Am Dienstag, 8. Oktober 2013, 11:05:45 schrieb Anton at the-wire:
> zapataz said the following on 10/07/2013 01:13 PM:
> > F600EXR and HS30EXR don't use Fuji X-Trans CMOS sensor. Camera using it
> > are X-Pro1, X-E1, X-M1 and X100s.
> > 
> > Darktable relies mostly on it's own algorithms to process RAW files and a
> > bit from extra libraries.
> Having Linux tools makes 'detective' work a joy :-)

If it's for the joy then ignore this mail, if it's for actually learning how 
darktable works read on. :)

> Ultimately those libraries get back to the sources for dcraw.

That might be true for many but not for all. Especially rawspeed is not 
inheriting from the dcraw sources and it's our main means to read raw files. 
For formats/cameras that are not supported by rawspeed we resort to libraw 
which uses dcraw's code internally.

> Running ldd on darktable shows one of its libraries is libGraphicsMagick

ImageMagick or GraphicsMagick or whatever flavour we happen to link against is 
used to load otherwise unsupported file formats like PNG, GIF, ... as a very 
last resort. I don't know if it would be able to load raw files using that 
mechanism, but I doubt it.

So you might wonder, if dcraw can read X-Trans files and many other weird Fuji 
stuff, and if libraw inherits its code from dcraw, and if darktable can use 
libraw to read raw files, then how come that darktable can't open these files? 
The reason is quite simple and also the cause why adding support to these files 
to darktable is hard and nothing that you should hold your breath for: we only 
use these libraries to get the raw bayer data out of the containers that these 
files are. Together with some metadata like black levels, white point, ... We 
however don't use these libraries to transform the raw data into real images, 
we are just interested in the sensor readings. The actual demosaicing step is 
done later inside our processing pipeline. And since all the modules up to 
that step need to know how to handle the raw data we really want to keep 
ourselves sane and limit this to just support regular bayer patterns. So all 
the abominations coming from Fuji are beyond our scope. At least until someone 
comes up with a solid design and the code to support these, too.

[unrelated findings since we don't use any of that to load raw files]

HTH,
Tobias

Attachment: signature.asc
Description: This is a digitally signed message part.

------------------------------------------------------------------------------
October Webinars: Code for Performance
Free Intel webinars can help you accelerate application performance.
Explore tips for MPI, OpenMP, advanced profiling, and more. Get the most from 
the latest Intel processors and coprocessors. See abstracts and register >
http://pubads.g.doubleclick.net/gampad/clk?id=60134071&iu=/4140/ostg.clktrk
_______________________________________________
Darktable-users mailing list
[email protected]
https://lists.sourceforge.net/lists/listinfo/darktable-users

Reply via email to