A little box.
A lot of possibility.
Make a rhythm. Give it a melody. Move your hand and hear it change. This is your friendly guide to making music with Sparsha.
Make your first pattern16 STEPS / 90 BPMSEQ
♩ = 90
Think in little musical loops.
Sparsha is a sequencer and a synthesizer in one instrument. You choose when a note happens, what it plays, and how it feels.
One sound.
A voice is one synth part. It might be a bass, a bell, a drum-like hit or a soft pad. You have four, each with its own pattern.
One moment.
The playhead moves through a row of moments. A lit step plays a note. An unlit step is a rest. At 16 steps, you have a familiar one-bar loop in 4/4.
One thing that changes.
Pitch, loudness, filter and envelope each have their own lane. Give lanes different lengths and your music starts to evolve.
Before you play
Connect the stereo output to your mixer, powered speakers or headphone amplifier. Start with the listening volume low. Power the instrument through the Pico’s USB connection.
On a fresh session, the sequencer starts at 90 BPM, with all steps off. If you have played before, your saved session may load instead.
This is a self-contained instrument. Its sounds come from its own synthesis engine. USB supplies power, serial diagnostics and the Preset Studio connection.
The current firmware has no MIDI input/output or external gate/clock support. The panel’s Gate Input label does not indicate a working input.
Touch. Slide. Turn. Wave.
The controls work together. The OLED tells you what you’re editing right now.
Two voices on the grid.
The 32 pads show a pair of 16-step voices. Read each bank left to right across its two rows. Selecting V1 or V2 shows voices 1 + 2. Selecting V3 or V4 shows voices 3 + 4.
Select a voice, then shape it.
Tap V1–V4 to choose the voice for the dial, hand recording and voice-specific menu actions. The pads still address whichever voice owns their bank.
The LED matrix mirrors the pad grid: lit steps are active, and a moving accent shows playback. When stopped, the matrix breathes.
The dial turns slowly for fine adjustments and quickly for larger changes. Touch a record button to aim it at that parameter.
Four faders. The screen decides their job.
Use these tabs as a quick reminder. Special pages take priority over the normal fader controls.
The SEQ / MENU switch
In SEQ, button 7 is Slide. In MENU, the OLED shows a utility menu and button 7 becomes OK. Buttons 1–5 still record the same lanes, button 6 still plays/stops, and the normal voice and fader controls continue working.
Turn the dial to scroll. Press 7 to run an action or start editing a value. Turn to change that value; press 7 again when done. Toggles flip on a press. Clear voice, Clear all and Load need a 0.4-second hold.
A more specific task—holding a record button, editing a step or opening a page—temporarily takes the screen and dial. The menu returns when that task ends.
Your first pattern, in five moves.
A simple pulse is enough to get going. Everything after that is listening and adjusting.
Pick Voice 1.
Tap V1. On a fresh session it uses the Square preset. Its 16 steps are the top two rows.
Turn on steps 1, 5, 9 and 13.
Tap those pads in the top bank. Each tap toggles a step on or off. Try this on a fresh pattern; on an existing one, choose which steps to keep.
Press Play if you need to.
Tap 6 to start or stop. The moving LED shows the current step. Fader 1 changes tempo; fader 3 changes master volume.
Give it a melody.
Hold 1 Note and move your hand over the distance sensor while the pattern runs. Your hand writes pitch into the active notes of the selected voice. Release the button to keep what you recorded.
Add a second part.
Tap V2. Put a few steps in the bottom bank. On a fresh session, Voice 2 is Bass. Your first two parts now play together.
A pattern starts with a few lights.
Tap the steps, then press Play. Optional sound is a simple browser synth illustration; it does not reproduce Sparsha’s engines or connect to your hardware.
One note can have its own personality.
Long-press a pad for about 0.4 seconds to enter Step Edit. You are now editing that one step, rather than the pattern as it plays.
Shape the note’s envelope.
The four faders become Attack, Decay, Sustain and Release for that step. The OLED shows an ENV page and the selected step.
- Attack
- How quickly the note rises. Short is sharp; long is a swell.
- Decay
- How quickly the peak settles down to the sustain level.
- Sustain
- The level held while the note’s gate is open.
- Release
- How long the sound fades after the gate closes.
The dial edits the chosen lane.
Hold a parameter button and turn the dial to change that parameter on the selected step. Otherwise it uses the active edit parameter or encoder target.
Follow the patch: fresh steps use the sound preset’s own envelope. The OLED puts those values in parentheses. Hold Shift and move an envelope fader to return that lane of the step to the patch value.
Tap a pad, or long-press the selected pad again, to leave Step Edit. Check that the ENV page has gone before recording a moving pattern.
Some sounds use the envelope controls differently
On the waveguide strings, the four faders edit pick hardness, ring time (T60), pick position and stiffness. Hypersaw, NoiseStorm and recipe sounds use the first two for their engine controls; faders 3 and 4 still edit Sustain and Release. Look at the OLED names and listen as you move.
Base value versus a recorded step value
The patch is the voice’s starting sound. Outside Step Edit, the dial changes a voice’s base value. Untouched steps follow it; a step with its own recorded absolute value keeps that value. Velocity, Filter, Attack, Decay, Sustain and Release use this model.
That is why moving a base may leave some steps unchanged. Reset those steps to follow the patch if you want them to move together. Hand recording and ENV faders write absolute values; the middle of the sensor range is not a special “neutral” point.
Repeat a rhythm.
Keep finding something new.
The notes can repeat every 16 steps while the filter repeats every 5. The two loops meet in a different way each time around.
Record a hand movement.
Select a voice. Hold 3 Filter, then move your hand about 55–700 mm above the sensor as the sequence plays. Lower hand positions write lower values; higher positions write higher ones. Release the button when you like the movement.
You can hold several record buttons together. Each writes at its own lane’s current position. Shift + a record button latches recording; repeat the chord to release the latch. Moving your hand out of range pauses writing and keeps the existing values.
Change a lane’s length.
Hold the lane button and tap a pad in the selected voice’s bank. Filter + step 5 makes a five-step Filter lane. It repeats while the other lanes carry on.
For the voice’s overall pattern length, hold its V button for 0.4 seconds and tap a step in its lit bank. This changes the Gate lane length, from 2 to 16 steps.
Loop a smaller slice.
Press a lane button to choose a lane. Hold one pad and touch a second in the same bank: the lane loops between those steps. For example, choose Velocity, then hold step 4 and tap step 8.
Either end can go first. The pads don’t toggle gates during this gesture. The minimum loop span is two steps.
Same clock. Different journeys.
This drawing shows independent lane positions. It is a timing illustration, not an audio or hardware simulator.
Copy a lane to another voice
- Hold the source voice button, then press one of the five record buttons. The OLED shows COPY LANE. Release both.
- Hold the destination voice button and press a lane button to paste. You can paste repeatedly; a different lane receives a scaled version of the shape.
- Press Shift to leave copy mode.
The lane’s values, length and loop start copy together. The destination voice keeps its own sound. A paste replaces the destination lane and has no undo. Exit copy mode before trying to change Gate length.
Slide, randomize and swing
Slide: in SEQ, button 7 toggles Slide mode. Tap steps to set their slide states instead of their gates. A slide glides into the note without retriggering the envelope. Use the encoder’s Slide Time target to set the glide duration; turn Slide mode off to edit gates again.
Randomize: Shift + V2, or the menu’s Randomize row, changes the selected voice’s parameter values while keeping Gate and Slide intact. Your rhythm stays recognizable.
Swing: choose the Swing row in MENU. The groove templates change timing around the beat. Try one against a straight rhythm and listen to where it leans.
Four parts. Pick their personalities.
Every voice can use any factory sound. Make a bass quartet, a cloud of strings, or four completely different parts.
Stop with 6 to open the preset browser, or hold 6 for about 0.4 seconds to open it while playback continues. Select V1–V4, then tap a lit preset pad. In this page, the grid is a sound selector.
The 29 factory sounds occupy the first 29 physical pads in row order. Tap 7 to switch between preset pads and voice-parameter pads. Starting playback closes the stopped browser.
Go deeper with Voice Edit.
Hold Shift, tap V4, then release both. Playback stops and audio mutes. Select a voice with V1–V4. Buttons 1/2 move between groups, 3/4 move between parameters, and the dial changes the displayed base.
Hold 5 for fine adjustment. Hold 6 for 0.7 seconds to reset that parameter to the loaded preset’s value. 7 shows the control guide; 8 Shift exits. Press Play to hear your changes. Save the project to keep them across power cycles.
Make your own presets with Preset Studio
Preset Studio is the Windows companion editor. Connect with a USB data cable and close other programs using the serial port. Start from a factory sound, edit its values, and enable Live or use Play now to hear it on the chosen Sparsha voice.
Name, colour and place the preset, then save the library on your computer. Send to Pico previews new, changed and removed presets before replacing the user bank. Uploading briefly pauses the clock.
User sounds appear on extra browser pages. The firmware’s pad index 31—the 32nd physical pad, bottom-right—is the page key. The two user pages each hold 31 presets. Keep your library file as a backup; the user bank is separate from your song.
Echo it. Surround it. Layer it.
The master effects treat the combined sound of all four voices. Think of them as the room around your pattern.
Delay: a repeating shadow.
In normal play, fader 2 adds delay. Hold Shift and move fader 2 for time (10–375 ms), or fader 1 for feedback. Feedback sets how much echo feeds back into the next repeat.
For repeats tied to the music, enable Delay sync in MENU. Shift + fader 2 then chooses a note division, including dotted and triplet options. The 375 ms buffer still limits the actual delay time. Free-time changes bend the repeats like tape. Delay controls reset after reboot.
Reverb: a space that keeps ringing.
Open Reverb from MENU, or hold Shift + 6, then press 2 and release everything. Playback continues. 2 switches MAIN/TONE, 1 freezes the tail, and Shift exits.
| Layer | Fader 1 | Fader 2 | Fader 3 | Fader 4 |
|---|---|---|---|---|
| MAIN | Mix | Decay | Damping | Unused |
| TONE | Low cut | Diffusion | Mod depth | Width |
Start with a little Mix and a moderate Decay. Damping softens the high end; Low cut clears the lows. Diffusion blends reflections together. Width spreads the wet sound from mono through natural stereo to exaggerated width.
Freeze holds the tail that is already there; a silent tail freezes silence. Reverb settings save with the project, but Freeze and the actual ringing audio do not. Very long tails add up, so raise Mix and Decay gradually.
The looper: catch a moment, build on it.
The loop button is a separate switch from the eight function buttons. It records the mono mix after delay and before reverb.
Choose a length.
Shift + loop opens Loop Settings. Fader 2 chooses 4, 8 or 16 steps for the next take. Shift exits.
Tap to record.
The take waits for the next loop boundary while the transport runs, then records one full pass and plays automatically. With the transport stopped, recording begins at once.
Tap again to add a layer.
The next pass blends live music into the loop. A tap during a take queues another layer; tap again before it starts to cancel that queued layer.
Hold to clear.
Hold the loop button for 0.8 seconds to erase it. There is no undo.
| Loop fader | Job | Try this |
|---|---|---|
| 1 · Loop volume | Recorded loop level | Bring the captured phrase forward. |
| 2 · Length | 4 / 8 / 16 steps | A shorter loop is a tighter refrain. |
| 3 · Sequencer volume | Live mix level after recording tap | Lower it to hear the loop by itself; recording still receives the live bus. |
| 4 · Regen | 10–100% retained each repeat | 100% keeps layers; lower values gradually let older material fade. |
Slower, longer loops have a lo-fi sampler character. Tempo changes retime and pitch-shift the loop. Stopping abandons an unfinished first take; restarting a completed loop plays it from the top. The loop audio and its settings are not saved.
Volume, compression and the live envelope page
Normal fader 3 is master volume. Shift + fader 3 moves the compressor macro from Warm, through Glue, to Punch. Master volume saves; the macro resets to 50% on reboot.
For the selected voice’s base envelope, choose Envelope in MENU or hold Shift + 6, then press a voice button. The four faders are Attack, Decay, Sustain and Release. This changes the voice’s base envelope; steps with their own values retain them. Shift exits.
Hold a chord.
Let it find a path.
The arpeggiator turns the grid into a scale keyboard and walks through the notes you hold.
Tap V1–V4 to pick the sound. Hold several pads to build a chord. The pads are now 32 successive scale degrees, starting at the root. Button 7 in SEQ toggles Latch so the chord keeps playing after your fingers lift.
The four step sequencers pause while Arp is active. Turn the dial for rate. Hold Shift and turn it for tempo. The hand sensor changes note dynamics; it is not recording lanes here.
| Control | Normal | With Shift |
|---|---|---|
| Fader 1 | Hits in the rhythm | Octave range, 1–4 |
| Fader 2 | Rhythm length, 1–16 | Gate, 5–95% of interval |
| Fader 3 | Rotate rhythm | Swing |
| Fader 4 | Accent depth | Filter |
| Buttons 1–5 | Up / Down / Up-Dn / Random / Order | All / Pulse / Tresillo / Five / Orbit rhythms |
| Button 6 | Play/Stop · hold for presets | Seven rhythm |
| Button 7 in SEQ | Latch | Restart the note walk |
The MENU list changes to Arp controls, including Chord pattern, octave range, rhythm, restart, random chord and clear chord. Arp mode is temporary; reboot returns to the step sequencer.
Same pattern.
A different musical world.
A scale chooses the notes you use. A tuning chooses the space between them. Sparsha offers 29 tunings and 47 scale tables across the library.
For a quick change of feel, use Shift + V3 to cycle scales within the current tuning. For a different tuning, select Tuning in MENU and press 7.
Choose by ear.
Tap a tuning pad or turn the dial to explore the library. The grid colours group Equal, Just, Temperament, Indian and Xeno families. A tuning automatically selects a compatible scale.
Buttons 1–6 select scales in the current tuning’s list. Faders 1–3 control tonic, A4 reference and scale. Shift exits.
Make quick comparisons.
Press 7 to A/B swap with the previous tuning. Each remembers its scale while you play.
Tap a voice button to recall a favourite tuning; hold it to store the current tuning. Hold again on that same tuning to clear the slot. In this page, voice buttons are favourites.
Save it. Come back to it.
The instrument stores one current project in its internal flash. Your patterns and sounds can be waiting when you power up again.
Save deliberately.
Flip to MENU, choose Save and press 7. The clock briefly pauses for the flash write, then resumes. Look for SAVED.
Know about autosave.
About one second after you stop transport, changed project data automatically saves. Stopping is not a way to preserve an older version.
Recall the last saved project.
Choose Load in MENU and hold 7 for 0.4 seconds. This replaces the live project with the stored one. On power-up, a valid save loads automatically.
Kept with your project
- All four patterns and lane lengths / loop starts
- Presets and your Voice Edit changes
- Tempo, master volume, scale and swing
- Tuning, reference pitch and favourites
- Reverb settings and LED theme
Temporary performance state
- The audio loop and loop settings
- Delay mix, time, feedback and sync choice
- Reverb Freeze and the ringing tail
- Compressor macro position
- Arpeggiator mode and per-tuning scale memory
User presets sent from Preset Studio live in a separate bank. Saving the song keeps the sound values on each voice; replacing that bank does not rewrite a saved song’s sound.
What happens after a freeze?
A retained-RAM copy is refreshed about once a second. A watchdog reboot can resume that recent live session. If recovery fails repeatedly or the copy is invalid, the instrument parks instead of cycling forever. A power cycle tries the flash save. This recovery mechanism does not save temporary loop audio.
If the instrument is unresponsive, reconnect power and check the startup messages. Save important work deliberately before unplugging.
The useful bits, at a glance.
These gestures are for normal step-sequencer operation. A dedicated page takes priority; check its OLED footer.
| You want to… | Do this |
|---|---|
| Choose a voice | Tap V1–V4. |
| Turn a step on / off | Tap its pad in the voice’s bank. |
| Edit one step | Long-press its pad, about 0.4 s. |
| Clear one step | Shift + pad. Gate goes off and values reset. |
| Play / Stop | Button 6, or Shift + V1. |
| Open preset browser | Stop transport, or hold button 6, about 0.4 s. |
| Record a lane | Hold 1 Note / 2 Velocity / 3 Filter / 4 Attack / 5 Release and move your hand. |
| Latch recording | Shift + record button. Repeat to unlatch. |
| Set a lane’s length | Hold its record button + tap an end step. |
| Set a lane’s loop range | Choose the lane, hold one pad, press another in the same bank. |
| Set voice pattern length | Hold voice 0.4 s + tap step in its lit bank. |
| Copy / paste a lane | Hold voice + press lane. Repeat on destination; Shift exits. |
| Randomize values | Shift + V2, or Randomize in MENU. Gates and slides stay. |
| Cycle scale | Shift + V3, or Scale in MENU. |
| Edit a voice’s sound | Shift + short tap V4, release both. |
| Enter / leave Arp | Shift + hold V4, about 0.4 s. |
| Open voice envelope page | Shift + hold 6 + voice button; release. Or Envelope in MENU. |
| Open Reverb | Shift + hold 6 + button 2; release. Or Reverb in MENU. |
| Open Tuning | Tuning in MENU, then button 7. |
| Record / layer a loop | Tap separate loop button. Hold 0.8 s to clear. |
| Open Loop Settings | Shift + loop button. Shift exits. |
| Save / Load | MENU → Save, tap 7. MENU → Load, hold 7 for 0.4 s. |
Menu map
| Section | Rows in step-sequencer mode |
|---|---|
| Transport | Play/Stop · Tempo · Swing |
| Sound | Presets · Voice edit · Scale · Tuning |
| Effects | Delay sync · Delay time · Reverb · Envelope |
| Pattern | Randomize · Slide · Arp mode · Clear voice · Clear all |
| System | Save · Load · Theme · Enc target · Serial dump |
Clear voice resets all of the selected voice’s lanes, gates, slides and lengths. Clear all does that for all four voices; presets and tempo stay. Both need a 0.4-second hold of button 7. There is no undo.
Something feels different?
Check the context.
Most surprises come from a page, a held button or an unlit gate. The OLED is your first clue.
I can’t hear anything.
Check the output connection, mixer level and master volume (normal fader 3). Confirm transport is running and some steps are lit. If you just left Voice Edit, press Play; it deliberately leaves transport stopped. In Arp, check that you have chord notes and Hits is not zero.
The pads choose sounds instead of switching steps.
You are in the preset browser. Start playback to leave the stopped browser, or hold button 6 to toggle the browser. Tuning, Arp and other pages also give the grid different jobs.
The faders changed envelopes instead of tempo or volume.
You selected a step for editing. The ENV page means the four faders are editing that step. Tap a pad or long-press the selected pad again to leave Step Edit. Reverb, Loop, Tuning and Arp also reassign faders.
I can’t record a pitch, or the hand seems to do nothing.
Turn the target gate on first. Select the right voice and hold or latch a record button. Keep your hand in the 55–700 mm window. An out-of-range or invalid reading pauses recording. Check for Step Edit; it directs writing to one selected step. Strong light can cause sensor jitter.
The dial scrolls a menu instead of changing my sound.
In MENU, the idle dial scrolls rows. Hold a record button to target a parameter, or select a step. You can also return the switch to SEQ. Special pages have their own dial actions.
Some notes go quiet after I edit the envelope.
A very long Attack can outlast a short gate. A very short Decay and zero Sustain can make a note almost disappear. On that step, hold Shift and move the affected envelope fader to restore the patch value. Then try smaller changes.
My echo or loop disappeared after power-off.
Delay settings and loop audio are temporary. Reverb settings save, but the ringing tail and Freeze do not. Patterns, voice patches and saved musical settings should restore from the project.
The screen says Voice 0. Is there a fifth voice?
No. Some screens and older diagnostics count from zero: 0, 1, 2, 3 correspond to the panel’s V1, V2, V3, V4.
What does NO LOOP mean?
The loop buffer could not be allocated at boot. The rest of the instrument remains available. Keep the exact startup diagnostics if reporting it; the loop memory depends on the current firmware and available RAM.
A small musical dictionary
- Gate
- The “hold this note” signal. Gate length changes duration, not pitch.
- Velocity
- A note’s strength, usually heard as loudness.
- Filter
- A tone control. Lower cutoff usually darkens a sound; higher opens it.
- Patch / preset
- A stored starting sound and its settings.
- Polymeter
- Loops of different lengths sharing the same clock.
- Portamento / slide
- A glide from one pitch to the next.
- Wet / dry
- Effect sound / original sound.
- Regeneration
- How much of a loop survives into its next pass.