I'm working on a raw convertor (dlRaw). It should become somehow an alternative between ufraw and rawtherapee, but that aside. It's as much for the fun of it.
Like nearly all of those projects I did an effort around dcraw (making it behaving more as a library, less global variables, re-entrant, you know the stuff much better than I do :)).
While this is manageable, it is an unnecessary burden and your library has drawn my attention for :
- it does all of this more or less for me.
- it closely tracks dcraw evolutions and as now projects as digiKam are building on libraw, I think libraw got an eternal life :)
- you started building in openmp.
Before diving in the effort of converting to libraw (which I'm considering strongly) there are a few things I would like to check or have your opinion upon. Can I refer you to http://dlraw.sourceforge.net/Processing.shtml (Only SubPhase1 and SubPhase2 should be of interest in this story). It indicates what I'm currently doing with dcraw.
What I would like to understand :
- Can I use libraw to do the loading, use the interpolation and colour scaling as it happens in dcraw, *but* with a cached image in between ? (playing with colour scaling should not invoke a reload of the image over and over again).
- Can I still obtain the multipliers that dcraw suggests (for instance from camera) , but without blowing them necessarily to minimum 1.0 ?
- Can I still use the whitebalancing based upon 'greybox' as in dcraw ? How do I calculate back from output image (that might be flipped etc) to coordinates of the original image ?
- Are there other gotcha's I should be aware of before starting the effort ?
Many thanks for your advice and support.