Aside from that new parameter "white point adjustment" In the development
version of darktable iop shadows&highlights,
There is also new color reconstruction iop:
https://www.darktable.org/2015/03/color-reconstruction/

Also, i'm working on 2 new highlight reconstruction iops (global (something
like inpainting) and local), they already kind of working,
but i do not know if they will make it into next release.

Roman.

On Sat, Jun 6, 2015 at 7:01 AM, Chris Siebenmann <[email protected]> wrote:

>  I've been using darktable for a few years now and processed a decent
> number of photos with it. Over all of that time, the single most
> difficult and troublesome thing for me to deal with in darktable is
> trying to recover artificially burned out highlights.
>
> (By 'artificially burned out' I mean highlights that are not burned out
> in the RAW before the base curve and white balance are applied.)
>
>  The obvious way to go is the highlights recovery setting on the shadows
> and highlights module. Unfortunately in practice this does little or
> nothing to the high end of highlights; they stay burned out even if you
> crank the highlights well down into unnatural territory. None of the
> other modules I've tried do a particularly good or easy job of this
> either.
>
>  It's my view that this situation is kind of absurd. Artificially blown
> highlights are quite common and it seems crazy that darktable cannot
> easily handle them, and especially that it cannot handle them in the most
> natural module for the task. If there is something unusually complicated
> about handling them in 'shadows & highlights' (which does come quite
> late in the module processing list), I'd suggest that darktable add a
> dedicated highlights recovery module at whatever point in the processing
> pipeline makes the most sense and/or makes it easiest internally.
>
>  As for modules I've tried:
>
>  The graduated density module? Sort of useful when I have eg burned out
> sky, but again it seems to do remarkably little to burned out highlights,
> far less than I would expect based on its nominal EV. In fact its nominal
> EV seems rather inconsistent; I have sometimes used +EV in the exposure
> module followed by the same -EV in graduated density and observed that
> the resulting grad'd area does not have the same exposure as before
> both modules.
>
>  In theory I should be able to fiddle with the applied base curve to
> tame the high end. In practice the interface of the base curve makes
> it extremely difficult to do this, because all of the highlights are
> crammed in at the top left and need extremely fine manipulation to get
> natural results that recover highlights. Turning off the entire base
> curve 'curse' highlights at the expense of everything else, which is
> not really an easy way to go.
>
> (I've advocated for better base curve control before, but yeah, here it
> is again. The 'logarithmic' display is not really it, although it's a
> bit of an improvement over the linear display.)
>
>  The tone curve suffers from the same precision issue and at least
> some of the time it seems to have the same problem of not really touching
> the burned out areas. It also requires a deft hand to only touch the
> highlights (or you get to go try to make it play well with parametric
> action restrictions).
>
>  The zone system module seems to have real problems getting to *just*
> the burned out highlights or even close to them in the highlights;
> I have to turn the number of zones up to the max to even get close to
> 'just the burned out area or really close'. And then it's once again
> awkward to select and act on just the highlights.
>
>         - cks
>
>
> ------------------------------------------------------------------------------
> _______________________________________________
> Darktable-users mailing list
> [email protected]
> https://lists.sourceforge.net/lists/listinfo/darktable-users
>
------------------------------------------------------------------------------
_______________________________________________
Darktable-users mailing list
[email protected]
https://lists.sourceforge.net/lists/listinfo/darktable-users

Reply via email to