# Hyperwave

> Hyperwave (hyperwave.app) is a browser-based node-graph wavetable generator: build tables visually and export packed WAV files that load into Serum, Vital, and wavetable-capable hardware. Not affiliated with Blamsoft Expanse Hyperwave or Roland SDZ004 Hyperwave.

- Browser-based node-graph wavetable generator: tables are designed and rendered entirely in the browser, then exported as packed wavetable WAV files.
- Export container: RIFF/WAVE WAV with "fmt " and "data" chunks, written with a canonical 44-byte header (16-byte fmt chunk).
- Sample format: IEEE float (format tag 3) by default — 32-bit samples, 4 bytes each, 1 channel (mono); 16-bit PCM WAV is a direct export option for classic hardware banks.
- Header sample rate: 48,000 Hz by default (44,100 Hz selectable).
- Frame layout: single-cycle frames packed end-to-end in one data chunk, in wavetable order.
- Documented import targets: 33 total (18 software, 11 hardware, 4 modular/Eurorack). The target registry records a source check for 32.
- Glossary: 44 plain-language wavetable synthesis terms.

## Docs
- [Hyperwave docs](https://www.hyperwave.app/docs): how the node graph shapes wavetable output and where exports go.
- [Export formats](https://www.hyperwave.app/docs/formats): the packed-WAV byte contract and the import format every documented target expects.
- [Node library](https://www.hyperwave.app/docs/nodes): generators, spectral tools, modulators and curves, shapers, and inspection, in outline.
- [Tutorials](https://www.hyperwave.app/docs/tutorials): Make a wavetable from a sample; Choosing a wavetable frame size; Import wavetables into Serum and Vital; Wavetables for hardware synths.

## Guides
- [Wavetable import guides](https://www.hyperwave.app/guides): one step-by-step import guide per documented target.
- [How to import wavetables into Serum](https://www.hyperwave.app/guides/serum): Drag-and-drop packed-WAV import at 2,048-sample frames
- [How to import wavetables into Serum 2](https://www.hyperwave.app/guides/serum-2): Wavetable Editor imports packed WAVs at 2,048-sample frames (up to 256)
- [How to import wavetables into Vital](https://www.hyperwave.app/guides/vital): Packed-WAV import at Vital's assumed 2,048-sample frames
- [How to import wavetables into Pigments](https://www.hyperwave.app/guides/pigments): Imports .wav wavetables as 2,048-sample positions, up to 256, via browser or drag-and-drop
- [How to import wavetables into Phase Plant](https://www.hyperwave.app/guides/phase-plant): Loads 256×2,048-sample WAV/FLAC wavetables; other lengths convert in the editor
- [How to import wavetables into DUNE 3](https://www.hyperwave.app/guides/dune-3): WAV import at 2,048-sample frames, up to 256 waveforms
- [How to import wavetables into Surge XT](https://www.hyperwave.app/guides/surge-xt): Drag-and-drop packed WAV import
- [How to import wavetables into Wavetable (Live)](https://www.hyperwave.app/guides/ableton-wavetable): Drag WAV/AIFF onto the display; Live re-slices to 1,024-sample frames
- [How to import wavetables into Icarus 2](https://www.hyperwave.app/guides/icarus-2): Serum-compatible packed WAV at 2,048-sample frames via drag & drop or menu
- [How to import wavetables into Avenger](https://www.hyperwave.app/guides/avenger): Drops packed WAVs onto a Resampler OSC; you set the frame length (128–4,096)
- [How to import wavetables into Hive 2](https://www.hyperwave.app/guides/hive-2): Loads 2,048-sample packed WAVs (up to 256 frames) from its wavetable folder
- [How to import wavetables into Bitwig Studio](https://www.hyperwave.app/guides/bitwig-sampler): Load compatible WAV wavetables in Polymer or The Grid
- [How to import wavetables into modwave native](https://www.hyperwave.app/guides/modwave-native): Imports WAV as wavetable straight from the Librarian page — same formats as the hardware
- [How to import wavetables into UVI Falcon](https://www.hyperwave.app/guides/falcon): Drag-and-drop import with a _{samples} filename suffix
- [How to import wavetables into HALion 7](https://www.hyperwave.app/guides/halion): Wavetable Editor imports packed WAVs via Fixed Size and Window Size
- [How to import wavetables into Vaporizer2](https://www.hyperwave.app/guides/vaporizer2): Free (GPL) synth importing Serum/Icarus packed WAVs by drag-and-drop
- [How to import wavetables into SynthMaster (2 / One / 3)](https://www.hyperwave.app/guides/synthmaster): Drag-and-drop WAV wavetables across the SynthMaster family
- [How to import wavetables into Codex](https://www.hyperwave.app/guides/codex): IMPORT button converts any WAV into a wavetable automatically
- [How to import wavetables into Blofeld](https://www.hyperwave.app/guides/blofeld): Import WAV or SysEx through Waldorf's plugin
- [How to import wavetables into Iridium / Quantum](https://www.hyperwave.app/guides/iridium-quantum): On-instrument WAV import at a set sample period
- [How to import wavetables into Peak / Summit](https://www.hyperwave.app/guides/peak-summit): Import audio through the Components editor
- [How to import wavetables into modwave](https://www.hyperwave.app/guides/modwave): Loads Serum-style tables via the Editor/Librarian
- [How to import wavetables into nanobox fireball](https://www.hyperwave.app/guides/nanobox-fireball): 2,048 × 256 WAVs from microSD
- [How to import wavetables into ELZ_1 play](https://www.hyperwave.app/guides/elz-1-play): Serum-format WAVs registered from SD, 32 user slots
- [How to import wavetables into Pro 3](https://www.hyperwave.app/guides/sequential-pro-3): 16 waves × 1,024 samples via the Wavetable Generator utility and SysEx
- [How to import wavetables into Waldorf M](https://www.hyperwave.app/guides/waldorf-m): Headerless 8-bit wavetable files from SD (conversion required)
- [How to import wavetables into Elektron Tonverk](https://www.hyperwave.app/guides/tonverk): Float-friendly hardware: 2,048-sample frames, up to 64 waves; _wt<number> suffix for other sizes
- [How to import wavetables into MicroFreak](https://www.hyperwave.app/guides/microfreak): MIDI Control Center imports WAV/AIFF into a 16-slot wavetable bank (firmware 4.0+)
- [How to import wavetables into Deluge](https://www.hyperwave.app/guides/deluge): Official firmware 4.0+ auto-detects wavetable WAVs from SD
- [How to import wavetables into E352 Cloud Terrarium / E370 Quad VCO](https://www.hyperwave.app/guides/e352-e370): microSD WAV banks built in WaveEdit
- [How to import wavetables into Piston Honda MK3](https://www.hyperwave.app/guides/piston-honda-mk3): Eight 16-bit WAV files on microSD
- [How to import wavetables into Disting EX](https://www.hyperwave.app/guides/disting-ex): Serum-format native; 2,048-sample auto-detection from SD
- [How to import wavetables into VCV Rack (Wavetable VCO)](https://www.hyperwave.app/guides/vcv-rack): Right-click Load wavetable on the free bundled VCO; any common bit depth

## Tools
- [WAV to wavetable converter](https://www.hyperwave.app/tools/wav-to-wavetable): convert an audio file into a packed wavetable WAV in the browser.
- [Wavetable inspector](https://www.hyperwave.app/tools/wavetable-inspector): parse a wavetable WAV and report its frame layout and metadata chunks.
- [Frame size calculator](https://www.hyperwave.app/tools/frame-size-calculator): sample/frame/byte math for every documented target.

## Reference
- [Export targets](https://www.hyperwave.app/wavetables-for): import steps and failure modes for every documented synthesizer and module.
- [Compare wavetable tools](https://www.hyperwave.app/compare): a source-linked comparison of 13 wavetable creation tools.
- [llms-full.txt](https://www.hyperwave.app/llms-full.txt): this document plus the complete glossary text, every export-target entry, the WAV format facts, and recent changelog sections.

## Changelog
- [Hyperwave changelog](https://www.hyperwave.app/changelog): available dated notes from the app changelog, newest first.
- [Changelog Atom feed](https://www.hyperwave.app/changelog.xml): machine-readable feed of those notes.

## Glossary
- [Wavetable glossary](https://www.hyperwave.app/glossary): all 44 definitions on one page.
- [Wavetable synthesis](https://www.hyperwave.app/glossary/wavetable-synthesis): Wavetable synthesis plays short, looping single-cycle waveforms from a table and sweeps across them to change timbre while pitch holds still.
- [Wavetable position](https://www.hyperwave.app/glossary/wavetable-position): Wavetable position is the index into a wavetable's frames that selects which single-cycle waveform, or blend of adjacent waveforms, the oscillator is currently playing.
- [Single-cycle waveform](https://www.hyperwave.app/glossary/single-cycle-waveform): A single-cycle waveform is one complete period of a periodic signal, stored as a fixed number of samples and looped by an oscillator to produce a steady pitch.
- [Wavetable frame](https://www.hyperwave.app/glossary/wavetable-frame): A wavetable frame is one single-cycle waveform occupying one slot in a wavetable; the number of frames sets how much timbral motion the table can hold.
- [Wavetable frame size](https://www.hyperwave.app/glossary/wavetable-frame-size): Wavetable frame size is the number of samples in each single-cycle frame, held constant across the whole table.
- [Interpolation between frames](https://www.hyperwave.app/glossary/interpolation-between-frames): Interpolation between frames is how a wavetable oscillator blends its two neighboring frames when position lands between them, so timbre changes continuously instead of in audible steps.
- [2048 × 256 table](https://www.hyperwave.app/glossary/2048x256-table): A 2048 × 256 wavetable (256 frames of 2048 samples each) is the de facto modern standard layout, popularized by Serum and adopted by many current wavetable synths.
- [256-sample table](https://www.hyperwave.app/glossary/256-sample-table): A 256-sample wavetable has frames that are 256 samples long. Several Eurorack wavetable modules use this frame size.
- [8×8 wavetable bank](https://www.hyperwave.app/glossary/8x8-wavetable-bank): An 8×8 wavetable bank lays 64 wavetables out on a grid, so two control signals, one per axis, scan the whole bank instead of one table's frames.
- [Morphing vs. crossfading](https://www.hyperwave.app/glossary/morphing-vs-crossfading): Crossfading is an amplitude blend between two wavetable frames, while morphing is the broader goal it serves: moving between frames so the result reads as one continuously evolving timbre.
- [Spectral morphing](https://www.hyperwave.app/glossary/spectral-morphing): Spectral morphing transitions between wavetable frames by interpolating partial amplitudes and phases instead of blending waveform samples directly.
- [Harmonics](https://www.hyperwave.app/glossary/harmonics): Harmonics are components of a periodic waveform at whole-number multiples of its fundamental frequency. The second harmonic is at twice that frequency and the third at three times it.
- [Partials](https://www.hyperwave.app/glossary/partials): Partials are the individual sine-wave components of a sound's spectrum. They include harmonics and, in sounds that contain them, inharmonic components.
- [Formant](https://www.hyperwave.app/glossary/formant): A formant is a resonant region in a spectrum that stays at roughly the same frequency as the played pitch changes. Formants help distinguish vowel sounds.
- [Phase](https://www.hyperwave.app/glossary/phase): Phase is each partial's offset within the waveform's cycle, where in the loop every sine component starts, measured as an angle rather than a time.
- [Phase distortion](https://www.hyperwave.app/glossary/phase-distortion): Phase distortion reshapes a waveform by varying the rate at which an oscillator reads it within each cycle.
- [FFT resynthesis](https://www.hyperwave.app/glossary/fft-resynthesis): FFT resynthesis analyzes a waveform with a fast Fourier transform (the algorithm that reports one magnitude and phase per frequency slot), edits that spectrum, and rebuilds a waveform from it with the inverse transform.
- [Resynthesis](https://www.hyperwave.app/glossary/resynthesis): Resynthesis is analyzing an existing sound and rebuilding it as a wavetable, converting spans of the source into single-cycle frames rather than playing back samples.
- [Waveshaping](https://www.hyperwave.app/glossary/waveshaping): Waveshaping passes a signal through a transfer curve. It computes each output sample from an input sample to change the waveform's harmonic content.
- [Wavefolding](https://www.hyperwave.app/glossary/wavefolding): Wavefolding is a waveshaping technique that reflects a signal back on itself once it passes a threshold, multiplying harmonics without changing the fundamental.
- [Hard sync](https://www.hyperwave.app/glossary/hard-sync): Hard sync is an oscillator arrangement in which a master oscillator resets a second, detuned oscillator every time the master completes its own cycle, producing a biting, formant-rich tone.
- [Through-zero FM](https://www.hyperwave.app/glossary/thru-zero-fm): Through-zero FM is frequency modulation in which the carrier's frequency is allowed to pass through zero and come out negative, continuing the phase in reverse rather than stopping.
- [Pulse-width modulation](https://www.hyperwave.app/glossary/pwm): Pulse-width modulation (PWM) varies a pulse wave's duty cycle. The duty cycle is the fraction of each cycle spent at its high level, and changing it changes the timbre.
- [Formula-generated wavetable](https://www.hyperwave.app/glossary/formula-generated-wavetable): A formula-generated wavetable has frames computed from expressions evaluated per sample and per frame.
- [Aliasing](https://www.hyperwave.app/glossary/aliasing): Aliasing is inharmonic distortion that appears when a digital signal contains frequencies above half the sample rate (Nyquist), which reflect back down into the audible range.
- [Band-limiting](https://www.hyperwave.app/glossary/band-limiting): Band-limiting restricts a waveform's harmonics to frequencies below the Nyquist limit, which is half the sample rate.
- [DC offset](https://www.hyperwave.app/glossary/dc-offset): DC offset is a nonzero average level in a waveform. The cycle sits above or below zero instead of being centered on it.
- [Wavetable bank](https://www.hyperwave.app/glossary/wavetable-bank): A wavetable bank is a grouped collection of wavetables that an instrument loads together and offers under a single browser or selector.
- [WAV file](https://www.hyperwave.app/glossary/wav-file): A WAV file is the everyday audio container: a header describing the format, a block of sample data, and room for extra metadata chunks.
- [Packed wavetable WAV](https://www.hyperwave.app/glossary/packed-wavetable-wav): A packed wavetable WAV is a single audio file that stacks every frame of a wavetable end to end, so a wavetable-capable synth can import the whole table from one file.
- [clm chunk](https://www.hyperwave.app/glossary/clm-chunk): The 'clm ' chunk is optional metadata in some Serum-format wavetable WAV files; plain WAV exports may omit it.
- [srge chunk](https://www.hyperwave.app/glossary/srge-chunk): The 'srge' chunk is Surge XT's own versioned metadata chunk inside WAV files, telling its importer the frame layout the same way Serum's 'clm ' chunk does.
- [wt file](https://www.hyperwave.app/glossary/wt-file): A .wt file is Surge XT's native wavetable format: a short tagged header followed by the table's frames as raw samples, optionally trailed by metadata.
- [vitaltable file](https://www.hyperwave.app/glossary/vitaltable-file): A .vitaltable file is Vital's native wavetable format. It is a readable JSON document describing how the table was built instead of storing only rendered audio.
- [uhm format](https://www.hyperwave.app/glossary/uhm-format): A .uhm file is a u-he Hive wavetable script: a readable text file of commands and formulas that Hive evaluates step by step to build a wavetable.
- [Additive synthesis](https://www.hyperwave.app/glossary/additive-synthesis): Additive synthesis builds a sound by summing sine-wave partials. It sets each component's amplitude and, optionally, its phase.
- [Subtractive synthesis](https://www.hyperwave.app/glossary/subtractive-synthesis): Subtractive synthesis starts with a harmonically rich waveform and removes or reduces parts of its spectrum with filters.
- [Granular synthesis](https://www.hyperwave.app/glossary/granular-synthesis): Granular synthesis builds sound from many short grains. Each grain is a brief, windowed slice of an audio buffer played at a chosen time and pitch.
- [Wave sequencing](https://www.hyperwave.app/glossary/wavesequencing): Wave sequencing plays a programmed series of waves within one note. The Korg Wavestation supports up to 255 steps with settings for duration, pitch, and crossfade.
- [PPG Wave](https://www.hyperwave.app/glossary/ppg-wave): The PPG Wave is the synthesizer line that introduced commercial wavetable synthesis, built by Wolfgang Palm's Palm Products GmbH from 1981 to 1987.
- [Waldorf Microwave](https://www.hyperwave.app/glossary/waldorf-microwave): The Waldorf Microwave (1989) is the rack synthesizer that carried PPG's wavetable lineage forward after PPG closed, pairing wavetable scanning with an analog filter.
- [Korg DW-8000](https://www.hyperwave.app/glossary/korg-dw-8000): The Korg DW-8000 (1985) is a hybrid synthesizer with 16 stored digital waveforms and an analog filter. Its DWGS oscillators select a fixed waveform rather than scan a wavetable.
- [Ensoniq Transwave](https://www.hyperwave.app/glossary/ensoniq-transwave): A Transwave is Ensoniq's variant of wavetable scanning. Instruments such as the VFX, TS, and Fizmo sweep stored waves with programmable start, end, and loop points.
- [Adventure Kid Waveforms](https://www.hyperwave.app/glossary/adventure-kid-waveforms): Adventure Kid Waveforms (AKWF) is Kristoffer Ekstrand's free library of several thousand single-cycle WAV files for samplers and wavetable synthesizers.

---

# Full content

## WAV format

The exact packed-WAV byte contract Hyperwave writes.

- Container: RIFF/WAVE WAV with "fmt " and "data" chunks; canonical 44-byte header with a 16-byte fmt chunk.
- Sample format: IEEE float, format tag 3 — 32-bit samples, 4 bytes each — the default; 16-bit PCM (format tag 1, 2 bytes per sample) is a direct export option for classic Eurorack loaders.
- Channels: 1 (mono); every frame is one single-cycle waveform.
- Header sample rate: 48,000 Hz by default (44,100 Hz selectable).

## Export targets

Every documented wavetable import target, in registry order (33 total). A source check records evidence from Hyperwave's export code or manufacturer documentation; it does not establish a hands-on device test.

### Serum (Xfer Records)

Category: software. Source check: recorded.

Import format: Packed mono WAV, 2,048 samples per frame, up to 256 frames (~11.9 s at 44.1 kHz); Serum's own wavetable exports are 44.1 kHz mono in 16- or 32-bit, and the sample-rate header doesn't affect frame layout. An optional 'clm ' metadata chunk records frame size and morph mode.

Sources (checked 2026-08-16): https://xferrecords.com/forums/general/file-types

### Serum 2 (Xfer Records)

Category: software. Source check: recorded.

Import format: Serum 2's Wavetable Editor imports audio files through its Import menu or file drag-and-drop. Its internal wavetable layout uses 2,048 samples per frame and up to 256 frames (2 MB of 32-bit sample data). Import methods include Dynamic Pitch, Frequency Estimation, Constant Framesize, and FFT; the selected method can change the source data.

Sources (checked 2026-09-23): https://xferrecords.com/manual/serum-2

Notes: Current Serum 2 documentation does not establish how it uses clm or uhWT metadata from an arbitrary third-party WAV. Choose an import method and check the resulting table.

### Vital (Matt Tytel)

Category: software. Source check: recorded.

Import format: Packed WAV (FLAC loads too when dragged in) imported frame-by-frame; Vital assumes 2,048 samples per frame and reads the left channel of stereo files. Its own wavetable exports are 256-frame, 2,048-sample WAVs.

Sources (checked 2026-08-16): https://github.com/mtytel/vital

### Pigments (Arturia)

Category: software. Source check: recorded.

Import format: Loads user wavetables from a single .wav parsed into consecutive 2,048-sample positions, up to 256 positions; only the first 524,288 samples are used. Bit depth and other frame sizes are not documented.

Sources (checked 2026-08-16): https://dl.arturia.net/products/pigments/manual/pigments_Manual_7_0_0_EN.pdf

Notes: Verified against §8.12 (pp. 103–104) of the Pigments 7.0 manual.

### Phase Plant (Kilohearts)

Category: software. Source check: recorded.

Import format: Loads user wavetables from WAV or FLAC files of exactly 256 frames × 2,048 samples (524,288 samples); the built-in wavetable editor converts any other sample into a wavetable. Bit depth is not specified.

Sources (checked 2026-08-16): https://kilohearts.com/docs/phase_plant https://kilohearts.com/docs/wavetables

### DUNE 3 (Synapse Audio)

Category: software. Source check: recorded.

Import format: Wavetable Editor imports WAV with 2,048 samples per waveform — a packed file loads frame-per-waveform via Import WAV (Adjust Wavetable), up to 256 waveforms; drag-and-drop onto the wavetable display since v3.6. Exports as WAV or DUNE's own WT format.

Sources (checked 2026-08-16): https://www.synapse-audio.com/DUNE-36-Manual.pdf https://www.synapse-audio.com/dune3.html

### Surge XT (Surge Synth Team)

Category: software. Source check: recorded.

Import format: Loads packed mono WAV — 16-bit PCM or 32-bit float — with power-of-two frame sizes from 2 up to 4,096 samples, plus its native .wt format. Its exporter can also write Serum-style 'clm ' metadata.

Sources (checked 2026-08-16): https://github.com/surge-synthesizer/surge

### Wavetable (Live) (Ableton)

Category: software. Source check: recorded.

Import format: Drag-and-drop WAV or AIFF onto the oscillator's wavetable display (Live 10.1+, Suite). Live slices imports into 1,024-sample wavetables — at most 256 per file, downmixed to mono — and post-processes them unless Raw mode is on; files created in Serum (2,048-sample frames) are downsampled to 1,024.

Sources (checked 2026-08-16): https://www.ableton.com/en/manual/live-instrument-reference/ https://help.ableton.com/hc/en-us/articles/360002719179-User-Wavetables

Notes: Verified against chapter 30.13 of the Live Instrument Reference and Ableton's User Wavetables help article.

### Icarus 2 (Tone2)

Category: software. Source check: recorded.

Import format: Loads packed WAV wavetables in the Icarus wav format — frames in series at exactly 2,048 samples each, 1–256 waveforms — including tables from Serum and Falcon; drag & drop onto the oscillator's 3D display or load via the oscillator/editor menu.

Sources (checked 2026-08-16): https://www.tone2.com/handbook/Tone2_Icarus2.pdf https://www.tone2.org/forum/index.php?topic=2417.0

Notes: Icarus's own wav format is documented 16-bit mono; accepted import bit depths aren't specified, though Serum-format tables (32-bit float) load per the handbook and Tone2's forum.

### Avenger (Vengeance Sound)

Category: software. Source check: recorded.

Import format: Two documented WAV routes: a one-cycle WAV converts to an OSC shape, and a multi-frame WAV dropped on a Resampler/Wavetable OSC's wavetable view is sliced into a wavetable using user-set Root (128–4,096 samples per waveform) and Increment spacing — 2–256 waveforms per table, then exportable. No automatic packed-WAV or frame-metadata detection is documented.

Sources (checked 2026-08-16): https://downloads.vengeance-sound.com/products/Avenger/Manuals/VPS_Avenger2_Manual_English_2_4_1.pdf https://vengeance-sound.com/plugins.php?sub=Vengeance+Producer+Suite+Avenger+2

### Hive 2 (u-he)

Category: software. Source check: recorded.

Import format: Mono or stereo WAV interpreted as 2,048-sample frames by default, with the frame count calculated from the file size up to 256; other frame sizes via a -WT64/128/256/512/1024 filename suffix (e.g. MyTable-WT512.wav). Also loads .uhm text-scripted wavetables.

Sources (checked 2026-08-16): https://u-he.com/downloads/manuals/plugins/hive/Hive-Wavetables.pdf

### Bitwig Studio (Bitwig)

Category: software. Source check: recorded.

Import format: Polymer and The Grid have a Wavetable oscillator that reads Surge-format WT files and Serum- or WaveEdit-compatible WAV wavetables. Hyperwave's default packed WAV uses the Serum-style 2,048-sample frame layout. Bitwig's separate Sampler Cycles workflow changed in 6.1 and needs an application check before giving exact Root Key steps.

Sources (checked 2026-09-23): https://www.bitwig.com/polymer/ https://www.bitwig.com/userguide/latest/grid_modules/ https://downloads.bitwig.com/6.1/Release-Notes-6.1.html

### modwave native (Korg)

Category: software. Source check: recorded.

Import format: Same two import formats as the hardware modwave: 32-bit float with exactly 2,048-sample waves (Serum-style, up to 64 waves end-to-end) or 16-bit with exactly 256-sample waves — also capped at 64 waves per file.

Sources (checked 2026-08-16): https://support.korguser.net/hc/en-us/articles/13236558460313-How-can-I-create-new-Wavetables https://storage.korg.com/korgms/support/modwave_native/modwave_native_OM_E.pdf

### UVI Falcon (UVI)

Category: software. Source check: recorded.

Import format: Drag-and-drop audio (or image) files onto the Wavetable oscillator's waveform display, or drop them in User Data/Wavetable for the User list. Sequential-slice convention: the samples-per-slice is appended to the filename after an underscore (e.g. MySweep_128.wav); one slice per channel for stereo. No fixed frame size or bit depth is documented.

Sources (checked 2026-08-16): https://support.uvi.net/hc/en-us/articles/360001260738-Falcon-Loading-Custom-Wavetables https://www.uvi.net/falcon

### HALion 7 (Steinberg)

Category: software. Source check: recorded.

Import format: Imports wavetables saved as .wav files — one wave or multiple single-cycle waves. Auto-imported when the WAV header carries wave-size information; otherwise the Wavetable Editor's Fixed Size mode and Window Size set the single-cycle length.

Sources (checked 2026-08-16): https://www.steinberg.help/r/halion/7.1/en/halion/topics/wavetable_synthesis/wavetables_importing_t.html

### Vaporizer2 (VAST Dynamics)

Category: software. Source check: recorded.

Import format: Wavetable editor imports standard Serum/Icarus-format packed wavetables, including via drag-and-drop.

Sources (checked 2026-08-16): https://github.com/VASTDynamics/Vaporizer2 https://www.vast-dynamics.com?q=Vaporizer2

### SynthMaster (2 / One / 3) (KV331 Audio)

Category: software. Source check: recorded.

Import format: Loads wavetables from WAV files by drag-and-drop (SynthMaster 2.9+, One, and 3); up to 256 waveforms; SynthMaster 3's wavetable editor imports and exports WAV with 2–256 frames. Frame size is not documented on official pages (community reports 2,048 samples).

Sources (checked 2026-08-16): https://www.kv331audio.com/synthmaster3.aspx https://kv331audio.com/synthmaster.aspx https://www.kv331audio.com/synthmasterone.aspx

### Codex (Waves)

Category: software. Source check: recorded.

Import format: The oscillator's IMPORT button converts user WAV files of any length, bit depth, and sample rate into wavetables automatically; 64 wavetable slots (internal or imported). 1–5 second files recommended.

Sources (checked 2026-08-16): https://www.waves.com/plugins/codex https://assets.wavescdn.com/pdf/plugins/codex.pdf

### Blofeld (Waldorf Music)

Category: hardware. Source check: recorded.

Import format: User wavetables reach the synth through Waldorf's official tools: the Blofeld plugin loads prepared SysEx/MIDI wavetables, or imports audio WAV with samples-per-wave selectable from 32 to 4,096 (the classic on-device table is 64 waves × 128 samples; the plugin thins larger wave counts on import).

Sources (checked 2026-08-16): https://waldorfmusic.com/faq-items/user-wavetables

### Iridium / Quantum (Waldorf Music)

Category: hardware. Source check: recorded.

Import format: On-instrument 'Import from .wav' builds wavetables from .wav/.aiff files with a constant sample period configurable from 64 to 4,096 samples (2,048 is called out as the common choice); user wavetables load, save, and exchange via microSD or USB storage.

Sources (checked 2026-08-16): https://waldorfmusic.com/iridium-en/ https://waldorfmusic.com/faq-iridium/

### Peak / Summit (Novation)

Category: hardware. Source check: recorded.

Import format: The Components Wavetable Editor imports audio files or third-party wavetables such as Serum's. Its import player lets you select a portion for one waveform or a wavetable; each user wavetable has five waveform slots and the synth has ten user wavetable slots (Peak firmware 1.3+ / Summit 1.1+). The documented import flow does not establish a one-to-one mapping from a 512 × 5 WAV to those slots.

Sources (checked 2026-08-16): https://support.novationmusic.com/hc/en-gb/articles/360009550480-Components-Summit-Peak-Guide https://support.novationmusic.com/hc/en-gb/articles/360013523460

### modwave (Korg)

Category: hardware. Source check: recorded.

Import format: The Editor/Librarian imports packed WAV in two formats: 32-bit float with 2,048-sample waves (the Serum-style layout) or 16-bit with 256-sample waves — up to 64 waves per file in either mode. Longer float tables are assumed to be generated crossfades and thinned to 64. The software modwave native shares both formats, with its own Librarian-page 'Import WAV As Wavetable' route.

Sources (checked 2026-08-16): https://www.korg.com/us/products/synthesizers/modwave_mk2 https://support.korguser.net/hc/en-us/articles/13236558460313-How-can-I-create-new-Wavetables

### nanobox fireball (1010 Music)

Category: hardware. Source check: recorded.

Import format: Wavetables use 2,048 samples per cycle and 256 cycles; loaded from microSD.

Sources (checked 2026-08-16): https://1010music.com/product/fireball https://1010music.com/wp-content/uploads/2022/02/nanobox-fireball-User-Guide.pdf

### ELZ_1 play (Sonicware)

Category: hardware. Source check: recorded.

Import format: Supports importing Serum-format wavetable files from SD card; up to 32 user wavetables can be registered. Built-in tables are licensed Waldorf/PPG data.

Sources (checked 2026-08-16): https://sonicware.jp/pages/elz_1-play https://www.sonicware.co.jp/DL/ELZ_1_play/DOC/ELZ_1_play_V2_manual_en.pdf

### Pro 3 (Sequential)

Category: hardware. Source check: recorded.

Import format: User wavetables via Sequential's Wavetable Generator web utility: individual waves 16-bit, 48 kHz, ideally 1,024 samples (shorter supported, all equal length), up to 16 single-cycle waves per table; delivered to the synth as SysEx (slots 33–64).

Sources (checked 2026-08-16): https://www.sequentialwaves.com/docs/Pro%203%20Wavetable%20Generator%20Utility.pdf https://sequential.com/product/pro-3/ https://sequential.com/new-pro-3-wavetables/

### Waldorf M (Waldorf Music)

Category: hardware. Source check: not recorded.

Import format: The M imports headerless 8-bit wavetable files from SD, with 256 samples per wave. Published manual copies and user guidance disagree on signedness, and the documented wave count and slot numbering need a current firmware check.

Sources (checked unknown date): https://waldorfmusic.com/m-en https://files.kraftmusic.com/media/ownersmanual/Waldorf_M_User_Manual.pdf https://faq.waldorfian.info/faq-browse.php?product=m

### Elektron Tonverk (Elektron)

Category: hardware. Source check: recorded.

Import format: Wavetables use the same audio file formats as regular samples — WAV/AIFF/M4A/CAF/FLAC at 16/24/32-bit (integer or float); frame size is a power of two from 64 to 4,096 samples (default 2,048), up to 64 waves (longer tables are evenly decimated). Append _wt<number> to the filename (e.g. _wt1024) only when the waves aren't 2,048 samples — 2,048-sample waves are assumed with no suffix.

Sources (checked 2026-08-16): https://www.elektron.se/en-us/support

Notes: Wavetable support arrived with OS 1.3.0.

### MicroFreak (Arturia)

Category: hardware. Source check: recorded.

Import format: Firmware 4.0+ WaveUser oscillator loads user wavetables via the MIDI Control Center — MFWZ, WAV, or AIFF — into a bank of 16 wavetables, with bidirectional drag-and-drop. Per press coverage, tables are 32 cycles × 2,048 samples (converted to 256 samples per cycle in the MIDI Control Center, not on the device).

Sources (checked 2026-08-16): https://support.arturia.com/hc/en-us/articles/4414446806290-MicroFreak-Oscillators https://cdm.link/microfreak-v4-adds-users-wavetables-accelerated-encoders-more-to-arturias-freaky-little-wonder/

### Deluge (Synthstrom Audible)

Category: hardware. Source check: recorded.

Import format: Official firmware 4.0+ (April 2022) added wavetable synthesis: WAV/AIFF files (16/24-bit 44.1 kHz preferred, up to 32-bit 96 kHz) loaded from SD card; wavetable-formatted files are recognized automatically and assigned oscillator TYPE 'WAVEtable'. Files under 20 ms are treated as single-cycle.

Sources (checked 2026-08-16): https://synthstrom.com/4-0-0-firmware-out-now https://synthstrom-audible-deluge.s3.us-east-2.amazonaws.com/Deluge-Guidebook-4p0-edits-only.pdf

### E352 Cloud Terrarium / E370 Quad VCO (Synthesis Technology)

Category: modular. Source check: recorded.

Import format: WAV banks loaded from microSD: 256 samples per wave, 16-bit, 64 wavetables per bank in an 8×8 array; WaveEdit is the official editor and the module holds 26 user banks.

Sources (checked 2026-08-16): https://synthtech.com/eurorack/E352 https://synthtech.com/eurorack/E370

### Piston Honda MK3 (Industrial Music Electronics)

Category: modular. Source check: recorded.

Import format: microSD load of eight mono 16-bit WAV files named 1.wav through 8.wav — the eight files are eight independent Z-axis banks, each holding an 8×8 X/Y table of 64 waves × 256 samples (together the 8×8×8 = 512-waveform cube). All eight files must be present or the load fails.

Sources (checked 2026-08-16): https://ime-assets.s3.amazonaws.com/uploads/manual/manual/24/phmk3manual11.pdf https://www.thomann.co.uk/industrial_music_electronics_piston_honda_mkiii.htm

### Disting EX (Expert Sleepers)

Category: modular. Source check: recorded.

Import format: Wavetable algorithms support 16-bit (24-bit since firmware 1.3) and 32-bit float (firmware 1.6.1+) WAVs — no 32-bit integer support is documented; single-file wavetables default to a 2,048-sample wavelength when the file is a multiple of 2,048 — Serum-format wavetables are natively supported. Folders of single-cycle WAVs are discovered automatically; playlists optionally order them, and the -wavelength playlist setting (default 600) applies when a file isn't a multiple of 2,048.

Sources (checked 2026-08-16): https://www.expert-sleepers.co.uk/distingEXfirmwareupdates.html https://www.expert-sleepers.co.uk/downloads/manuals/disting_EX_user_manual_1.15.pdf

### VCV Rack (Wavetable VCO) (VCV)

Category: modular. Source check: recorded.

Import format: The bundled Wavetable VCO loads packed WAVs up to 524,288 total samples in 8/16/24/32-bit integer or 32-bit float; frame size is taken from the sample-rate header if it is a power of two (e.g. 1024 for 1,024-sample frames), otherwise set via the module's Wave points menu. WaveEdit banks work.

Sources (checked 2026-08-16): https://vcvrack.com/Free

## Glossary

Complete glossary text, one block per term.

### Wavetable synthesis

Wavetable synthesis plays short, looping single-cycle waveforms from a table and sweeps across them to change timbre while pitch holds still.

A wavetable holds a series of single-cycle frames. An envelope, LFO, or direct control can move the playing position through them, changing the oscillator's harmonic content while its pitch stays the same. Each frame loops in the oscillator, allowing evolving tones with modest processing cost. Instruments such as Serum and Vital use this technique.

Hyperwave note: Hyperwave is a browser node-graph wavetable generator: you build the table visually and export it in the formats these instruments import.

See also: wavetable-position, wavetable-frame, single-cycle-waveform

### Wavetable position

Wavetable position is the index into a wavetable's frames that selects which single-cycle waveform, or blend of adjacent waveforms, the oscillator is currently playing.

Position usually runs from the first frame to the last, and fractional values interpolate between neighbors so the sweep is continuous rather than stepped. Modulating position with an envelope or LFO is what turns a static table into evolving, vocal, or growling timbres. How that motion sounds depends heavily on how related adjacent frames are.

See also: wavetable-frame, interpolation-between-frames, morphing-vs-crossfading

### Single-cycle waveform

A single-cycle waveform is one complete period of a periodic signal, stored as a fixed number of samples and looped by an oscillator to produce a steady pitch.

Because the loop repeats exactly, the waveform defines a static harmonic spectrum: its shape fixes which partials are present and at what amplitudes and phases. Every frame in a wavetable is a single-cycle waveform, so the raw material of wavetable design is the shape of that cycle, its harmonics and their phase.

See also: wavetable-frame, wavetable-frame-size, harmonics

### Wavetable frame

A wavetable frame is one single-cycle waveform occupying one slot in a wavetable; the number of frames sets how much timbral motion the table can hold.

Frames are ordered so that adjacent frames usually sound related, which lets a position sweep read as one evolving timbre instead of a sequence of unrelated waves. Tables range from a handful of frames to 256 or more, with the frame count fixed for a given wavetable.

See also: wavetable-position, single-cycle-waveform, wavetable-frame-size

### Wavetable frame size

Wavetable frame size is the number of samples in each single-cycle frame, held constant across the whole table.

Common sizes are powers of two (256, 512, 1024, 2048), which work efficiently with table lookup and FFT processing. Larger frames can represent more partials but use more memory. Importers slice packed WAVs by frame size, so the size must match the target instrument's expectation.

Hyperwave note: Choose an export preset in Hyperwave to set its recommended frame size and frame count. You can also set those values manually for other destinations.

See also: single-cycle-waveform, packed-wavetable-wav, band-limiting

### Interpolation between frames

Interpolation between frames is how a wavetable oscillator blends its two neighboring frames when position lands between them, so timbre changes continuously instead of in audible steps.

Fractional positions weight the two adjacent frames, and most engines use a simple linear crossfade, which is cheap and transparent when neighboring frames are closely related. Higher-quality modes interpolate in the spectral domain, and some newer engines, Serum 2 for example, render intermediate frames on the fly so position behaves as a continuous control rather than a blend of two stored frames.

See also: wavetable-position, morphing-vs-crossfading, wavetable-frame

### 2048 × 256 table

A 2048 × 256 wavetable (256 frames of 2048 samples each) is the de facto modern standard layout, popularized by Serum and adopted by many current wavetable synths.

Serum popularized this pairing, and Vital can export tables as 256-frame, 2048-sample WAVs. Many wavetable synthesizers accept packed WAVs with these dimensions, though import requirements vary. A 2048-sample frame can represent 1023 positive harmonics strictly below Nyquist, and 256 frames can represent changes in timbre across the table. At 44.1 kHz, the packed mono audio lasts about 11.9 seconds.

Hyperwave note: Hyperwave's authoring default is 2048 samples per frame and 256 frames. Check the destination's import requirements before export.

See also: wavetable-frame-size, packed-wavetable-wav, wavetable-frame

### 256-sample table

A 256-sample wavetable has frames that are 256 samples long. Several Eurorack wavetable modules use this frame size.

Synthesis Technology's E352 and E370 and the Piston Honda MK3 load banks of 256-sample waves from microSD, and hardware import paths such as the Korg modwave's 16-bit mode expect the same frame length. The compact frames cap each cycle at 128 partials, which suits bright synth tones and simple shapes but limits how detailed a low, harmonically rich frame can be.

See also: wavetable-frame-size, 8x8-wavetable-bank, band-limiting

### 8×8 wavetable bank

An 8×8 wavetable bank lays 64 wavetables out on a grid, so two control signals, one per axis, scan the whole bank instead of one table's frames.

The Synthesis Technology E352 and E370 expose their 64 tables of 256 samples this way, and the Piston Honda MK3 fills an equivalent 8 × 8 × 8 cube of waves from eight WAV files on its card. It turns the bank itself into a two-dimensional playing surface: timbre can jump or glide between unrelated tables, not just between the frames inside one table.

See also: 256-sample-table, wavetable-bank, wavetable-position

### Morphing vs. crossfading

Crossfading is an amplitude blend between two wavetable frames, while morphing is the broader goal it serves: moving between frames so the result reads as one continuously evolving timbre.

A straight crossfade sums two waveforms at complementary levels, which can thin the sound at the midpoint when phase differences cause partial cancellation, and it always keeps both frames' spectra audible. Phase-aligned and spectral techniques instead reshape partials between frames, giving sweeps that stay full and coherent where a plain crossfade would dip or smear.

See also: spectral-morphing, wavetable-position, interpolation-between-frames

### Spectral morphing

Spectral morphing transitions between wavetable frames by interpolating partial amplitudes and phases instead of blending waveform samples directly.

Working per-partial lets the morph follow the harmonic structure of the frames, so brightness, formant shape, and phase relationships travel together instead of canceling. It is the engine behind smoothly evolving, vocal-like sweeps that would sound hollow if the same frames were only crossfaded.

See also: morphing-vs-crossfading, resynthesis, fft-resynthesis

### Harmonics

Harmonics are components of a periodic waveform at whole-number multiples of its fundamental frequency. The second harmonic is at twice that frequency and the third at three times it.

An exactly repeated single-cycle waveform contains harmonic components. A sawtooth's harmonic amplitudes fall roughly as 1/n, while a square wave contains odd harmonics. Changing the harmonic amplitudes across frames changes the table's timbre.

See also: partials, single-cycle-waveform, additive-synthesis

### Partials

Partials are the individual sine-wave components of a sound's spectrum. They include harmonics and, in sounds that contain them, inharmonic components.

An FFT of a wavetable frame reports magnitudes and phases for its frequency bins. An additive editor lets you set component levels directly. A frame of N samples has N/2 bins up to the Nyquist frequency, so shorter frames can represent fewer harmonics.

See also: harmonics, fft-resynthesis, wavetable-frame-size

### Formant

A formant is a resonant region in a spectrum that stays at roughly the same frequency as the played pitch changes. Formants help distinguish vowel sounds.

Vocal-style wavetables work by moving formant peaks across frames while the fundamental holds still, which is why scanning them feels like a mouth opening and closing. Because formants live in the pattern of partials rather than in the time-domain shape, formant motion is far cleaner when done per-partial, as spectral morphs do, than by crossfading raw frames.

See also: partials, spectral-morphing, harmonics

### Phase

Phase is each partial's offset within the waveform's cycle, where in the loop every sine component starts, measured as an angle rather than a time.

Two frames can share identical partial amplitudes yet look completely different because their partials are phased differently; the ear keys mostly on the spectrum, but the shape differences matter the moment frames interact. Phase agreement is what lets a morph stay loud and coherent, and out-of-step phases are the usual reason a plain crossfade dips or turns hollow halfway through.

See also: single-cycle-waveform, spectral-morphing, phase-distortion

### Phase distortion

Phase distortion reshapes a waveform by varying the rate at which an oscillator reads it within each cycle.

Casio's CZ series used phase distortion to read a sine table through a variable phase curve. Steep parts of that curve add harmonics as the control changes. Wavetable tools can use the same calculation in warp modes and formula-generated frames.

See also: phase, waveshaping, hard-sync

### FFT resynthesis

FFT resynthesis analyzes a waveform with a fast Fourier transform (the algorithm that reports one magnitude and phase per frequency slot), edits that spectrum, and rebuilds a waveform from it with the inverse transform.

In wavetable work, the cycle is transformed into magnitudes and phases for each frequency bin. An editor can then remove partials, shift formants, or reset phase before the inverse transform builds the stored frame. Spectral morphing and harmonic editors use this process.

See also: resynthesis, spectral-morphing, partials

### Resynthesis

Resynthesis is analyzing an existing sound and rebuilding it as a wavetable, converting spans of the source into single-cycle frames rather than playing back samples.

A typical analysis tracks amplitude and phase per partial over time and encodes each short span of the source as a frame, so the source's timbral evolution becomes position motion you can edit, stretch, or exaggerate. Because the result is a real wavetable it can be swept, morphed, and exported like any hand-drawn table, at the cost of the exact time detail that pure sampling preserves.

See also: spectral-morphing, wavetable-synthesis, fft-resynthesis

### Waveshaping

Waveshaping passes a signal through a transfer curve. It computes each output sample from an input sample to change the waveform's harmonic content.

The curve determines the result: symmetric curves can add odd harmonics, while asymmetric curves can add even harmonics. Designed curves such as Chebyshev polynomials can target specific harmonics. Increasing the drive across wavetable frames stores that change in the table. Nonlinear processing can introduce aliasing unless it is band-limited or oversampled.

See also: wavefolding, phase-distortion, aliasing

### Wavefolding

Wavefolding is a waveshaping technique that reflects a signal back on itself once it passes a threshold, multiplying harmonics without changing the fundamental.

Each fold maps amplitude beyond a threshold back toward zero, adding upper partials to the waveform. Applying more folding across successive frames changes their harmonic content. Band-limiting or oversampling can reduce the resulting aliasing.

See also: aliasing, band-limiting, waveshaping

### Hard sync

Hard sync is an oscillator arrangement in which a master oscillator resets a second, detuned oscillator every time the master completes its own cycle, producing a biting, formant-rich tone.

Each reset interrupts the second oscillator's waveform and adds harmonics. Detuning changes those harmonics. A wavetable can store successive frames that approximate different sync settings, making wavetable position a control for the stored sweep.

See also: pwm, phase-distortion, wavetable-frame

### Through-zero FM

Through-zero FM is frequency modulation in which the carrier's frequency is allowed to pass through zero and come out negative, continuing the phase in reverse rather than stopping.

The carrier's phase reverses direction when its instantaneous frequency passes through zero. This can produce a symmetrical response around the zero crossing. Simple carrier waveforms such as sines produce fewer sidebands than harmonically rich carriers.

See also: hard-sync, phase, waveshaping

### Pulse-width modulation

Pulse-width modulation (PWM) varies a pulse wave's duty cycle. The duty cycle is the fraction of each cycle spent at its high level, and changing it changes the timbre.

A pulse wave's harmonic amplitudes depend on its width. At 50%, it contains odd harmonics; other widths change that balance. A wavetable can store successive pulse widths as frames, allowing position to control the stored width sweep. Similar widths in adjacent frames usually produce a smooth transition.

See also: hard-sync, wavetable-position, single-cycle-waveform

### Formula-generated wavetable

A formula-generated wavetable has frames computed from expressions evaluated per sample and per frame.

Formulas can sum harmonics, shape phase, add noise, and vary parameters by frame index. One expression can generate a set of related frames. Script formats include u-he's .uhm and Surge XT's Lua .wtscript; notebooks and signal-processing tools can also generate frames this way.

Hyperwave note: Hyperwave takes the same idea visually: instead of typing formulas you wire operations as a node graph, and the table is computed from it.

See also: uhm-format, additive-synthesis, waveshaping

### Aliasing

Aliasing is inharmonic distortion that appears when a digital signal contains frequencies above half the sample rate (Nyquist), which reflect back down into the audible range.

In oscillators it shows up as metallic, out-of-tune artifacts that worsen as pitch rises, because higher notes push a waveform's partials past Nyquist. Wavetable engines control it by band-limiting playback per pitch, and nonlinear processing such as wavefolding re-introduces it unless it is oversampled.

See also: band-limiting, wavefolding

### Band-limiting

Band-limiting restricts a waveform's harmonics to frequencies below the Nyquist limit, which is half the sample rate.

Wavetable oscillators can store several progressively low-pass filtered versions of a table and select among them as pitch rises. This reduces aliasing in high notes. Frame size also limits the number of harmonics a cycle can represent; shorter frames can limit brightness at low pitches.

See also: aliasing, wavetable-frame-size

### DC offset

DC offset is a nonzero average level in a waveform. The cycle sits above or below zero instead of being centered on it.

In a looping wavetable frame, DC offset uses headroom and adds energy at 0 Hz. Different offsets in adjacent frames can cause clicks during a position sweep. Center each frame before export to remove its offset.

See also: single-cycle-waveform, wavetable-frame

### Wavetable bank

A wavetable bank is a grouped collection of wavetables that an instrument loads together and offers under a single browser or selector.

A bank can group factory tables, a sound designer's series, or other related tables for selection in an instrument. In some instruments, a bank has another selection axis for choosing or blending between tables independently of position within one table.

See also: wavetable-synthesis, packed-wavetable-wav, 8x8-wavetable-bank

### WAV file

A WAV file is the everyday audio container: a header describing the format, a block of sample data, and room for extra metadata chunks.

Hyperwave's standard WAV header is 44 bytes. Its fmt chunk specifies the PCM or IEEE float format, channel count, and sample rate; the data chunk holds the samples. WAV files can include additional chunks containing metadata. An importer interprets a packed wavetable WAV as consecutive frames.

Hyperwave note: Hyperwave exports mono 32-bit IEEE float WAV by default and can also export 16-bit PCM. Its audio importer uses the browser's decoder and mixes channels to mono before slicing.

See also: packed-wavetable-wav, clm-chunk, wavetable-frame-size

### Packed wavetable WAV

A packed wavetable WAV is a single audio file that stacks every frame of a wavetable end to end, so a wavetable-capable synth can import the whole table from one file.

An importer divides the audio into blocks of its expected frame size and maps them to table slots. If the length or frame size is wrong, it may read the frames incorrectly. Serum and Vital both accept packed WAV wavetables, though their import settings and limits differ.

Hyperwave note: Hyperwave exports packed WAV tables. Choose a target preset for its recommended frame size and count, or set them manually. The default format is 32-bit float WAV; 16-bit PCM is also available.

See also: wavetable-frame-size, wavetable-frame, 2048x256-table

### clm chunk

The 'clm ' chunk is optional metadata in some Serum-format wavetable WAV files; plain WAV exports may omit it.

A 'clm ' payload can contain a marker such as '<!>2048' and flags for interpolation. Xfer's Serum 2 manual distinguishes plain 16- and 32-bit WAV exports without extra interpolation headers from a Serum-format selection export with extra header information. A chunk's presence identifies metadata; its absence does not prove that a packed WAV cannot be imported.

See also: wav-file, srge-chunk, packed-wavetable-wav

### srge chunk

The 'srge' chunk is Surge XT's own versioned metadata chunk inside WAV files, telling its importer the frame layout the same way Serum's 'clm ' chunk does.

The payload contains a version number and a frame-length value. Surge XT writes 'srge' into its wavetable WAV exports and 'srgo' into sample exports. It also reads 'clm ' metadata on import. These chunks can tell an importer how to divide audio into frames.

See also: wt-file, clm-chunk, wav-file

### wt file

A .wt file is Surge XT's native wavetable format: a short tagged header followed by the table's frames as raw samples, optionally trailed by metadata.

The header starts with a 'vawt' tag and records the frame size, frame count, and format flags. Frames are stored as raw 32-bit floats or 16-bit integers at power-of-two lengths from 2 to 4096 samples. Surge XT can use a trailer for additional wavetable metadata, including scripts used to generate a table.

See also: srge-chunk, packed-wavetable-wav, wav-file

### vitaltable file

A .vitaltable file is Vital's native wavetable format. It is a readable JSON document describing how the table was built instead of storing only rendered audio.

It stores the wavetable's sources, modifiers, and keyframes. Opening it in Vital restores an editable table instead of a finished stack of frames. Vital can also export packed WAVs with 2,048-sample frames for other synths. Inspect the exported WAV when the destination requires metadata.

See also: packed-wavetable-wav, 2048x256-table, wav-file

### uhm format

A .uhm file is a u-he Hive wavetable script: a readable text file of commands and formulas that Hive evaluates step by step to build a wavetable.

Commands can fill frames from time-domain formulas, edit the spectrum or phase per partial, import and reorder wavetables, and write buffers. A script can define 1 to 256 frames of 2048 samples. Hive loads .uhm scripts and .wav tables. WAVs default to 2048 samples per frame unless the filename specifies another size, such as -WT512. Its .h2p presets reference tables by name rather than embedding them.

See also: formula-generated-wavetable, packed-wavetable-wav, wav-file

### Additive synthesis

Additive synthesis builds a sound by summing sine-wave partials. It sets each component's amplitude and, optionally, its phase.

Because the spectrum is the instrument, additive control means drawing partials: a precise sawtooth, a bell with stretched inharmonic partials, or a vowel-shaped formant stack are all just partial recipes. It pairs naturally with wavetables, where one additive recipe per frame turns the table into a spectrum you can sweep.

Hyperwave note: Hyperwave treats additive drawing as one way to author a frame: set the partials, and the frame is computed from them.

See also: harmonics, partials, subtractive-synthesis

### Subtractive synthesis

Subtractive synthesis starts with a harmonically rich waveform and removes or reduces parts of its spectrum with filters.

An analog subtractive patch often sends a bright oscillator through a low-pass filter controlled by an envelope. A wavetable oscillator can supply the source waveform in the same setup. Filtering changes only the frequencies present in that source.

See also: additive-synthesis, wavetable-synthesis, single-cycle-waveform

### Granular synthesis

Granular synthesis builds sound from many short grains. Each grain is a brief, windowed slice of an audio buffer played at a chosen time and pitch.

A granular engine can play grains at different times and pitches to produce changing textures. Serum 2 has a granular oscillator. A wavetable oscillator loops a single-cycle frame, while a granular oscillator plays brief windowed slices that may span multiple cycles.

See also: wavesequencing, wavetable-synthesis, resynthesis

### Wave sequencing

Wave sequencing plays a programmed series of waves within one note. The Korg Wavestation supports up to 255 steps with settings for duration, pitch, and crossfade.

Where a wavetable scan morphs between similar single cycles to change timbre, a wave sequence can jump between unrelated waves, sync its steps to a clock, and loop the route, yielding the Wavestation's signature rhythmic pads. The idea reads as wavetable thinking applied to time: the list of waves, not the stack of frames, is the instrument.

See also: granular-synthesis, wavetable-position, wavetable-synthesis

### PPG Wave

The PPG Wave is the synthesizer line that introduced commercial wavetable synthesis, built by Wolfgang Palm's Palm Products GmbH from 1981 to 1987.

The 1978 Wavecomputer 360 scanned 30 wavetables of 64 waves each. The later Wave series paired digital wavetable oscillators with analog filters and amplifiers. Many later wavetable instruments also use ordered single-cycle waves, position scanning, and interpolation between adjacent waves.

See also: waldorf-microwave, wavetable-synthesis, wavetable-position

### Waldorf Microwave

The Waldorf Microwave (1989) is the rack synthesizer that carried PPG's wavetable lineage forward after PPG closed, pairing wavetable scanning with an analog filter.

The Microwave was a rack-mounted instrument with its own wavetables. Later models included the DSP-based Microwave II and XT. Waldorf also released wavetable instruments such as the Blofeld, Quantum, and Iridium.

See also: ppg-wave, wavetable-synthesis, wavetable-bank

### Korg DW-8000

The Korg DW-8000 (1985) is a hybrid synthesizer with 16 stored digital waveforms and an analog filter. Its DWGS oscillators select a fixed waveform rather than scan a wavetable.

Each oscillator selects one DWGS waveform, including saw, piano-like, and bell-like shapes. The analog circuitry then filters the sound. The DWGS stores single-cycle waves, but it does not scan a series of frames with a position control.

See also: single-cycle-waveform, subtractive-synthesis, wavesequencing

### Ensoniq Transwave

A Transwave is Ensoniq's variant of wavetable scanning. Instruments such as the VFX, TS, and Fizmo sweep stored waves with programmable start, end, and loop points.

Because a patch can define which slice of the wave bank it scans and how the scan loops, one stored bank serves many different evolving behaviors. Ensoniq built whole instruments around the idea, and the Fizmo's slowly swirling pads are still cited as classic Transwave sound design.

See also: ppg-wave, wavetable-position, wavesequencing

### Adventure Kid Waveforms

Adventure Kid Waveforms (AKWF) is Kristoffer Ekstrand's free library of several thousand single-cycle WAV files for samplers and wavetable synthesizers.

The original waveforms are 600 samples long. Community conversions resize them to 256, 512, 1024, or 2048 samples for different instruments. AKWF files can be used to test single-cycle import and assemble custom wavetables.

See also: single-cycle-waveform, wavetable-frame-size, wavetable-bank

## Changelog (latest 10 sections)

Release notes, newest first.

### Version 0.2.38, 2026-08-20

- Unchanged canvas frames now reuse cable geometry, reducing repeated drawing work while keeping routes stable during movement.
- The serial and threaded sound engines were rebuilt and checked together.

### Version 0.2.37, 2026-08-19

- The homepage now opens with a full-width, slow-moving wave field, a clearer one-line promise, and one prominent Open Hyperwave action. The animation stays lightweight, pauses when it is not visible, and becomes a still image when reduced motion is preferred.
- Play Keyboard now uses the four physical typing rows, proper Volume and Filter ADSRs with draggable curve diamonds, and the Wavetable MSEG. Key, scale, octave, every synth control, both ADSR shapes, and wavetable motion now save with local, versioned, Cloud, and exported presets.
- Preset loading now rejects malformed graph data earlier and avoids reopening verified Cloud archives during import.
- Public pages now share one consistent site shell, and every account dashboard page applies the same sign-in protection and safe return path. The public landing page is now the homepage, while the browser app opens at /app and existing /landing links continue to work.
- UI mono text now uses the bundled Departure Mono typeface across the transport bar, history, preset panels, consoles, and technical pages; graph nodes retain their established face for hierarchy and scanability. The note picker now matches the active console palette and uses edge-to-edge rows for faster scanning.
- Node creation shortcuts now use a compact, fine-outlined keyboard with joined keys and quiet assignment colors. The complete writing key set fits narrow windows without clipping the description or actions, while keeping the current theme and assignments unchanged.

### Version 0.2.36, 2026-08-18

- The canvas now retires old nodes, cables, nested previews, and scene resources as one protected lifecycle when a graph is replaced or the workbench reloads. Interrupted cable-routing updates also return to the last complete frame instead of leaving stale hover or geometry behind.
- Press Tab on the canvas to open Play Keyboard: an 8×8 scale grid you can play with your typing keyboard, pointer, or a selected MIDI device. It includes Key and Scale choices, a selectable filter, and three compact MSEG slots for volume, filter, and wavetable position while graph shortcuts stay safely off. The Volume curve marks its Release stage; filter and wavetable motion continue through that fade, and each note's three playheads fade away together when it finishes. The modal now uses square pads with root-note corner markers and compact dithered rails, including a snapping filter-mode rail. Each MSEG has one duration and one Loop control, and the layout reflows cleanly without coordinate axes, duplicate controls, or clipped text.
- Opening Play Keyboard with Tab now rests on an outline-free modal shell instead of jumping straight to the Key selector. The controls are ready when you deliberately navigate into the modal.
- Repeating the same note in Play Keyboard now lets the previous Release tail ring out underneath the new press instead of cutting it off.
- Play Keyboard now keeps every sounding note, including dense chords and repeated pitches, without an artificial voice limit. New voices fade in smoothly, the mix stays controlled, and silent voices are cleaned up only after their sound has finished.
- The preset browser's Load and Loaded action now fills its whole footer, removing the unnecessary box-within-a-box appearance.
- Session history now shows the complete undo tree instead of only edits currently available to redo. Your current path is marked HEAD, alternate paths stay separated at each fork, and everything remains visible as you undo, make a different edit, and choose a branch.
- Getting Started walkthroughs now explain what you’ll learn directly in the preset preview, without showing controls for a waveform they do not load. Each description now names the lesson’s topics in one short sentence, so it is easy to choose or skip at a glance. Preview and Cloud status messages are centered and easier to read when they wrap.
- Error diagnostics no longer interrupt the workbench with a reproduction dialog. When diagnostics are configured, errors are reported quietly in the background without attaching graph contents to analytics.
- Cloud folders now use the compact preset format for save, preview, load, and verified download. Hyperwave stores the graph only and still creates all rendered sound locally.
- Cloud saves now handle simultaneous graph and folder/name edits safely, clear private Cloud state on sign-out, and clean up interrupted uploads.
- You can now drag selected nodes into a nested graph preview. The source graph window lights up and says Drop node here; hovering shows the complete graph with those nodes added and dissolves that instruction so it cannot cover the preview. Their connected cables dissolve away from the dragged nodes and return in reverse when you leave. Dropping inside still moves everything as one undoable edit.
- Nested graph previews no longer show Click to open over nodes that are already visible inside the preview.
- The transport Note menu now shows one C per octave, from C0 through C6, so changing the audition pitch no longer requires scrolling past every chromatic note.
- New workspaces now start auditioning at C2, while your saved note choices continue to be preserved.
- The preset browser button now comes first in the transport controls for quicker access.
- Node creation shortcuts now use a compact, layout-aware keyboard for letters, numbers, and punctuation. Select a key to open the same node picker used in the graph, then choose what that key should create. Hover a key to instantly see its current assignment in the familiar tooltip, with only its corner accent switched to the secondary theme color; assigned keys are accented, unassigned keys stay gray, and your existing shortcuts remain the default. Right-click a key to unassign it, or use Reset all in the keyboard's lower-right corner to restore every default. The assignment tooltip now lets pointer hover pass through to the key, and its dither reveal is 30% faster without changing other tooltips.
- Mouse-wheel scroll now zooms the canvas again, while two-finger trackpad movement keeps panning smoothly. Hold Shift while scrolling to pan left and right, or Command/Ctrl to pan up and down.
- Fixed smooth or fractional mouse-wheel movement still being mistaken for a trackpad gesture, including on high-resolution Mac displays.
- Restored trackpad pinch zoom while keeping Command/Ctrl plus a physical mouse wheel mapped to vertical canvas movement.

### Version 0.2.35, 2026-08-17

- Export target pages now include per-target import guidance.
- Hyperwave now has a shared account entry, an account dashboard, and an always-visible Cloud library area.
- Copy, paste, cut, duplicate, undo, and redo now stay inside Hyperwave without clipboard prompts. Copied nodes remain available when you start or load another preset. The History drawer can recover alternate redo paths, and new portable preset files use a smaller, versioned format.

### Version 0.2.34, 2026-08-16

- Added export target guides, comparisons, tutorials, and browser tools.

### Version 0.2.33, 2026-08-15

- Rebuilt the save dialog with clearer controls and consistent labels.

### Version 0.2.32, 2026-08-14

- Connections now keep their preferred clearance from node bodies, including when routing rails are nudged around tight layouts.

### Version 0.2.16, 2026-07-24

- Node composition changes now animate cleanly, skip stale evaluation after repulsion, and keep canvas pan overlays synchronized in the same frame.

### Version 0.2.15, 2026-07-23

- Workbench previews, tour motion, worker scheduling, cache retention, and deployment validation were tightened for smoother editing and safer releases.

### Version 0.2.14, 2026-07-22

- Preview rendering now reuses shared textures and bounded caches while transport and inspector scheduling stay stable during rapid scrubbing.