sustain pedal. On sustained instruments, this — not velocity — is the main expression channel.

12. Which programmed part violates its instrument's idiom?

A) A monophonic bassline with varied note lengths B) A piano chord voiced bass note low, three notes mid-register, melody on top C) A "bass guitar" patch sustaining three-note chords through the chorus D) A string line using a legato patch with slightly overlapping notes

Verify **C.** Real bassists are monophonic the vast majority of the time, and low chords pile harmonics into the mud zone. The other three are exactly the chapter's prescriptions.

13. The chapter's verdict on gridlocked, velocity-flat programming is:

A) Always a mistake to fix B) A legitimate aesthetic — chiptune, techno, hyperpop — when chosen on purpose; a failure only when accidental C) Acceptable only on drums D) Obsolete since MIDI 2.0

Verify **B.** Stiff vs human is an aesthetic axis, not a quality axis (theme three). Kraftwerk's precision is meaning; the crime is the doorbell that *thinks* it's a pianist. Commit per part, on purpose, in writing.

14. "It sounds so MIDI" as an insult actually describes:

A) A flaw in the MIDI protocol's timing B) The thin sound of 1990s General MIDI playback engines plus flat, gridlocked programming — not the protocol, which carries every DAW production today C) The 31,250 bits/sec data rate distorting audio D) Low sample rate in the MIDI file

Verify **B.** General MIDI (1991) standardized 128 program names so files played "correctly" anywhere — through whatever cheap sample-playback engine the device had. The protocol carries no sound at all (and MIDI files have no sample rate — see C and D).

15. MIDI 2.0 (ratified 2020) improves on MIDI 1.0 with:

A) Audio streaming inside MIDI messages B) Higher resolution (16-bit velocity, 32-bit controllers), per-note controllers, and bidirectional capability negotiation — while staying backwards-compatible C) A replacement 5-pin cable standard D) Automatic humanization

Verify **B.** Resolution (no more stair-stepped sweeps), per-note expression as a native feature, and two-way conversation between devices. Backwards compatibility was the non-negotiable lesson of the forty years. Intention remains unsupplied — that's still your job.

Section 2 — True/False with Justification (3 points each: 1 for the call, 2 for the why)

16. Changing your session's tempo after recording a MIDI performance will chipmunk the pitch of the notes.

Verify **False.** MIDI is instructions, not audio: notes simply replay at the new tempo, pitch and timbre untouched, because the virtual instrument re-performs them. Audio is what chipmunks (or needs warping, with artifacts — [Chapter 15](../../part-03-recording/chapter-15-editing/index.md)). This is the mechanism behind the half-tempo recording trick.

17. On a sustained string patch, velocity is the primary tool for shaping the dynamics of a held line.

Verify **False.** Velocity fires once, at note-on; a sustained note's life happens *during* the note. CC1 (dynamic layer crossfade) and CC11 (expression) are the primary tools — the hairpin rule: every held note gets a shape. Velocity-only string writing is a leading cause of the uncanny valley (Case Study 2).

18. A library with eight velocity layers sounds no better than a one-layer library if you program everything at velocity 100.

Verify **True.** Flat velocity selects exactly one layer, every time — one timbre, rubber-stamped. The depth you paid for is triggered by the *data you program*; the library's realism is rented through your MIDI.

19. Good humanization should make a part's timing deviations as unpredictable as possible.

Verify **False.** The opposite: deviations should be *patterned* — consistent leans per voice, directional chord rolls, phrase-correlated dynamics. Unpredictable jitter is precisely what the randomize button does wrong. Intention reads as a ghost; noise reads as a tremor.

20. When quantizing a played take, you should generally quantize note starts but not note ends.

Verify **True.** Snapped endings clip releases and create robotic gaps; real notes end where the phrase breathes. Starts carry the rhythmic information worth tightening; ends carry phrasing worth protecting.

Section 3 — Short Answer (5 points each)

21. Explain "the ghost in the grid" in two or three sentences: what exactly got added to Jaylen's piano line, given that not a single pitch changed?

Verify A programmed *performance*: velocity shaped into phrase arcs with melody-note hierarchy, chords rolled bottom-to-top, consistent micro-timing leans (late second hits), note lengths that breathe between phrases, and a sustain-pedal CC lane. The pitches are the skeleton; the structured deviations are the intention — and intention is what listeners read as "somebody playing." (Full credit mentions at least three concrete ingredient families and the idea of intention/structure.)

22. Your friend's expensive orchestral library "still sounds fake." Give a four-item diagnostic checklist, in order of likely payoff, drawn from this chapter.

Verify (1) Flat or velocity-only dynamics → shape velocities *and* ride CC1/CC11 hairpins on every sustained line; (2) no legato — gaps or hard restarts between notes → use the legato patch with slight overlaps; (3) machine-perfect simultaneity → humanize starts with small, structured offsets (sections aren't sample-accurate); (4) idiom violations → re-voice piano-style block chords into line-based, breath-length string writing. (Accept also: no round-robins on repeats; everything equally loud/close with no phrase shape.)

23. Why does a bassline's note length matter as much as its pitches? Name the chapter's term for what length creates, and give the two contrasting examples.

Verify Length is the groove dimension of bass: where notes end shapes the pocket — the space before the next downbeat is the funk. Short, detached notes *drive*; notes ringing into each other *flow*. Identical robotic lengths read as a metronome regardless of pitch choice. (The bass-vs-kick relationship deepens in Chapter 13.)

24. "A library's realism is rented through your MIDI." Unpack this sentence using two specific library features and the programming that activates each.

Verify Velocity layers — multiple recorded dynamics per note — only speak when your velocities *vary and shape phrases*; flat 100s play one layer forever. Round-robins only matter when repeats occur, and pairing them with alternating velocities completes the illusion of consecutive human strikes. (Also acceptable: release samples activated by realistic note ends; legato transitions activated by overlapped notes; CC1 crossfades activated by ridden controller lanes.) The instrument supplies recordings; the *performance data* decides whether they breathe.

Section 4 — Applied Scenario (10 points)

25. A reader sends you a ballad: programmed piano, sampled strings, and a synth bass, all clicked in with a mouse. Their complaint: "It's technically correct and completely lifeless — but when I hit Humanize, it got sloppy AND stayed lifeless." You have one hour to fix it and ten minutes to explain. Write your response: diagnose why both the original and the randomized version fail, then lay out an ordered, instrument-by-instrument repair plan using this chapter's tools (be specific about velocity, timing, lengths, CC, and idiom checks), and end with the one aesthetic question they must answer before touching anything.

Verify Strong answers include: **Diagnosis** — the original fails because deviation is absent (gridlocked, flat velocity, simultaneous chords, no CC = no intention to read); the randomized version fails because deviation is *unstructured* (jitter has no correlation with the music — drunk, not human). **The aesthetic question first** — is any part meant to be machine-stiff on purpose? A ballad usually wants human, but commit per part in writing before editing. **Piano** (highest exposure, biggest payoff): velocity pass — phrase arcs, top-note +5–10 hierarchy, varied repeats; roll chords 5–20 ms bottom-up, left hand slightly leading; vary note lengths so phrases breathe; add CC64 legato pedaling (down after the chord, up at the change); idiom check — voicings two hands could play. **Strings**: switch to the legato patch, overlap consecutive notes; move dynamics from velocity to performed CC1 with CC11 micro-phrasing; hairpin every held note; humanize line starts with small consistent offsets; check parts are line-shaped and breath-length, not block chords. **Bass**: monophonic; consistent slight lean (e.g., a touch late, all song); note-length pass for pocket — daylight before downbeats; steady velocities (basses want consistency) with a few ghosts. **Order and verification**: work velocity → timing/lengths → CC; small structured moves; A/B against the gridlocked original *and* a reference track at matched, calibrated level; stop when intention reads — don't fiddle past done (bounce to audio if needed). Scoring: 3 pts diagnosis (structure vs randomness), 5 pts plan quality/specificity across all three instruments, 2 pts for the commit-on-purpose framing and verification discipline.

Scoring

Section Points available
Multiple choice (15 × 2) 30
True/False + justification (5 × 3) 15
Short answer (4 × 5) 20
Applied scenario 10
Total 75
Score Verdict
68–75 The ghost is in your grid. Program the checkpoint and move to Chapter 10 with confidence.
56–67 Solid — re-skim the sections behind your misses (most people drop points on CC vs velocity and on structured vs random deviation), then run DAW Lab C1.
41–55 You have the vocabulary but not yet the model. Re-read "Velocity," "Humanization," and "Why Good Libraries Breathe," then retake in two days — spaced beats crammed.
Below 41 Honest reset: re-read the chapter with the C1 drill open in your DAW, doing each move as you read it. This chapter is hands-first; the score will follow the hands.