heya, On Tue, Jan 31, 2017 at 1:56 AM, Marcus Sundman <[email protected]> wrote: > On 30/01/17 22:01, johannes hanika wrote: >> >> hi, >> >> On Tue, Jan 31, 2017 at 5:42 AM, Marcus Sundman <[email protected]> wrote: >>> >>> On 30/01/17 10:13, johannes hanika wrote: >>> >>>> hi, >>>> >>>> i suppose you've seen this >>>> >>>> >>>> http://www.darktable.org/2016/05/colour-manipulation-with-the-colour-checker-lut-module/ >>>> ? >>> >>> >>> "to interact with the colour mapping, you can change both source and >>> target >>> colours. the main use case is to change the target colours however, and >>> start with an appropriate palette" >>> >>> However, I want to do the opposite. I want to start with an appopriate >>> target palette (the xrite color checker card) and then pick the source >>> colors from the picture. >> >> more specifically i meant the second half of it, about the darktable-chart >> tool. > > > That is insanely complicated and time-consuming, involving multiple GUI > tools as well as command line commands, generating ICCs etc. Imagine setting > the whitebalance by first rotating and cropping the picture, setting various > color settings, then exporting the picture, importing it into another tool, > running various command line tools on the intermediate files, then importing > the final results into darktable and access your custom made (for this > picture) ICC to apply to the picture. Now do the for every picture. You'd > have died of old age by the time you make it past your 10th photoshoot. > > Now, imagine if you instead could simply select the source colors in a > picture (which you can!) and choose which target palette they map to (which > I think you cannot). This would be almost as quick as the normal white > balance tool.
yes, i agree setting the precise colour target instead of the offset would be a nice addition to this tool. if i find time i might try a stab at it. i'd be curious to see whether this does something useful with a 24-checker thing. i /think/ the linear part of the spline should be equivalent to using a matrix for white balance, which is one up from a 3-float white balance multiplier. the fitter will, however, not approximate a linear matrix in the least squares sense, but compute a mapping that precisely goes through the target points you give it. choosing few such points may result in overshoots and such in between the sample points (not much, these are thin plate splines, but still..). j. > And if there was a way to choose the whole palette at once (just mark the 4 > corners of the palette in the picture, like you do in the xrite tool) it > would be virtually as quick as the normal white balance. And if this tool > would manage to find the palette by itself (like the xrite plugin does) then > it'd be even faster than the normal white balance tool! > > Cheers, > Marcus ____________________________________________________________________________ darktable user mailing list to unsubscribe send a mail to [email protected]
