deeply messed up negative

As suggested by @age in another post, I’m bringing this deeply messed up negative with the potential not only to exercise negative inversion, but also local color corrections (via curves as @age mentioned, or darktable color balance module).
There’s an ongoing and extremely promising effort in refurbishing negative inversion in darktable. No, actually, in making a brand new and better tool, negadoctor, by @aurelienpierre. I risk making bad marketing of the new tool with this messy digitezed negative and careless edit.
Here’s my try.

IMG_7291.cr2.xmp (22.0 KB)
(works only on negadoctor branch)
IMG_7291.cr2 (11.5 MB)

Since some tools need parts of unexposed film to properly invert the image, follows another image from the same film strip for that sole purpose.
IMG_7286.cr2 (11.8 MB)

This file is licensed Creative Commons, By-Attribution, Share-Alike.

1 Like

IMG_7291.cr2.xmp (30,9 Ko)

The best I got so far.


Is that Kodak?

IMG_7291rt.cr2.pp3 (12.2 KB)


Without addressing individual parts (the vignetting and the person for example) I came up with this:

deeply.messed.up.negative.cr2.xmp (8.7 KB)

Darktable: Future negadoctor version.

BTW: I’m not able to load the xmp by @gadolf (version mismatch 1!=1), @aurelienpierre’s xmp load nicely though. I freshly (re-)installed this morning due to missing colour pickers, so I might have a newer version in comparison with gadolf.


This is normal, I have added 2 parameters yesterday, without taking care of old parameters import.

I imported the raw image into imageJ using bioformats. The histograms look very, very strange to me:

Note the spikes in R and G!


1 Like

My try in darktable 3.0

IMG_7291.cr2.xmp (13.9 KB)


Very unexpected edit and very nice indeed!

OK, I need to be honest and say that I was only expecting darktable posts that use the negative modules, be it the current invert or the future negadoctor. I was not expecting a neat trick primarily based on the rgb curve module!

Sorry, I was just messing around. That was be no means meant to be “scientific” or the usual way to develop negatives :wink: .

Here is one where I also tried to remove the color casts in the lower corners. Darktable is a great an flexible tool!

IMG_7291.cr2.xmp (19.7 KB)


@Thomas_Do Colours still odd but feels the most natural thus far.

I think I have a pretty decent edit here with RT (dev branch, but should be okay to open in RT5.8)

IMG_7291.jpg.out.pp3 (12.3 KB)


Here’s my second try with the latest negabranch release, which I have to say, is far more easier to set up.
Given the high dynamic range of the scene, I chose to use tone equalizer to recover the shadows, which is a bit trickier because at that point in the workflow you’re still working with the negative image. The rest is only negadoctor workflow.

IMG_7291.cr2.xmp (5.4 KB)

No, Fuji 400 ASA (does code w16b mean anything?)
Is this a problem, since the module is based on Kodak Cineon densitometry model?

EDIT: Here’s another scan from the same film. If I use only the color pickers, I end up with a somewhat greenish image. Because of that, I used color balance on the highlights. If I recall, standard Fuji prints always presented this kind of greenish look, but I may be wrong.

IMG_7290.cr2.xmp (8.0 KB)
IMG_7290.cr2 (12.0 MB)
This file is licensed Creative Commons, By-Attribution, Share-Alike.

Just move it after the inversion (Shift + Ctrl + drag).

Kodak Cineon is just a densitometry model to produce prints. It’s pretty generic.


IMG_7291.cr2.xmp (8.9 KB)

Sorry for the late reply, but i couldn’t resist :slight_smile:
Here’s my attempt, using RT with the filmneg_stable_mult PR (the profiles don’t work with the 5.8 release) :

IMG_7290.cr2.pp3 (11.7 KB)

IMG_7291.cr2.pp3 (11.7 KB)