Closed mkuemmel closed 2 years ago
The documentation states that the scaling for the modified chi-square fit (https://astrorama.github.io/SourceXtractorPlusPlus/Model.html#modified-least-squares) involves the rms of the pixel (u_0 = kappa * sigma_i). In contrast the implementation uses a fixed scaling of value 10: https://github.com/astrorama/SourceXtractorPlusPlus/blob/develop/ModelFitting/ModelFitting/Engine/AsinhChiSquareComparator.h That does not seem to be a big deals otherwise we would have realized that inconsistency earlier on. But it would be good to settle this.
We had a long discussion on that one. @ebertin will look at the equations and decide what is correct.
Equations fixed in 9d8bccb. Code matches the equations
The documentation states that the scaling for the modified chi-square fit (https://astrorama.github.io/SourceXtractorPlusPlus/Model.html#modified-least-squares) involves the rms of the pixel (u_0 = kappa * sigma_i). In contrast the implementation uses a fixed scaling of value 10: https://github.com/astrorama/SourceXtractorPlusPlus/blob/develop/ModelFitting/ModelFitting/Engine/AsinhChiSquareComparator.h That does not seem to be a big deals otherwise we would have realized that inconsistency earlier on. But it would be good to settle this.