I'm experiencing some weird white balance problems with the current git
version of Darktable and D7100 RAWs, and I'm wondering if anyone else is
seeing this (with a D7100 or another camera) and whether other people
can reproduce this.
It's easiest to explain what I'm seeing with a sample NEF and its
embedded preview:
http://www.cs.toronto.edu/~cks/tmp/darktable/DSC_4429.NEF
http://www.cs.toronto.edu/~cks/tmp/darktable/DSC_4429_preview4.jpg
This picture was taken in overcast conditions shortly before (local)
sunset using the camera's AWB. Conveniently the large amount of snow
makes it easy to both white balance the picture and to assess the white
balance.
If you bring up this NEF and JPEG preview pair up side by side in a way
where what's displayed for the NEF is the darkroom-processed version
you can see that by default, simply applying the embedded white balance
information, darktable gives the NEF version a pronounced magenta
cast relative to the JPEG preview (the JPEG preview version's white
balance is a bit too magenta, but not anywhere near as much as the NEF
version). That's the first problem. I know that tone rendition varies
between a DT-processed NEF and the preview image, but I assume that the
white balance should be the same.
(I don't know where this issue comes from; the actual RGB multipliers
darktable claims to be applying are what the EXIF information records to
three decimal places. I don't know if DT is rounding these off even for
internal processing or just for display.)
The next problem is that darktable has the (NEF) colour temperature
pegged at '23000K', its maximum; among other effects this means that you
can no longer usefully manipulate the colour temperature (you have to
play with the r/g/b multipliers by hand). Using manual white balance on
the snow brings down the red and blue multipliers somewhat but does not
bring it under 23000K as darktable computes things.
(Clearly this white balance is very, very off. I built a test version of
DT that allowed the white balance kelvin numbers to go higher and had to
go up to absurd numbers before it didn't max out.)
A current version of Rawtherapee produces a similar default magenta
rendering of the NEF but reports the white balance as 10302 K on the
base version and 11130 K on a spot white balanced version (and thus
allows you to still manipulate the white balance further). This seems
much closer to what's likely the actual reality. It would be nice if
DT matched that white balance number and thus allowed one to still
manipulate things.
- cks
------------------------------------------------------------------------------
Rapidly troubleshoot problems before they affect your business. Most IT
organizations don't have a clear picture of how application performance
affects their revenue. With AppDynamics, you get 100% visibility into your
Java,.NET, & PHP application. Start your 15-day FREE TRIAL of AppDynamics Pro!
http://pubads.g.doubleclick.net/gampad/clk?id=84349831&iu=/4140/ostg.clktrk
_______________________________________________
Darktable-users mailing list
[email protected]
https://lists.sourceforge.net/lists/listinfo/darktable-users
