On 18/05/17 21:31, Jean-Luc CECCOLI wrote:

Maybe because the effect produced by the module affects the final result, whatever the values are set to.

Setting them to "zero" will enhance the result in a neutral way, which does not mean "with no effect".

Moving the sliders will then increase or decrease the initial effect.

I feel like having written martian... very bad feeling... did someone understand what I mean ?


What I see it could be happening is the following: the contrast brightness saturation module will normally change the L,a,b components of the image, and then will guarantee "reasonable" value. This probably means that (as an example) L is capped at 0 --- because negative L make no sense.

Now, if the image arriving from the pipeline is not gamut clipped, and so has *illegal* values, the CBS module even if you chose 0-0-0 will still clip to reasonable value.

In my opinion the problem is that the two modules are not agreeing on a basic idea: either you let the values pass thru even if they have no physical meaning, hoping that some future processing can use the information, or you clip them so that you always transmit "legal" values. The surprise arise, I think, because the input color profile (at least by default) and the CBS module do not agree on this policy.

Or maybe I am completely wrong...

Romano




    > Message du 17/05/17 20:11
    > De : "I. Ivanov" <[email protected]>
    > A : [email protected]
    > Copie à :
    > Objet : Re: [darktable-user] contrast brightness saturation
    changes image even without changing values
    >
    > Hi Tobias,
    >
    > The concern is actually not that the blue cast exists. The
    concern is
    > that it changes from blue to purple by just enabling the "contrast
    > brightness saturation module" - without populating any values in it.
    >
    > My expectation is that if I am to enable the module and do not
    modify
    > any of the values - brightness / contrast or saturation - all of
    them
    > remain zero - then it should not alter the image (but it does).
    >
    > Regards,
    >
    > B
    >
    >
    > On 2017-05-17 01:32 AM, Tobias Ellinghaus wrote:
    > > Am Dienstag, 16. Mai 2017, 18:42:23 CEST schrieb I. Ivanov:
    > >> Hi All,
    > > Hi.
    > >
    > >> Stumbled on the following.
    > >>
    > >>
    https://www.dropbox.com/s/nwq5lt3r8d1b2ju/Screenshot%20from%202017-05-16%201
    > >> 8-33-53.png?dl=0
    > >>
    https://www.dropbox.com/s/ef1un9w1oca8rl0/Screenshot%20from%202017-05-16%20
    > >> 18-33-10.png?dl=0
    > >>
    > >> The only difference between the 2 images is turning ON or OFF
    Contrast
    > >> Brightness Saturation. However - the values remain 0,0,0
    > >>
    > >> I expect that there should be no change as even if the module
    is active
    > >> - I have not changed any values within it
    > >>
    > >> Even the histogram changes by just enabling the module and
    the exported
    > >> images are certainly different.
    > >>
    > >> Is this expected? I did not expect it.
    > > What you are seeing is a problem with very saturated blue. Try
    enabling the
    > > gamut clipping in the input color profile (2nd dropdown
    there). Set it to
    > > "linear Rec2020 RGB" for a start.
    > >
    > >> DT 2.2.4 on Ubuntu 16.04.
    > >>
    > >> Regards,
    > >>
    > >> B
    > > Tobias
    >
    >
    ____________________________________________________________________________
    > darktable user mailing list
    > to unsubscribe send a mail to
    [email protected]
    >
>

____________________________________________________________________________ darktable user mailing list to unsubscribe send a mail to [email protected]

--
Romano Giannetti
http://www.rgtti.com/


____________________________________________________________________________
darktable user mailing list
to unsubscribe send a mail to [email protected]

Reply via email to