Hi Francisco
I think the darks have to be very dark in the image to see artifacts --
if they even happen. When you choose the eyedropper from tone curve and
click in the image, the L levels in the darkest (and artifact-ed) bits
of the X-Trans image have L values of 1 or 2 (and are changed to about
11 or 12 by the curve). In the Nikon image they're not that dark
(generally about 15, a few are as dark as 5, which gets scaled up to
approx. 12), so there's not as disproportionate an amplification of the
chroma. It would be interesting to see an image with bright reds in
deep shadow.
I'm talking through my hat here, but it's also possible that the input
profile for the Nikon image does a better job keeping the chroma values
in check? Or there could just be a difference in how the sensors record
darker colors?
Best, Dan
On Mon, Jan 5, 2015, at 09:56 AM, Francisco Cribari wrote:
> Hi Dan:
>
> Here is a photo I've just made (for testing purposes) of my wife
> wearing a red shirt (.NEF, Nikon D800).
>
> RAW (.NEF) file:
> https://dl.dropboxusercontent.com/u/2171814/_F1C6322.NEF
>
> XMP file: https://dl.dropboxusercontent.com/u/2171814/_F1C6322.NEF.xmp
>
> Best, Francisco
>
>
> On Mon, Jan 5, 2015 at 11:22 AM, Dan Torop <d...@pnym.net> wrote:
>> __
>> Hi Francisco,
>>
>> Do you have an example NEF file? One with shadows with saturated
>> colors would make this most visible. The Nikon sensor could also be
>> less noisy in the shadows.
>>
>> I'm not sure if these artifacts are just a cost of automatic chroma
>> scaling, or should be considered a bug for a corner case.
>>
>> Best, Dan
>>
>>
>> On Mon, Jan 5, 2015, at 06:03 AM, Francisco Cribari wrote:
>>> Hi Dan:
>>>
>>> Thank you for your message. I have been noticing color artifacts
>>> when raising the lower point of the tone curve when editing
>>> X-Trans RAW images but not when editing standard Bayer RAW images
>>> (e.g., Nikon D800 .NEF). Is there anything specific to X-Trans RAW
>>> images here?
>>>
>>> Thanks again. Best, Francisco
>>>
>>> On Mon, Jan 5, 2015 at 12:56 AM, Dan Torop <d...@pnym.net> wrote:
>>>> __
>>>> Hi Francisco,
>>>>
>>>> This is interesting...
The trouble seems to come via the tone curve module's automatic scale
chroma setting. (The noise doesn't seem to depend upon the demosaicing
algorithm -- it's there with VNG or Markesteijn.)
>>>>
>>>> When
"scale chroma" is set, each pixel processed has its saturation ("a" and
"b") adjusted proportionally to how its luminance ("L") is changed.
That is, if the luminance for that pixel is increased by 5%, its
saturation is increased by 5%. For the tone curve you drew, which
pushes up the black point, this breaks down. The very dim pixels in
this curve have their luminance increased by approx. 10x, and hence
"scale chroma" then correspondingly saturates them by about 10x. As the
shirt is so red, this makes the saturation for the shirt especially
garish in its darker areas (indeed, the "a" value becomes greater than
100). Furthermore, the saturation changes are jumpy because slight
luminance changes in the shadows produce radically amplified saturation
changes -- and this makes visible the noise inherent to the shadows.
All this adds up to the horrible artifacts which you are seeing.
>>>>
>>>> There
probably would be a way to rewrite the tone curve's "scale chroma" to
handle this case, but it wouldn't be backwards compatible, so doubtful
dt devs would go for it? (There already seems to be some code there to
better handle very low luminance values.)
>>>>
>>>> What
if instead you use "scale chroma" set to "manual"? There's no artifacts
then, and the results appear pleasing. Alternately, in the tutorial you
link to, it looks as though the tone curve change is happening in RGB
colorspace, not Lab. You could simulate this in dt via enabling the
color balance module and increasing the "factor" for its"lift"
parameter. Or even by lowering the "black" in the exposure module?
>>>>
>>>> Best, Dan
>>>>
>>>>
>>>> On Sun, Jan 4, 2015, at 06:22 PM, Francisco Cribari wrote:
>>>>> My second camera is a Fuji X100S, which has an X-Trans sensor. I
>>>>> am getting color artifacts in several images when I raise the
>>>>> lower end of the tone curve. The problem does NOT happen with my
>>>>> Nikon D800 RAW images, only with X-Trans RAW images. I would also
>>>>> like to point out that it is common to raise the lower end of tone
>>>>> curve and to lower its upper end. This is done to give the image a
>>>>> "vintage/classic" look. See, e.g.,
>>>>>
>>>>> https://photographylife.com/that-classic-vintage-look
>>>>>
>>>>> For some reason, that tends to cause color artifacts in X-Trans
>>>>> RAW images processed using Darktable.
>>>>>
>>>>> Here is an example:
>>>>>
>>>>> RAW (.RAF) file:
>>>>> https://dl.dropboxusercontent.com/u/2171814/_FFX7093.RAF
>>>>>
>>>>> XMP file:
>>>>> https://dl.dropboxusercontent.com/u/2171814/_FFX7093_01.RAF.xmp
>>>>>
>>>>> Zoom in the image and take a look at the man's shirt (near his
>>>>> right arm). (I have already reported another case in this mailing
>>>>> mailing list.)
>>>>>
>>>>> I run DT version 1.6.1+30~g6b508d0 on an Ubuntu notebook. Any
>>>>> suggestions? Thank you. Francisco
>>>>>
>>>>> --
>>>>> Francisco Cribari - http://about.me/cribariFrancisco
>>>>> CribaraaahhhFrancisco Cribari - http://about.me/crihhhrancisco
>>>>> Cribari - http://about.me/cribariffaaaaFrancisco Cribari -
>>>>> http://www.cribari.com.br - "All theory, my friend, is grey, but
>>>>> green is life's glad golden tree." --Goethe (Faust)
>>>>>
>>>>> ------------------------------------------------------------------------------
>>>>> Dive into the World of Parallel Programming! The Go Parallel
>>>>> Website, sponsored by Intel and developed in partnership with
>>>>> Slashdot Media, is your hub for all things parallel software
>>>>> development, from weekly thought leadership blogs to news, videos,
>>>>> case studies, tutorials and more. Take a look and join the
>>>>> conversation now. http://goparallel.sourceforge.net
>>>>> _________________________________________________
>>>>> Darktable-users mailing list Darktable-users@lists.sourceforge.net
>>>>> https://lists.sourceforge.net/lists/listinfo/darktable-users
>>>>>
>>>>
>>>
>>>
>>>
>>> --
>>> Francisco Cribari - http://about.me/cribariFrancisco
>>> CribaraaahhhFrancisco Cribari - http://about.me/crihhhrancisco
>>> Cribari - http://about.me/cribariffaaaaFrancisco Cribari -
>>> http://www.cribari.com.br - "All theory, my friend, is grey, but
>>> green is life's glad golden tree." --Goethe (Faust)
>>
>
>
>
> --
> Francisco Cribari - http://about.me/cribariFrancisco
> CribaraaahhhFrancisco Cribari - http://about.me/crihhhrancisco Cribari
> - http://about.me/cribariffaaaaFrancisco Cribari -
> http://www.cribari.com.br - "All theory, my friend, is grey, but green
> is life's glad golden tree." --Goethe (Faust)
------------------------------------------------------------------------------
Dive into the World of Parallel Programming! The Go Parallel Website,
sponsored by Intel and developed in partnership with Slashdot Media, is your
hub for all things parallel software development, from weekly thought
leadership blogs to news, videos, case studies, tutorials and more. Take a
look and join the conversation now. http://goparallel.sourceforge.net
_______________________________________________
Darktable-users mailing list
Darktable-users@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/darktable-users
