Wavetable file formats
Most wavetables travel as packed WAV files that stack every frame end to end. Some synths add metadata or keep native formats such as .wt and .vitaltable.
A packed WAV is the format nearly every wavetable synth imports, so it is the safe way to move a table between tools. Metadata chunks inside the WAV can tell an importer the frame size, and native formats keep extra information, such as how the table was built, that only their own synth reads.
Packed WAV
A packed wavetable WAV is one mono audio file with every frame laid end to end. The importer divides the sample count by its frame size to get the frame count, so the file length should be an exact multiple of the frame size. If the importer assumes the wrong size, it slices across cycle boundaries and the table plays out of tune.
Importers learn the frame size in one of three ways: a metadata chunk such as 'clm ' or 'srge', a convention such as Hive's -WT512 filename suffix, or a default such as 2048 samples.
The clm chunk (Serum)
The 'clm ' chunk is optional metadata in Serum-format wavetable WAVs. Its payload can carry a marker such as '<!>2048' and interpolation flags. Xfer's Serum 2 manual distinguishes plain 16- and 32-bit WAV exports from a Serum-format export with this extra header information.
A missing chunk does not stop a packed WAV from loading in importers that use a default frame size or let you choose one. Its presence identifies metadata; its absence proves nothing about whether a file will import.
The srge chunk (Surge XT)
Surge XT writes its own versioned 'srge' chunk, holding a version number and a frame length, into its wavetable WAV exports, and 'srgo' into sample exports. On import it reads 'clm ' as well, so Serum-format tables keep their frame size.
.wt (Surge XT)
Surge XT's native format starts with a 'vawt' header recording the frame size, frame count, and format flags, followed by the frames as raw 32-bit floats or 16-bit integers at power-of-two lengths from 2 to 4096 samples. A trailer can hold extra metadata, including the script that generated the table. Surge XT reads packed WAVs too, which is usually the simpler hand-off from another tool.
.vitaltable (Vital)
A .vitaltable is a readable JSON document describing how a table was built: its sources, modifiers, and keyframes. Opening it in Vital restores an editable table rather than a finished stack of frames. To move tables between Vital and other synths, use packed WAVs: Vital exports them with 2048-sample frames, and dropping a WAV onto its wavetable display loads one.
.uhm (u-he Hive)
A .uhm file is a text script of commands and formulas that Hive evaluates step by step to build a table of 1 to 256 frames of 2048 samples. Commands can fill frames from time-domain formulas, edit the spectrum or phase per partial, and import or reorder tables. Hive loads .uhm scripts and WAV tables, reading WAVs at 2048 samples per frame unless the filename names another size. Its .h2p presets reference tables by name rather than embedding them.