Hi again,

We've now figured out what's different about this camera and that
explains why a few others are also broken. Unlike most bayer cameras
these don't have 2 green pixels in each 2x2 block. Instead they have a
Cyan/Yellow/Green/Magenta pattern[1][2]. There is at least also a
Red/Green/Blue/Emerald camera, the Sony F828[3]. Our demosaic process
currently assumes there are two identical colors in each 2x2 bayer
pattern and returns a 3 color image in the end. To be able to fix this
the demosaic code would have to return 4 colors sometimes so we could
use a 4x3 color matrix to convert to RGB. That's a difficult thing to
implement and there are very few and only very old cameras that need
it so it's unlikely we'll be doing it soon if at all.

Sorry for the bad news,

Pedro

[1]https://en.wikipedia.org/wiki/CYGM_filter
[2]http://www.dpreview.com/reviews/nikoncp5700/2
[3]http://www.dpreview.com/reviews/sonydscf828/

On Sun, Dec 20, 2015 at 7:04 PM, Pedro Côrte-Real <pe...@pedrocr.net> wrote:
> Thanks for the file, it would be good if you could submit it as a bug
> report though instead of sending it to everyone on the mailing list.
> The camera is currently broken in darktable and we don't know how to
> fix it. It apparently uses a different kind of bayer pattern we've
> never seen in another camera.
>
> Sorry we can't be more helpful,
>
> Pedro
>
> On Sun, Dec 20, 2015 at 6:46 PM, Venu Gopala Rao Arakali
> <garak...@gmail.com> wrote:
>> Dear Mr. Pedro Côrte-Real,
>>
>>
>> Thanks for your kind response to my query. As you observe, Nikon E 5700 is
>> on the list of cameras supported by Darktable. I have a Nikon Coolpix 5700.
>> Again, as you did, I too presume that the names Nikon E 5700 and Nikon
>> Coolpix 5700 refer to the same Nikon model and DT supports Nikon Coolpix
>> 5700. To follow up your suggestion to open a bug-report and include a sample
>> NEF file in it (and as I do not know how to do it formally),  I have
>> attached a (compressed) NEF image file from my camera Nikon Coolpix 5700
>> below, and observe that this file gets corrupted when opened in Darktable
>> 1.6.9. Incidentally, this file DSCN1663.NEF opens up normally with an viewer
>> such as Shotwell.
>>
>> With regards,
>>
>> --
>> Dr.A.V.Gopala Rao
>> Professor of Physics (Retired)
>> Department of Physics, University of Mysore
>> Manasagangotri, Mysore 570006,INDIA

------------------------------------------------------------------------------
_______________________________________________
Darktable-users mailing list
Darktable-users@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/darktable-users

Reply via email to