Afterglow

A patch that plays itself and sounds finished: a soft, plucked arpeggio rolls through a four-chord progression over a warm pad. Its notes bounce between the speakers on dotted-eighth tape echoes, and the whole mix plays back off tape. A Step Sequencer plays the arpeggio. A Chord Sequencer moves it from chord to chord and plays the same chords on the pad: Fmaj9, Cadd9, G6, Am9.

Load it: choose 📚 Examples → Afterglow in the toolbar and press ▶ Play. It needs no keyboard. The patch file is patches/afterglow.json.

Or play it here: press ▶ Play in the corner. The full app is a click away under Open in Soba. The arpeggio voice and the effects run along the top, ending on tape, and the pad runs along the bottom. The cable from the Chord Sequencer's Root up to the arpeggio's oscillator is what moves the arpeggio from chord to chord.

What it teaches

• Transposing with a second sequencer. Pitch inputs add up, so one sequencer can shift another's notes. Here the Chord Sequencer's Root moves the arpeggio over four chords.
• Clocking from end of cycle. A sequencer's EOC output can clock a second sequencer, so the second one moves once per bar.
• Chords on one cable. The Chord Sequencer sends each chord down a polyphonic cable, voiced and voice-led, with the bass underneath. One pad voice plays the whole progression.
• Accents that ring. Velocity stretches the pluck's decay and release, so the loud notes ring on a little after they're struck, as a harder-plucked string does.
• Chords without a third. Sus2 shapes are neither major nor minor, so the same arpeggio shape works on any root.
• Two voices, one space. A bright lead and a dark pad share one chorus, delay and reverb, so they sound as if they're in the same room.
• A mix on tape. A Tape module last in the chain softens the peaks and lets the pitch sway a little, for the warmth of a record rather than a computer.

Modules

How it's built

One shape, four chords

[Clock Gate] ──> [Step Sequencer Clock]
[Step Sequencer Pitch] ──> [Oscillator 1 V/Oct]
[Step Sequencer EOC] ──> [Chord Sequencer Clock]
[Chord Sequencer Root] ──> [Oscillator 1 Exp FM]

The Step Sequencer plays a sixteen-step arpeggio at one note per sixteenth. It uses four notes: C, G, D and the octave C, which are a chord's root, fifth, ninth and octave. They're spread over two octaves.

The Chord Sequencer holds the chord. It's clocked by the Step Sequencer's EOC output, which pulses each time the arpeggio wraps around, so it steps once per bar. Its Root output carries each chord's root as a single pitch: F3, C4, G3, A3. That goes into Oscillator 1's Exp FM input, and with Exp FM at 1.0 octave per volt that input works exactly like a second V/Oct. The oscillator adds the two pitches together. C4 is zero volts, so a chord on F3 moves the whole arpeggio down seven semitones, and C4 leaves it where it is.

The arpeggio's shape has no third. Without one, a chord is neither major nor minor (it's a sus2). So moving the shape doesn't change its mood, and on these four roots every note it plays stays on the white keys, in C major. The pad decides what the chords are (below). Fmaj9, Cadd9, G6 and Am9 each contain the root, fifth and ninth the arpeggio plays over them, so the two parts always agree.

The pluck

[Oscillator 1 Out] ──> [Ladder Filter In]
[Ladder Filter LP24] ──> [VCA 1 In]
[Step Sequencer Gate] ──> [ADSR 1 Gate]
[Step Sequencer Velocity] ──> [ADSR 1 Velocity]
[ADSR 1 Out] ──> [VCA 1 CV]
             ──> [Mix In 1]
[LFO Out] ──> [Mix In 2]
[Mix Out] ──> [Ladder Filter Cutoff]

Two slightly detuned saws go through the ladder filter's 24 dB lowpass. A fast envelope opens both the VCA and the filter, so each note starts bright and dulls as it fades, like a plucked string. Each step has its own velocity. The loud ones fall on steps 1, 4, 7, 11 and 14, a 3-3-4-3-3 pattern that pushes against the even sixteenths, so the line grooves instead of ticking.

[Step Sequencer Velocity] ──> [Attenuverter In]
[Attenuverter Out] ──> [ADSR 1 Decay]
                   ──> [ADSR 1 Release]

Velocity also sets how long each note rings. A string plucked harder swings wider and takes longer to settle, so the accents should ring on where the soft notes are gone. The ADSR's Decay and Release jacks stretch their stages in octaves of time: +1 doubles the knob's time, and 0 leaves it alone. The Attenuverter turns a step's velocity into that many octaves. With Offset at −0.5, a velocity of 64 (half) adds nothing, so the soft steps ring for the knobs' 350 ms as before. The loudest step, 112, rings about 30% longer, and every step in between rings a little longer the harder it's played.

Velocity goes to both jacks because the gate is shorter than the decay. At 60% of a sixteenth, each note's gate closes after about 100 ms, partway through Decay, and the rest of the note is Release. Velocity in Decay alone would only shape the first 100 ms. Watch the ADSR's display: the D and R are tinted while their jacks are patched, and the curve widens on the accents.

The Mix adds a very slow LFO to the envelope before both reach Cutoff. Over 25 seconds the whole arpeggio gets darker and then brighter again. One cycle of the LFO lasts nearly ten bars, so each pass through the progression sounds a little different.

The pad

[Chord Sequencer Pitch] ──> [Oscillator 2 V/Oct]     (poly, 4 channels)
[Oscillator 2 Out] ──> [SVF Filter In]
[SVF Filter LowPass] ──> [VCA 2 In]
[Chord Sequencer Gate] ──> [ADSR 2 Gate]
[ADSR 2 Out] ──> [VCA 2 CV]

The Chord Sequencer plays the bass and the chords on one cable. With Bass voice on and Voices at 4, its Pitch carries the root on its first channel and three chord tones above it. The oscillator, filter, envelope and VCA are all polyphonic, so each channel gets its own five detuned saws, an octave down. The Mixer hears the four channels summed.

The bass is the root: F2, C3, G2, A2. Because the bass already has it, each chord leaves out its own root and plays its colour tones instead. For Fmaj9 that's its 3rd, 7th and 9th (A, E, G), and for Am9 the same (C, G, B). With Voice lead on, each chord takes the inversion nearest the last, so from bar to bar each of the pad's voices holds its note or moves a third at most. The SVF filter at 420 Hz keeps it all dark and round: a glow under the arpeggio rather than a second part.

VCA 2's CV Amt of 60% means the envelope controls only part of the level, and the other 40% is always there. So the pad never stops. On each chord change, the sequencer's short gate (4% of the bar, about 100 ms) fires ADSR 2, and the pad swells, then settles back over two seconds. The very first bar plays over Fmaj9 before the Chord Sequencer has had a clock. The sequencer already sends that first chord, with its gate low, so the pad comes in quietly and blooms on the first change.

Echoes and space

[VCA 1 Out] ──> [Mixer Ch 1]       (Pan R 35)
[VCA 2 Out] ──> [Mixer Ch 2]       (Pan C)
[Mixer Out L/R] ──> [Chorus In L/R]
[Chorus Out L/R] ──> [Stereo Delay In L/R]
[Stereo Delay Out L/R] ──> [Reverb In L/R]
[Reverb Out L/R] ──> [Tape In L/R]

The Mixer sets the arpeggio a little to the right and keeps the pad, which carries the bass, in the middle. Its Master at +3 dB wins back the 3 dB that panning takes from a centred sound on each side. The Chorus widens both parts. The Stereo Delay is synced to a dotted eighth, which is three sixteenths. Each echo therefore falls between the notes of the arpeggio, and Ping-Pong moves it to the other speaker. The arpeggio sounds busier than it is, the echoes stay in time with it, and the rhythm keeps moving. Tape gives the repeats a little wobble and wear. LoCut keeps the bass out of the echoes, so the low end stays clean. The first echo of each note leaves from the arpeggio's side and lands on the other, so the echoes answer the notes across the room. A medium reverb with a five-second tail puts everything in the same room.

On tape

[Reverb Out L/R] ──> [Tape In L/R]
[Tape Out L/R] ──> [Audio Output Left/Right]

Last of all, the whole mix is recorded onto tape and played back. The settings are gentle, a well-kept machine at 15 ips, and each knob adds one thing:

• Sat at 35% pushes the mix into the record head. The plucks' attacks round off, so the peaks come down about 1.5 dB while the mix gets slightly louder (about 0.5 LU). The arpeggio sits into the pad instead of on top of it, and the head bump adds weight to the bass.
• Wow at 20% sways the pitch by about ±1.5 cents, too little to hear as out of tune. The chords seem to breathe. Flut adds a faint shiver, and Width lets the two sides sway a little apart, which widens the pad.
• Age at 20% takes a little off the top, and Hiss at 8% leaves a quiet bed of tape noise that you only notice in the gaps.

To hear what the tape adds, select it and press Ctrl+B to bypass it, then press it again. For more of it, turn up Wow first. Around 40% the chords start to go seasick, which can be lovely for a few bars. Then try Speed at 7½ ips, which sways deeper and dulls the top, like a cassette. For less, set Speed to 30 ips and Wow to 0, so only the glue of Sat is left.

Variations

A new progression. Right-click a Chord Sequencer step to change its root and type. Roots on C, D, F, G and A keep the arpeggio in C major, and types like Dm9 or G6 suit them. E and B roots add an F♯ or a C♯ to the arpeggio, which can sound lovely, but the arpeggio and the pad won't share a key.

A longer progression. Set the Chord Sequencer's Steps to 8 and fill in four more chords, for example Fmaj9 Cadd9 G6 Am9 Fmaj9 Cadd9 Dm9 G6.

A bigger pad. Turn Voices up to 5 or 6, for the 5ths and doubled notes, and set Voicing to Spread. Lower VCA 2's Level to keep the arpeggio on top.

The original sus2 pad. Switch Bass voice off and set Voices to 1. The pad then plays only the roots, as Afterglow did before the Chord Sequencer.

A different arpeggio. Set the Step Sequencer's Dir to P-P to rock back and forth, or Rnd for a new order every bar. Any order of the four notes still fits the chord.

Triplet shimmer. Set the delay's Sync to 1/8T. The echoes now land in threes against the sixteenths.

Glassier. Switch Oscillator 1's Wave to Tri and raise the Ladder Filter's Cutoff to 2 kHz. The pluck goes from felt to glass.

Struck harder. Set the Attenuverter's Offset to 0. Every note now rings longer, and the accents nearly twice the knobs' time, so they hang over the notes after them and the line turns legato on its loud steps. Set Offset to −0.9 for the opposite: the accents ring for the knobs' time and the soft notes are about a fifth shorter, like strings damped by the side of the hand.

A worn cassette. On the Tape, set Speed to 7½ ips, Age to 45%, Hiss to 25% and Wow to 55%. The arpeggio sways like a tape that's been played every day for a summer.

Slower and wider. Turn the Clock down to 70 BPM. The delay follows the new tempo. Then raise the Reverb's Decay to 9 s and its Mix to 50%.

Related

• Chord Sequencer – chord types, voicing and the bass voice
• Step Sequencer – editing steps and the EOC output
• Oscillator – V/Oct and Exp FM
• Tape – what each knob does to a mix, and more starting points
• Generative Ambient – one sequencer, and a second voice that wanders in key
• Rhythmic Sequence – the same dotted-eighth delay under an acid bassline