I recently posted a tutorial on conversion of colour negatives with GIMP 2.10 and RawTherapee 5.8, using RawTherapee's Colour Toning tool to remove the film base colour cast.

Fellow forum member @jbinpg (Jack) informed me that Darktable has the same or similar colour adjustment capability in the form of its color correction module. I've now figured out how to apply this such that it achieves the exact same result as in RawTherapee. The method is basically the same, but with one small wrinkle, which is easily dealt with... Good news for forum members who'd like to try this negative conversion approach but prefer Darktable as their raw converter.

Here, then, is a condensed tutorial for the Darktable variation of the process. Please read it with reference to the original tutorial, replacing the RawTherapee steps with those below.

---

GIMP should be configured to use Darktable for importing and converting raw files.

After opening GIMP and loading the negative DNG file, the Darktable plugin is loaded. De-activate all optional modules except for white balance (output color profile, input color profile, demosaic and raw black/white point are permanently activated):



In the input color profile module, set input profile to standard color matrix; working profile should be left at the default of linear Rec2020 RGB; gamut clipping set to off:



Set white balance to camera:



Increase the reproduction ratio to 100% and navigate to an area of the photo's border. Select the color picker drop-down and click on the eye-dropper:



Place the colour picker on the border area. The current (R)ed, (G)reen and (B)lue values will be shown in the color picker drop-down:



As in the previous tutorial, the aim is to adjust the image until these R, G and B values are equal or close to equal.

Activate the color correction module. Click and drag the dot to adjust the balance of R, G and B values. Moving the dot LEFT increases the (G)reen value and decreases the (R)ed simultaneously, while moving it RIGHT increases (R)ed and decreases (G)reen. Moving the dot UP increases (G)reen and decreases (B)lue, and moving it DOWN increases (B)lue and decreases (G)reen:



Darktable's color correction module doesn't have the range of adjustment offered by RawTherapee's Colour Toning tool. Here, I've equalised the G and B channels at 207, but I've run out of range for R at 245 - I can't move the dot further to the left, and hence can't reduce the R value

The solution is to create a new instance of the color correction module, and continue the adjustments in this second instance, which has a cumulative effect on the first:



The R, G and B values are now identical. When Darktable is closed, the image will be transferred to GIMP for the rest of the workflow, as per the previous tutorial.

I hope this is helpful to those who prefer Darktable. The adjustments are just as effective and accurate, and in fact a little easier to make, since the color correction module is less sensitive to mouse movements than RawTherapee's Colour Toning tool.

Both RawTherapee and Darktable variants of the process produce near-identical and wondefully-accurate colours in final positive images

Have fun

[clickclick]

I really need to sit down with both of these programs and get to know them better.

[Medex]

What light (in Kelvins) do you use for picture taking ("negative scanning using DSLR)? I suggest to use 6000-6500 K

[BigMackCam]

I really need to sit down with both of these programs and get to know them better.

Both have good online documentation these days, and whist it's sometimes a little technical in nature, it's good support when you're experimenting with the the various modules in each one...

What light (in Kelvins) do you use for picture taking ("negative scanning using DSLR)? I suggest to use 6000-6500 K

I've read differing advice - some that suggests 5000K is optimal, some suggesting 6500K. Kaiser's panels produce ~5000K light (mine measures 5301K, based on the readings from my K-3).

Having said that, the colour temperature at which the camera profile was created is also important. Ideally, you'd use the the same light source as for profile creation, as otherwise, you're relying on interpolation for colour accuracy. So, with the Camera Standard profiles taken at 6500K, using a light panel with 6500K CCT should be ideal. Practically speaking, I doubt it makes a huge difference.

CRI is far more important than CCT, in my view. White balance can deal with different CCT rather well, but with a low or unknown CRI light source, it's almost certain there'll be poor reproduction of certain colours. Reds are often problematic with lower-CRI light sources, and even some moderately high-CRI ones (CRI R9 measurements, if available, can be very enlightening)...

---

Last edited by BigMackCam; 02-21-2022 at 01:43 AM.

[Medex]

Both have good online documentation these days, and whist it's sometimes a little technical in nature, it's good support when you're experimenting with the the various modules in each one...



I've read differing advice - some that suggests 5000K is optimal, some suggesting 6500K. Kaiser's panels produce ~5000K light (mine measures 5301K, based on the readings from my K-3).

Having said that, the colour temperature at which the camera profile was created is also important. Ideally, you'd use the the same light source as for profile creation, as otherwise, you're relying on interpolation for colour accuracy. So, with the Camera Standard profiles taken at 6500K, using a light panel with 6500K CCT should be ideal. Practically speaking, I doubt it makes a huge difference.

CRI is far more important than CCT, in my view. White balance can deal with different CCT rather well, but with a low or unknown CRI light source, it's almost certain there'll be poor reproduction of certain colours. Reds are often problematic with lower-CRI light sources, and even some moderately high-CRI ones (CRI R9 measurements, if available, can be very enlightening)...

Yes, LEDs have very different CRI. I did some "film negative scanning using DSLR" job couple of years ago and you just did remind me that I always set custom WB using LED light and clear film frame (that part of the film in the beginning of the roll that is not exposed to light). IMO this should eliminate at least in part color cast that is determined by film. I did not try 5000 K light, but using 3700K and 4000K light always was associated with more blue cast that was impossible to correct using Temperature slider in Lightroom. 6000-6500 K light gives more blue for negative, but after conversion to positive - more red colors ant they can be easily cooled down if needed by Temperature slider.

[BigMackCam]

I did some "film negative scanning using DSLR" job couple of years ago and you just did remind me that I always set custom WB using LED light and clear film frame (that part of the film in the beginning of the roll that is not exposed to light). IMO this should eliminate at least in part color cast that is determined by film.

Setting custom camera white balance against the unexposed portion of the film doesn't work very well, unfortunately. Nor does using the raw converter's white balance eye-dropper to sample and white balance against the border. Both approaches remove the colour cast from the border, but they affect all the other colours in the exposed photo. The result is a positive with a colour cast that's very difficult to remove effectively. The photo may still look OK after tweaking, but it won't be colour-accurate.

This is really the whole point of my previous tutorial and this Darktable-specific update. The film base doesn't change the colour temperature of the light shining through it... rather, it adds a tint. By using RawTherapee's colour toning tool or Darktable's color correction module, we can subtract that tint in much the same way, I believe, that acetates or a dichroic enlarger head would in darkroom wet-printing. The result is much greater colour accuracy...

[clickclick]

Both have good online documentation these days, and whist it's sometimes a little technical in nature, it's good support when you're experimenting with the the various modules in each one...

Especially when I think of our own Squirrel Mafia's New and Improved Raw Therapee thread! I really have no excuses.

[mir]

With apologies to the OP for the slight detour, and not wishing to "pimp" my own threads, but if you haven't already reviewed the below articles, you might wish to take a look. I spent a lot of time developing a workflow for colour negatives with GIMP and either RawTherapee or Darktable, that gives consistently better results than the built-in modules in either of those raw converters. There are more steps involved, but it's easy and - as I said - consistent. I now use this approach in preference to the built-in modules, and even in preference to Lightroom 6 with NegativeLabPro...

Colour-Accurate Conversion of Negatives using GIMP 2.10 and RawTherapee 5.8 - PentaxForums.com

Colour-Accurate Conversion of Negatives - Darktable variant - PentaxForums.com

I have now tried your step guide in Darktable an the results is impressive and way better and simpler (for me at least) than the negadoctor module working with color negatives. I have a suggestion and a question:
Question: Why do you not activate the lens correction module when in Darktable mode in Gimp? (it cannot be applied afterwards in Darktable)
Suggestion: For RawTherapee you refer to storing the adjustment in a template. In Darktaple it is called a style an can be created via the little button besides the big button 'compress history stack'

[BigMackCam]

I have now tried your step guide in Darktable an the results is impressive and way better and simpler (for me at least) than the negadoctor module working with color negatives.

Thanks for the nice feedback! I'm so pleased you like the process and results

... I have a suggestion and a question:

Question: Why do you not activate the lens correction module when in Darktable mode in Gimp? (it cannot be applied afterwards in Darktable)

You're quite right, I should have included lens corrections for completeness. My intention was to demonstrate the elements of processing that are essential to the negative conversion process, rather than dictate the entirety of Darktable processing (which might differ for each user). However, depending on the lens used for capturing the negatives, it may indeed be important to apply corrections. I use my K-5 or K-3 APS-C cameras for digitising, with either a D FA100/2.8 Macro WR for 35mm negatives or Sigma 50/2.8 EX DG Macro for 120. Both are full frame lenses, and on APS-C they show little appreciable distortion or vignetting at f/8... so for my own purposes, I haven't been applying lens corrections - but really I should. Thanks for pointing this out

Suggestion: For RawTherapee you refer to storing the adjustment in a template. In Darktaple it is called a style an can be created via the little button besides the big button 'compress history stack'

Excellent - thanks for letting me know! That's one of the reasons I enjoy these forums... I learn so much from other members

I asked the mod team to move your last post and my reply to this thread, as it's helpful to other members and improves upon the process I'd documented. I hope that's OK...