> So, currently the best option is to use auto-correction and disable > TCA in the lensfun module.
Agreed, it is my preferred workflow. > has changed. The fitting is now computed based on better estimations, > with proper weighting. For example, a large pseudo-uniform blue sky > does not have a bad influence on shift estimations. I would have expected CA fitting to work on edges and instead of area's, Looks like patch seems to be more like that. > About the fitting itself, the applied red and blue color shift are > resolved as 2-dimensional polynomials of degree 4 (hard-coded but can > be modified in the source). That is, a sum of x^n*y^m components with > n + m <= 4. The previous implementation used max(n,m) < 4, which is This seems indeed a better way. In fact the simpeler the polynome the better (more stable) the fitting result. >> And would you consider adding a manual slider option or do >> you use SVD for the fitting? > > I like the zero-configuration approach. If you refer to the > possibility to fine-adjust the amount of correction with a fixed > geometry, like with the lensfun approach, maybe wait a little, it > might become reasonable to use lensfun earlier in the pipeline. zero-config is really nice and auto-ca works great for most situations. However I was talking about the cases where the fitting can go wrong. Most simple fitting algorithms find the local minimum fitting error and that works well for quadratic polynomes. However for more complex higher order polynomes there might be more local minima and code like single value decomposition can find the global minima even if there are local minima. It's most likely the source of the issue I have of not getting perfect denoising profiles. For the lensfun module you suggestion for moving it to RAW would be great. I would suggest doing only the CA correct of lensfun on the raw bayer data, the lens distort at the end of the pipeline simultataneous with crop and rotate and vignetting after any post demosaic denoising. Putting all in one position of the pipeline seems counter productive. Best, Gert -- Panta Rhei -- Life is a dance ------------------------------------------------------------------------------ Learn Graph Databases - Download FREE O'Reilly Book "Graph Databases" is the definitive new guide to graph databases and their applications. Written by three acclaimed leaders in the field, this first edition is now available. Download your free book today! http://p.sf.net/sfu/NeoTech _______________________________________________ darktable-devel mailing list darktable-devel@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/darktable-devel
