colour.XYZ_to_msds

colour.XYZ_to_msds(XYZ: ArrayLike, method: Literal['Gaussian', 'Smits 1999'] | str = 'Gaussian', *, as_array: Literal[False] = False) → MultiSpectralDistributions

colour.XYZ_to_msds(XYZ: ArrayLike, method: Literal['Gaussian', 'Smits 1999'] | str = 'Gaussian', *, as_array: Literal[True]) → NDArrayFloat

Recover the multi-spectral distributions from the specified CIE XYZ tristimulus values using the specified method.

Parameters:

• XYZ (ArrayLike) – CIE XYZ tristimulus values to recover the spectral values from. The last dimension must be size 3.

• method (Literal['Gaussian', 'Smits 1999'] | str) – Computation method.

• as_array (bool) – Whether to return raw spectral values as a numpy.ndarray of shape (*XYZ.shape[:-1], wavelengths) instead of a MultiSpectralDistributions instance. Defaults to False.

Returns:

Recovered multi-spectral distributions, or the underlying spectral values when as_array=True.

Return type:

MultiSpectralDistributions or numpy.ndarray

Notes

Domain	Scale - Reference	Scale - 1
XYZ	1	1

• Both methods will internally convert specified CIE XYZ tristimulus values to RGB colourspace array assuming sRGB primaries and equal energy illuminant E.

References

[Smi99]

Examples

Gaussian reflectance recovery:

>>> import numpy as np
>>> XYZ = np.array(
...     [
...         [0.20654008, 0.12197225, 0.05136952],
...         [0.14223761, 0.23042375, 0.10498415],
...         [0.07820260, 0.06157595, 0.28106183],
...     ]
... )
>>> XYZ_to_msds(XYZ, method="Gaussian", as_array=True).shape
(3, 421)
>>> float(XYZ_to_msds(XYZ, method="Gaussian", as_array=True)[0, 300])
...
0.3785...

Smits (1999) reflectance recovery:

>>> XYZ_to_msds(XYZ, method="Smits 1999", as_array=True).shape
(3, 10)
>>> float(XYZ_to_msds(XYZ, method="Smits 1999", as_array=True)[0, 6])
...
0.3207...