Skip to content

Keep physical units

By default, Cutana converts pixels to Jansky and then normalises them for display, which discards the physical scale. This page shows how to keep calibrated values.

Keep pixels in Jansky

Flux conversion to Jy is on by default and uses the MAGZERO header keyword (configurable via config.flux_conversion_keywords.AB_zeropoint). Normalisation stretches pixels into the data_type range, so turn it off and write floats:

config.normalisation_method = "none"  # or config.do_only_cutout_extraction = True
config.data_type = "float32"  # uint8 rescales to 0-255, losing the scale

To keep the parent tile's own units instead, disable the conversion:

config.apply_flux_conversion = False

Conserve flux when resizing

For photometry, resize with flux conservation:

config.flux_conserved_resizing = True
config.data_type = "float32"  # preserve numerical precision
config.normalisation_method = "none"  # keep the original flux values

This preserves total flux through the resize. It uses drizzle, which is slower than the default resizing.

Extract raw cutouts

Set do_only_cutout_extraction = True with output_format = "fits" to skip resizing, normalisation and channel combination. Only the flux conversion is applied, and you can disable that too. The output data_type follows the input. This mode does not support Zarr output.

Check the result with the BUNIT, FLUXAPPX, UNIT and CONSVFLX header keywords, described in the output formats reference.