Making music, explained

Notes as data, not recordings

A MIDI note stores a handful of numbers: pitch (which key), velocity (how hard it was hit), start time and length. That's it. When the song plays, an instrument — a synth, a sampled piano, a drum kit — reads those numbers and produces the actual sound.

Why record MIDI instead of audio?

Because instructions are editable and audio mostly isn't. Played a wrong note? Drag it to the right pitch. Timing sloppy? Nudge or quantize it. Wrong instrument entirely? Point the same notes at a different one — the take survives every change of mind. An audio recording of a piano is forever a piano; a MIDI clip is a performance waiting for an instrument.

One language for all gear

MIDI (Musical Instrument Digital Interface) was agreed on by rival synth makers back in 1983, and it's still how music hardware and software talk: your keyboard controller sends MIDI into SANDYNE, SANDYNE sends MIDI to instruments and plug-ins. Forty-year-old gear and yesterday's soft-synth speak the same language.

This is why a “MIDI keyboard” makes no sound on its own — it only sends instructions. The sound comes from whatever receives them.

Instructions age well; recordings don't bend

A MIDI clip transposes perfectly, re-times perfectly, and will play any instrument you point it at — it's just numbers. An audio clip can be pitch-shifted or time-stretched, but the DAW has to mathematically rebuild the sound, and past a few semitones or percent you hear it: smeared transients, warbly “underwater” artifacts. And some edits — like swapping the instrument — simply don't exist for a recording.

So why record audio at all?

Because audio captures what MIDI can't: a voice, a real guitar, a room, a vinyl crackle — the actual sound of a performance, with every nuance no instrument plug-in can regenerate. The rule of thumb: keep things MIDI while you're still deciding, record audio when the sound itself is the point.

Common workflow: write and arrange in MIDI, then record (or “bounce”) parts to audio once they're final — audio is lighter on the CPU and locks the sound in place.

Squeeze, stretch, repeat

The molecules barely move — they jiggle in place and pass the push along, like a crowd doing the wave. What travels is the pattern of pressure, at about 343 metres per second. Take away the air (space!) and there's nothing to squeeze: no sound at all.

Amplitude — how hard the squeeze is

Drag loudness down and the bands fade: the pressure swings are smaller, and you hear the tone get quieter. Same wave, gentler push. Amplitude is what a volume fader changes — nothing else about the sound moves.

Frequency — how often it squeezes

Drag pitch up and the bands crowd together: more squeezes per second, and the tone rises. Squeezes per second are measured in hertz — and everything about what we can hear hangs off that one number.

Why 440 Hz is “A”

Pure convention — and a surprisingly recent one. Orchestras spent centuries tuning anywhere from 400 to 460 Hz until A = 440 Hz was standardised in the 20th century. There's nothing magical about the number; what matters is that everyone agrees, so every instrument — and every DAW — lands on the same pitch.

The octave — doubling sounds the same

220, 440, 880: physically very different frequencies, yet your brain files them all as “A”. A doubled frequency lines up perfectly with the original — every second peak matches — so they blend almost into one sound. That's the octave, and it's why note names repeat instead of going on forever. Men and women singing “the same melody” are usually an octave apart without anyone noticing.

Seven letters, twelve keys

Between one A and the next, Western music places twelve equal steps — the repeating pattern of seven white keys (C D E F G A B) and five black ones (the sharps ♯ and flats ♭). The number after the letter is just which octave you're in: C4 is “middle C”, A4 the tuning note. Why twelve? That's a story about the steps themselves — a topic of its own, coming soon.

Equal steps means multiply, not add

Pitch works in ratios: an octave is ×2, so twelve equal steps means each semitone is × 2^1/12 ≈ ×1.0595 — about 6% up, every step. That's why the rungs above crowd together at the low end: 6% of 220 Hz is a smaller jump than 6% of 400 Hz. Divide the octave into equal Hz instead (the “Equal Hz” mode) and every note lands between the cracks — your ear notices instantly.

Why twelve?

Notes sound good together when their frequencies form simple ratios — 3:2 is the sweetest of all (the “perfect fifth”). Twelve happens to be the smallest number of equal steps where the ladder lands almost exactly on those ratios: seven semitones give ×1.498, a hair from 3:2. It's a compromise — every interval is a tiny bit off, but the same amount in every key, which is what lets one instrument play in all of them. Other traditions chose differently: Arabic music famously divides further, into quarter tones.

Cents — the fine print between steps

Each semitone splits into 100 cents — too fine to hear one at a time, but exactly the scale of the Det knob you met on the oscillator: detuning two voices by a few cents makes a sound shimmer, and a supersaw is just seven copies spread across a few dozen cents.

Tempo — beats per minute

120 BPM means two beats every second. Genres cluster around tempos: ballads drift at 60–80, hip-hop leans 80–95, house sits near 120–128, drum'n'bass flies at ~174. In a DAW the tempo is set once and everything snaps to it — the piano roll grid, the arpeggiator, tempo-synced delays. The metronome click is that grid, made audible.

Bars — beats come in groups

Beats aren't equal: every few beats, one feels heavier — count along and you'll say ONE-two-three-four without trying. That heavy beat is the downbeat, and one group is a bar. Music is built in bars the way text is built in sentences: chords usually change on the bar, sections last 4, 8 or 16 of them.

Time signatures — how many, and what kind

Read the two numbers like a recipe: the top says how many beats each bar holds, the bottom says what kind of note one beat is — 4 means quarter notes ♩, 8 means quicker eighth notes ♪. 4/4 — four beats per bar — is the default of pop, rock and nearly all electronic music; DAWs open in it for a reason. 3/4 is the waltz: ONE-two-three, ONE-two-three. 6/8 looks like six but feels like two big pulses each split in three — a rolling lilt you'll know from ballads and folk. Switch between them above and count along.

If you're unsure what a song is in, tap the pulse and listen for where the “heavy one” lands — every 4 taps, every 3, or in two rolling threes.

Note lengths — a family of halves

Note lengths are fractions of the bar: a whole note fills it, a half note takes half, then quarters, eighths, sixteenths — each generation exactly half the one before. That's the entire naming system. The glyphs just count flags: ♩ quarter, ♪ one flag = eighth, ♬ two flags = sixteenth.

Subdivision — the grid under everything

Choosing a note length to think in is called subdividing, and it's literally the grid in your DAW: the piano roll's snap setting is you picking a subdivision. Musicians subdivide out loud: “1 e & a, 2 e & a…” — the count row above. The numbers are the quarters, the “&”s are the eighths between them, “e” and “a” fill in the sixteenths.

Resolution is a feeling

Quarters feel solid — the walking pace of a song. Eighths add drive; sixteenths bring energy and urgency. Real patterns layer them, exactly like the rows above: a drummer's kick lives on quarters, the snare on twos and fours, the hi-hats ticking away in eighths or sixteenths on top.

Try it above: quarters only — calm. Add the “&” eighths — it starts to walk. Sprinkle a few sixteenths — suddenly it grooves.

The number — how a pair splits

Swing is measured as how each pair of notes divides its time. At 50% the pair splits evenly — dead straight. At 58% the first note holds a little longer and the second lands late — the classic drum-machine sweet spot. At 66.7% the split is 2:1 — true triplet swing, the lilt of jazz and blues shuffle. Past that you're limping on purpose.

Groove — swing plus everything else human

Swing is the biggest lever, but groove is the whole family of small deviations: hats slightly quieter here, a snare a hair early, velocity breathing from note to note. Genres live at addresses on this map: house hats shuffle around 54–58%, classic hip-hop swings its sixteenths (the same trick, one subdivision finer), and jazz rides at the triplet end. The legendary MPC sampler's swing knob is exactly the slider above — it's been selling records since the '80s.

How much to use

Start at 54–58% — enough to feel, not enough to name. Swing the hi-hats first; they carry the groove while kick and snare hold the anchor (notice the demo's kick never moves). In a DAW, swing is usually a project-wide or per-clip setting, so the arpeggiator, drums and bassline can all lean the same way together.

If the groove sounds obviously swung, you've probably gone too far — the best settings are felt in the shoulders, not heard.

Built from thirds — two sizes of skip

A triad is root + two stacked thirds, and a third comes in exactly two sizes: minor = 3 semitones, major = 4. Every flavour above is just a different pairing: 4+3, 3+4, 3+3, 4+4. Four moods from one recipe — that's the entire trick.

Major and minor — the middle note decides

The outer notes of major and minor are identical; only the middle note moves, by one semitone — and the mood flips from bright to melancholy. That one semitone carries more emotional weight than any other note choice in music. Flip between C major and C minor above and watch just one key change.

Diminished and augmented — the unstable two

Squeeze both thirds small (3+3, diminished) and the chord turns tense and brittle; stretch both wide (4+4, augmented) and it goes dreamlike and unresolved. Neither wants to stay — which is precisely their job: they're spice and transition chords, pulling the music toward somewhere more stable.

Most songs are almost entirely major and minor triads. Reach for dim and aug when a moment needs to feel like it's leaning somewhere.

Roman numerals — chords by job, not by name

Build a triad on each step of a scale and number them: I ii iii IV V vi vii°. Uppercase = major, lowercase = minor. The point of the numerals is that they name a chord's role in the key, not its letter — so “I–V–vi–IV” describes the same feeling whether you play it in C or in F♯. Learn progressions as numerals and you own them in every key at once.

Home, away, tension — what a progression does

I is home — stable, resolved. V is tension — it leans hard back toward home. IV is away, and vi is home's melancholy cousin (the relative minor). Every progression is a little journey: leave home, wander, feel the pull, return. Try ending the loop above on V — it refuses to sit still; end on I and it settles.

Four chords really are enough

I–V–vi–IV — the “axis” progression — underpins hundreds of hits across every decade and genre; start it on vi and the same four chords turn wistful. This isn't a limitation to outgrow: the chords are the canvas, and melody, rhythm and sound design do the painting.

Write the loop first, then change one chord and listen to what it does to the mood — that's the fastest ear-training there is.

What it's actually for

Pick your song's key on the circle and its best chords are already sitting next to it: the two neighbours (IV and V) plus the three minors just inside — that's the whole I–V–vi–IV toolkit at a glance. Want the song to lift or drift somewhere new? Keys one step away share six of their seven notes, so moving there feels smooth; jumping across the circle feels dramatic. And the moment someone says “it's in E”, the circle tells you: four sharps.

Why fifths close into a circle

The perfect fifth (+7 semitones, a 3:2 ratio) is the friendliest jump in music after the octave. Here's the small miracle: keep jumping by it and you visit all twelve notes exactly once before landing back where you started. No other simple interval does that — jump by octaves and you stay on one note forever. Twelve fifths, twelve notes, one perfect loop.

Distance on the circle = distance in sound

The circle is a map where nearby means similar. C and G differ by a single note; C and F♯ share almost nothing and sound like different worlds. That's why DJs and producers check it when blending tracks (harmonic mixing — the “Camelot wheel” on DJ software is this exact circle, renamed), and why a chorus that jumps one step clockwise feels like a gear-shift up.

The inner ring — every major has a minor twin

Inside each major key sits its relative minor — Am inside C, Em inside G. The twins use exactly the same notes; they just call a different one home, which flips the mood from bright to melancholy for free. If your track feels too happy, try treating the relative minor as home — same chords, new feeling.

Producer shortcut: writing in Am? Everything touching C on the circle is safe material. Same notes, zero clashes.

Harmony that moves

A chord answers “what notes?”; an arpeggio adds “in what order, and how fast?”. Because it never plays more than one note at a time, an arpeggio keeps the harmony present while taking up almost no space — thin as a melody, rich as a chord. Pianists call the gentle version a broken chord; synth players call the fast version an arp.

Where arpeggios live in a song

The intro: a lone arp sets mood and tempo before anything else enters — countless songs open exactly this way. Texture behind a vocal: where a held pad might be boring and strummed chords too busy, an arp fills the space with motion that never competes for attention. The bassline: arpeggiate the low notes of each chord and the bass plays itself — the engine of synthwave and classic italo. The hook itself: fast, bright arps are the lead in trance and much of EDM.

Make it evolve

An arp that never changes turns into wallpaper. The classic moves: widen it (1 → 2 octaves) into a chorus, speed it up (1/8 → 1/16) through a build, sweep a filter over it with an LFO or automation, or drench it in delay — a dotted-eighth delay on a simple arp creates patterns more intricate than anything you'd program by hand.

Change exactly one thing per section — direction, octaves, or rate. Listeners hear “something lifted” without knowing what.

The arpeggiator — hold a chord, get a pattern

You rarely play arpeggios by hand anymore. An arpeggiator is a MIDI effect: hold down a chord and it generates the pattern live — notes in, notes out, sitting at stage 2 of the signal chain, before the instrument. The demo above is exactly that machine. In SANDYNE it lives in the MIDI effects rack, locked to your song's tempo.

The four basic shapes

Sine is the purest sound possible — one frequency, nothing else. Soft, round, great for sub-bass. Triangle adds a few gentle overtones: still mellow, a little more present — think soft flutes and chip-tune leads. Saw contains every overtone — bright, buzzy, brassy; the go-to for big synth sounds because filters have so much to chew on. Square skips every other overtone, which makes it sound hollow and woody — classic video-game and clarinet territory.

Octaves — same note, different register

The Oct knob shifts the oscillator in whole octaves — doubling or halving its frequency. On its own it just changes register, but the real trick is layering: two oscillators an octave apart sound huge, and a sine one octave down is the oldest sub-bass recipe in the book.

Detune & the supersaw

Two oscillators tuned almost together drift in and out of phase, and the sound shimmers — the same reason a 12-string guitar or a string section sounds rich. That's detune, measured in cents (hundredths of a semitone). Push the idea further and you get the supersaw: seven saws spread around the note, a sound so effective it's carried trance and EDM leads for thirty years. Select SUPERSAW above and sweep Det while it plays.

PWM — a square that never sits still

A square wave doesn't have to be symmetric: pulse width is how long the wave spends up versus down. 50 % is the classic hollow square; narrow it and the sound gets thinner and more nasal. PWM (pulse-width modulation) means wiggling that width automatically — usually with an LFO — so the tone continuously shifts. One oscillator suddenly sounds alive, almost like two detuned ones. Hit Animate PW above with the square selected.

Attack — how fast the sound starts

The time from pressing the key to full loudness. A short attack (milliseconds) snaps like a drum or a plucked string; a long attack fades in like strings or a soft pad. It's the single biggest clue your ear uses to identify an instrument.

Decay — the fall after the peak

Right after the attack peak, most sounds settle down a little — a piano hammer strike is loud for an instant, then the note relaxes. Decay is how long that fall takes, dropping from the peak to the sustain level.

Sustain — the level while you hold

The odd one out: sustain is a level, not a time. It's how loud the note stays for as long as you hold the key. An organ sustains at full volume forever; a plucked string has zero sustain — it just keeps decaying no matter how long you hold.

A, D and R are measured in time. S is a volume. That trips everyone up exactly once.

Release — the fade after you let go

When you release the key, the sound doesn't have to stop dead. Release is how long it takes to fade from the sustain level to silence. Short release: tight and dry. Long release: notes ring out and overlap, like holding a piano's sustain pedal.

Type — which side gets cut

Low-pass (LP) lets lows through and cuts highs — the workhorse. It's what makes a synth bass sound "closed" or "muffled", and opening it up is the classic sweep in electronic music. High-pass (HP) is the mirror: cuts lows, keeps highs — used constantly in mixing to remove rumble and mud from anything that isn't bass. Band-pass (BP) keeps only a slice around the cutoff and cuts both sides — think telephone voice or a wah pedal.

Cutoff — where the filter starts working

The cutoff frequency is the hinge point: on one side the sound passes untouched, on the other it's pushed down. Sweep it and you hear the filter move — the single most expressive knob on any synth. Automating cutoff with an envelope is how a synth note goes "wow" instead of just "beep".

Resonance — a boost at the edge

Resonance emphasises the frequencies right at the cutoff, creating a peak just before the drop. A little adds presence and character; a lot makes the filter whistle and honk; push it to the extreme and many filters self-oscillate — they sing a pure tone at the cutoff frequency. Try maxing it in the demo above.

Slope — how steep the cut is

Past the cutoff, the level falls by a fixed amount per octave: 12 dB/octave is gentle and natural, 24 dB/octave is steeper and more surgical — the sound beyond the cutoff disappears faster. Synth folks say "2-pole" and "4-pole" for the same thing. Neither is better; they're different flavours.

Too slow to hear — and that's the point

A normal oscillator vibrates hundreds of times a second and you hear a tone. An LFO cycles maybe 0.1 to 20 times a second — below hearing range. Wired to a parameter, it becomes pure motion: the same shapes you met in the oscillator topic (sine, triangle, square, saw), just repurposed as gestures. A sine sways; a square snaps between two settings; a saw ramps and resets.

The target gives the effect its name

The same LFO, aimed at different knobs, produces effects so familiar they have their own names: on volume it's tremolo, on pitch it's vibrato (a singer's natural wobble is about 5–6 Hz), on a filter's cutoff it's the wah / wobble that powers dubstep basses, and on pan it sweeps the sound between your speakers.

Rate and depth — the only two knobs that matter

Rate is how fast the LFO cycles; depth is how far it pushes the target. Slow + deep = seasick sweeps; fast + shallow = gentle shimmer. In a DAW the rate is usually synced to the tempo — a quarter-note wobble stays locked to the beat at any BPM.

In SANDYNE they're called Modulators

SANDYNE generalises the idea: a Modulator is anything that moves a parameter for you — an LFO shape, an envelope, or a random generator — and it can be routed to almost any knob in the app. External synths and VST plugins will say “LFO” on the panel; it's the same machine. Learn it once, recognise it everywhere.

Modulators turn knobs in a loop; automation turns them along the timeline. Loops for texture, automation for storytelling — most tracks use both.

Fix — remove what shouldn't be there

Every recording picks up something unwanted: rumble below 60 Hz, boxy mud around 200–500 Hz, harshness at 2–5 kHz. A narrow cut at the offending frequency cleans it up without touching the rest. This is EQ as repair — and it's mostly cuts.

Fit — give every instrument its own space

When two sounds occupy the same frequencies, the louder one hides the other — engineers call it masking. It's why a mix of great-sounding solo tracks can still sound like soup. The classic move: cut the guitars a little where the vocal lives, dip the bass where the kick punches. Each part gets its own shelf on the spectrum, and suddenly everything is audible at once.

Flavour — shape the tone you want

Wide, gentle moves change a sound's character: a broad low-shelf boost makes it warmer, a lift around 10 kHz adds “air”, a soft dip in the low mids makes room feel bigger. Taste, not surgery — a dB or two goes a long way.

Rule of thumb: cut narrow, boost wide — and if you find yourself boosting everything, you're really just turning the volume up.

Gain, frequency, Q — the three knobs of every band

Each band asks three questions. Frequency: where? Gain: up or down, and how much? Q: how wide? High Q is a narrow scalpel for fixing; low Q is a broad brush for tone. Bands at the spectrum's edges often become shelves — they lift or lower everything beyond their frequency, like the bass and treble knobs on a stereo.

Threshold & ratio — when, and how much

Threshold is the level where the compressor wakes up: everything below it passes untouched. Ratio is how firmly it pushes back above that line — at 4:1, sound that goes 4 dB over the threshold comes out only 1 dB over. Gentle ratios (2–4:1) even things out; high ratios (10:1+) act as a limiter, a hard ceiling nothing gets past.

Attack & release — how fast it reacts

Attack is how quickly the compressor clamps down once the sound crosses the threshold. A slow attack lets the first few milliseconds — the drum hit, the pick, the consonant — sneak through before clamping, which reads as punch. A fast attack catches everything and sounds tighter but flatter. Release is how quickly it lets go; set it too slow and the track audibly “breathes” or pumps with the beat — a flaw that electronic music adopted as a signature sound.

Makeup gain — the loudness trade

Compression only turns things down, so the result is quieter. Makeup gain raises the whole squashed signal back up — and because the peaks were tamed, you can raise it further than before. That's why compressed tracks feel louder and denser at the same meter reading.

Fair comparison tip: when you A/B a compressor, match the loudness first — “louder” almost always wins even when it isn't better.

Where you'll use it

Vocals, first and always — a singer's natural dynamics are wider than a mix can hold. Then: bass (evenness), drum bus (glue and punch), and the master. A special trick called sidechain compression lets one track's level duck another — the kick pushing the bass down on every hit is the heartbeat of house and EDM.

What panning actually does

A pan knob doesn't move sound — it changes the balance of level between the two speakers. Full left means only the left speaker plays; center means both play equally, and your brain constructs a phantom source in the middle. It's the same trick your ears use in the real world: whichever ear gets the sound louder (and sooner) wins.

An unwritten seating chart

Decades of mixing have settled into a convention: kick, bass and lead vocal stay center — they carry the weight and need both speakers' power. Everything else spreads outward: hats and shakers off to one side, keys and guitars opposite each other, backing vocals wide. Symmetry matters less than balance — if something sits left, give the right side a counterweight.

The stereo image — width as a feeling

The stereo image is the overall picture all those placements add up to: how wide, how deep, how organized the mix feels. Width also comes from differences between the two channels — a doubled guitar take panned hard left and right, a stereo reverb, a detuned synth voice per side. The more the sides differ, the wider it feels.

Always check your mix in mono once — clubs, phones and Bluetooth speakers often are. If a part vanishes when mono'd, its two sides were cancelling each other out.

Dry / wet — the effect runs beside the sound

Unlike EQ or compression, space effects don't replace the signal — they add a processed copy next to it. The original is the dry signal, the effect's output is the wet, and the mix knob sets the balance. That's why both demos above keep the dry hit untouched and only change what follows it — and why in a real mix these effects usually live on a shared send track that many instruments feed into.

Delay — distinct echoes, on the grid

A delay records the signal and plays it back after a set time. Feedback routes the output back into the input, so each repeat spawns another, quieter one — the bars in the chart shrink by the feedback amount every hop. Short times (60–120 ms) give the classic slapback; longer times are usually synced to the tempo (an eighth note, a dotted eighth) so the echoes land on the beat and feel like part of the groove.

Reverb — thousands of echoes become a room

Reverb is what a real space does: sound bounces off every surface, and the reflections arrive so fast and so densely that you stop hearing echoes and start hearing a room. Decay is how long that wash takes to die away — from a small booth (0.3 s) to a cathedral (5 s+). Pre-delay is the gap before the wash starts: a longer gap reads as a bigger room, and it keeps the dry attack clean before the tail blooms behind it.

Depth — wet is far, dry is close

Panning places sounds left-to-right; space effects place them front-to-back. A dry vocal sits right at the listener's face; drench it in reverb and it steps to the back of the hall. Mixing depth means choosing: lead elements stay drier and closer, pads and backing parts sit wetter and further away.

Less is more: reverb is the easiest way to drown a mix. If you can clearly hear the reverb, you probably have too much — turn it up until you notice it, then back it off a touch.

The decibel — a logarithmic ruler

Ears don't hear linearly, so audio isn't measured linearly. A decibel is a ratio: +6 dB means double the signal amplitude, −6 dB means half. Roughly +10 dB sounds twice as loud to a human. That's why faders are marked in dB — equal fader moves feel like equal loudness changes, whether the track is whisper-quiet or roaring.

0 dBFS — the ceiling, and what clipping is

dBFS means “decibels relative to Full Scale” — the largest number a digital sample can hold. 0 dBFS is that maximum, and every level below it is negative: −6, −12, −60. A signal that tries to go past zero simply can't — the samples are clamped at the maximum, the smooth waveform gets a flat top, and you hear it as harsh, crackling distortion. Analog tape saturated gracefully when pushed; digital just breaks. That's the whole reason gain staging exists.

Headroom — the space you deliberately don't use

Headroom is the gap between your loudest peak and the ceiling. Good practice: keep individual tracks peaking around −12 to −6 dBFS. It looks “too quiet” on the meter — that's fine. Quiet costs nothing (you can always turn the monitors up); clipping costs everything. The headroom is where the mix bus, effects, and the mastering stage get room to work.

Mixes are built with cuts

When a part is buried, the instinct is to turn it up — then the next part is buried, so that goes up too, and soon everything is slamming the ceiling. This is the classic “everything louder than everything else” spiral. The fix is to work the other way: turn the masking tracks down. Since only relative levels matter inside the mix, cutting is free — it buys clarity and headroom at the same time.

Watch each stage of the signal chain: a track can be clean at the fader but clip inside an effect that's being fed too hot. Stage by stage, keep levels sane.

The journey, stage by stage

A MIDI clip holds the notes — pure data, no sound yet. MIDI effects transform that data first: arpeggiators, chord generators, scale snappers — notes in, different notes out, still silent. The instrument (an oscillator shaped by an envelope and filter) performs the result. The track's effects rack — EQ, compression, anything else — shapes that sound further. The mixer sets its level and pan and sums it with every other track. The master is where that sum gets its final polish and leaves for your speakers.

Order matters

Signal flows left to right, and each stage only hears what the previous one produced. A filter before a compressor behaves differently than after it; automation recorded on the instrument sounds different than the same move on the mixer fader. When something sounds wrong, trace the chain — the problem is almost always at (or before) the first stage where you can hear it.

One chain per track, one master for all

Every track runs this pipeline in parallel — dozens of private chains, all pouring into one master bus. That's the whole architecture of a DAW. Audio tracks simply skip stages 1–3: the recording already is sound, so it enters at the effects rack.