[EN] Emulating film look on digital from Kodak Portra scans

[@ankipruthi]

Just yesterday I was trying to find resources on Kodak Portra 400 NC and how to replicate that look inside of Darktable, and today you upload this. Thanks a bunch!! Looking forward to watching the whole thing. I am curious about learning the differences between a digital raw and a negative film scan converted and developed. And also thanks for pointing out how Darktable has become so bloated and starting the R&Darktable fork. Even as an amateur photographer, I dont want to spend a lot of time in post processing each and every raw. Honestly, these camera manufacturers are skipping the biggest advantages of digital photography by having crappy color science and completely dropping out all the knowledge of color that humanity collected over several decades of developing various films. Photography should be about spending more time taking lots of images and less time fixing things in post and reading research papers, writing scripts and what not.

[@tobiasandersen172]

Thank you for these videos! Watch them all. A great help.

[@pedrorrodriguez1]

Thank you for the video! Full of very interesting tips on color grading, halation and grain. I'd personally be really interested in an improved grain module for darktable. Congratulations and keep up the great work!

[@deansonneborn7134]

The master class as usual. Thanks

[@laurimela118]

Hi Aurelien and many thanks for your efforts with Darktable! With the negadoctor inverted image you mention about how the input color profile after negadoctor module is illogical. I think it makes perfect sense. With linear scans as input to Darktable we have RGB density of the color negative. The color negative is not for direct viewing by humans and thus not applicable for use with an input profile. What I mean is that we should not care what perceptual color the scanned film negative itself has. This is because C-41 color negative is designed to be "seen" by RA-4 paper, not by our vision. Of course, at some point in the Darktable pipeline we need the input color profile to go from RGB values to perceptual color. With negadoctor inverted film negatives this should probably be a profile that converts input RGB light intensity to the color (reflectance) as produced by an RA-4 paper. Using something like AdobeRGB or SRGB there is quite arbitrary, even if it can produce reasonable looking results.

It doesn't make sense because the scanner records an RGB image of the negative. Any scan is an image of an image. But the scanner adds some color distortions that need to be corrected first, which is the purpose of the input color profile. Once this is applied, you get the theoritically perfect scan of the negative, which you can then invert and correct for the negative issues. What I did here doesn't take the scanner flaws into account, which is a shame because it's the whole selling point of Silver Fast.

[@laurimela118]

@ I agree that there can/should be a linearizing input profile to use with a color negative film scanner, but under no circumstances can it be a profile that converts the RGB to perceptual color. Converting film negative to perceptual color and then inverting is nonsense. C-41 with RA-4 does not work like that, neither does Kodak Cineon scanning system.

@@laurimela118 I don't know what you call perceptual color, since color is by definition perceptual. Color profiles embed the cooking recipe to go from any color space to a connection space, usually CIE XYZ 1931. Since anyway the scanner records an RGB image of the negative, it can perfectly be linearized against XYZ.

[@drendelous]

where are you my friend? pls come back

[@josefsmolik5991]

Thanks for the nice video, but unfortunately the subtitles don't work in other languages.

[@josefsmolik5991]

Please, can you check why your video only has English subtitles. If I switch to Czech, the subtitles are no longer displayed. Isn't it because you lowered the video quality? It's a shame, I don't speak English, so the video isn't very useful to me. I wish you much success and thank you for your contribution to darktable.

I have no idea what KZbin does to subtitles. Anyway, the English subtitles are auto-generated so we are dealing with an artificial intelligence here and those are delicate. I have no control over the auto-translations of the auto-subtitles.

[@josefsmolik5991]

@ "This video was recorded in HD 1080p at 10fps with mono audio to save up to 5% bandwidth and 35% processing power for decoding compared to an average video (HD 1080p, 30fps, stereo)." Namá Does this setting affect the translation of the subtitle in other languages? Previously, the subtitles from you were fine. This change may have affected the possibility of translations. It's a shame, this way your videos will be useless for non-English speakers.

@@josefsmolik5991 My videos have been recorded like that for 2 years or so, it's not a change. Again, I'm sorry but I have no control over the translations, it's all Google AI.

[@josefsmolik5991]

@ Thanks for the reply, and I'll hope it goes well again next time.