kBradfordInverse

const float fl::detail::kBradfordInverse

Initial value:

= {
    {0.986992955f, -0.14705427f, 0.159962654f},
    {0.432305276f, 0.518360257f, 0.049291227f},
    {-0.0085286675f, 0.0400428213f, 0.968486726f},
}

Its inverse, precomputed.

This used to be invert3x3(kBradford, ...) on every profile bind – a float determinant, a reciprocal, three column-scale divisions and nine multiplies, run to recompute a matrix that cannot change. The operand is a compile-time constant, so the answer is one too.

The values are what invert3x3 produces from kBradford in float32, to the last bit, so the adaptation matrix this feeds is unchanged. test_bradford_inverse_matches_the_general_solver pins that against the solver rather than against a transcription, which is what catches a typo here – a wrong digit in a plausible-looking matrix is exactly the kind of constant nothing else would notice (FastLED #4043).

Definition at line 32 of file chromatic_adaptation.cpp.hpp.

                                     {
    {0.986992955f, -0.14705427f, 0.159962654f},
    {0.432305276f, 0.518360257f, 0.049291227f},
    {-0.0085286675f, 0.0400428213f, 0.968486726f},
};