Re: [darktable-user] White balance Spot Selection & History stack copying

2018-08-12 Thread Jochen Keil
Hi Bernhard,

sorry for the late reply, I've been on vacation.

I looked briefly at the resources you provided, but I think they don't
solve the problem of flickering white balance / color temperature.

However, in the mean time I cloned darktable's source code repository
and skimmed through the code. I more or less understand how modules
work and I think (?) that the color picker needs mouse focus to work,
which is why it cannot work in the background. I wonder if it's
possible to get the color information for all pixels in the image and
calculate an intermediate color temperature, maybe using a bias value.

If someone can help me on

1) how to get all the necessary color information (e.g. using the
color picker module without GUI interaction(?))
2) the algorithm on how to calculate the color temperature using a bias value

then I think I might be able to write a module for auto white balance.

Btw. Rawtherapee has an AWB mode which works exactly the way described above. :)

Cheers,

  Jochen



On Wed, Aug 1, 2018 at 7:21 PM Bernhard  wrote:
>
> I remembered there was something that reminded me at lrtimelaps - but using 
> darktable - in the past.
> A quick search showed this up:
> https://discuss.pixls.us/t/how-do-you-create-a-time-lapse-on-darktable/8094/14
>  is the newer approach
> https://www.youtube.com/watch?v=C-0bCAIJR0c is the older one.
> Perhaps this might give useful hints (?).
>
> --
>
> regards
> Bernhard
>
> https://www.bilddateien.de
>
>
> Jochen Keil schrieb am 01.08.2018 um 15:05:
>
> Hi,
>
> On Wed, Aug 1, 2018 at 1:31 PM,  wrote:
>
> Jochen Keil (2018-Aug-01, excerpt):
>
> Is there a way to tell darktable that it should analyze each image on its
> own using the selected spot area for white balance?
>
> I doubt it.  I've been asking for the same thing with respect to
> exposure correction (metering a certain region on all images for
> neutral gray).  While “exposure” has an automatic mode (which totally
> did not work for me), there's no such thing in “white balance”.
>
> Implementing this would require darktable to offer two choices when
> copying a setting to another image: Do you want to copy *how* the
> change was determined and apply the result (your use case), or copy
> the change itself (e.g., copy white balance form one photo of a white
> reference to all photos from the same shoot)?  It would be nice to
> have this, but it's probably quite nasty to implement in the
> details...
>
> that's sad to hear. I'd have figured that it might be easy to implement,
> but then I don't know anything about the inner workings of darktable.
>
> Well then, I can only hope that this feature makes it to one of the future
> versions of darktable.
>
> Any suggestions on alternatives or how to fix this?
>
> Thank you,
>
>   Jochen
> 
> darktable user mailing list
> to unsubscribe send a mail to darktable-user+unsubscr...@lists.darktable.org
>
>
>
>  
> darktable user mailing list to unsubscribe send a mail to 
> darktable-user+unsubscr...@lists.darktable.org

darktable user mailing list
to unsubscribe send a mail to darktable-user+unsubscr...@lists.darktable.org



Re: [darktable-user] Philosophical question

2018-08-12 Thread Timothy Spear
In theory, the data is converted to a standard, the unknown elements are
supposed to be added as maker's notes.
You have to look at the specific DNG conversion software to determine how
good a job it will do.

The second most common reason to not convert is because it adds another
step to the process flow and uses computing power; which may affect you
depending on your workflow. Since I import at the end of the day/trip and
rarely start editing the same day I have not had an issue letting the
computer chug away converting images while I go eat, sleep, check email
Since I take pictures for personal/hobby I am not time constrained and the
extra step does not bother me.

The third most popular reason I have seen is that at some point a brilliant
idea in image processing will occur, and it depends on the original raw
files.

The most common reasons to convert are to deal with longevity of the data;
since we are already seeing problems recovering/viewing data in other
industries in many other fields. However, I am not concerned about this one
yet. Since I generally lag on updating my software except for security
patches, when support for an older format is dropped I expect will have
time to convert any file I have not done so. I would also expect Adobe,
Microsoft, Apple, dt etc... to provide a warning message that the next time
we upgrade it will no longer be able to read X. Lastly, worst case, I just
load a VM with an older version of the software, do the conversion and
re-import the data.

The primary reason I have converted from CR2 and JPEG to DNG was for the
inherit hash embedded in the file which can help prevent/notify you of bit
rot. Since I have previously experience with bit rot I consider this worth
it.

Which reminds me, I need to put in a new ticket issue to suggest a
validation function to dt which will store a hash for both the XMP and the
original image, and can run in a background process.

Tim

On Sun, Aug 12, 2018 at 12:39 AM, Mike Strock 
wrote:

> What do I lose converting from CR2 RAW files to DNG? I always shoot raw
> but is there a benefit or a degradation to converting to DNG for future
> proofing my images?
>
> If this is off topic I apologize.
>
> Mike.
>
> Sent from ProtonMail Mobile
> 
> darktable user mailing list to unsubscribe send a mail to
> darktable-user+unsubscr...@lists.darktable.org


darktable user mailing list
to unsubscribe send a mail to darktable-user+unsubscr...@lists.darktable.org