upgrade_h5
upgrade_h5(src_h5, dst_h5, basis_size=_container.BASIS_SIZE)Rewrite a format-1 archive in format 2, without re-compressing.
Every format-1 scale is repacked into the flat codec group with a shared spatial basis; the temporal coefficients are kept as they are (minus those that only reconstruct the guard bands), so the only loss is the basis projection (< 0.01 dB SNR at m = 32). Everything else (meta, sync datasets, saturation tables, format-2 scales) is copied as is.
Parameters
| Name | Type | Description | Default |
|---|---|---|---|
| src_h5 | path - like | Source multi-recording archive (/<recording>/<scale>/... layout). |
required |
| dst_h5 | path - like | Output archive (always created fresh). | required |
| basis_size | int | Number of shared spatial basis vectors. Default 32. | _container.BASIS_SIZE |
Returns
| Name | Type | Description |
|---|---|---|
| Path | Resolved path to dst_h5. |