demosaicing white balance method camera

[S. Burmeister]

Hi!

If I use dcraw with -w (camera white balance) to demosaic a RW2 file it shows
the following:

dcraw -w -v P1030225.RW2

Scaling with darkness 15, saturation 4095, and
multipliers 1.901141 1.000000 1.745247 1.000000

If I open the same image in digikam's image editor (2.0) with "white balance
method" set to "camera" the debug output shows:

Scaling with darkness 0, saturation 4080, and
multipliers 1,000000 0,526000 0,918000 0,526000

Should those not be the same numbers as above?

Sven
_______________________________________________
Digikam-users mailing list
[hidden email]
https://mail.kde.org/mailman/listinfo/digikam-users

[Gilles Caulier-4]

Good question. libraw used by libkdcraw and finally digiKam is the dcraw API (but not only).

In libkdcraw, there are some dcraw like command line tool to check functionalities.

Try to process your RAW file with it and look traces. It's the same than dcraw ?

Gilles Caulier

2011/8/2 Sven Burmeister <[hidden email]>

Hi!

If I use dcraw with -w (camera white balance) to demosaic a RW2 file it shows
the following:

dcraw -w -v P1030225.RW2

Scaling with darkness 15, saturation 4095, and
multipliers 1.901141 1.000000 1.745247 1.000000

If I open the same image in digikam's image editor (2.0) with "white balance
method" set to "camera" the debug output shows:

Scaling with darkness 0, saturation 4080, and
multipliers 1,000000 0,526000 0,918000 0,526000

Should those not be the same numbers as above?

Sven
_______________________________________________
Digikam-users mailing list
[hidden email]
https://mail.kde.org/mailman/listinfo/digikam-users

_______________________________________________
Digikam-users mailing list
[hidden email]
https://mail.kde.org/mailman/listinfo/digikam-users

[Remco Viëtor]

On Tuesday 02 August 2011 06:44:49 Gilles Caulier wrote:

> Good question. libraw used by libkdcraw and finally digiKam is the dcraw API
> (but not only).
>
> In libkdcraw, there are some dcraw like command line tool to check
> functionalities.
>
> Try to process your RAW file with it and look traces. It's the same than
> dcraw ?
>
> Gilles Caulier
>
> 2011/8/2 Sven Burmeister <[hidden email]>
>
> > Hi!
> >
> > If I use dcraw with -w (camera white balance) to demosaic a RW2 file it
> > shows
> > the following:
> >
> > dcraw -w -v P1030225.RW2
> >
> > Scaling with darkness 15, saturation 4095, and
> > multipliers 1.901141 1.000000 1.745247 1.000000
> >
> > If I open the same image in digikam's image editor (2.0) with "white
> > balance
> > method" set to "camera" the debug output shows:
> >
> > Scaling with darkness 0, saturation 4080, and
> > multipliers 1,000000 0,526000 0,918000 0,526000
> >
> > Should those not be the same numbers as above?

Yes and no:
the difference in darkness and saturation are strange,
the white balance multipliers still have the same ratio though,
so the applied white balance correction is identical.
digikam apparently has chosen to always keep the factors <= 1,
this has the advantage that you won't blow a channel due to the white balance
correction. but of course the image will appear darker.

Remco
_______________________________________________
Digikam-users mailing list
[hidden email]
https://mail.kde.org/mailman/listinfo/digikam-users

[S. Burmeister]

Am Dienstag, 2. August 2011, 06:44:49 schrieb Gilles Caulier:
> In libkdcraw, there are some dcraw like command line tool to check
> functionalities.
>
> Try to process your RAW file with it and look traces. It's the same than
> dcraw ?

I used dcraw_emu and got the same "wrong" values for darkness and saturation
as digikam shows in its output.

For the multipliers it shows the same values dcraw shows, i.e. other values
than the digikam debug output.

Sven
_______________________________________________
Digikam-users mailing list
[hidden email]
https://mail.kde.org/mailman/listinfo/digikam-users

[S. Burmeister]

Am Dienstag, 2. August 2011, 07:27:00 schrieb Remco Viëtor:
> the difference in darkness and saturation are strange,
> the white balance multipliers still have the same ratio though,
> so the applied white balance correction is identical.
> digikam apparently has chosen to always keep the factors <= 1,
> this has the advantage that you won't blow a channel due to the white
> balance correction. but of course the image will appear darker.

Shouldn't this be an option rather than forced? Channels are not always blown
due to white balance correction and in those cases the darker image is yet
another obstacle in processing the picture. Though it is certainly useful to
be able to enable this feature.

Sven
_______________________________________________
Digikam-users mailing list
[hidden email]
https://mail.kde.org/mailman/listinfo/digikam-users

[S. Burmeister]

Am Dienstag, 2. August 2011, 08:41:05 schrieb Sven Burmeister:
> Shouldn't this be an option rather than forced? Channels are not always
> blown due to white balance correction and in those cases the darker image
> is yet another obstacle in processing the picture. Though it is certainly
> useful to be able to enable this feature.

I am not sure but one might argue that currently digikam does not use the
camera's white balance (since the camera delivers absolute values) but only
the camera's white balance ratio – so strictly speaking it does not obey the
user's settings – even if for a reason.

Sven
_______________________________________________
Digikam-users mailing list
[hidden email]
https://mail.kde.org/mailman/listinfo/digikam-users