Re-Scapegrace

A randomising stereo delay · Re-series · Vulpus Labs

Re-Scapegrace records the last sixteen seconds of audio. Each time it gets a trigger, it picks a random slice of that recording and plays it back. Up to sixteen slices can play at once, each with its own length, stereo position and chance of playing backwards, an octave higher or slightly out of tune.

The slices can also be fed back into the recording memory. That means a slice might contain something played several seconds ago, or even an earlier slice, creating layers of sound that build on one another.

This isn't a conventional delay with echoes at regular intervals. It's a way of taking whatever you've played into it and turning it into a changing, unpredictable texture.

Re-Scapegrace is a complete rebuild of Scapegrace. It retains the original's features and adds an internal random clock, control over how far back it looks for slices, a freeze function, trigger output, CV control over every knob, and a display showing the recording and the slices playing from it.

For existing Scapegrace users: see the final section for a list of changes.

To get started: connect a sound to IN L, and connect OUT L and OUT R to your outputs. Turn DENSITY about halfway up, so slices start roughly once a second on average, and increase WIDTH. Bits of the incoming sound will start coming back at different positions in the stereo field. Turn MIX towards WET to hear more of the slices, and raise FEEDBACK to let them build on one another.

How it works

Re-Scapegrace continuously records the input into a sixteen-second memory buffer. Each trigger, whether it comes from the TRIG input or the built-in DENSITY clock, starts a new voice.

That voice picks a section of the recording, plays it once with a fade at either end, and stops.

When a slice starts, its characteristics are chosen at random using the current control settings:

• Length: a random duration within the range set by SIZE.
• Position in the recording: anywhere within the last AGE seconds.
• Direction: forwards or backwards, according to the REVERSE setting.
• Speed: normal or double speed, according to the DOUBLE setting. Double speed raises the pitch by an octave.
• Pitch: a random amount of detuning within the range set by DETUNE.
• Stereo position: a random position within the range set by WIDTH.
• Fades: the fade-in and fade-out durations set by FADE.

Each slice keeps those settings for its entire playback. Turning a knob changes the slices that start next, not the ones already playing. The controls therefore determine the range of possible results, rather than specifying exactly what each slice will do.

Every slice comes from audio already recorded, so playback can begin immediately. Triggers are ignored until at least 100 ms of audio has been recorded.

SLICES

SIZE sets the maximum slice length, from 0.1 seconds at the minimum to sixteen seconds at the maximum. The range changes on a ratio scale, so the steps are proportional rather than evenly spaced. Each slice is then assigned a random length between one tenth of that maximum and the maximum itself.

SIZE starts at the centre, as it did in Scapegrace.

Hover over SIZE to see the current range, for example Size: slices 0.13 to 1.26 s.

A slice cannot be longer than the available recording or the time range allowed by AGE. Shortly after starting the module, slices are limited to the audio recorded so far.

SIZE sets how long a slice plays. A slice at double speed covers twice as much of the recording during that time.

FADE sets the fade-in and fade-out times for each slice. Both fades are at least 100 ms, preventing clicks when a slice starts or ends in the middle of a sound.

At the default centre setting, each fade occupies a quarter of the slice's duration. Turn FADE fully right and the fade-in takes up the first half, with the fade-out taking up the second half. The slice reaches full level only momentarily.

The fades use equal-power curves and never overlap. For slices shorter than 200 ms, the two fades divide the slice equally.

Hover over FADE to see the setting, for example Fade: 50% of each half of the slice, at least 100 ms.

AGE sets how far back in the recording a slice can come from. At minimum, it can use only the last second. At maximum, it can reach back through the full sixteen seconds. The default is fully right.

A low AGE setting makes Re-Scapegrace work with the sound you're playing now, creating something like a stuttering delay. A high setting lets it bring back material from much earlier in the performance, even after the music has moved on.

Hover over AGE to see the range, for example Age: from the last 16.0 s.

WIDTH sets how far from the centre of the stereo field a slice can be placed. Each slice gets a random position within the selected range.

At zero, the default, slices stay in their original stereo position and at their original level. A mono input therefore stays mono. At maximum, slices can land anywhere from hard left to hard right.

The panning uses an equal-power curve, so a slice stays roughly the same loudness wherever it lands. WIDTH changes the slice's centre position but preserves its original stereo spread. Even when a stereo slice is placed hard left or right, some of its sound remains in the other channel.

Hover over WIDTH to see the range, for example Width: voices within +/-60%.

VOICES

DENSITY is Re-Scapegrace's built-in random clock. It starts slices at random times without needing an external trigger.

At minimum, it is off and only TRIG can start a slice. This is the default. Turn it up to increase the rate, from about one slice every ten seconds just above zero to around thirteen slices per second at maximum.

The timing is random, not evenly spaced. Two slices might start close together, followed by a longer gap, rather like raindrops falling.

DENSITY and TRIG work independently, so both can start slices. If they trigger at the same instant, only one slice starts.

Hover over DENSITY to see the average rate, for example Density: 1.1 voices a second on average.

REVERSE sets the probability that a new slice will play backwards. At minimum, no slices are reversed; at maximum, every slice is. It starts at minimum.

Hover over REVERSE to see the probability.

DOUBLE sets the probability that a new slice will play at double speed, raising its pitch by an octave. At minimum, no slices are doubled; at maximum, every slice is. It starts at minimum.

Double-speed playback uses a band-limited recording to avoid aliasing. See Notes below.

Hover over DOUBLE to see the probability.

DETUNE sets how far each slice's pitch can differ from the original, by a random amount in either direction. The maximum is 50 cents (half a semitone) either way. It starts at zero, so slices play at their original pitch unless DOUBLE is active.

A little detuning thickens the sound, much like a chorus, because slices that originally played the same note no longer quite agree. DETUNE also applies to doubled and reversed slices.

Hover over DETUNE to see the range, for example Detune: up to +/-25 cents.

Sixteen voices can play at once. If another slice starts while all sixteen are busy, it takes over from the oldest one. The old slice fades out over 10 ms rather than being cut off abruptly.

With DENSITY at maximum and long slices, the module will generally be playing its sixteen most recently started slices.

LOOP

FEEDBACK sends the slices back into the recording memory along with the live input. Future slices can then contain earlier slices, creating layers of sound that build up over time.

A reversed slice can be recorded and later played forwards. A doubled slice can be doubled again, raising its pitch by another octave.

FEEDBACK starts at zero. At maximum, slices are fed back at full level. Even a modest amount can build up a dense texture; high settings can produce a continuous wall of shifting sound.

The feedback signal is filtered on each pass to prevent excessive low-frequency rumble and high-frequency fizz. Frequencies below 40 Hz and above 9 kHz are rolled off, and DC is removed. The signal is also gently limited at 5 V to keep the recording memory from running away, regardless of the FEEDBACK setting.

Hover over FEEDBACK to see the amount being fed back.

MIX blends the dry input with the slices, from DRY to WET. It uses an equal-power curve to keep the perceived level reasonably consistent as you turn it.

MIX starts at 0.3, so the output is mostly dry, as in Scapegrace.

At the DRY end, the output is the input signal exactly, provided it stays within ±5 V. See the output description below.

Hover over MIX to see the wet proportion.

TRIG starts a slice whenever its input rises above 1 V. It must fall below 0.5 V before it can trigger another slice, preventing a noisy signal around the threshold from producing a burst of triggers.

A gate, clock or trigger signal will work. If TRIG and DENSITY happen to trigger at the same instant, they start just one slice.

START sends out a +5 V trigger lasting 1 ms whenever a slice starts, whether from TRIG or DENSITY. Use it to trigger another module in time with the slices. With DENSITY running, START becomes a random clock.

FREEZE stops the recording memory from being written to. While freeze is active, neither the live input nor the feedback signal is recorded. New slices continue to play from the existing memory, with AGE and SIZE determining which parts they use.

The position in the memory is held relative to the moment FREEZE was activated, so the slices keep working from the captured material rather than continuing to chase the live recording.

The dry input still passes through, allowing you to play new material over the frozen sound.

Switch FREEZE off to resume recording. You can use the button or the GATE input. The gate activates freeze above 1 V and releases it below 0.5 V.

MODULATION

Every knob has a CV input and an AMT knob.

AMT is off at its centre. Turn it right to let the CV move the control; at maximum, a ±5 V signal can sweep the control across its full range. Turn AMT left of centre to reverse the effect.

The CV is added to the knob's setting, and the result is limited to the control's minimum and maximum values.

Most controls are read only when a slice starts. Modulating them therefore changes future slices, not those already playing. For example, an LFO on AGE makes slices draw from further back in the recording at some points in its cycle than at others.

DENSITY, FEEDBACK and MIX are different: these controls respond continuously to their CV inputs, so their settings change as the voltage changes.

I/O

IN L and IN R are the stereo inputs. If IN R is unpatched, IN L feeds both channels.

OUT L and OUT R, on the coral block, are the stereo outputs. Signals pass through unchanged up to ±5 V. Above that, the output bends gradually towards ±10 V and never exceeds it.

Sixteen loud slices playing together can easily push the output into this limiting range. If you hear the sound becoming compressed or distorted, try turning MIX down.

The recording memory is not saved with the patch. Loading a patch starts with an empty memory. The knob settings and FREEZE button state are saved as usual.

The display

The display across the top shows the sixteen-second recording memory, with the oldest material on the left and the newest on the right.

The thin light line marks the newest point in the recording. Faint grey bars show the level of the recorded audio, with one bar representing each tenth of a second and a scale reaching 5 V.

The recording scrolls left as new audio comes in. When FREEZE is active, it stops moving.

Each slice currently playing appears as a band over the section of memory it is using. The band's height shows its stereo position: left at the top, right at the bottom and centre in the middle.

A line across each band shows how far playback has progressed. Forward-playing slices are coral, while reversed slices are violet and their progress lines move from right to left. A slice being replaced by a newer one remains faintly visible while it fades out.

The display shows both what has been recorded and what is being played back. With AGE low, the bands cluster near the right edge. With SIZE high, they can stretch across much of the display.

Jacks

All unpatched inputs read 0 V.

Notes

Double-speed slices avoid aliasing. Doubling the playback speed raises every frequency by an octave. Without filtering, frequencies above 12 kHz would fold back into the audible range as aliasing.

When DOUBLE is active, Re-Scapegrace maintains a second copy of the recording memory with frequencies above 12 kHz removed. Doubled slices use this filtered copy.

The copy is only maintained while DOUBLE is above zero. A slice taken from before DOUBLE was activated may therefore use the unfiltered memory and produce some aliasing. Once DOUBLE has been active for as long as AGE reaches back, all possible doubled slices come from the filtered copy.

Slices at their original pitch use the original samples. A slice that is neither detuned nor doubled plays back the recorded samples directly, without interpolation. Detuned slices use smooth Hermite interpolation to calculate the values between samples.

CPU use depends on how many slices are playing. Re-Scapegrace uses little processing power when idle, more as voices are added, and more again when those voices are detuned. The filtered memory used for DOUBLE has a fixed processing cost while DOUBLE is active, regardless of how many slices use it. There is no additional cost for this copy when DOUBLE is at zero.

The randomness doesn't repeat on a cycle. Unlike Re-Speccy, Re-Scapegrace has no seed setting. Each slice is chosen independently, with no repeating pattern. For slices triggered on a regular beat, connect a clock to TRIG and leave DENSITY off. The timing will be regular, but the content of each slice will still be random.

For Scapegrace users

• Renamed controls. REV is now REVERSE, DBL is DOUBLE, MOD is DETUNE, and FBCK is FEEDBACK. Their functions and default settings are unchanged.
• CV now has an AMT control. In Scapegrace, CV swept a control between its minimum and the knob setting as the voltage moved from -5 V to +5 V. Re-Scapegrace adds the CV to the knob setting instead, with AMT controlling the depth. At full AMT, ±5 V can cover the control's entire range. At the centre setting, AMT disables the CV.
• AGE, DENSITY, START, FREEZE and the display are new. With AGE fully right and DENSITY off, slices can come from anywhere in the full memory and start only when TRIG fires, as in Scapegrace.
• TRIG has defined thresholds. It triggers above 1 V and re-arms below 0.5 V. Scapegrace triggered on any rise above 0 V.
• A seventeenth slice replaces the oldest. Scapegrace ignored new triggers when all sixteen voices were busy. Re-Scapegrace fades out the oldest voice and starts the new slice.
• WIDTH at zero preserves the original level. Stereo slices keep their original stereo spread even when positioned to one side.
• MIX uses an equal-power blend. At the DRY end, the output is the input signal exactly.
• Double-speed playback is band-limited. The outputs also limit signals that rise above ±5 V, bending gradually towards ±10 V.

Starting points

A drifting cloud. Feed a pad or slow melody into IN L. Set DENSITY to the centre, SIZE to three-quarters, FADE fully right, WIDTH fully right, MIX halfway and FEEDBACK to a quarter. Long, soft-edged slices of the incoming sound drift around the stereo field.

Stutters on the beat. Connect a clock to TRIG and leave DENSITY off. Set SIZE to a quarter, FADE fully left, AGE fully left and MIX halfway. Each beat picks a fragment from the last second. Add a little REVERSE to turn some fragments backwards.

Octave shimmer. Set DOUBLE halfway, DETUNE to a quarter, FEEDBACK about halfway, SIZE at the centre, DENSITY just below halfway and MIX halfway. Doubled slices feed back into the memory and can be doubled again, building layers an octave and two octaves above the original sound.

Remembered backwards. Set REVERSE fully right, AGE fully right, SIZE to three-quarters, FADE fully right and DENSITY low. Material from earlier in the performance returns backwards, sometimes long after the original sound has disappeared.

A held memory. Play for sixteen seconds, then press FREEZE. Keep DENSITY running and WIDTH wide. The module continues picking slices from the captured recording while you play new material over it. A slow LFO connected to AGE's CV moves the slices between recent and older material.

Tape smear. Set DETUNE fully right, DENSITY fully right, SIZE to a quarter, FADE fully right, WIDTH halfway and MIX to a third. Many short, slightly detuned slices overlap to create a wavering, chorus-like texture.

A random clock. Set DENSITY to the centre and connect START to an envelope or drum trigger input. Each new slice triggers the other module, so it plays in time with the randomly timed slices.