Help with Creating DNG Spec Forwarding Matrix

[Adam Bonn]

Advertisement (gone after registration)

Hi everyone,

I was hoping to get some advice..

The Premise

I'd like to create forwarding matrices for use with the 'embedded' profiles that Leica include with their cameras.

My Problem

I've read (and re-re-re-re-read!!) the adobe DNG spec document and I just can't get my head round what I need to do...

I think I'm ok with the maths.... 

I'm lead to believe that the Forward Matrices (hereon in 'FM' for brevity) can be calculated using the data found in the Colour Matrices (CMs)

The maths for this seems fairly simple....

From Adobe:

              First, invert the XYZtoCamera matrix. 

             CameraToXYZ = Inverse (XYZtoCamera)

            Second, the white balanced transform is computed
            CameraToXYZ_D50 = CA * CameraToXYZ

(CA is an equation found on Bruce Lindbloom's site) 

I'm using excel to perform the maths...

(MINVERSE to inverse matrices and MMULT to multiple them)

I'm not managing to get values that work (hell I'm not even managing to get values that look right!!)

• I'm fairly sure that the my problem is the data I'm using to base the XYZtoCamera inversion on (but don't take my work for it !!)

Originally I thought that this was simply the CMs inversed, but I guess not....

I think I have to apply a matrix calculation to the CMs in order to obtain the correct values to invert

I was hoping that someone here could help walk me through it....

There are a few blogs etc that detail this process, however they start from a base colour target and build a profile from scratch... what I want to do is piggyback from Leica's existing CMs

Thanks in advance for any help

 

 

[Jeff S]

Maybe some of the information presented here might help...

https://www.odelama.com/photo/Developing-a-RAW-Photo-by-hand/Developing-a-RAW-Photo-by-hand_Part-2/
 

Jeff

[Advertisement]

The Leica Q3 is available for pre-order at B&H Photo Video and Adorama

[Adam Bonn]

Thanks Jeff

I’ve looked at that site before and it is very informative, but it starts from the base of independent data collection 

 

 

[Adam Bonn]

Perhaps I need to phrase my question differently 

using nothing more than the two CMs and the ‘as shot neutral’ tag value that Leica embed into the dng 

how do i created a single matrix using mired interpolation?

I would like to do this in excel. I don’t have matlab.

Anyone?

[jaapv]

I am wondering: what would be the advantage of this exercise over profling existing raw conversions?

[Adam Bonn]

I wish to add the 'missing' forward matrices to leica's embedded profiles 

[jaapv]

Advertisement (gone after registration)

I see. Would it not be far simpler to create colour profiles in your raw converter?

[Adam Bonn]

42 minutes ago, jaapv said:

I see. Would it not be far simpler to create colour profiles in your raw converter?

Yes (and I've done this in the past)

But I want to learn how to do the thing that a member did to some degree of fanfare.

See here

 

Seems like we were both convinced!

[Adam Bonn]

So @jaapv do you know how to do this?

[jaapv]

Prevent a yellowish cast under LED light?
Set the camera colour balance by a greyscale reading off a white piece of paper in the relevant light. Then warm up the image to taste using the yellow-blue slider in the colour balance dialog.
Otherwise, in Photoshop, change the contrast curve of the *B channel in LAB..

[Adam Bonn]

4 hours ago, jaapv said:

Prevent a yellowish cast under LED light?
Set the camera colour balance by a greyscale reading off a white piece of paper in the relevant light. Then warm up the image to taste using the yellow-blue slider in the colour balance dialog.
Otherwise, in Photoshop, change the contrast curve of the *B channel in LAB..

No not that

Calculate the correction forwarding matrices from the embedded colour matrices using chromatic adaptation

[jaapv]

I wouldn't...

[Adam Bonn]

30 minutes ago, jaapv said:

I wouldn't...

Wouldn’t or couldn’t? And if wouldn’t, why not?

Edited April 1, 2021 by Adam Bonn

[jaapv]

Possibly couldn't too, mainly due to laziness in adding brain clutter, but wouldn't because I see no advantage over profiling and editing in a colour-managed workflow